Washington, D.C.
The Subcommittee on Standards and Security of the National Committee on Vital and Health Statistics (NCVHS) held hearings on May 29-30, 2002, at the Hubert H. Humphrey Building in Washington, D.C. The meeting was open to the public. Present:
May 29-30, 2002
On May 29-30 the Subcommittee on Standards and Security held the third of several hearings planned for 2002 on code set issues in an ongoing process focused on administrative simplification provisions of the Health Insurance Portability and Accountability Act (HIPAA). The first day's focus was on the HIPAA diagosis codes; the second day they returned to the procedure code issues identified in April. During the first day, the Subcommittee heard 17 presentations and held discussions concerning the possible replacement of ICD9 clinical modification (ICD-9CM) vol. 1 and 2 (diagosis) with ICD-10-CM, and issues related to transistions to new code sets. The second day, five presenters testified on procedure code issues.
Ms. Trudel announced that three HIPAA regulations would be published in the Federal Register on Friday: one related to the DSMO addenda, another addressed the National Drug Code and other changes related to the retail pharmacy transactions, and the final rule for the employer identifier that adopted the IRS's employer identification number. The Subcommittee will develop a letter in response to the notices of proposed rule making and vet it with the full Committee in June.
Possible Replacement of ICD-9-CM vol. 1 & 2 (diagnosis) with ICD-10-CM: Dr. Ed Sondik, Director, NationalCenter for Health Statistics
Dr. Sondik emphasized that classification formed the core for producing health statistics and that it was critical that public health and epidemiological databases were founded on the most current internationally recognized versions of ICD. ICD-10-CM provided additional classification detail, included logical restructuring in the relative placement of certain entities, and remedied cumbersome classification dilemmas that impaired ICD-9-CM. The structure and conventions in ICD-10-CM were similar to ICD-9-CM; anyone familiar with ICD-9-CM could move to ICD-10-CM without extensive retraining. Dr. Sondik said enhanced ease of coding, benefits of added specificity, improved clinical accuracy and consistency offset conversion costs. Dr. Sondik expressed confidence that ICD-10-CM was well-positioned to replace ICD-9-CM and would provide the data necessary for statistical reporting and administrative transactions. The National Center for Health Statistics (NCHS) believed ICD-10-CM should replace the diagnosis portion of ICD-9-CM as soon as possible, with a target date no later than October 2005.
Overview of ICD-10-CM: Donna Pickett, National Center for Health Statistics
Ms. Pickett said the DSMO worked closely with medical specialty surgical societies, professional organizations, and other federal agencies in creating ICD-10-CM. Codes were restructured to provide additional room for expansion and more enhancements were added. As a sixth character and laterality were added, some of the more complex structures of ICD were modified to make it a true diagnosis classification, and relationships with procedures and procedure codes were removed from the classification. Trimester information was added to the OB codes and diabetes codes were revised to be consistent with American Diabetes Association's new categorization of diabetes. Injury, post-operative complications, alcohol and substance abuse codes were expanded and code extensions were added for injuries and external causes of injury. A new concept added to ICD-10 by the World Health Organization (WHO) was “activity”:what the person engaged in when injured. Work continued on an alpha version of the database and further guidelines to make use of the classification adequate, accurate and consistent over time. Parts of ICD-10-CM were harmonized with DSM-4, the cancer registries, and diagnosis in the nursing classification.
Summary of April 15-16, 1997, NCVHS Hearings on Code Sets: Lizabeth Fisher, NCHS
Ms. Fisher summarized the pivotal April 1997 meeting of the Subcommittee on Health Data Needs, Standards and Security. The theme for reporting diagnoses was to continue to use ICD-9-CM and move to ICD-10-CM following development and testing. Many requested public domain access to code sets; open input and access to the maintenance process; uniform coding guidelines, reporting requirements, and code acceptance by all users; cost/benefit analysis done prior to moving to ICD-10-CM; and linking ICD to other systems. Other recommendations and feedback included: changes must be compatible to all payment systems (e.g., DRGs, APCs); one system might not serve all data users' purposes; consider all users, rather than just hospitals, in selecting and implementing code sets; use existing/proven coding systems; and consider the impact of these systems in migrating to the computerized patient record.
Panel 1: Possible replacement of ICD-9-CM, Vol. 1 & 2 (Diagnosis) with ICD-10-CM
Dr. Powers said ICD-9-CM couldn't keep up with the expansion of knowledge and changes in neurologists' diagnostic classifications. The number of available fifth-digit codes was almost nil, and codes and commonly used diagnoses were scattered in sundry places. Tension headache was coded with “psychalgia” and essential tremor was listed under a non-specific code, creating access issues. Transverse myelitis was given an inappropriate code that implied an infectious etiology. Older classifications were a huge problem: e.g., mitochondrial diseases were essentially non-existent. Noting the American Academy of Neurology's own extraction of the neurologic codes from ICD-9-CM used a crosswalk with ICD-10-NA's neurologic adaptation for research coding for stroke, Dr. Powers said the Academy was pleased to proceed to ICD-10-CM for stroke diagnoses. She said ICD-10 was far more similar to the language physicians used to describe diagnoses. The Academy supported proceeding with implementation of ICD-10-CM as a national classification system binding to all third party payers, with rules binding to all payers.
Dr. Mirin said ICD-10 could reflect DSM-4 compatible names for mental disorders and allow adoption of an alphanumeric code set that provided space for identifying more scientifically reliable, valid disorder definitions--facilitating documentation of prevalence and treatment rates of mental disorders and the public health burden caused by emerging new disorders and drugs of abuse. However, ICD-10-CM had no operationalized definitions or diagnostic criteria and only an introductory statement referring to DSM-4 for definitions of psychiatric terms and conditions. Noting that continued use of two separate code sets for mental disorders seemed inconsistent with HIPAA's goal of administrative simplification of diagnostic reporting, Dr. Mirin asked NCVHS to approve: adoption of DSM-4 criteria as the official descriptors for ICD-9-CM; designation of American Psychiatric Association (APA) as DSMO for these criteria; and approval and implementation of ICD-10-CM for HIPAA reported requirements at the earliest possible date.
Dr. Strunk said the American College of Obstetricians and Gynecologists (ACOG) believed ICD-10-CM offered greatly enhanced capabilities for identifying and tracking important obstetric conditions and interventions aimed at reducing adverse pregnancy outcomes. He noted addition of a fifth digit to report the trimester of pregnancy would help monitor the provision of prenatal care and occurrence of complications. ICD-10-CM permitted identification of individual fetuses in a multiple gestation; a specific condition could finally be linked to a specific fetus. It included expanded codes for reporting abnormal findings on antenatal screening and additional episodes of care. ICD-10-CM also included a category for care provided to the fetus, due to abnormalities. ACOG believed that the benefits offered by ICD-10-CM were substantial enough to justify the costs and disruption adoption would inevitably entail.
Dr. Warner said adoption of ICD-10-CM would establish compatibility with US mortality data, provide greater detail and details more appropriate for a wider range of users, and enable future expansions allowing for better data describing external causes and nature of injury. Injury codes in ICD-10-CM enabled users to differentiate between initial and subsequent visits and the sequelae of injury; the level of detail available to code the nature of injury and external causes was also greatly improved and kept up with the current knowledge engine in injury prevention and control. Dr. Warner emphasized the importance of having comparable codes for datasets to facilitate data analysis: incompatibility between ICD-9-CM and ICD-10 had made analysis of data more difficult, time consuming and results less precise. ICD-10-CM offered improvements to the collection of injury morbidity data and the surveillance of injury and poisoning in the U.S. and allow for comparability with mortality data in this country and morbidity data in others.
Mr. Muldoon said the National Association of Children's Hospitals and Related Institutions (NACHRI) found limitations with the specificity of the information in in-patient and population-based ICD-9 classification systems running comparative reporting programs and policy analyses. He reiterated NACHRI's 1997 recommendation to move forward with ICD-10-CM. The many generic codes in ICD-9 were extremely frustrating and one lost sight of the nature and status of an injury. Areas of specific interest for pediatric, the congenital anomaly and perinatal codes, were particularly underdeveloped in terms of specificity. Often very different conditions were in the same code. Many codes were clustered and had to be broken out. Mr. Muldoon reiterated NACHRI's strong support for the movement from ICD-9-CM to ICD-10-CM.
American Health Information Management Association (AHIMA) believed that replacement of the 9-CM diagnosis codes was necessary and that final regulations promulgating adoption of 10-CM should be issued no later than 2003, with an implementation date no later than 2005. Ms. Prophet said the definition of the ICD-10-CM coding standard should include system rules and guidelines in a process allowing public input to codes and that maintenance should be continued with implementation of 10-CM. An advisory group comprised of representatives of stakeholders should provide input into the maintenance and guideline development processes. AHIMA espoused six principles for medical code set maintenance processes. Ms. Prophet cautioned against neglecting the magnitude of the training issue. She said AHIMA was ready to take the lead in national workforce retraining and user education for the new procedural and diagnosis coding systems.
Ms. Frazier said American Academy of Professional Coders' (AAPC's) written testimony concurred with many of Ms. Prophet's comments. AAPC agreed that the current ICD-9 made for imprecise coding and inaccurate documentation for reimbursement. AAPC coders heartily supported the uniform coding guidelines' reporting requirements and code acceptance by all users. Ms. Frazier said AAPC would work with NCHS in disseminating training materials.
The American Hospital Association's (AHA's) Central Office on ICD-9-CM served as the U.S. clearinghouse for issues related to ICD-9-CM. AHA considered ICD-9-CM diagnosis codes outdated and insufficient to provide detail for non-acute conditions and AHA supported migration to ICD-10-CM. AHA advised that the timeline for implementation should be carefully orchestrated to minimize administrative burden to providers. AHA believed the cost of implementing significant new regulations should be worked into the Medicare prospective payment rate update and that Congress should establish grants to help hospitals with the enormous cost of complying with the HIPAA rules, including conversion to ICD-10-CM. ICD-10 had been implemented internationally in diagnosis related groups (DRGs). AHA believed hospitals were capable of taking on the challenge of implementing ICD-10-CM and stood ready to provide ICD coding advice and training.
Ms. Smith conveyed McKesson Information Solutions' belief that the 10-CM structure was logically organized and provided greater clarity for coding diagnosis. Increased specificity enabled use of fewer codes to describe a condition (reducing coding errors) and facilitated DRG assignments. Greater exactness in linking patient diagnosis to out comes facilitated more accurate collection of outcomes data, leading to a greater ability to measure the effectiveness of services provided and management of disease processes, safety, and efficiency of care. Migrating to ICD-10-CM also allowed for expansive comparison of medical data internationally. Although McKesson believed the advantages of ICD-10 would prove beneficial for clinical usage, as an information solution provider, McKesson's interests rested mostly with the timing of ICD-10-CM.
Panel 2: Systems Issues Related to Transitions to New Code Sets (Vendor Perspective)
Tenet HealthSystems (THS) believed five things needed to be looked at: process, technology, technology's impact on other systems, training, and the perception of the health information management (HIM) departments in THS's facilities. Ms. Johnson predicted that these HIPAA changes wouldn't significantly impact the processes within HIM, but, she expressed concern that the data collected via the diagnosis and procedure codes was also used for credentialing, cost accounting, case and disease management. Those systems also needed to be ready to handle the new six-digit data format. She cautioned that institutions that coded manually faced a more labor-intensive effort and needed additional remedial education. Ms. Johnson emphasized that it was critical to develop a transition plan for old and new data.
Mr. Riopelle said Gambro worked diligently for two years to respond to HIPAA changes. Implementation of the transaction standards alone involved some 30,000 hours of professional time and hardware purchases totaling 15 percent of Gambro's IT capitol annual budget. The transition from local to standard code sets was a substantial effort for Gambro, which had over 13,000 potential payer relationships and over 23,000 local codes. The implementation process impacted other areas within IT and required a significant re-engineering of business processes. Another consideration for Gambro was the timing and interconnection of privacy, security and transactions. Mr. Riopelle encouraged the Subcommittee to continue considering all points of view as it evaluated the impact of code set changes and the timing of things so interconnected.
Mr. Hardesty noted the industry already struggled to achieve compliance with changing requirements and anticipated rules and didn't see how it could do this necessary and inevitable migration. If compliance with major components of HIPAA was achievable by the end of 2003, Cerner supported a moratorium during 2004, so the industry could assimilate necessary changes and mitigate difficulties. Implementation of ICD-10 and ICD-10-PCS should be tandem. And he urged the Secretary to issue industry guidance regarding the intent to migrate to ICD-10 and the Department's overall plans and directions. Cerner supported a three-year time frame for conversion after implementation of HIPAA's major components. Mr. Hardesty commented on implementation issues with both proprietary and legacy systems throughout the industry and the need for significant planning, testing, and coordination of efforts between vendors.
Ms. Smith reported that, from an information systems perspective, McKesson's interest rested mostly in the volume of changes that needed to occur to software applications and supporting documentation. She described some of those changes and illustrated the importance of the timing of the ICD-10-CM implementation. McKesson's preference was to implement ICD-10-CM and PCS concurrently. McKesson also believed that public maintenance of the 10-CM and PCS code set best served the vendor, provider and payer industry. Ms. Smith suggested that a group within HHS respond to questions as software developers crafted functional specifications for the required changes.
Mr. Mitchell noted the current code sets were embedded into virtually every portion of the healthcare system; changing code sets meant each system in place had to be reviewed. All interfaces between systems that transmitted coded patient data had to be reviewed to assess the impact of the new code sets and, as needed, changed. Data formats also would change to support characteristics of the new codes. While the effort to implement changes would be substantial, Mr. Mitchell said benefits derived from improved coding sets would outweigh the cost of implementation. 3-M-HIS encouraged the Committee to set a clear direction and timeline to allow providers and vendors an opportunity to effectively prepare for the change. Mr. Mitchell said Canada's and Australia's experiences indicated that with sufficient clarity and notice, necessary changes could be made in a timely manner, but full notice and a non-changing specification were key to implementing smoothly
Subcommittee Discussion of the ICD-10-CM Testimony
Based on what they'd heard, members discussed recommending to the Secretary that the Department publish a notice of proposed rule making (NPRM) proposing to move simultaneously to 10-CM and PCS for the uses ICD-9 currently addressed two years after publication of the final rule, which would be no later than October 2003, to meet an implementation date of October 1, 2005. Noting the absence of any payer on the panels and that they'd only heard from large vendors, members reflected that this was a major rule that would have significant impact on everyone and assessing and explaining that impact was critical. They noted that comments might lead them to recommend other actions.
Members discussed whether they should ask the Secretary to notify the APA that a crosswalk between DSM-4 and ICD-9 was compatible with the HIPAA standard. Ms. Pickett noted that, given what they'd heard that morning, they had to start anew on a white paper distinguishing between coding guidelines and diagnostic criteria. Ms. Humphreys concurred with assuring APA that mental health professionals could still use DSM. Ms. Pickett noted ICD-9-CM was the coding classification hospitals used, but mental health providers billed on a professional claim using DSM-4. APA tried to harmonize the numbers but some descriptors were different. Another issue AHA presented regarded diagnostic criteria and clinical descriptions linked to DSM's numerical “equivalent” of ICD-9-CM code. Descriptors didn't go into the HIPAA procedures and payers could misinterpret the DSM CM codes because of subtle distinctions between descriptors in DSM-4 and ICD-9-CM. Ms. Trudel suggested the Subcommittee wait for the white paper and a sense of all the issues.
A request was noted from APA that, if DSM was adopted as the HIPAA code set, the association be designated as a code set maintainer. Members agreed that after hearing tomorrow's testimony they'd draft a letter to the Secretary to bring to the June meeting. The staff will draft a letter on the NPRMs for the Subcommittee to edit.
Panel 3: Procedural Code Set Follow-up Reports -- HCPCS and AMA Update Processes
Noting the Subcommittee had expressed concern about the proposed semi-annual early release of CPT Category III codes for new and emerging technology and the lack of coordination between CPT and the Health Care Procedural Coding System's (HCPCS's) newly adopted quarterly update schedule, Mr. Beebe provided further information. The CPT editorial panel discussed the Category III code release and implementation dates with payer groups and agreed on a July 1 release of the new Category III codes for January 1 implementation. Mr. Beebe said AMA and the CPT editorial panel agreed on the need for greater communication and cooperation. AMA recommended that the CPT editorial panel be provided a seat on the HCPCS editorial panel. Mr. Beebe assured the Subcommittee that AMA had no intention of changing its licensing prices and policies. He said AMA was exploring ways of using the Internet to post agenda material and allow public comment.
Mr. Gustafson said three levels of codes were used for ambulatory procedures: Level I--the CPT 4 codes the AMA editorial panel maintained, Level III--local codes (used largely by state Medicaid agencies and local carriers) that were being phased out, and what HCPCS usually referred to, Level II--e.g., codes for drugs, durable medical equipment, medical and surgical supplies, outpatient pass-through items. Mr. Gustafson said payer systems permitted use of unlisted codes when a specific code wasn't available. Medicare established G codes to quickly add a code for professional services used within its system. Noting provisions for quarterly updates of C codes used in Medicare's hospital outpatient perspective payment system to identify newly arrived drugs and devices, he commented on the trade off between the speed one could move and the administrative burden on stakeholders. And he noted that information found in an implementation guide was in the operating instructions from payers. Mr. Gustafson said a schedule of public hearings gave manufacturers three opportunities annually to make presentations to CMS.
Follow-up Report from Alternative Link and AMA
Michael Beebe, Director, CPT Editorial and Information Services, American Medical Association
Mr. Beebe said AMA and the editorial panel recognized that, as a designated national HIPAA code set, CPT might be required to develop codes to describe alternative therapy services and expand HCPCS's membership to include professional organizations representing additional alternative therapy providers. But he emphasized that any attempt to expand or change the CPT code set had to follow a well established, publicized and documented CPT editorial process. Mr. Beebe said the panel appointed an Alternative Therapy Workgroup that included Alternative Link to evaluate and develop descriptions for alternative therapy services for possible inclusion in CPT. Mr. Beebe said AMA made serious attempts to work with Alternative Link, but the meetings hadn't produced agreement. Mr. Beebe delineated the difference between creating codes for a service and importing codes for that service developed by another entity that could erode a critical level of consistency in procedure coding, code application and clinical decision making about what constituted appropriate descriptions of healthcare services. Mr. Beebe contended that AMA didn't intend to utilize Alternative Link's pre-existing codes, but to create new codes that met CPT editorial panel standards and were an integral part of a coherent work of coding. Once the Alternative Therapy Workgroup established a functional framework and alternative therapy codes were developed by AMA through the editorial panel process, AMA offered to collaborate on development of a crosswalk from CPT alternative therapy codes to alternative therapy codes in Alternative Link's ABC coding system.
Ms. Gianinni said AMA's newly proposed alternative therapy CPT coding initiative established AMA as a direct competitor, rather than collaborator, and that during those interactions AMA was unreliable, self-serving and predatory. She said the role AMA proposed for itself would position it to: delay code set availability and keep alternative medicine, nursing and other non-allopathic physician practitioners disenfranchised. Ms. Gianinni said AMA members were politically active and strategically placed to maintain control over this critical component of the healthcare infrastructure, despite well known weaknesses in the CPT code set and editorial process. She noted that in 1993 NCVHS reported that AMA's code development process was flawed and widely viewed as favoring AMA membership. AMA's CPT editorial panel and CPT advisory committee were almost exclusively conventional medical doctors. Despite the advisory bodies within AMA, she said final decisions were made by a decision making body that was almost exclusively allopathic physicians. Ms. Gianinni used an analogy from the grocery industry to illustrate why stakeholders benefiting from an existing supply chain infrastructure shouldn't have control over coding for new technologies. And she cautioned that public health might suffer if AMA was granted coding authority over alternative medicine, nursing and other non-allopathic physician interventions.
Impact of Elimination of Local Codes on Claims Processing for Nursing Services:
Carol Bickford, Registered Nurse and Senior Policy Fellow, Department of Nursing Practice and Policy, American Nurses Association.
The American Nurses Association (ANA) didn't considered the removal of local codes a significant issue because nursing had never used them, but Ms. Bickford noted substantial coding issues remained. Nursing's holistic healthcare considered the patient and his or her environment in context and was a very different practice model from the pathology-based, diagnostic and procedure HIPAA code sets. ANA hadn't focused on reimbursement, but on the patient and describing reality. ANA strove to integrate nursing languages and code sets into information systems environments that supported their description of clinical practice. Standardized code sets were being developed that permitted linkage of the diagnosis or problem, expected outcomes, interventions and actual outcomes for evidence-based practice. Nursing languages were incorporated into the UMLS. SNOMED CT First Release included NANDA Taxonomy II and the nursing diagnoses/problems from Home Health Care Classification (HHCC), Omaha System, and the Perioperative Nursing Data Set (PNDS). Ms. Bickford noted ANA had been a long-time participant in standards and code set development with ANSI, ASTM, HL7 and at the international level. ANA had been working on ICD-10-CM and had participated in CPT-4 and the CPT-5 discussion for the proposed revision. Ms. Bickford identified substantive coding issues confronting the nursing specialties. ANA recommended: implementing ICD-10-CM for diagnoses as soon as possible, reflecting nurses and other non-physician provider's work in codes and systems, including attachments to HL7 messages as an automated solution to tease out clinical details that better informed decision-making about quality and cost, and designating permanent nurse seats on the Committee and CPT editorial panel.
Subcommittee Discussion on Issues Presented During Morning Sessions
Dr. Zubeldia said they'd heard again that second morning the recommendation to move forward with ICD-10 for the same uses ICD-9 diagnosis and procedures codes now served. Mr. Beebe noted the Committee's interest in collecting additional data on the impact of PCS. Members summarized yesterday's discussion. The Committee might recommend to the Secretary that they move forward with an NPRM related to ICD-10-CM and ICD-10-PCS for in-patient only. Implementation would be simultaneous. A two-year-plus implementation period would follow the final rule. The effective date of no sooner than October 2005 would be on-cycle (October or January). There was a need for a date certain, a recommendation for corresponding coding guidelines, and acknowledgement that both CM and PCS were broken. Noting they'd heard that proposing a rule followed by implementation two years later, without a strong statement about a specific time frame, might just “put the monkey on somebody else's back” while sustaining the difficulty of describing what went on in hospitals, members considered including an end date in the recommendation. Recalling they'd heard a lot about implementation planning, Dr. Cohn suggested they needed to inform the Secretary that, as this moved forward, HHS carried a responsibility to develop a national implementation plan and assist with successful implementation. Members noted they'd heard that WEDI-SNIP should be involved and milestones had to be a part of the implementation plan. A draft letter to the Secretary will be circulated to the Subcommittee and the full committee prior to the next Committee meeting, and time will be provided on the June breakout agenda for final wordsmithing.
Subcommittee Discussion on Additional Carryover Issues
Ms. Pickett asked to move the open issue about DSM being a crosswalk to the June or subsequent agenda. Based on APA's testimony, the Subcommittee's additional discussions, and what they'd heard yesterday, a number of issues had to be rolled into the white paper. Version control was another open issue. Today's protocol was that the codes were effective on the date service was rendered or, in the case of in-patient discharge, the date of discharge. But some payers had grace periods and continued to submit old codes because the provider or payer system hadn't been updated to reflect them and some state DRG programs used two-to-three year old groupers that didn't accept new codes. Members noted that standard use of code set guidelines was an across-the-board issue with various payers interpreting use of the code sets differently. AHA or AHIA could provide details about this issue and about statutes of limitation in various segments of the industry. Ms. Greenberg noted that the issue was whether payers actually grasped that, in the post-HIPAA world, the whole intent was to get rid of local variations. Dr. Cohn suggested they might need a letter for the September full Committee meeting stating concern about version control, guidelines, the standard use of codes, and other issues and recommendations the Secretary had to highlight and communicate so the industry was aware of them. Ms. Gilfoy noted testing had to begin by April, assuming everybody filed for an extension, and changes would have to accommodate whatever the letter said. She also noted that requiring the 837 institutional and professional formats and contents for the same service by different payers was a major issue that, in some cases, made payer-to-payer coordination of benefits impossible. Members noted they could hear in August from WEDI, the DSMOs, UBC, and UCC and others about the problem and possible solutions. Hopefully their letter could offer solutions.
Ms. Leon-Chisen suggested that the Subcommittee split the issues of version control and guidelines. The guidelines issue was more complex and probably would take longer to work out. AHA will work with CMS, AMA and others, providing examples and helping develop guidelines. Dr. Cohn clarified that this issue was already divided into situations without mandated guidelines and situations with guidelines that people didn't realize they needed to follow. Ms. Humphreys recommended that CMS involve itself more directly with the Medicaid group in terms of old groupers. Dr. Cohn said the Subcommittee would talk with the DSMOs about their timeline and availability and whether they wanted to begin discussing these and other issues in August or October.
Subcommittee Discussion of Future Agendas for HIPAA Code Set Issues
Dr. Cohn acknowledged receipt of letters from the National Association of Community Pharmacists and another group regarding a single or dual format for retail pharmacies related to professional supplies and services. The issue will be part of the Subcommittee's October discussion about the yearly process. He noted they had to ask whether the DSMOs had other issues to discuss in August.
Dr. Cohn noted the full Committee will meet on June 26-27; the Subcommittee scheduled hearings for August 28-29, October 22-23, and December 10-11. Ms. Trudel reported the NPRM was posted on the CMS Web: www.cms.hhs.gov/hipaa. Dr. Cohn noted that after the thirty-day comment period the Subcommittee would be reviewing NPRMs. He noted there would be barely enough time to draft a letter for discussion at the June Subcommittee and full Committee meeting. There wouldn't be time for further revision after those meetings, but they could fix it up then and get it approved. He anticipated the letter wouldn't be lengthy and would be supportive, mirroring recommendations and aspects they'd heard. Dr. Cohn said a major discussion at the June meeting would be the Administrative Simplification Compliance Act and the Subcommittee's review of a sample of the compliance delay requests. Mr. Augustine was developing an approach and methodology. Hopefully, members would receive a draft before the June meeting. Ms. Trudel said requests were being filed electronically seven days a week; slightly over a thousand had been filed so far.
Dr. Cohn said the main focus of the August meeting hadn't been finalized. Noting some discussions around carryover issues were pertinent to HIPAA implementation, Dr. Cohn said they'd get back to them. They also sought an understanding from the industry about other issues. If the DSMOs were prepared, they'd renew those discussions in August. The second half of the year, the Subcommittee hoped to assess lessons learned so far from the implementation and consider suggestions for how they might learn more and make recommendations to the Secretary on how to improve the process. Ms. Greenberg noted the Subcommittee would be reviewing the sample and recommending best practices and ways to overcome problems. They'd also begin looking at vocabulary for patient medical record information (PMRI) standards. Dr. Cohn said he was still determining how all the pieces fit in the puzzle; hopefully in June they could lay out plans for the fall hearings. Mr. Blair remarked that it would help to know the industry's response to any letter HHS sent out concerning incentives for guidance, based on the Committee's initial recommendations on PMRI message format standards.
Day One
Dr. Zubeldia introduced the meeting as a continuation of the April 9-10 hearings that looked at the code sets and applicability of ICD-10. The April hearings had focused on the procedure codes. That day they looked at the diagnosis codes. The next day they would return to the procedure code issues. Dr. Zubeldia noted that these were very controversial coding issues and that the Committee's recommendations potentially could have considerable impact on the industry
Ms. Trudel announced that three HIPAA regulations would be published in the Federal Register that Friday: one related to the DSMO addenda, another addressed the National Drug Code and other changes related to the retail pharmacy transactions, and the final rule for the employer identifier that adopted the IRS's employer identification number. All had a 30-day comment period. Dr. Cohn congratulated Ms. Trudel and the Department on getting those regulations out and said the Subcommittee would develop a letter in response to the notices of proposed rule making and vet it with the full Committee at the June meeting.
Possible Replacement of ICD-9-CM vol. 1 & 2 (diagnosis) with ICD-10-CM: Dr. Ed Sondik, Director, National Center for Health Statistics
Dr. Sondik said NCHS's mission was to provide statistical information to guide actions and policies that improve the health of Americans and that the International Classification of Diseases and Related Health Problems (ICD) and a variety of other classifications comprised the core of what NCHS did. ICD was developed and maintained collaboratively between WHO and 10 international centers. NCHS housed the WHO collaborating center for the family of international classifications for North America and had responsibility for implementation of ICD. ICD was the core of the family of WHO classifications that include ICF, ICD-02 for oncology, and various specialty adaptations. In 1989, the 43rd World Health Assembly adopted ICD-10. The United States implemented ICD-10 beginning in 1999.
The ICD-9-CM was used in the U.S. since 1979. Adapted from ICD-9 through extensive modifications including the use of fifth-digit subclassifications, ICD-9-CM's structure permits statistics to be collapsed back to ICD-9 for comparability between mortality and morbidity statistics. ICD-9-CM was adapted and adopted by the federal government and the private sector for a number of purposes including statistical reporting, data collection, quality of care analysis, resource utilization, research and reimbursement.
ICD-9-CM was updated annually to accommodate changes in medical technology and the need to provide greater specificity in classifying diagnosis and external causes of injury. The ICD-9-CM Coordination and Maintenance Committee handled modification requests. The committee, co-chaired by NCHS and the Centers for Medicare and Medicaid Services (CMS), provided a public forum to discuss possible updates and revisions to ICD-9-CM. The committee discussed a variety of topics including the need: to update due to changes in medical technology, correct inaccuracies in classification, and add greater specificity in classifying diagnoses. Copies of the meetings' topics and summaries were on NCHS's Web site. Dr. Sondik noted ICD-9-CM was rapidly losing its capacity to incorporate additional specificity. Few areas remained in ICD-9-CM for expansions needed to enhance classification details. In 1993, NCVHS recommended that the Department commit resources to assess the applicability of ICD-10 for coding morbidity, identify problem areas, make necessary modifications, and develop implementation plans. The Center for Health Policy Studies subsequently evaluated ICD-10 as a statistical classification for morbidity reporting in the U.S. and developed recommendations to improve ICD-10 and correct identified problems. Dr. Sondik noted ICD-10 had greatly expanded and included significant improvements in coding primary care encounters, external causes of injury, mental disorders, neoplasms and preventive health as well as advances in medical technology that occurred between revisions. Other advances in ICD-10 included new histologic types and more space to accommodate expansions. By 1996, 28 countries had implemented ICD-10 for mortality or morbidity applications and WHO authorized publication of ICD-10 in 37 languages. Every G-7 nation (except the U.S. and Italy which use ICD-9-CM) and Australia moved to ICD-10 or a clinical modification for its morbidity data.
Stating that ICD-10-CM was a significant improvement in clinical specificity, ease of use, and accessibility over ICD-10 and ICD-9-CM, a technical advisory panel convened by the NCHS contractor strongly recommended that NCHS implement an improved clinical modification version and that ICD-CM's tabular list and alphabetic index be adopted and implemented as the standard US classification as soon as possible. Further work on enhancements to ICD-10-CM were carried out through the review of the Coordination and Maintenance Committee minutes, collaborations with surgical and medical specialty groups, and in response to comments.
Dr. Sondik said NCHS believed ICD-10-CM provided additional classification detail and included logical restructuring in the relative placement of certain entities. He said ICD-10-CM remedied many cumbersome classification dilemmas that impaired ICD-9-CM (e.g., a major expansion in the chapter dealing with factors influencing health status, contact with health services, and the musculo-skeletal chapter). Dr. Sondik expressed confidence that ICD-10-CM was well-positioned to replace ICD-9-CM and would provide the data necessary for statistical reporting and administrative transactions. He noted, however, that some were concerned that implementation costs, including the expense of retraining and converting computerized information reporting systems based on ICD-9, might outweigh the advantages. The structure and conventions in ICD-10-CM were similar to ICD-9-CM: Dr. Sondik said anyone familiar with ICD-9-CM could move to ICD-10-CM without extensive retraining. But the alphanumeric structure of ICD-10 required software modifications and there was the cost of developing new edits, lists of valid codes, and other support packages. However, Dr. Sondik said enhanced ease of coding, benefits of added specificity, and improved clinical accuracy and consistency offset conversion costs.
Dr. Sondik said the availability of educational materials and training programs along with the HIPAA-mandated 24-month lead time before implementation further allowed for a smoother transition. It also facilitated integration of ICD changes with other system changes and new system development and enabled spreading the cost of implementation over several years.
The states struggled to maintain comparability between mortality statistics captured in ICD-10 and morbidity statistics captured in ICD-9-CM. NCHS will make available crosswalks between ICD-9-CM and -10-CM and planned to offer ICD-10-CM in both book and electronic formats. Noting some stated they'd prefer to continue updating ICD-9-CM, Dr. Sondik cautioned that that decision would make national and international comparisons and aggregations difficult. He said this was particularly true because parts of ICD-10 were already implemented in some fields. ICD-10's neoplasm chapter was comparable with ICDO-2, used in cancer registries since 1995. Chapter five (mental health disorders) was comparable to the fourth edition of the Diagnostic and Statistical Manual for Mental Disorders (DSM-4).
Dr. Sondik said continued use of ICD-9-CM in administrative transactions would be ill-advised beyond 2004: the ability to update ICD-9-CM beyond 2002 was limited and might eventually lead to a loss of data, and negative program impacts for federal agencies and other organizations already engaged in international collaborative efforts or using mortality and morbidity data together. Dr. Sondik noted many in the U.S. and the international community awaited ICD-10-CM and questioned why implementation lagged behind other countries; the U.S. was considered a leader in mortality and morbidity classification and several countries based their clinical modifications of ICD-10 on U.S. work. In 1999, the Institute of Medicine's Committee on Injury Prevention and Control recommended that high priority be given to insuring uniform and reliable coding of the external cause and nature of injury using ICD on all health systems data, particularly hospital and emergency departments' discharge records. The committee said special efforts should be directed at training to ensure optimal use of ICD-10. Dr. Sondik expressed NCHS' belief that the U.S. had an international obligation through treaty with WHO to implement ICD-10.
Noting some cited the need to conduct cost benefit or return-on-investment analysis, Dr. Sondik encouraged efforts to minimize costs with appropriate tools and cross blocks. But he observed that quantifying benefits to public health and healthcare of better information could be challenging. In order to improve quality of life and health status, it was critical that public health and epidemiological databases be founded on the most current internationally recognized versions of ICD. The creation, review and revision of health policy relied on the availability of accurate and timely healthcare data generated by providers and payers, statistical surveys, and other health research efforts. He reiterated that classification formed the core for producing health statistics and that the most current classifications available were needed. NCHS believed ICD-10-CM should be implemented as soon as possible, with a target date no later than October in 2005.
Discussion
Ms. Trudel confirmed that DSM-4 and ICD-10-CM, chapter five were almost identical. Dr. Sondik clarified that he advocated implementing ICD-10-CM by October 2005, noting that provided latitude. Asked about his sense of the costs and final benefits, Dr. Sondik said benefits were more difficult to address in a comprehensive way. He depicted them as providing a different language that enabled everyone to determine things and better understand qualities of care. He predicted the greatest benefit would come from what was gleaned from the research that could be done with this new classification. Because the classification was more specific in terms of morbid diagnoses, they might be able to parse out a diagnosis and better understand a variety of conditions within it and the costs of care and allocation of attendant resources. He noted an ability to share data internationally (increasingly a problem without this change) would be enormously useful.
Ms. Humphreys commented that she'd attempted to map more current clinically specific terminologies to ICD-9-CM and it hadn't work well. Mapping current, updated clinically specific terminology to the classification and diagnosis systems was extremely difficult because of the outdated organization within ICD-9-CM. She'd tried but couldn't get a current view of oncology or even breakdowns in various types of diseases with a worldview nearly 30 years old. She cautioned that this was a major problem with ICD-9-CM, if the Committee was looking toward the time when clinically specific information could be captured and higher-level aggregations generated accurately for statistical and billing purposes.
Dr. Sondik suggested that a way to quantify the need to move forward in dollar terms would be to compare differences between ICD-9-CM and ICD-10-CM. He predicted that they'd find that the differences were so significant and so likely to grow that they required a framework more attuned to medicine today and the way it would be over the decade.
Dr. Fitzmaurice said he resonated with two things Dr. Sondik said: needing to understand the health system better and the unknown value of research. He noted the $1.3 trillion health system accounted for about 14 percent of the gross domestic product and needed to be understood better. Whatever part of that was spent on human genome research would help them understand diseases now and in the future. If the human genome was mapped and related to diagnoses coded in ICD-9, it would be less specific. ICD-10 would give more power to the research investment on the human genome and efforts to look at that relationship. As five-to-20 years of ICD-10 data accumulated, it could be combed for greater specificity in the diagnoses, going back to tissue and blood samples collected, and learning more about revelations with the human genome and its relationship to diseases.
Mr. Augustine remarked that, as a biostatistician, more accurate data meant he could draw better inferences. But, being in the private sector, he was concerned--not only about cost and lead time--but also about HHS's, CMS's and contractors' abilities to meet these demands. Dr. Sondik noted a need to be more specific about what it would take to implement ICD-10. He said the coding wouldn't require extensive re-training because the systems were relatively alike, but computer work and edits were needed. He said NCHS and its sister agencies would help.
Overview of ICD-10-CM: Donna Pickett, NationalCenter for Health Statistics
Ms. Pickett noted a number of G-7 nations used ICD-10 for reimbursement and/or case mix systems. She said implementation wasn't an impossible task and they could draw upon the experience of other countries. She discussed an example of the expanded detail in both ICD-10 versions. ICD-9-CM had one three-digit category for diabetes with expansion at the fourth-and-fifth digit level for juvenile or adult onset. ICD-10 added five new codes conveying insulin and non-insulin dependent, malnutrition related, other specified, and unspecified. In ICD-10-CM revised diabetes categories reflected the American Diabetes Association's recent revisions to the classification. Codes were restructured to provide additional room for expansion. Expansion also occurred in the external cause of injury chapter. External cause of injury (E-800 through E-999) translated to V-00 through Y-98 in ICD-10 and more enhancements were added there in ICD-10-CM. The DSMO worked closely with medical specialty surgical societies, professional organizations, and other federal agencies in creating ICD-10-CM. Those who'd had problems and realized changes couldn't be accomplished within ICD-9-CM contributed to the changes. As a sixth character and laterality were added, some of ICD's more complex structures were modified to make it a true diagnosis classification, and relationships with procedures and procedure codes were removed. Trimester information was added to the OB codes. Injury, post-operative complications, alcohol and substance abuse codes were expanded, and code extensions were added for injuries and external causes of injury.
In ICD-9-CM, a number of categories were grouped under open or superficial wounds. Relating that information to outcomes research or any additional detailed analysis was practically impossible without going back to the medical record, something researchers working with abstracted data couldn't do. Expansions at the fourth-, fifth-, and sixth-digit character indicate expansion for laterality, body site, and breaking out specifics under Superficial. Ms. Pickett noted that expansions for post-operative complications were important for work going on in medical errors and patient safety.
In ICD-9-CM, some codes in the complications chapter didn't provide body system information or the exact type of complication unless paired with another code. Codes within ICD-10-CM's body system chapter specifically identified different types of complications. A combination of codes was also created for poisoning and external cause of injury.
Extensions were created for the injury and external cause chapter at the request of many who had struggled to use the injury codes to identify types of injuries and look at healthcare services utilization for repeat encounters for fractures or open wounds (e.g., ICD-9-CM had a code for healing fracture, but it was in the After-Care section--there was no site specificity or way to aggregate the data to determine true costs). Similar extensions were added to the fractures to denote an initial encounter for a closed fracture, open fracture, or encounter for routine healing.
Occurrence codes were expanded beyond five digits in ICD-10-CM and had details. WHO added a new concept: activity. (What the person engaged in when injured). A three-digit category with sub categories and fifth-digit codes for expansion was created for activity. An X indicated there was additional detail with that particular code. Adding activity as a fourth- and/or fifth-digit character eliminated some future expansion, but many indicated they needed detail at that level.
Ms. Pickett said public comments from the 1998 posting were incorporated. Tabular list revisions were finalized and there was a revised index and crosswalk. Reformatted guidelines were easier to use. And with the additional detail, further guidelines were being added to make use of the classification adequate, accurate and consistent over time. Work continued on an alpha version of the database and revised guidelines.
Ms. Pickett summarized the disadvantages of remaining with ICD-9-CM. The classification was old, outdated, with almost no room for modification. Data was being lost. Much happening in the healthcare system couldn't be captured. And the states were challenged in comparing ICD-9-CM collected data for morbidity with ICD-10 mortality data.
Ms. Pickett noted a number of ICD-10-CM's expansions were relevant for epidemiological research and decision support services. This data would also be used in other settings including grouping, case mix, reimbursement, patient safety, and surveillance and prevention activities. Parts of ICD-10-CM were harmonized with other classifications. Chapter Five was harmonized with DSM-4. Chapter two was already in use by the cancer registries by virtue of their use of ICDO-2. Some 90-95 percent of what was described by way of diagnosis in the nursing classification was also harmonized. The ICD-10-CM classification, tabular, and alphabetic index will be posted on the NCHS Web site: www.cdc.gov./nchs/icd9.htm.
Discussion
Dr. Zubeldia expressed concern that the “place of occurrence” in added extensions was unrelated to a diagnostic classification and that activity probably wasn't conveyed best this way. He asked if the intent was to create a coding system that satisfied all needs or just diagnostic's. Ms. Pickett clarified that place of occurrence was already in ICD-9-CM's External Cause of Injury Code. Users requested additional specificity so that they didn't have to reference the clinical record. “Activity” was also a concept included in ICD-10; ICD-10-CM expanded on what WHO already had in place and used internationally.
Noting ICD-10-CM wasn't currently on the Web site, Dr. Cohn asked whether the complete version had been reviewed publicly or if the Subcommittee was considering something unfinished. Ms. Pickett explained that the public had been invited to comment on a draft version in 1998. About half the comments had focused on the injury and external cause of injury chapters. Based on those comments, the classification was expanded. Collaboration with the American Academy of Orthopedic Surgeons and other groups provided additional detail. A near-final draft will soon be posted. The classification would be retested to ensure that the alphabetic index was complete enough that concepts could be found and documentation enabled health information management (HIM) professionals to use the codes based on these added details. Ms. Pickett noted many of the changes (e.g., diabetes and some cardiovascular changes) had been adopted or adapted by other countries and were already in use.
Mr. Augustine asked if the directions stated that the place of occurrence and activities codes (which didn't describe a disease state or condition of injury) were add-on codes that shouldn't be entered alone. Ms. Pickett said the external cause of injury codes were always used in combination with an injury or other code. The ICD-9-CM guidelines were written that way and would be applicable to ICD-10-CM. Additional instruction on the application of the activity code needed to be added to the ICD-10-CM guidelines.
Noting that an important bit of information for patient safety was whether an inpatient had a condition on admission or acquired it subsequently, Dr. Fitzmaurice asked if any part of the code in ICD-10 relayed that information or if it would be a separate variable. Ms. Pickett said that, based on discussions with National Uniform Billing Committee (NUBC), at this time that information was deemed separate from the classification. Ms. Greenberg noted this was in the next version of the 837 standard and would be in the HIPAA implementation guide.
Recalling that both Ms. Pickett and Dr.Sondik mentioned the 1994 study, Dr. Yasnoff asked if any methodologies were used or identified or any data developed that would support the current recommendation for conversion. Ms. Pickett said she would review the 1994 report and follow up with the Subcommittee. The staff pointed out that a code in ICD-10-CM for nosocomial infection could be used in combination with the disease entity or condition to make that link. Dr. Fitzmaurice remarked that there wasn't anything like that for a broken limb.
Mr. Blair said he imagined that, as they'd developed the latest version of ICD-10-CM for diagnostic codes, they'd considered that increased understanding of the human genome and DNA would result in an exponential increase in the need for certain sections of ICD-10-CM. Agreeing that that possibility loomed large, Ms. Pickett said unused codes in ICD-10-CM offered room for expansions.
Summary of April 15-16, 1997, NCVHS Hearings on Code Sets: Lizabeth Fisher, NCHS
Ms. Fisher summarized the pivotal April 1997 meeting of the Subcommittee on Health Data Needs, Standards and Security, noting the testimonies and transcript were available on the Web site. The meeting's purpose was to: (1) receive input from users of classification systems, code sets, vocabularies, and nomenclature; (2) discuss implementation of HIPAA standards for code sets; (3) and consider diagnoses and procedure code sets. Ms. Fisher noted that day's summary had concentrated solely on the diagnosis code sets.
Testifying panels represented different aspects of classification use including: professional healthcare providers, healthcare facilities, special data users, providers and health plans, public health and research, federal agency data users, and developers of coding and classification systems. Presenters provided input on: (1) systems in use and how the code sets were utilized; (2) a recommendation for an initial standard, given HIPAA's timeframe; (3) whether migration to ICD-10-CM was possible by 2000; (4) how to achieve simplification and retain medically meaningful information; (5) how to maintain the code sets; and (6) resource implications of changing to a new classification/code set. Other considerations included: (1) whether one classification system could serve most users, (2) whether the same disease classification could be used for administrative transactions and statistical reporting, (3) whether current systems represented everyone's discipline, (4) problems encountered linking coded data, (5) implications of these code sets and systems with the computerized patient records (6) the market acceptance for recommended systems, (7) how NCVHS could assist the U.S. in developing international systems, (8) whether single procedure classifications should be considered, and (9) requiring universal acceptance of code sets.
The primary system for diagnosis reporting was ICD-9-CM for a variety of uses including: disease indexing, statistical reporting (public health, state/federal), provider billing/reimbursement, disease staging, performance indicator measurement, longitudinal research, and HEDIS measures. Other systems used included DSM (mental health diagnosis), SNOMED, LOINC, READ, National Drug Codes, and the Nursing Diagnosis Classifications (NANDA).
ICD's strengths included: (1) it was well-known and used by hospitals, physicians, other providers, payers, and data system users; (2) training/education methods were in place; (3) it had international recognition, comparability, and could be mapped back to the WHO version; (4) data systems were set up to accept it; (5) it was an open system in the public domain; and (6) the comfort level was high. Weaknesses also were noted. Some commented that ICD wasn't specific enough for all data users needs, couldn't classify severity of illness, and didn't provide enough clinical detail-- preventive medicine, home health, and nursing users especially needed more detail. Participants also commented on the lengthy timeframe to update/new codes. ICD-9-CM was running out of space for new codes. And payers' use of ICD wasn't consistent.
Most testifiers recommended continuing to use ICD-9-CM for diagnosis reporting with a plan to migrate to ICD-10-CM. Some recommended using a combination of ICD-(n)-CM with other systems, such as DSM. Participants suggested using ICD with crosswalks/links to other systems such as SNOMED, LOINC, and READ. Many dealt with Y-2K issues in 1997; their systems were enveloped and they felt that implementing ICD-10-CM by 2000 was too ambitious. ICD-10-CM wasn't yet posted on the Web site and some who hadn't seen the entire classification needed to evaluate it. Whatever standard was implemented, many called for thorough field testing and cost/benefit analysis. Citing the lack of space for expansion and comparability internationally, participants cautioned not to delay implementation of ICD-10-CM.
Asked how to achieve administrative simplification and still attain medically meaningful information, testifiers suggested included having a single classification or combining classification systems to meet all users' needs. Many called for a single set of guidelines adhered to by all. Other recommendations included thorough testing and cost/benefit analysis before implementation and using modifiers to enhance data collection. A theme throughout the testimony was (1) one set of guidelines for all healthcare encounters (instead of inpatient versus outpatient), (2) uniform adherence to coding rules by those assigning the codes as well as users of data, (3) requiring all payers to abide by one set of rules published by an official source, and (4) all payers must use and accept all codes in the code set and its guidelines without modification.
Most said having a code set in the public domain, without copyright restrictions, and using the coordination maintenance committee process with annual/regular updates was the best way to maintain the code sets. They recommended allowing physicians, professional groups and other users of the classification to volunteer input to the updated classification to ensure that maintainers of the code set were responsive to the needs of all users. They also noted the maintenance process should be readily available to the public and government funded.
Asked about resource implications of implementing new code sets, participants noted: (1) the need to train all users (coders, physicians and data users), (2) the need for comprehensive system programming and conversions, (3) a proficiency level and learning curve before people became fully comfortable with the changes, (4) a need for crosswalks and reporting to compare data across systems, (5) resources to remap rules, tables and existing systems, and (6), to always weigh resource implications against the benefits of change.
Ms. Fisher summarized the common themes brought out during two days of testimony. There were many different classification/data needs by different users. For reporting diagnoses, the theme was to continue to use ICD-9-CM and move to ICD-10-CM following development and testing. Many requested public domain access to code sets (no copyrights); open input and access to the maintenance process; uniform coding guidelines, reporting requirements, and code acceptance by all users; cost/benefit analysis done prior to moving to ICD-10-CM; and possibly linking ICD to other systems. Other recommendations and feedback included: any change must be compatible to all payment systems (e.g., DRGs, APCs); one system might not serve all data users' purposes; consider all users, not only hospitals, in selecting and implementing the code sets; use existing/proven coding systems; and consider the impact of these systems in migrating to the computerized patient record.
Discussion
Ms. Fisher reiterated that the discussion about development and testing stemmed from the classification not yet being posted on the Web site for open comment. Testifiers were still unfamiliar with the clinical modification changes and wanted to make sure that there was open comment as well as testing of the classification before migrating to ICD-10-CM. She noted there hadn't been any discussion about methodology. Dr. Yasnoff reflected that this sounded like what the Subcommittee heard that day. Ms. Humphreys observed that many people had had an opportunity to review and attempt to apply ICD-10-CM since 1997.
Panel 1: Possible replacement of ICD-9-CM, Vol. 1 & 2 (Diagnosis) with ICD-10-CM
Dr. Powers recalled that she was in practice when neurologists started using ICD-9-CM codes for office billing and in the early 1990's she began working from the Academy with NCHS to improve neurologic diagnoses in ICD-9-CM. She noted that 13,652 of the Academy's 18,040 members were board-certified neurologists and the diseases they treated, including stroke, multiple sclerosis, epilepsy, and Alzheimer's disease, had a significant effect on U.S. healthcare dollar expenditures. Finalizing the identification of the human genome had been a boon to the field, leading to the explosion in neurologic treatment over the last 15 years. With this expanded knowledge, neurologists' diagnostic classifications changed and ICD-9-CM no longer kept up. The number of fifth-digit codes available was almost nil, and codes and diagnoses were scattered in sundry places in order to include now commonly used diagnoses. Sleep disorders were in the symptom chapter (in ICD-10-CM, they were restored in the neurology chapter). Tension headache (seen by neurologists as an extension of vascular headaches and the whole headache spectrum) was coded as 307.81 with “psychalgia,” resulting in denial of access to neurologists. In ICD-10, tension headache was back with the other headache diagnoses in the neurology chapter.
Older classifications were a huge problem for neurologists. Many diseases previously diagnosed otherwise were now known to be of mitochondrial origin, but mitochondrial diseases were essentially non-existent in ICD-9-CM. ICD-9-CM listed essential tremor as “essential and other specified forms of tremor,” creating another access issue because many carriers considered that code too non-specific to use. Dr. Powers noted about ten percent of the population had essential tremor. Transverse myelitis was given a fairly non-specific code (323.9) that implied an infectious etiology, whereas most transverse myelitis patients neurologists saw were non-infectious. Transverse myelitis was appropriately placed with demyelinating diseases in ICD-10. Dr. Powers pointed out that the language involved in the epilepsy codes in ICD-10-CM was quite similar to the language neurologists and other physicians used in describing these diagnoses.
Noting the Academy's own extraction of the neurologic codes from ICD-9-CM used a crosswalk performed with stroke codes, Dr. Powers said the Academy had looked at several stroke classifications and decided that WHO and ICD-10-NA's neurologic adaptation for research coding for stroke were most appropriate. She said the Academy was pleased to proceed to ICD-10-CM for stroke diagnoses.
While remarking that physicians were taught to code themselves because only a physician knew some of the details (whether a headache or epilepsy was intractable or if a hemisphere stroke was dominant), Dr. Powers pointed out the many nuances to knowing how to use each code and the difficulty for anyone not trained to code in the office. She noted ICD-10 was far more similar to the language physicians used when describing diagnoses everyday and would eliminate some of these problems. Dr. Powers cautioned that difficulty choosing a code resulted in less than accurate coding and unreliable data.
Dr. Powers said she understood from the morning's discussion that HIPAA called for implementation within a year after it was decided that ICD-10-CM would be placed. She noted neurologists had done that with CPT-4 and that ICD-10-CM had the same concept and structure and was familiar because of the language. She emphasized that the Academy supported proceeding with implementation of ICD-10-CM as a national classification system binding to all third-party payers, with rules binding to all payers. She requested and said she understood there would be an adequate transition period in which to train members and challenge arbitrary computer “edits,” which she said would inevitably occur early in the implementation.
Panel 1: Possible replacement of ICD-9-CM, Vol. 1 & 2 (Diagnosis) with ICD-10-CM
Dr. Mirin noted that psychiatrists, physicians and all mental health professionals used DSM-4 to describe patient diagnoses with sufficient specificity: clear communication was possible for clinical management and record keeping. DSM-4 was also the official mental disorder code set for Medicaid in all states, the federal employee's health benefit program, the Food and Drug Administration, the National Institutes of Health and most private insurance companies. Failure to designate DSM-4 code descriptors under HIPAA would mean all of these practitioners, agencies and institutions had to abandon the diagnostic criteria in favor of ICD-9-CM and a reflection of medical science stunted in the late 1960's.
Dr. Mirin recommended having DSM-4 designated as the descriptor code set and delegating responsibility to its developer, APA, for convening scientific experts needed to periodically update these descriptors. Dr. Mirin said the overall structure and lack of operationalized definitions, particularly in the mental disorders sections, rendered ICD-9-CM inadequate for use by mental health professionals. He noted the glossary of definitions in ICD-9-CM had been developed for ICD-8, which was parallel to DSM-2--the operational code set in the late 1960s when he was a resident in psychiatry.
In 1980, concerns that specificity and subtyping provided ICD-9 weren't reflective of contemporary scientific thought led APA to proceed with DSM-3, departing from the ICD-9 format. DSM became the de facto system of classification of mental disorders in the US for all mental health professionals, healthcare facilities, federal and state government agencies. Dr. Mirin said DSM was overwhelming accepted because of its reliable, clear definitions of mental disorders; its reliance on descriptions of observable signs and symptoms; and its use of a documented evidence base derived from a comprehensive review of empirical literature and input from a broad array of professionals and healthcare organizations. Today, DSM-4, its operational criteria, and definitions were coupled with ICD-9-CM code numbers, enabling the insurance industry, government agencies and institutions to apply these criteria in determining diagnoses, need for treatment, and appropriateness of treatments provided.
Dr. Mirin cautioned that continued use of ICD-9-CM posed serious problems. ICD-9-CM's archaic terminology carried negative connotations for patients: e.g., “psychosis” was used in its 1960's sense as an indicator of severity, rather than as a designation indicating specific signs and symptoms, and could adversely effect patients insurability. And he noted that in psychiatry a number of DSM disorders shared the same ICD-9-CM code. Some 29 substance abuse disorders were designated under code 292.89, creating a profound problem for research on the incidence and prevalence of those disorders.
A number of categories in ICD-9-CM originally constructed in the early 1970's had no empirical foundation and weren't recognized as mental disorders in the United States (e.g., “misery,” “unhappiness disorder,” and “academic under achievement disorder”). Dr. Mirin said use of archaic subtypes and wording forced medical coders to make diagnostic distinctions that had no basis in current science and lacked any precise definition. Mental health professionals and institutions used the codes provided in DSM-4 as surrogates for the ICD-9-CM categories for diagnosis, record keeping, and billing. But coders in acute care hospitals continued to use ICD-9CM to determine the most appropriate diagnostic code to add to the medical record and billing statements. Dr. Mirin noted this was inconsistent with the HIPAA mandate to adopt single uniform coding systems for appropriate specialties and that adoption of ICD-10-CM wouldn't address the lack of diagnostic criteria. Only DSM-4 fulfilled this need.
Dr. Mirin expressed APA's support for implementing ICD-10-CM, which provided greater freedom for adding new categories and subtypes and reflected current usage of medical terminology. He noted that NIMH, with APA's close cooperation, provided over a decade of research and support to WHO as it developed ICD-10-CM. He said it had been expected that ICD-10-CM would be adopted for use for morbidity, as well as mortality reporting, but ICD-9-CM had remained the standard for Medicare because of HCFA's (now CMS) concern over the cost to re-program computers and accommodate conversion to an alphanumeric system. Most the developed world had approved ICD-10 for both mortality and morbidity reporting; but Dr. Mirin said the U.S. lagged behind and held back countries that looked to the U.S. for leadership in healthcare.
Dr. Mirin summarized APA's perspective. For over 50 years, APA provided leadership in the development of scientifically based diagnostic criteria for mental and addictive disorders for DSM and ICD similar to what NCVHS provided other sections of ICD. The ICD-9-CM system was archaic for most disorders and even more dated for mental disorders, with terminology and glossary definitions created in the 1960's. ICD-10 could reflect DSM-4 compatible names for mental disorders and allow adoption of an alphanumeric code set that provided space for identifying more scientifically reliable, valid disorder definitions--facilitating documentation of prevalence and treatment rates of mental disorders and the public health burden caused by emerging new disorders and drugs of abuse. However, ICD-10-CM had no operationalized definitions or diagnostic criteria and only an introductory statement referring the reader to DSM-4 for definitions of psychiatric terms and conditions. Dr. Mirin questioned what effect, if any, this informal reference would have with respect to compliance with HIPAA regulations and said continued use of two separate code sets for mental disorders seemed inconsistent with HIPAA's goal of administrative simplification of diagnostic reporting. Noting continued reliance on the current informal dependence of ICD-9-CM and ICD-10-CM on DSM-4 would no longer be legally permissible, Dr. Mirin proposed that APA could fulfill the HIPAA requirement to maintain the code sets pursuant to a formal regulatory review and updating process. To avoid needless destruction of mental health service delivery, APA asked NCVHS to approve: adoption of DSM-4 criteria as the official descriptors for ICD-9-CM; designation of APA as DSMO for these criteria; and approval and implementation of ICD-10-CM for HIPAA reported requirements at the earliest possible date.
Panel 1: Possible replacement of ICD-9-CM, Vol. 1 & 2 (Diagnosis) with ICD-10-CM
Dr. Strunk said Medical knowledge and the range of treatments offered in obstetric care had expanded greatly since ICD-9-CM was introduced almost three decades ago. Developments in sonography revolutionized examination of the fetus. The ability to make genetic diagnoses had changed dramatically and would change exponentially over the next 10 years. Understanding and clinical use of maternal and fetal physiology and management of labor also improved significantly. Knowledge and classification of important maternal conditions such as diabetes mellitus and hypertensive diseases of pregnancy also changed. Dr. Strunk said ICD-9-CM was woefully inadequate to capture these changes. NCHS staff and the Coordination and Maintenance Committee were responsive to ACOG's recommendations for changing ICD-9-CM's obstetric chapter to meet current needs. However, ICD-9-CM's ability to change and expand was stretched to its limit and further expansion would be difficult. ICD-10-CM offered greatly enhanced capabilities for identifying and tracking important obstetric conditions and interventions. These data were needed for high-priority public health efforts to reduce adverse pregnancy outcomes.
Dr. Strunk highlighted four specific aspects of ICD-10-CM that facilitated reporting issues of growing importance in obstetric care. He said the addition of a fifth digit to report the trimester of pregnancy should prove valuable in efforts to monitor the provision of prenatal care and the occurrence of complications. ICD-10-CM also responded to the growing prevalence of multiple pregnancies, permitting identification of individual fetuses in a multiple gestation. For the first time, a specific condition could be linked to a specific fetus. In addition, ICD-10-CM included expanded codes for reporting abnormal findings on antenatal screening. The range of screening tests recommended or offered to pregnant women continued to expand, generating additional episodes of care that required appropriate diagnosis codes. None were available. Finally, ICD-10-CM would include a category for care provided to the fetus, due to abnormalities.
Dr. Strunk noted several challenges and potential disadvantages to adopting ICD-10-CM. Physicians, hospitals, coders and payers had to learn a new system. And the six-digit ICD-10-CM codes dictated computer systems and software would need reconfiguring. Noting the cost associated with training and computer changes wouldn't be trivial, Dr. Strunk said that in an environment of declining reimbursement and increasing regulatory demands, physicians, hospitals, and healthcare providers couldn't be expected to bear this entire burden alone. ACOG believed that the benefits offered by ICD-10-CM were substantial enough to justify the costs and disruption adoption would inevitably entail.
Panel 1: Possible replacement of ICD-9-CM, Vol. 1 & 2 (Diagnosis) with ICD-10-CM
Dr. Warner, an injury epidemiologist, reiterated the other panelists' views that the classification framework had to keep up with a subject's knowledge area. She discussed motor-vehicle-related injury and death as an example of how injury prevention and control had changed, noting parallels in burn injury research. Over two decades, they'd gained air bags, car seats, and the third rear brake light. Car seats, in particular, were shaped by data collected using ICD-9 for fatalities that indicated infants were over-represented and led to the realization that they were flung through the windshield. Now data indicated, along with the decrease in fatalities, an increasing number of lower limb fractures--and the subsequent need for more classifications for fractures, particularly in the lower limbs. Dr. Warner noted that chapter 19 of ICD-10-CM, “Injury, poisoning and certain other consequences of external causes,” and chapter 20 “External causes of morbidity and mortality data,” resulted from extensive consultation and review with emergency room physicians, epidemiologists, statisticians and other public and private sector experts in the field of injury prevention and control.
Dr. Warner said adoption of ICD-10-CM would establish compatibility with US mortality data which used ICD-10 to code fatalities, provide greater detail due to the extension codes, and enable more future expansions that would allow for better data describing external causes and the nature of injury. The classification system also would provide details that made it more appropriate for a wider range of users (e.g., potentially workers' compensation). Dr. Warner emphasized the importance of having comparable codes for datasets to facilitate data analysis. She noted that incompatibility between the structures of ICD-9-CM and ICD-10, made analysis of data more difficult, time consuming and results less precise.
The U.S. had coded mortality data using ICD-10 since 1999. While ICD-10 and ICD-10-CM were compatible, Dr. Warner pointed out that, on many levels, there were incompatibilities. ICD-9-CM was organized by type of injury; ICD-10 was organized by body region. ICD-9-CM used external cause codes E-800 to E-999 and nature of injury codes 800 to 999; ICD-10 used external cause codes B, W, X, Y and nature of injury codes S, M, T. Probably most important, there wasn't a one-to-one relationship between the codes in ICD-9-CM and ICD-10. Dr. Warner emphasized it was important for the U.S. to compare injury and poisoning morbidity data between countries in order to better understand causes and circumstances of injuries. Because many countries (including the European community and Australia) already used a version of ICD-10, U.S. data was outdated and incomparable. Dr. Warner cautioned it would become very difficult to compare between countries where it already was cumbersome using the same classification system but different applications of the rules.
Many injury codes in ICD-10-CM had far greater specificity than ICD-9-CM's codes. Dr. Warner said capturing data in injury and poisonings and toxic effects would be greatly enhanced using ICD-10-CM. ICD-10-CM included the intent of the poisoning or toxic effects as a final character in the diagnosis codes. External cause codes in ICD-9-CM weren't specific to the agent involved. ICD-10-CM allowed room for expansion at the fifth character. Dr. Warner noted it was essential to include room for expansion of the poisoning code because the number of biological and medicinal substances rapidly increased. She said classification systems needed modification in order to accommodate emerging public health needs. ICD-10-CM was designed to accommodate expansion, thus eliminating the situation ICD-9-CM currently faced.
Noting Ms. Pickett had pointed out that injury codes in ICD-10-CM enabled users to differentiate between initial and subsequent visits for the same injury as well as the sequelae of injury, Dr. Warner said this would permit more detailed analysis of people injured than the current situation that only allowed analysis of healthcare utilization patterns. (ICD-9-CM coding guidelines required that a fracture be coded only once at its initial episode). Any further care was coded as “after care” and information on the type of fracture was lost during the course of healing. Neither the guidelines nor codes in ICD-9-CM could be changed to correct this. With the extensions that apply to all injury codes, the episode of care was identified and the fracture code used continuously throughout the treatment. Data on fractures as well as other injuries would be far superior with ICD-10-CM, because of the ability to differentiate between initial and subsequent visits for injury.
Dr. Warner said the level of detail available to code injuries was greatly improved in ICD-10-CM. This was true for the nature of injury codes (e.g., fracture and traumatic spinal cord injury) and for external causes (e.g., the transportation related injuries). Dr. Warner said this kept up with the current knowledge engine in injury prevention and control. She noted one more example of the level of detail provided was the addition of laterality to all ICD-10-CM codes, enabling identification of the exact site of an injury. Another example was the distinction between degrees of open fractures provided in ICD-10-CM.
Panel 1: Possible replacement of ICD-9-CM, Vol. 1 & 2 (Diagnosis) with ICD-10-CM
NACHRI worked with children's hospitals around the country, developing both in-patient and population-based ICD-9 classification systems, running comparative reporting programs, and doing various policy analyses. Mr. Muldoon said NACHRI found limitations with the specificity of the information in all these activities and he reiterated NACHRI's 1997 recommendation to move forward with ICD-10-CM.
Mr. Muldoon said the many generic codes were extremely frustrating and one lost sight of the nature and status of an injury. Noting the prevalence of mental health and neurologic conditions when children interact with the healthcare system, Mr. Muldoon said he'd never met anyone in the children's hospital world or mental health field who worked with ICD-9-CM diagnosis code. Cancer was less common in pediatrics, but the title alone indicated nothing about the particular cancer (e.g., whether neuroblastoma was in the adrenal gland or thorax), and without histological information, ICD-9-CM provided little. Mr. Muldoon said areas of specific interest for pediatric, the congenital anomaly and perinatal codes, were particularly underdeveloped in terms of specificity. In a number of instances, the congenital anomaly codes didn't distinguish between atresia, stenosis, insufficiency, hypoplasia and other malformations (particularly heart) and some renal conditions. Distinctions weren't made between bilateral and unilateral.
Mr. Muldoon noted that often very different conditions were in the same code. This was most common with the “not elsewhere” classified codes: e.g., about 150 different terms were indexed under 759-89, “Other congenital anomalies affecting multiple systems.” He acknowledged that most were quite rare, and didn't need to be separate codes. But he said ICD-10-CM struck a reasonable balance by breaking them out into about half-a-dozen syndrome types predominantly affected in certain body systems.
Cranial-facial anomalies were clustered in 756-0 and had to be broken out. Infections were very important in the newborn period, but there was little specificity amongst the pneumonia diagnoses or sepsis. Newborn sepsis had just been broken out from perinatal infection, NEC, but not the specific organism. Mr. Muldoon noted similar shortfalls with certain hemorrhage, pulmonary, and noxious substances codes. Noting that muscle tone disorders were “thrown together” with cardiac and renal failures under 7709.8, “Other conditions not elsewhere classified,” Mr. Muldoon reiterated NACHRI's strong support for the movement from ICD-9-CM to ICD-10-CM.
Discussion
Dr. Mirin clarified that with ICD-9, ICD-10, or DSM-4, APA advocated descriptors from DSM-4. The mental disorders section of ICD-10-CM stated, "see DSM-4." Leaving the ICD-10, including guidelines, under HIPAA would satisfy APA's need for the DSM-4 descriptors. Dr. Cohn asked if a letter from HHS acknowledging the validity of the crosswalk between ICD-9 and DSM-4 would be helpful. Dr. Mirin said it would be enormously beneficial to have that crosswalk acknowledged.
Noting they'd heard and read testimony about the lack of definitions in ICD-9, Dr. Cohn asked how important definitions were in a diagnostic coding system. Dr. Mirin said that in an emerging field with such an enormous difference between earlier iterations and current and anticipated DSMs, an ability to be as specific as possible about diagnostic criteria was crucial, particularly from a public health standpoint, in understanding the prevalence of disorders. APA's goal in developing the DSM and subsequent iterations was a single worldwide classification of mental illness. Dr. Mirin confirmed that he believed this would result in common definitions as well as common titles. Dr. Cohn noted Mel Ingobar from CMS also commented on the lack of clear definitions in the ICD about chronic renal failure. Ms. Greenberg clarified that the purpose of the disease classification system was to assign a code based on the narrative in the medical record, not to diagnose patients. APA had incorporated it into a classification; other specialties relied on a whole literature to guide in assigning narrative. Classification wasn't considered a diagnostic tool.
Ms. Greenberg remarked that her entire staff had developed 10-CM and when people said they appreciated collaborations within NCHS, it was really this team piloted by Amy Bloom, David Berglund, and Beth Fisher, under Donna Pickett's leadership, with whom they'd collaborated. Ms. Bloom explained that the team specifically chose not to include definitions in the statistical classification and official coding guidelines required the coder to assign a code based on the physician's documentation. Coders were to thoroughly review medical records, but not to consider whether a patient's lab or test results confirmed a diagnosis. At APA's request, the glossary of mental disorders had been removed from ICD-10-CM and the reference (which hadn't been updated in 23 years) added so people knew the definitions could be found in the DSM. Ms. Bloom noted DSM did have some definitions (e.g., acute MI was defined as a four-week time period; coders were to continue to code an MI for four weeks after it was diagnosed). There were also specific definitions for trimester requested by ACOG and for premature ruptured membrane. But otherwise, ICD-10-CM utilized instructions on how to properly assign a diagnosis code (e.g., sequencing of paired codes, combination codes), rather than definitions.
Dr. Cohn said that beyond hearing how antiquated ICD-9 had become, partly due to the lack of room, he'd heard the worldview had changed (e.g., headaches no longer were a psychiatric illness). He noted this was evident in diabetes and genomics, and that he'd heard everyone might suddenly be thinking differently about neurology and OB-GYN. Dr. Powers concurred. She commented on how Washington University's Web site for neuromuscular disease (which initially used traditional classifications of neuromuscular disease) completely changed over the last several years with more genetic classifications and direct links to OMIM as more was learned about the genetic classifications of neurologic disease.
Dr. Cohn acknowledged that ICD-10 had more space for additions. But he noted that when the worldview changed around diabetes, it hadn't been a question of adding extra numbers-- things had to be re-ordered. Whatever changed ICD-10 would be expensive. Dr. Cohn said he doubted that anyone would be happy hearing in eight years about moving to ICD-11. Ms. Greenberg pointed out that an important distinction between ICD-9 and ICD-10 was that ICD-9 never changed. When ICD-10 was implemented and approved by the World Health Assembly, there was agreement on an updating process so that changes could be made to try to keep up with medical science (e.g., ICD-9 never had codes for HIV-AIDS). Codes were developed, but there was no mechanism for changing ICD-9. ICD-9-CM began an annual update in 1985, but was always based on core ICD-9 at the three-digit level, first published in 1979. Ms. Greenberg said it was possible that whole sections of ICD-10, perhaps even a whole chapter, could be updated without going to an entirely new classification. International committees were set up for modifying ICD-10 and a number of modifications already had been made.
Dr. Kolodner noted intent to build in the evergreening process and suggested that, hopefully, ICD-10 could be extended another 23 years. He asked the testifiers to elaborate on a common theme heard from all the end users, particularly those who coded in ambulatory care settings where the vast majority of codes were assigned: there was a distortion trying to fit current clinical understanding into the old codes, and changing to a new system more consistent with clinical care and concepts wouldn't be so difficult for end users because it aligned with how they practiced. Dr. Mirin explained that the clinicians essentially assigned the codes in ambulatory settings and private psychiatric hospitals. Acute care general hospitals had a somewhat different situation, but he cautioned there would be utter confusion for everyone who delivered mental healthcare and was accustomed to entering diagnostic code from DSM-4 into the notes that enabled everyone to hopefully arrive at the DSM-4 diagnosis that fit the clinical record. Once HIPAA was implemented, Dr. Mirin said people would be searching through the ICD-9-CM codes to try to find something that fit their understanding of the patient. He cautioned that, in some instances it would be extraordinarily difficult, particularly for young people trying to adopt an evidence-based approach to diagnosis and treatment, to be thrown back into an obsolete world with a wholly different theoretical view of many illnesses, completely devoid of the genetics determinants of illness, biochemistry and psychopharmacology. Dr. Powers agreed. When physicians were able to use their own language it was less stressful. Dr. Strunk said the Committee on Coding and Nomenclature and its staff had been fortunate to have very good collaboration and communication with NCHS staff and, while nothing was absolutely seamless, it would be easier to talk about an instance (e.g., infections of the genital-urinary tract in pregnancy), not as a single category, but as one involving, in this case, the kidney, bladder, or urethra. Similarly, talking about abnormal findings on an ante-natal screen of a mother (which currently didn't include any breakdown in ICD-10-CM) would have hematologic, biochemical, ultrasonographic, and chromosomal genetic findings. Dr. Strunk said he didn't anticipate a great deal of difficulty in making this transition, but he saw tremendous benefit in the added specificity.
Dr. Warner commented that, as people came to understand that you could prevent and heal injuries, they realized the importance of classifying types and levels of fractures and the science of injury prevention and control grew. There were sonograms and other ways to look at and diagnosis traumatic brain injuries. Mr. Muldoon said he spent a lot of time in classification system development work translating what ICD-9 codes really were to the physician members of the team. He said he anticipated needing to spend far less time explaining ICD-10-CM codes.
Noting Dr. Powers mentioned the payers ought to follow uniform work rules, Dr. Zubeldia asked if the impact of changing to ICD-10 could be a catalyst for all the payers to follow the same rules. Dr. Powers said she wasn't sure who legislated how third-party payers behaved. She suggested this was an opportunity to go to a single rule system. She said she didn't know of a way to force insurance companies to do that, but she advised that it would help the entire medical community if everyone followed the same rules. Ms. Humphreys pointed out that the published rule for ICD-9-CM diagnosis gave the coding rules and guidelines and required everyone to use them. She said personally she favored moving to 10-CM, but the requirement already was in place once the transaction rule went into effect.
Dr. Powers clarified that her issue was not that the insurers didn't accept the claim, but that they didn't pay on it or would disallow use of this code. Ms. Humphreys said that with implementation of the transaction rule for ICD-9-CM, an official national coding set was part of the standard and everyone was explicitly required to accept all valid codes. Dr. Zubeldia noted the fine distinction between rejecting a claim and accepting but not paying it. HIPAA required them to accept the claim, but didn't require payment. Ms. Humphreys countered that, based on the testimony, there were cases where a claim with four or five codes on it was sent to a payer who was supposed to pay for a couple procedures and the whole claim came back with instructions to delete a specific standard code that payer didn't want in its system. She said requiring everyone to recognize rather than reject a claim based upon the presence of a standard code would smooth out the payment process, even if the payer didn't intend to pay for that code. Ms. Humphreys noted this also was a major issue with implementation dates. Updates and new editions came out and some payers lagged behind others in accepting valid additions; people coded diseases in their local environments according to a currently valid set of codes and found that for months claims were rejected because those payers hadn't updated. Ms. Humphreys remarked that there was no need to wait for ICD-10-CM to move on this issue, but she said 10-CM was desirable if they wanted data that reflected what people actually knew about diseases. Dr. Powers noted that this also happened when coding for certain procedures or laboratory tests; some insurance companies required rule-out codes that weren't used officially.
Asked about the pricing structure for DSM codes, Dr. Mirin made a distinction between the codes themselves (which were the ICD-9-CM codes) and descriptors. The codes were free and available to anyone. APA developed a set of descriptors in the form of a textbook (DSM-4-R) for clinicians. The definitions that went along with the codes were part of the copyrighted textbook. Dr. Mirin said APA felt an obligation to continue to update the classification of disease that was the hallmark of the field and every year incurred the significant cost of developing DSM's next iteration. APA derived some revenue from selling this text (the majority of purchasers weren't psychiatrists, but college students learning the classification of mental disorders). The costs of maintaining DSM came from the textbook revenue and APA members' dues.
Asked how much lead time the healthcare industry needed to move to ICD-10 diagnosis, Dr. Mirin noted DSM-4 was totally compatible with ICD-10-CM. Dr. Powers concurred that it wasn't a new concept. It was a matter of learning different numbers and categories and the language made everything easier. She said she didn't know how long it would take the healthcare industry to change their computers, but that physician could learn fairly quickly. Two years would be enough time to educate members. Dr. Strunk said ACOG members would be able to do it within that time frame. ACOG's greater concern was what would happen with software and computers in other elements of the industry. Mr. Muldoon agreed that conceptually it wasn't difficult. He noted that in the afternoon testifiers would address software and related implementation issues.
Panel 1: Possible replacement of ICD-9-CM, Vol. 1 & 2 (Diagnosis) with ICD-10-CM
Ms. Prophet said AHIMA believed that replacement of the 9-CM diagnosis codes was absolutely necessary and 10-CM should be adopted. Many countries already had or were in the process of implementing ICD-10, and continued use of 9-CM caused significant difficulty in comparing international data used in public health, research and development, and the study of issues related to quality, safety and efficacy of medical care.
As they'd heard, the ability to compare mortality and morbidity data within the U.S. was compromised. When 9-CM was implemented, ICD had been primarily used for indexing purposes in the hospital in-patient settings. The need for greater coding accuracy and specificity had heightened considerably since then. Ms. Prophet said this system was woefully inadequate for meeting the needs of non-inpatient hospital settings or serving the many other purposes for which ICD data were currently used.
The 9-CM codes weren't used for reimbursement purposes until well after the system's development and implementation. The consequences of inaccurate claims data in a fee-for-service environment hadn't been nearly as critical. Once prospective payment systems (PPS) came into existence, concerns about data quality, coding education, and medical record documentation heightened. Today, several PPS for different healthcare settings depended heavily on 9-CM diagnosis codes, and many non-prospective payment methodologies required complete, accurate, detailed coding in order to negotiate or calculate appropriate reimbursement rates, determine coverage, and establish medical necessity. Providers, payers, researchers and other users daily faced problems caused by continued use of an obsolete diagnosis coding systems.
Ms. Prophet noted AHIMA had been involved in implementation of ICD-10 since 1987, when it established an internal task force whose conclusions raised many points mentioned that day. As they'd heard, 10-CM's greater specificity, clinical detail, and alphanumeric structure represented a significant improvement over 9-CM and ICD-10 and met the criteria for adoption and maintenance set forth by HIPAA. 10-CM provided much better information for non-hospital, in-patient encounters, clinical decision making, and outcomes research. AHIMA didn't believe reimbursement considerations should drive code set revisions; however Ms. Prophet said good, specific coded data should be used and could determine and support appropriate reimbursement. The payment computation system, not the coding system, should define payment allowing for cost, service, or location data; cost analysis; and research. The level of specificity in 10-CM would provide payers, policy makers and providers with more detailed information for establishing appropriate reimbursement rates, evaluating and improving the quality of patient care, improving efficiencies in healthcare delivery, reducing healthcare costs and effectively monitoring resource and service utilization.
AHIMA believed final regulations promulgating adoption of 10-CM as the replacement for the ICD-9-CM diagnosis coding system should be issued no later than 2003, with an implementation date no later than 2005. Ms. Prophet said a decision to implement 10-CM couldn't be delayed until all current HIPAA rules were fully implemented, nor could the healthcare industry afford to hold off reimbursement and other considerations impacted with any coding system delay. She said the definition of the ICD-10-CM coding standard should include system rules and guidelines in a process allowing public input to codes and that maintenance should be continued with the implementation of 10-CM. An advisory group comprised of stakeholders should be established to provide input into the maintenance and guideline development processes.
In February, AHIMA indicated that maintenance was a key issue for any complete, flexible medical coding system and espoused six principles for medical code set maintenance processes. Existing official coding guidelines needed to be revised for use with 10-CM. Current development, maintenance and approval processes for developing the 9-CM official coding guidelines must continue for 10-CM and these guidelines, rules and conventions must be included in the 10-CM code set standard, insuring proper and consistent applications by all users.
A standard effective date for 10-CM implementation should be established and all affected payers and providers should be required to abide. A formal implementation plan encompassing all effected entities should be established well in advance of the scheduled implementation date. Ms. Prophet emphasized it was imperative that the NCVHS and HHS Secretary, with industry consultation, initiate the plan immediately.
AHIMA surveyed several countries that had already implemented ICD-10 or a modification for morbidity purposes. A number of countries expressed consternation about the U.S.'s failure to adopt 10-CM, noting the detrimental impact on international healthcare data comparability. Australia and Canada had developed a modification of ICD-10, and both countries used ICD for reimbursement purposes. A major impetus to adopt ICD-10 was the need for an up-to-date classification for a more clinically credible case mix system and improved applicability in non-hospital settings. Canada was the first country to move to an entirely electronic product for ICD-10 codes. Lack of vendor readiness, despite ongoing communication from the start of implementation planning, was a significant challenge for both countries.
Ms. Prophet emphasized it was extremely important to determine the strategic progression of 10-CM implementation as soon as possible. Development of 10-CM educational materials should begin as soon as the coding system was finalized and publicly available. Ms. Prophet cautioned everyone about neglecting the magnitude of the training issue. ICD-10-CM retained the traditional ICD format and conventions and wouldn't be difficult for users trained in 9-CM diagnosis coding to learn. But implementation of any new coding system required training for coders, clinicians responsible for documentation, and the growing number of data users. The range of users and settings requiring training programs was far wider for 10-CM and 10-PCS. Work force shortages and variable skill levels among those responsible for coding in diverse healthcare settings meant re-training would be more demanding for many on-the-job trained coders. Experienced coding professionals would primarily need to be educated on changes in structure, disease classification, definitions and guidelines. The greater clinical specificity in 10-CM also required coding professionals to have a better understanding of medical terminology and human anatomy. And clinicians had to change their way of documentation so that more specific, appropriate codes were selected.
Two-day training workshops for experienced coding professionals were conducted in Australia. In Canada, training was provided utilized coding software in three phases: a 21-hour self-learning package, a two-day practical application in a computer laboratory, and a self-learning package of ten case studies. Both countries offered post-implementation refresher workshops for coding professionals. Special educational sessions were provided to data users.
Australia and Canada reported that coding professionals found ICD-10 no more or less difficult to learn than ICD-9. The average learning curve was four-to-six months. Ms. Prophet counseled that productivity would be temporarily affected during the transition period until proficiency and confidence were achieved. She advised being cautious when conducting longitudinal data analysis. Diagnoses might be classified differently or code definitions could have changed, causing misinterpretation; users needed to understand that changes in trends or case volume might simply reflect changes in data reporting.
AHIMA considered interactive, Internet-based training an excellent methodology for training coding professionals and end users of coded data relatively quickly and specifically. While face-to-face seminars remained an effective training vehicle, Ms. Prophet said new technologies offered probably superior alternatives. Audio seminars could be delivered at low cost to a large audience and Web-based training offered accessible, flexible and highly effective training opportunities. AHIMA anticipated that 10-CM training for entry-level and advanced coding programs (and the coding component of HIM academic programs) would be incorporated into existing ICD coding curricula. AHIMA also expected that for some time students enrolled in coding programs would learn both 9- and 10-CM diagnosis codes. Affirming AHIMA's commitment to training the industry in HIM, Ms. Prophet said education would be accessible through the Internet and a network of 52 component state associations and coding professionals. She cautioned that training shouldn't be conducted so much in advance of implementation that its effectiveness was diminished. Because the scope of HIM included coding, she said AHIMA was also prepared to address appropriate coding documentation, management (including development and coordination of implementation strategies), and related data analysis issues. Ms. Prophet emphasized that coding leadership, professional development, and coding consistency were central, strategic mission issues for AHIMA.
Noting information systems personnel had to understand the logic and hierarchical structure of 10-CM, and data users needed to understand the definition and composition of categories in the classifications, Ms. Prophet said AHIMA understood that NCHS would provide a crosswalk between 9- and 10-CM. AHIMA also understood that NCHS planned to conduct a comparability study to assist users in discriminating between real changes and utilization by diagnosis and those changes that were artifacts of coding system changes.
While advising that, ideally, 10-CM and 10-PCS should be adopted at the same time in order to limit the transition period (and disruption of data quality, coding accuracy, coding productivity and prompt reimbursement of claims), Ms. Prophet acknowledged the systems were distinctly different and that it was possible to implement separately. She said the question needed to be considered in the context of the environment and a way that insured appropriate timetables and strategic planning to accommodate the needs of the industry, and not just a coding system.
Although significant costs were associated with adopting 10-CM as a standard code set, Ms. Prophet noted the high price being paid for retaining a hopelessly outdated and limited system. She said ultimately reductions in cost would be realized with the availability of better data. Further delay wouldn't reduce direct costs and indirect costs couldn't be ignored any longer. AHIMA believed 10-CM had the capacity to grow along with medical science, serving healthcare diagnostic data needs for many years, and that gains with the improved data were well worth the cost and difficulty encountered during transition. Benefits of implementing ICD-10 noted by Australia and Canada included greater detail and specificity, a more clinically credible and easier to use and understand classification system, improved coding system structure and code presentation, and a capability for international comparisons.
Ms. Prophet reiterated that AHIMA and its national network of coding professionals were uniquely capable of assisting in researching the best implementation strategy and taking the lead in national workforce retraining and user education for new procedural and diagnosis coding systems. She noted that the Federation of American Hospitals endorsed and recommended the content and approach of AHIMA's testimony.
Panel 1: Possible replacement of ICD-9-CM, Vol. 1 & 2 (Diagnosis) with ICD-10-CM
Ms. Frazier said AAPC's written testimony concurred with many of Ms. Prophet's comments. AAPC agreed that the current ICD-9 made for imprecise coding and inaccurate documentation for reimbursement. AAPC coders heartily supported the uniform coding guidelines' reporting requirements and code acceptance by all users.
The Subcommittee clarified for Ms. Frazier that, after the NPRM and public comment, the publication of the first transactions in the code set rule would be followed by a 24-month implementation period. Noting AAPC received many queries from members about the procedural and diagnostic coding, Ms. Frazier asked if the draft version to be posted on the NCHS Web site would be close enough to the final version that AAPC could use it for developing training materials. Ms. Pickett said guidelines and, at some point, training materials would be posted along with ICD-10-CM. Groups collaborating with ICD indicated, as Dr.Powers did, that they'd tailor material to accommodate their constituencies. Ms. Frazier said AAPC would like to work with NCHS in disseminating this information.
Panel 1: Possible replacement of ICD-9-CM, Vol. 1 & 2 (Diagnosis) with ICD-10-CM
Ms. Leon-Chisen explained that AHA's Central Office on ICD-9-CM served as the U.S. clearinghouse for issues related to ICD-9-CM. The central office also published Coding Clinic for ICD-9-CM, the official publication for ICD-9-CM coding guidelines and advice, as designated by the Cooperating Parties (CMS, NCVHS, NCHS, AHA, and AHIMA).
Ms. Leon-Chisen said AHA considered ICD-9-CM diagnosis codes outdated and insufficient to provide detail for non-acute conditions and AHA continued to support migration to ICD-10-CM. She noted Dr.Sondik had referred to the NCVHS 1993 report and said AHA continued to support the Committee's decision that, “because the International Classification of Diseases 9th revision (ICD-9) is rapidly becoming outdated and worldwide support and maintenance of the classification will terminate, and because the International Classification of Diseases, 10th revision (ICD-10) represents significant improvements in coding of primary care encounters, external causes of injury, mental disorders, and neoplasms, we have recommended that the Department immediately commit resources to assess the applicability of the ICD-10 for coding morbidity; to identify problem areas and make modifications necessary; and to develop implementation plans.” She noted that since then, ICD-10 was reviewed and critiqued by hospitals, coders, medical special societies and researchers. Areas for improvement had been identified and modifications made. The sole NCVHS recommendation left to accomplish was development of implementation plans.
Noting that accurate coding translated to accurate and more precise reimbursement for services delivered, Ms. Leon-Chisen cautioned that the limitations of ICD-9 were likely to constrain submission and payment of electronic claims. Ms. Leon-Chisen said the more detailed code assignments ICD-10-CM made possible would greatly reduce administrative burden for hospitals; reduce requirements for submission of additional documentation to support claims; allow the capture of accurate data on new ways of describing diseases due to advances in medical knowledge; provide data to support performance measurement, out comes analysis, cost analysis, and monitoring resource utilization; and increase the sensitivity of the classification when making refinements in applications such as grouping methods.
She noted many issues related to the implementation of ICD-10-CM (e.g., timing, cost, and testing) had previously been raised in testimony to the Subcommittee about ICD-10-PCS and were issues that needed to be considered in changing any clinical coding system. AHA advised that the timeline for implementation should be carefully orchestrated to minimize administrative burden to providers. Hospitals already faced numerous regulatory changes and their ability to absorb all these changes had to be carefully considered. The vast majority of hospitals were dependent on their hospital information system (HIS) vendors for programming changes. AHA supported the HIS industry in requesting implementation for diagnosis coding be carried out in tandem with the migration to the ICD-10-PCS procedure codes. AHA also supported the recommendation to implement ICD-10 three years after implementation of the HIPAA electronic transaction standards.
Lessons learned from the international implementation of ICD-10 diagnosis coding indicated it took a minimum of two years to implement. The U.S. had a greater level of automation and more pervasive use of clinical coding across different healthcare delivery systems. AHA believed an additional year would ensure more thorough testing of the information systems and electronic transactions utilizing the codes.
Remarking that implementation of ICD-10 diagnosis and procedure coding was a complex, costly process, Ms. Leon-Chisen urged Medicare and other health plans to be sensitive to increased regulatory costs and address current and future payments accordingly. AHA believed the cost of implementing significant new regulations (which weren't factored into the current prospective payment rates) that affect provider operations should be worked into the Medicare prospective payment rate update.
AHA believed Congress should establish grants to help hospitals with the enormous cost of complying with the HIPAA rules, including conversion to ICD-10-CM. For hospitals, the bulk of the cost associated with adoption of ICD-10-CM was related to personnel training. Coders and billers would require training to familiarize themselves with new coding guidelines, rules and definitions. Ms. Leon-Chisen said the training cost could be quite high for hospitals; even though the principles underlying ICD-10-CM were similar enough to ICD-9-CM that experienced coders could quickly learn the new system. Experienced coders were able to follow the test draft of the ICD-10-CM codes and identify how the new codes fit in with existing ICD-9-CM diagnosis codes. AHA believed that learning ICD-10-CM didn't have to be a monumental task, if sufficient time and planning efforts were undertaken.
Ms. Leon-Chisen said changes to the coding system would also require extensive and costly modifications to information systems. Software applications and functions such as code assignment, medical records abstraction, aggregate data reporting, utilization management, clinical systems, billing, claims submission, and groupers required modification. In essence, every electronic transaction requiring an ICD-9-CM diagnosis code had to be changed. Changes included software interfaces, field length formats on screens, report formats and layouts, table structures, holding codes, expansion of flat files, coding edits and significant logic changes. Hospitals had to bear the financial burden associated with software changes as well as possible hardware upgrades. The hurdle for AHA members was the initial investment needed to prepare for these changes.
During the transition period, information systems had to support both ICD-9-CM and ICD-10-CM coding systems, requiring additional data storage space. Smaller rural healthcare providers, many with less sophisticated information systems and facing serious financial challenges, were further handicapped in their ability to accommodate such changes and might require additional resources and support to acquire information and coding system support programs.
Although ICD-10-CM hadn't been thoroughly tested in the U.S., Ms. Leon-Chisen pointed out that the WHO and Australian and Canadian clinical modification versions were already successfully implemented. And ICD-10 had been implemented internationally in DRGs. ICD-10 had been used for a variety of functions, from management of utilization resources to clinical management and research, as well as reimbursement.
AHA believed hospitals were capable of taking on the challenge of implementing ICD-10-CM. A 1998 review of the ICD-10-CM draft found the changes logical, coder friendly and an improvement over ICD-9-CM. The greater level of detail had already corrected previously confusing areas. And greater detail in specifying conditions and the addition of unique codes for conditions previously allocated to a catch-all category would result in more specific data collection. A review evaluated whether the system would be capable of answering questions ICD-9-CM users submitted to the central office clearinghouse or that a hospital coder would have. Ms. Leon-Chisen noted further testing was needed to ensure that the alphabetical index was as complete and useful as possible. AHA and its members welcomed the opportunity to conduct further testing, including medical records review, across different healthcare settings. Ms. Leon-Chisen said testing should also consider comparability of the new system with prospective and other current payment systems. Based on the assumption that use of ICD-10-CM wouldn't result in lower reimbursements relative to levels received using ICD-9-CM, AHA supported migration to ICD-10-CM after testing took place and start-up funding was established. She reiterated that Medicare and other payers should recognize the increased cost of these regulations and adjust payments accordingly. She said AHA would be pleased to assist NCVHS and CMS in identifying the nature and magnitude of additional costs.
AHA supported well-defined implementation and maintenance classes. The process should be broad based, predictable, and take into account the needs of all users (including establishment of routinely scheduled meetings to review coding changes and a date certain for using approved codes) and their capabilities to adapt to coding changes. AHA believed the ICD-9-CM coordination and maintenance process was well-positioned to reach the broadest audience possible and supported that process for ICD-10-CM. AHA also believed clear, unambiguous instructions and consistent official coding and reporting guidelines should be readily available, recognized and accepted by all payers, preferably as part of the HIPAA standard code set. AHA supported the cooperating parties' continuing role in the development of guidelines and clarification on the application of ICD-10-CM and ICD-10-PCS, as was done for ICD-9-CM.
AHA, acting through a long standing memorandum of understanding with HHS stood ready to continue to provide ICD coding advice and training under ICD-10. Ms. Leon-Chisen said AHA would be happy to work with NCVHS and CMS in developing a detailed implementation timeline with milestones. Lessons learned from international implementations of ICD-10 should be applied. And backwards and forwards electronic crosswalks between ICD-10-CM and ICD-9-CM diagnosis code should be available free or at reasonable cost.
Panel 1: Possible replacement of ICD-9-CM, Vol. 1 & 2 (Diagnosis) with ICD-10-CM
Ms. Smith conveyed McKesson Information Solutions' belief that the structure of 10-CM was logically organized and provided greater clarity for coding diagnosis. McKesson's review of the 1998 draft of the endocrine, nutritional and metabolic diseases chapter, showed the assignment of separate codes for Type I and Type II diabetes allowed for greater specificity with a broader range of diagnosis codes to more accurately and precisely describe a patient's diabetic conditions. Increased specificity enabled fewer codes to describe a condition (reducing coding errors) and facilitated DRG assignments. Ms. Smith noted that greater exactness in linking patient diagnosis to outcomes facilitated more accurate collection of outcomes data. In turn, this could increase the ability to measure the effectiveness of services provided and management of disease processes, safety, and efficiency of care. Migrating to ICD-10-CM also allowed for expansive comparison of medical data internationally.
Although McKesson believed the advantages of transitioning to ICD-10 would prove beneficial for clinical usage, as an information solution provider, McKesson's interests rested mostly with the timing of ICD-10-CM. Unlike a medical record coder or clinician, McKesson's use of the codes was centered in the area of software applications. Ms. Smith said she'd discuss application software more thoroughly in the second vendor session.
Discussion
Noting they'd heard repeatedly that the government should fund implementation, Dr. Zubeldia asked if not having funding would change any of their views. Ms. Prophet replied that AHIMA hadn't taken a firm position on who should fund it, but believed (as some presenters indicated) that, although quantifying benefits could be tough, many cost reductions and financial benefits otherwise might not be realized. They had to go forward with 10-CM, even without government funding. Ms. Leon-Chisen said government funding would ease the pain; but ICD-9-CM was broken and moving to ICD-10-CM was necessary.
Asked their perspective on the government preparing training materials for ICD-10, Ms. Frazier said AAPC thought government-developed training materials would be helpful. Ms. Prophet agreed, but added that AHIMA also would develop customized educational materials for members and other users. Ms. Leon-Chisen concurred. Government-produced training materials would start everyone out with the same instructions. But usually more detail was needed and there should be room for private options.
Mr. Blair asked about how often classes had to be run because of attrition. Ms. Frazier said it took two years of education and experience to teach the I-9 and procedural coding; people earned continuing education units every 12-24 months depending on the level of coding. AAPC estimated two years for I-9 users to transition and longer to introduce new coders into the system. Training seminars at the different levels were conducted all the time. Ms. Frazier said certified coders tended to stay in the profession. Most coders at the national conference had over five years experience; some had 10-15 years experience. Ms. Leon-Chisen noted the educational program produced a constant supply of AHIMA-certified coders. ICD-10-CM was close enough to ICD-9-CM that it would only take a short time to train them. It would mostly be a matter of going over any differences in the guidelines and instructing on the code structure for ICD-10-CM. One to two days training would be sufficient. She concurred with Ms. Prophet that there were many ways of reaching coders (e.g., audio seminars, Web casts, video tapes). Ms. Prophet said AHIMA constantly developed new educational programs and had a variety of Web-based coding products for experienced coders. AHIMA also certified coders and required ongoing continuing education to maintain skill levels. Ms. Prophet reminded the Subcommittee of the coding industry's severe work force shortage. Coders could often name their job and career path and many did stay in the profession.
Noting that Ms. Prophet mentioned that Canada implemented ICD-10 codes electronically, Mr. Blair asked if electronic implementation of ICD-10-CM wouldreduce the requirement for coders over the next three to six years. Ms. Prophet replied that the coder friendliness and use of clinical terminology consistent with physician's terminology could help. Productivity would increase, reducing the coder requirement.
Recalling that Ms. Prophet had indicated a four-to-six month period during transition when people didn't code as well or as fast, Dr. Cohn asked if, given the work force issue, hospitals were concerned about potential revenue implications. Ms. Prophet reiterated that the code set's specificity was dependent on the specificity of the medical record documentation. Part of the learning curve involved the clinician's documentation. Although both Australia and Canada indicated it wasn't massive, there was a recognizable slip in productivity and quality until people gained proficiency. Both countries offered post-implementation refresher courses because in actual practice things came up that weren't encountered in the initial training. Ms. Leon-Chisen agreed. In creating new codes, they tried to eliminate as much potential confusion as possible: explaining why the code was created, what it was intended for, and anticipating questions coders might have. AHA was considering a clearinghouse and official guidelines so there was a way to get a consensus opinion on how codes should be applied and information disseminated and consistently applied. Ms. Frazier observed that coding wasn't country-bound: third-party payers sent documents overseas to be coded and that exchange might increase with implementation of ICD-10 internationally.
Ms. Burke-Bebee noted that a white paper prepared a while ago by a WEDI ICD-10 subcommittee couldn't endorse recommending ICD-10. She asked members of that workgroup if that morning's testimony differed from the group's conclusion. Ms. Leon-Chisen said she'd co-chaired the two-year-old effort. AHA had already told the Committee in 1997 that it supported ICD-10-CM but the WEDI paper, which was the work of many different organizations, came short of endorsing ICD-10-CM. Ms. Prophet added that AHIMA also had been in favor of implementing 10-CM.
Dr. Fitzmaurice asked how realistic it was to count on computer parsing of narrative text (e.g., hospital discharges) to assign diagnostic codes to the inpatient in order to ease the burden on coders. Did a move from ICD-9 to -10 and a more common vocabulary make that more realistic or should they count on structuring information entered by the person supplying the narratives (e.g., the physician)? Ms. Prophet said the burden on the production line coder attaching codes to the claim might be reduced because of the greater specificity in ICD-10-CM. But she noted the role of educated coding professionals was in the direction of clinical data management and the translation of code sets and the logic behind them. Computers lent themselves to the production line, but hadn't taken the place of the human brain. Canada implemented ICD-10 electronically, but much of their coding process was still done by coding professionals. Ms. Greenberg said the main example of automated coding wasn't in morbidity but in mortality where the level of clinical detail wasn't as great. The U.S. automated mortality coding system was used by a number of countries world-wide. The system coded about 80 percent of what the physician wrote on the death certificate. Ms. Greenberg noted the system didn't do well on external cause of injury information, but associated software systems assigned the underlying cause of death according to WHO rules and this was moving the world towards more standardized mortality data. While considerably less detailed than the clinical modification, this provided the experience and a foundation in automated systems for mortality. The ability to migrate that to morbidity remained to be proven.
Asked how hospitals would cope with the shortage of coders, Ms. Leon-Chisen remarked that ICD-10-CM probably lent itself better than ICD-9-CM to electronic coding. She noted they were seeing the application of codes with many more clinical settings, and more emphasis was placed in out patient coding (e.g., home health coding). She said there were hopes that, as advances were made in artificial intelligence, physicians might learn to document using terms a computer understood. Ms. Greenberg reported that with the automated systems for mortality there was less need for basic coders, but greater need for advanced coders and nosologists who could modify and update systems and understood the rules. As a result, the international community was training and credentialing mortality and morbidity coders worldwide. Ms. Prophet noted that the whole work force issue for the HIM profession was a major strategic focus of HIMA. Ms. Leon-Chisen added that for AHA, the work force shortage was an issue across all areas of healthcare delivery.
Mr. Augustine asked if the more accurate reimbursement ICD-10 provided would increase reimbursements. Ms. Leon-Chisen explained that in a budget neutral situation one wouldn't necessarily get more reimbursement for a particular case but overall reimbursement would be more accurate. At this point, with so many un-specific codes, she said people almost guessed as they grouped cases.
Asked what the implementation plan should look like, Ms. Prophet said AHIMA envisioned a strategy that encompassed everyone in the system. A lesson learned from other countries was that players involved from the beginning were more prepared. The strategy had to encompass the different categories of data users, how they'd be affected, and first steps. Ms. Prophet emphasized that many preparatory steps (e.g., readying information systems and developing educational materials) didn't have to wait for publication of the final regulation. Even the original HIPAA rule on electronic transactions advised that information systems should start to build flexibility into their systems in preparation for new code sets.
Dr. Zubeldia asked if the implementation plan put the transition burden more on payers than providers. Ms. Prophet said AHIMA recommended a standard effective date for everyone to start accepting and submitting the ICD-10 codes. Ms. Leon-Chisen agreed. It would be more burdensome and costly to maintain and manage two sets of codes. Ms. Prophet noted there'd be claims and services from encounters prior to the implementation and they would still have to maintain two sets of codes in their systems for some time.
Ms. Smith clarified that McKesson didn't support a rapid move. She anticipated that implementation would take McKesson three years. McKesson saw it as a natural, necessary progression, but she expressed concern about the time frame. McKesson believed in a drop-dead date. Otherwise, there would be system issues with having to identify codes without sufficient logic built into the system. Dr. Zubeldia remarked that what he'd heard wasn't a drop dead date, but to use ICD-9 for diagnosis up to a set day and then switch to ICD-10 for services provided after that day. Ms. Prophet concurred. Starting on October 1, discharges or service date encounters would be coded with 10-CM, but claims submitted for earlier dates would still use ICD-9-CM. Procedure code was tied to the date when the service was performed. The discharge date drove the hospital in-patient side. Ms. Leon-Chisen noted this was similar to what was done when the diagnosis codes changed.
Panel 2: Systems Issues Related to Transitions to New Code Sets (Vendor Perspective)
Ms. Johnson conveyed providers' thoughts about operationalization of the change to ICD-10 and what it required. THS believed five things needed to be looked at: process, technology, technology's impact on other systems, training, and the perception of the HIM departments in THS's facilities.
Ms. Johnson remarked that the process being considered wasn't significantly different from what was done today. ICD-9 and -CM had been around since the 1970's and were out of codes. Going forward wasn't a hard decision to support. ICD-10 would provide a significant advantage in the ability to define the procedure or diagnosis.
Ms. Johnson observed that the people who process and release records were part of a different function within the medical records department than those handling coding. THS (which represented thousands of entities and 116 provider organizations, ranging from very small to very large) believed that these HIPAA changes wouldn't significantly impact processes within HIM. However, she expressed concern that the data collected via the diagnosis and procedure codes was also used for credentialing, cost accounting, case and disease management. Ms. Johnson noted those systems also needed to be ready to handle the new six-digit data format. And she cautioned that institutions that used a code book and coded manually faced a more labor-intensive effort and needed additional remedial education.
Ms. Johnson advised that it was critical to develop a transition plan for old and new data. Codes changed when credentialing information was collected very specifically (e.g., how many gall bladders the general surgeon did). That information was now captured electronically and bridges had to be built to capture both chains of information in order to have a complete set of data. Ms. Johnson pointed out that this happened every year, but had to be taken into account. She said THS expected that the cost of major software changes in the supporting systems and increased field links would be incorporated in the annual license fee that covered updates in coding.
Ms. Johnson said most organizations that would be impacted by the implementation were anticipating and ready for the change. But THS had witnessed provider organizations in denial and unprepared for the coming changes. She emphasized that transition planning was critical. THS had begun talking about HIPAA early in 2000 and had inventoried its software and all its decision support systems, knew what changes had to be made, and had already begun to talk with vendors about how to handle those changes. Other than concerns about vendors not being able to handle the digit change, THS was confident the implementation would work for them.
Panel 2: Systems Issues Related to Transitions to New Code Sets (Vendor Perspective)
Representing Gambro, Inc., parent company for Gambro Health Care, a large dialysis provider of hemodialysis, perineal dialysis and related services to more than 40,000 patients in the U.S., Mr. Riopelle discussed the impact of code set changes on a large provider who utilized proprietary information systems. Gambro captured data on and submitted claims for over 6 million patient treatments each year on systems developed internally. In response to code set changes in HIPAA, Gambro organized its systems activities around two main platforms: Renal Information Management System (RIMS) and Billing and Accounts Receivable (BAR). A key component in generating clinical activity reports, RIMS captured clinical data and treatment level information, including coding input. BAR extracted appropriate clinical data from RIMS to create and submit claims to payers. BAR served as the fundamental basis for running necessary financial and administrative reports on Gambro's operations.
Mr. Riopelle said Gambro worked diligently for two years to respond to the changes HIPAA required. Gambro was in the midst of responding to both the privacy and transactions components of HIPAA and awaited publication of the final security regulation. The planning and assessment phase comprised several thousand hours of professional time and a significant six-figure total cost. Mr. Riopelle reported that the implementation process for the transaction standards would involve some 30,000 hours of professional time and hardware purchases totaling a mid-seven figure cost (about 15 percent of Gambro's IT capitol annual budget).
Mr. Riopelle noted the transition from local to standard code sets would be a substantial effort for Gambro, which had over 13,000 potential payer relationships in its system and over 23,000 local codes. The implementation process impacted many other areas within IT: e.g., training, reports, data warehousing, planning, testing, electronic medical records, billing documentation, translators and general infrastructure. The transactions compliance effort also required a significant re-engineering of business processes. Gambro ran some 300 reports on a regular basis within BAR and RIMs. As the content of those reports changed, the report running, interpretation process, and staff training changed.
Another consideration for Gambro was the timing and interconnection of privacy, security and transactions. Mr. Riopelle said managing the different compliance dates and their attendant systems impact was challenging, especially when the regulation hadn't been interpreted by the industry and final compliance dates and requirements were subject to change. He encouraged the Subcommittee to continue soliciting input from affected organizations on how to effectively time the modifications and to consider all points of view on the timing of things so interconnected.
Mr. Riopelle also commented on the effort required to support old formats for retrospective review. In some cases, this required operating two systems; at a minimum it meant having the data warehousing capacity for a number of years. Another consideration was compliance. Following Medicare billing and documentation requirements was a critical skill for anyone who had Medicare as their biggest customer. Medicare billing was complicated. To the extent coding and related documentation requirements changed, retraining Gambro's work force to continue to submit accurate, compliant claims could become a big undertaking with potentially severe penalties and consequences.
Mr. Riopelle urged the Subcommittee to consider the differences between a commercial software vendor and a proprietary user as it evaluated the impact of code set changes. A proprietary user's IT talent necessarily spent its time responding to ongoing business issues and it was forced to react to legislative change, rather than proactively search out potential changes. A commercial vendor had a significant financial incentive to proactively look for legislative changes and the human capital to study them and develop new products in response. Mr. Riopelle noted another difference was how each bore the cost of change. For a commercial vendor obligated to update through a licensing agreement for a fee, the cost of those changes was born by hundreds, perhaps thousands, of customers. Proprietary users didn't necessarily recognize any new revenue from that change and bore the net cost of a change alone. Because of its significant budget and business resource impact on a proprietary user, Mr. Riopelle said at least three years were needed to effectively prepare, budget, plan, and execute a change to ICD-10. Final comprehensive changes needed to be released with detailed action plans supporting them. He said the lead time should begin when the change was final, tested, and endorsed by key constituents impacted by it, and when possible funding to support the cost of implementation was identified.
Although reimbursement levels occasionally increased, Mr. Riopelle noted they didn't typically increase at a rate adequate to support wide-scale IT change. Gambro encouraged evaluation of a funding course to help underwrite these costs. Mr. Riopelle also supported establishment of a multi-disciplinary team that included proprietary users to evaluate all of the impacts of changing the code sets.
Mr. Riopelle emphasized that Gambro was supportive of coding changes that provided more effective capture of clinical data and more effective reimbursement. However, he said it was critical that such changes were required only after appropriate lead time, planning, testing, and evaluation emphasizing the impact to proprietary users.
Panel 2: Systems Issues Related to Transitions to New Code Sets (Vendor Perspective).
Mr. Hardesty, who was responsible for compliance efforts surrounding transactions and code sets at Cerner Corporation, a leading supplier of clinical management information and knowledge systems, said the industry's initial reaction to converting from ICD-9 to ICD-10 was one of incredulity. Already struggling to achieve compliance with changing requirements for transactions and code sets, privacy, and with anticipated rules governing identifiers and security, the industry didn't see how it could absorb the sweeping change required by conversion to ICD-10. But, given the inevitability of the change, Mr. Hardesty acknowledged that migration to ICD-10 was necessary and inevitable.
Mr. Hardesty said it was difficult to believe that there wouldn't be fallout and difficulties with these efforts, particularly with the issue of reimbursement. If compliance with major components of HIPAA was achievable by the end of 2003, Cerner supported a moratorium on major changes during 2004 so the industry could assimilate necessary changes and mitigate difficulties.
Cerner believed it would be beneficial for the industry to know that the transition to ICD-10 was definitely going to happen and would occur by a specific date. Mr. Hardesty said the implementation should be carried out in tandem with the implementation of ICD-10-PCS and any other contemplated changes to coding classification systems and procedures. He urged the Secretary to issue guidance that alerted the industry to the Department's overall plans and directions. Cerner supported a three-year time frame for conversion after implementation of HIPAA's major components.
Mr. Hardesty commented on several implementation issues with both proprietary and legacy systems. For some systems, the stored diagnosis element link presented the most significant remediation issue. Diagnosis information was stored in a variety of application product areas throughout a hospital: e.g., scheduling, registration, medical records, patient accounting, utilization reviews, and decision support. Databases that needed to be modified to accommodate ICD-10 would require significant conversion to accommodate necessary changes involving significant planning, testing, and down time during implementation.
Mr. Hardesty noted that other systems that didn't have the issue of database remediation might still have an issue with the display and/or reporting diagnosis. Vendor analysis of all affected system areas along with corresponding software modification was required. Mr. Hardesty predicted that most changes would be relatively minor. However, significant redesign might be necessary on particularly crowded displays or reports. Vendors also needed to analyze their internal storage of diagnosis element links. Those without issues within their own systems might be using third party software (e.g., for grouping or pricing) that used areas for internal storage that needed to be modified, tested and remediated. He noted that application and usage of an ICD-9/ICD-10 crosswalk would be necessary during the transition period and would require significant analysis. Mr. Hardesty agreed that there was no reason why many changes couldn't be implemented over time using existing ICD-9, a gradual transition could begin immediately rather than wait for ICD-10.
Mr. Hardesty said an even greater concern than each individual vendor's own system modification was the coordination of efforts between vendors. The challenge for any institution would be coordination of timelines between multiple vendors, testing of internal and external interfaces, and coordination of conversion outputs. Mr. Hardesty cautioned that, even with an absolute compliance date, there would be varying degrees of readiness (as currently with transactions) and coordinating testing with payers would be a challenge.
Given the direct correlation between diagnosis information and reimbursement, Mr. Hardesty said there needed to be a study and published guidelines that provided an implementation road map. And he noted that managed care contracts were also based on diagnosis and might require re-negotiation.
Mr. Hardesty said Cerner didn't charge additional software license fees for design development of software modifications related to government regulatory change. Such changes were covered under an ongoing agreement for maintenance support, a fairly common approach in the vendor community. However, given the broad impact of this change, Mr. Hardesty said there would be professional service fees for activities involving planning, testing, training, conversion and implementation. The costs would vary widely, depending on the size of the institution, the number of vendors involved internally and the overall number of external trading partners. He predicted a cottage industry would arise, similar to Y-2K, devoted to the coordination of implementation efforts.
Mr. Hardesty said vendors should unit test each affected application product area and systems test their entire product offering. Testing should be required for any necessary conversion programs and for all supported interfaces affected by the change. Vendors should also be required to test third party software utilized as product components. On-site testing would be required to ensure vendor-to-vendor interface integrity. Testing should be done with each external trading partner.
Remarking that he didn't believe any significant planning would occur until there was a date, Mr. Hardesty reiterated Cerner's request that the Subcommittee urge the Secretary to issue industry guidance regarding the intent to migrate to ICD-10. He encouraged the Subcommittee to call on ASAK and other organizations to help coordinate vendor efforts. Cerner also urged the Committee to ask WEDI to organize a workgroup to address various implementation issues.
Panel 2: Systems Issues Related to Transitions to New Code Sets (Vendor Perspective)
Ms. Smith reported that from an information systems perspective, McKesson's interest rested mostly in the volume of changes that needed to occur to software applications and supporting documentation. Software applications that captured, stored, maintained, received, sent or edited diagnosis would require modifications in every location containing a diagnosis code. Hospitals used software applications from multiple vendors and transition to 10-CM necessitated modifications to interfaces between the various internal and foreign systems. Hospital information systems often used third-party grouper and encoder systems. Generally, there was a sequence in which software vendors received file layouts, specifications and groupers; made system changes and tested changes in the software and interface. Because the final version of 10-CM hadn't been released, software vendors and developers of groupers and encoders probably would be designing and programming software changes simultaneously. While some overlap was expected, Ms. Smith noted the grouper and encoder systems needed to be completed and released in advance to allow smoother development cycles for providers and payers.
The state reporting files (e.g., cost containment, discharge reporting) and other state-specific files that utilized diagnosis data or the output of such codes required modifications. In payment systems, diagnosis codes often served as the foundation to other applications. In addition to field expansion and redefinition, system modifications would include revisions to the logic for fee schedule calculations, authorization screens, DRG assignments, pricing screens, data abstracts, auditing functionality and statistical analysis. Decision support systems contained databases for modeling purposes that addressed the current-versus-historical-data issue. For a while during the transition, a growing percentage of the population would be defined under the 10-CM code set while the remainder continued to be defined under the 9-CM code set. System logic had to change in order for decision-support systems to use the data for modeling or analysis (e.g., in defining when to invoke mapping programs).
Systems had to support dual code sets during the transition due to reporting purposes and the need to validate older claims. The maintenance and use of dual coding systems could impact programming and testing time (e.g., systems generally maintained ICD-9-CM codes by effective date and payer; factoring in ICD-10-CM potentially doubled the number of codes that had to be maintained).
Clearinghouses supported a variety of input formats into multiple forms that were executed as non-standard claim information was translated into standardized formats and the claims processed. Ms. Smith said these formats and programs required modifications to the field length and type. Field changes needed to be passed onto other modules within applications. Audit trails to validate that all the data was accounted for might need to be redesigned. Practice management systems used for claim processing already accommodated a six-character alphanumeric code, and modifications would be minimal.
Ms. Smith emphasized that the timing of the ICD-10-CM implementation was critical. Information system vendors already focused on meeting the compliance requirements of HIPAA transactions and code sets. McKesson recommended an implementation date of at least three years after that initial compliance date. Ms. Smith said this would allow the industry time to design, code and test software changes and implement system upgrades into payer and provider sites. She added that a drop-dead date and detailed implementation guidelines were critical.
As an information systems solution provider, McKesson's preference was to implement ICD-10-CM and PCS concurrently, avoiding the burden of maintaining version control. Ms. Smith noted the same areas in the software had to be modified for diagnosis and procedures. She also advised that DRG calculations would require grouper logic that recognized both codes and defined parameters to determine which code sets to use. With grouper changes generally made once a year, mixing the code sets would add further complexity to the system functionality, design, coding, and provider usage. A mapping between them would be required; forward and backward mapping would be desirable.
McKesson's experience in Canada involved building tables and setting parameters for users to convert to the Canadian version, ICD-10-CA. Code conversion occurred in abstracting applications. 10-CA wasn't used for billing purposes. In Australia, McKesson modified tables for alphanumeric codes in expanded fields as well as storage of multiple codes for patients. The scope of the changes was less complicated, partly due to the types of system applications the company had in those countries; McKesson hadn't dealt with groupers or DRGs.
Ms. Smith said training in ICD-10-CM would be required in order for software developers and programmers to understand the new system and facilitate applicable coding changes. She said resources needed to make those changes would vary, depending on the type of system and the extent of changes. McKesson estimated between 24 and 350 man-months.
McKesson believed that the existing process for public maintenance of the 10-CM and PCS code sets worked well and best served the vendor, provider and payer industry. Ms. Smith suggested that a group be identified within HHS that could respond to questions as software developers crafted functional specifications for the required changes.
Panel 2: Systems Issues Related to Transitions to New Code Sets (Vendor Perspective)
Mr. Mitchell remarked that the current code sets were embedded into virtually every portion of the healthcare system: vocabulary, software and processes. The code sets were used for clinical edits, care pathways, classification systems, compliance audits, as well as medical records coding. Logic, databases and software had been developed to support processes to leverage the clinical relationships and provide efficiencies and specific solutions to the industry. Changing to a new code set meant every system in place had to be reviewed. Changes to screens, algorithms and prompts had to be considered. All aspects of the current systems needed to be reviewed and potentially updated.
The issue of electronic interchange of data and interfaces, a core focus of HIPAA, was of particular relevance. All interfaces between systems that transmitted coded patient data must be reviewed to assess the impact of the new code sets and, as needed, changed. Interfaces consisted of two main components: a communication method and data formats (e.g., HL7, X12, XML). Code sets had little, if any, effect on communication methods. Data formats would change to support characteristics of the new codes being sent. For instance, current procedure code sets (ICD-9-CM and CPT) contained numeric characters only. New procedure code sets probably would contain alphanumeric characters and longer codes to increase specificity and support future expansion. These changes would require different data storage methods within a database and adjustment to any interface that depended upon type or length of field for accurate transmission of data. 3-M Health Information Systems (3-M HIS) encouraged and supported use of industry standard data formats (e.g., X-12 and HL7) to simplify the collaboration process.
Mr. Mitchell said changing code sets was an industry-wide challenge, but doable. Every vendor that elected to support the new code sets would need to systematically review every facet of their systems to identify dependencies on the current code sets. Necessary changes must be designed, developed and installed within customer systems. The new code sets would inevitably provide opportunities for new functionality within systems and processes; new code sets would not only be retrofitted to support existing system functionality, but potentially systems also would be enhanced. While the effort to implement changes would be substantial, Mr. Mitchell said the benefits derived from improved coding sets would outweigh the cost of implementation. 3-M HIS encouraged the Committee to set a clear direction and timeline to allow providers and vendors the opportunity to effectively prepare for the change.
Mr. Mitchell said 3-M HIS had expended significant effort to use the current code sets to address fundamental issues its customers faced in the business and delivery of healthcare. He noted that the process for assigning DRGs using diagnosis and procedure codes was first defined by Dr.Robert Fetter to improve the tools available for the administration of acute care hospitals in the area of analysis and process control. Subsequently, CMS used DRGs to address payment. These activities were replicated multiple times around the world using different code sets and algorithms.
While the new code sets required extensive changes to systems, Mr. Mitchell said the reasons customers purchased 3-M HIS's solution hadn't changed. Encoding, editing for accuracy, process improvement, training and education were still required. Unless business problems significantly changed due to the new code sets, 3-M HIS's practice was to update and upgrade its solutions within existing customer support contracts. Mr. Mitchell anticipated new services would be needed to fully implement any new code set. He noted that support of historic code sets (important for trending of data over time, use of public datasets, and provider and hospital comparisons) might require multiple solutions. Until new code sets were identified and documented, implications to the market couldn't be fully determined.
The implementation dates for HIPAA transaction formats and code sets were linked. Other portions of HIPAA (e.g., privacy and security) potentially required significant investment and infrastructure and process changes that tended to be time consuming and costly. While there was an advantage to bundling changes and minimizing the number of change cycles for providers, 3-M HIS recommended that the Committee be vigilant to the risks of combining the code set issue with other HIPAA issues requiring significant changes to infrastructure and processes as well as potential delays.
Mr. Mitchell recommended that notice be provided at least one year in advance of implementation, noting two years' lead-time was preferred. At the time of notice, all details related to new codes should be fully documented and publicly available. Further changes prior to implementation should be kept to the barest minimum. Mr. Mitchell noted that Canada and Australia provided vendors one-to-two-years advance notice with final ICD-10-CM modifications made as little as three months before final implementation. Even with late changes, vendors successfully updated products and providers made necessary changes and system upgrades in order to comply with government regulations. Mr. Mitchell said these experiences indicated that with sufficient clarity and notice, necessary changes could be made in a timely manner--but that full notice and a non-changing specification were key to implementing smoothly. Mr. Mitchell also noted that financial consequences of converting to a new code set might not be as great in countries that didn't contend with the U.S.'s PPS.
Discussion
Assuming security and provider ID came out before the end of 2002 for implementation at the end of 2004, Subcommittee members asked about an appropriate time to require the subsequent code set change. Ms. Johnson said THS had already taken the security regulations, written policies and procedures and put them in place. THS examined all applications and security functionality, including the databases, data warehouses and data centers and recognized probable changes. Not many were expected. THS also looked at the field links to determine whether THS could house additional IDs going forward. Ms. Johnson said THS could be ready in 2004, so long as the vendors were ready. THS' vendors believed they would be ready. Although there were system impacts relative to privacy and security, Mr. Riopelle said the code set change impact would be married closely with the transaction code sets. Noting the value of breathing room and the need to be comfortable with significant change, he said Gambro recommended three years from the final compliance date--October 2006. Mr. Hardesty said Cerner advocated between January and October 2006. Ms. Smith still looked to three years after the initial compliance date; McKesson could be ready in October 2005. Mr. Mitchell said 3-M HIS might be less sensitive to the transaction set issues, since it fed information into many systems as opposed to implementing them; 3-M HIS could be ready one-to-two years from when the decision was made.
Mr. Blair asked if some of the testifiers weren't focused only on their financial and administrative transaction systems. He noted many also had physician order entry, results reported, case management, and physician certification systems. This was a big task, but he pointed out that it was also a tremendous business opportunity that enabled richer clinical applications. Considering this broader perspective, he asked if testifiers still answered the same. Mr. Riopelle replied that, in that case, he'd like ten years. Mr. Blair noted that meant delaying the business opportunity ten years. Mr. Riopelle agreed there was a tremendous opportunity for vendors to provide more effective systems and recognize more revenue. A proprietary user gained a better system of capturing data and submitting claims but Mr. Riopelle pointed out he didn't have an interim financial benefit. Mr. Mitchell said the scenario Mr. Blair laid out was consistent with 3-M HIS' testimony. It wasn't just system changes, but new services--they'd heard about an emerging cottage industry. Mr. Mitchell reiterated that 3-M HIS could be ready one-to-two years after the Committee's decision to implement. Ms. Smith said McKesson wanted to make sure that the systems worked correctly and that it had enough time to do it right. The business opportunity didn't change McKesson's date of October 2005. Mr. Hardesty said that with Cerner's 1,500 clients, the issue was how to support and coordinate the implementation, conversion, and interface efforts within a condensed time frame. It wasn't a software issue, but a consulting and services issue. January 2006 was close to October 2005 and a reasonable time frame.
Mr. Augustine commented that IT spending, which had been significantly lower in healthcare than other industries, was getting up to speed. But he noted those extra resources were spent on HIPAA, so there was no seizing of an opportunity by the operations side to look at business opportunities. The decision support mechanisms and framework for healthcare systems were way behind other industries and things like Six Sigma enveloped the management culture. In other areas, the phrase was barely heard. Mr. Augustine surmised that this was shaped by healthcare's background as a QA, rather than QI or process-controlled type culture. He said this was something they had to work to change.
Noting some testifiers asked for early guidance and that the Administrative Procedure Act was dependent on public comment and “not done until it's done,” members discussed how the Secretary could provide firm guidance. Ms. Humphreys said what they heard across the board was, whatever the date give us a “sign on the wall.” She said publishing a final rule was the only way to declare a certain date. Even then, Congress could change it. Ms. Trudel and Ms. Humphreys explained that a notice of intent had been considered for the personal identifier because it was such a complex and controversial issue. The purpose was to open the issue to public discussion without the Secretary proposing a specific approach. But what they'd heard from everyone was only a date issue. The Committee could come up with a proposed rule and pick a date and hear comments. This wasn't a vexed issue that required a notice of intent. Ms. Humphreys clarified that this modification to the code set didn't have to go through the process mentioned in the final rules for transactions about exceptions from standards to permit testing of proposed modifications. That exception was to enable people who wanted to test to do so, but it wasn't a requirement for a modification to an existing standard.
Mr. Blair reflected that a compliance date for a vendor was different than a compliance date for the industry. He said he imagined that vendors had to have systems that would be compliant with the new ICD-10 codes a year and a half or two before providers and payers could be expected to meet the compliance date. He asked if the panelists had factored that in. Dr. Zubeldia added that the compliance date for the receiver of the transactions was different from the sender's compliance date. The sequence was the vendor, then receiver, and then the sender.
Noting that they all struggled with being left in limbo because of the lack of the final rules for transactions and any sense of what would happen with privacy, Mr. Hardesty remarked that what each vendor did with its product was the lesser problem. A greater problem was how to coordinate that at a given site where four or five vendors exchanged diagnosis information in one form, synchronizing that effort and making it work. Ms. Johnson said she routinely worked with 950 vendors. She depended on them. But she said everything done at THS required interface changes. THS dealt with those problems every day. Because of integration, nothing THS did anymore was unidirectional. Ms. Johnson said she wasn't diminishing the change, but she didn't want the industry crippled by waiting three-five years when they couldn't accurately reflect what was being done for patients today.
Recalling that Mr. Riopelle and Mr. Mitchell mentioned a resting period and that once the codes were in place they couldn't change, Dr. Zubeldia said he assumed that they understood that the codes were expected to change once each year. Mr. Riopelle said for Gambro the concept was finality. The industry was in a constant state of change from updated codes, but finality was what would happen on a fixed date. Annually adding to and updating the code was anticipated. That was finality because changes were made in small increments. Large system changes were disruptive. Vendors couldn't have a moving target.
Asked to describe a transition plan, Ms. Smith said McKesson believed everybody should switch over at the same time. It would be easier for payer and provider systems to pass information back and forth with everyone on the same system. She noted everyone also had to test with vendors and payers, but Ms. Smith couldn't say how long testing would take. Mr. Hardesty suggested that the WEDI workgroup could develop a study and a plan with milestone dates.
Dr. Cohn said he heard that most testifiers wanted, relatively soon, a notice of proposed rule laying out a direction. And then, in the not too distant future, they wanted a final rule providing absolute guidance and an absolute date followed by a three-year implementation. Ms. Smith affirmed that McKesson could start, once there were guidelines and a final rule. Mr. Hardesty and Mr. Mitchell concurred.
Mr. Blair said he'd heard the date linked to two different things: some certainty and implementation of other HIPAA rules. The final rule was certainty, whenever that happened. They knew it probably wouldn't be October 2003. Mr. Riopelle said his three years was simply digesting HIPAA and focusing on new efforts, allowing breathing time to create an implementation plan with milestones. Noting that the code changes came out about the same time every year, Ms. Johnson said it would help if ICD-10 came out before that, so they weren't maintaining both. She reiterated the need for the final rule, but asked to not delay the industry until 2007-2008. Instead, she urged everyone to think through the process, digest it, and give vendors time to figure out what they would do in response. She reiterated that THS couldn't do anything until it had vendors. Mr. Mitchell noted there were cycles within the industry (e.g., in-patient and out-patient cycles) and a rhythm to when people expected change. He recommended using them. He agreed that the benefit (increased accuracy of coding) shouldn't be delayed to 2006-2008. The sooner they got into the cycle of using the new code sets, the sooner they'd have better data and accurate information to make business and policy decisions.
Dr. Zubeldia asked again about the implementation plan. Ms. Smith said they needed to know the lead time, the drop dead date, and the guidelines for all the payers and how to implement. Dr. Zubeldia asked if vendors should be required to implement changes before the payers and providers. Noting they were talking about rolling ICD-10 out to the customers in that three-year time period, Ms. Smith said she didn't think having the vendors ready early would do much. Mr. Hardesty said Cerner's clients already required them to be ready. Mr. Riopelle said he was a vendor and a sender and the three years he talked about needed to incorporate internal milestones (if not provided as guidance by the government) of what had to happen. Part of that was creating as a vendor and testing as a sender with Gambro's receivers. He said guidance and direction in terms of milestones would be valuable.
Dr. Yasnoff pointed out that the Committee didn't have control over two time periods: the time between when it made a recommendation and there was an NPRM, and the time between issuance of the NPRM and the final rule. The only time period the Committee could make a definitive recommendation about was the period for implementation. If they assumed that the time period between making a recommendation and the issuing of a final rule was at least a year, and recommended that the time period for implementation after the final rule was two years, on that day they'd essentially be saying that implementation would be no sooner than 2005. Dr. Yasnoff asked the panel what it would like the Subcommittee to do with this one thing it could recommend.
Mr. Mitchell said two years from the final rule; 3-M HIS felt the absolute shortest time was one year. He said that was pushing it and assuming pre-work had been done. Two years from the final rule gave time to develop, coordinate, and have the rhythms in place. Ms. Smith reflected that if they basically had the final code set and it was a year before the final rule was published, and then they had two years on top of that, they'd have the three years they asked for. As a proprietary user contending with the challenges HIPAA created, Mr. Mitchell advocated a three- year implementation post-final rule, irrespective of how long it took for the final rule.
Alluding to Mr. Mitchell's idea about rhythms, Mr. Kolodner reflected that if it were a flat two years they might end up in the scenario described where they had an implementation date in the middle of August. Ms. Humphreys suggested a better algorithm would be the second October after the final rule was published. Mr. Mitchell said they might want to qualify it as fiscal years. Ms. Greenberg said they had the gist and would work out the algorithm's fine points. Given everything they'd heard about the inadequacies of 9-CM and that it would be a miracle if they had a final rule out in a year, she asked if anyone felt waiting two years from the final rule (making October 2005 unlikely) caused a problem. Remarking that, in her opinion, it would be an unmitigated disaster for the healthcare and data if this slipped beyond October 2006, Ms. Humphreys suggested another approach was to publish the NPRM with an implementation date certain of 2005. She noted they'd heard from a number of people that that was pushing too hard in terms of doing it right, so that the data was exchanged and interpreted correctly.
Mr. Riopelle agreed it wasn't valuable to put implementation off indefinitely. He acknowledged there were intrinsic and real beneficial impacts to doing this, but he pointed out that the folks that testified previously and today were heavily weighted towards large organizations who assimilated change fairly effectively. They hadn't heard yet from small rural providers. Noting many asked to implement CM and PCS together, members asked if the added complexity and the fact that the procedure part was more controversial and could take one-to-two years to resolve affected their thoughts about timing and sequencing. The testifiers' reiterated their preference to implement simultaneously and they said they'd gauged their time frames around that.
Dr. Cohn recalled Ms. Smith had talked about the need to have ICD-10-PCS to do groupers. Noting groupers could work on a procedure, diagnosis or both, he asked why that was compelling. Mr. Mitchell said there was a relationship, not a hard and fast dependency. OPCS procedures worked with I-10 diagnoses; I-9 diagnoses worked with I-9-CM procedures. It took different clinical research efforts to make the DRGs clinically coherent. Dr. Zubeldia suggested that if the switch happened at the same time, the groupers would be doing grouping based on the new procedure codes and new diagnosis codes, but if the switch occurred in two steps, there would have to be an intermediate grouper that used the new diagnosis code and ICD-9-PCS. He asked if it was worth waiting and only having to do it once, or should it be done because ICD-10-CM was needed now. Mr. Mitchell encouraged the Subcommittee to come to a resolution on the procedures, which was where the industry felt the most pain. Ms. Burke-Bebee pointed out that there would still be maintenance for the prior ICD-9 when ICD-10 was implemented. Ms. Johnson agreed it was a question of how long they'd have to run parallel systems; she cautioned that if there was going to be a significant delay, they needed to rethink it. Mr. Augustine said he didn't see how they could push this into 2008-2009. Ms. Leon-Chisen reminded the Subcommittee of their discussion at the last meeting. ICD-9-CM volume III was broken. Whether or not they decided to implement 10-CM and 10-PCS together, from the last hearing there was strong support to change to ICD-10-PCS to replace volume III procedures for the inpatient setting. If they weren't ready to decide on a single procedure classification system, at the very least they could take care of the hospital in-patient side.
Subcommittee Discussion of the ICD-10-CM Testimony
Ms. Humphreys said she thought the Committee should recommend to the Secretary that the Department publish an NPRM proposing to move simultaneously to 10-CM and PCS two years after publication of the final rule. Then they would know from the comments what the community thought about divorcing 10-CM from 10-PCS. Dr. Yasnoff agreed. But he added that the sense of the group was that the final rule should be issued no later than October 2003, to meet an implementation date of October 1, 2005. Mr. Blair remarked that they'd heard the vendors saying two years but also not earlier than October 2005. This covered both. Members also noted the Secretary had the option to make a one-year or six-months implementation between the second and third year--i.e., the second October after the final rule.
Asked how difficult it would be to get out another NPRM, Ms. Trudel noted the biggest concern was that this was a major rule that would have significant impact on everyone they did business with. Being able to assess and explain that impact was the critical factor. Ms. Burke-Bebee asked if two years out and 2005 was a realistic recommendation. Dr. Yasnoff said Ms. Humphreys and he recommended a two-to-three-year implementation after the final rule, with the hope that the Secretary would make every effort to issue the final rule before October 1, 2003, so that the implementation date would be 2005. But, in making the recommendation that way, they recognized it might take until October 2004 to issue the final rule. Ms. Humphreys agreed. In making a recommendation, she said the Committee had to be able to say to the Secretary that they'd brought people from all over to a series of hearings and this was what they'd heard. She said the way to get feedback so they could make these decisions was to publish the NPRM.
Mr. Blair said he'd heard Ms. Humphreys suggest that the Secretary put out the NPRM with a trial balloon, saying the compliance date would be two years from the final rule--and a caveat that in no event would it be earlier than October 2005. Ms. Humphreys said her preference was 2005. She emphasized that the Committee had to convey to the Secretary the fact that the code sets were broken in the worst possible way. Dr. Yasnoff said they also needed to communicate to the Secretary a decade of history. This wasn't a precipitous move.
Dr. Zubeldia asked about changes to the codes that might result from the comments to the NPRM. Ms. Humphreys said people were concerned that the procedures were broken more than the diagnoses. Some pressed for an alternative. She said she didn't see one, but she noted Ms. Leon-Chisen's suggestion that PCS be a replacement for the third volume, which wasn't required for out-patient. Ms. Humphreys said she could imagine a huge influx of people with various ways to solve the problem related to procedures only agreeing that there wasn't any option and that they had to go to 10-CM on the diagnosis side.
Dr. Zubeldia said he believed the vendors and providers wanted two-to-three-years to implement from the finality point, not necessarily the final rule. If the proposed rule said October 2005 and the final rule didn't come out until December 2003, they wouldn't have the full two years but Dr. Zubeldia said it wouldn't have a negative impact so long as they targeted for the October 2005 date.
Noting they had only heard from large vendors, Ms. Humphreys said others might comment that 2005 couldn't be done. Dr. Zubeldia countered that if the date was nailed down in the proposed rule, rather than left to float, everyone would know that was the target date and that they had three or four years. Remarking that some in the community didn't feel that was long enough, Ms. Humphreys said nonetheless a date should be mentioned. They'd heard that waiting forever wasn't an option. Dr. Cohn remarked that HIPAA taught them people didn't act until there was a final rule. He pointed out a noticeable absence of any payer on the day's panels. They'd been invited, and the Subcommittee had received written testimony from Blue Cross/Blue Shield Association (BCBSA) stating that they shouldn't do anything until all the transactions were implemented. But Dr. Cohn said that, based solely on testimony heard that day, they shouldn't feel comfortable that one year, 11 months would work. Ms. Humphreys suggested that Dr. Yasnoff's formulation might take care of both issues. Getting out some firm proposal to move to them at the same time, giving a minimum of two years to implement and take care of these issues, was probably the only logical step. Though she noted there was a chance that comments would lead the Committee to recommend a different course of action.
Recalling that a number of people said with reasonable definitiveness that they needed at least two years after the final rule, Dr. Yasnoff suggested a recommendation that covered that and still conveyed the testimony's sense of urgency that they shouldn't wait until 2007-2008. Both systems were broken and they needed to do something, but the reality was that took time. As Ms. Humphreys had pointed out: the comments would reveal any major problems with the recommendation. Mr. Augustine agreed that the proposal and floating of a date advanced the best compromise and conveyed urgency of action. Ms. Humphreys said recommending the replacement in the context of the current rule (which basically limited use of volume three of ICD-9 to inpatient) along with movement to PCS would fix the most broken situation and limit concerns about implementing PCS in other environments. Members reached consensus to migrate to ICD-10-PCS and CM for the uses that ICD-9 addressed that day.
Members discussed whether the recommendation to use the DSM-4 descriptors along with ICD-10-PCS needed to be expressed separately in this letter or if it was already adequately addressed in the PCS. They also discussed whether the Subcommittee should ask the Secretary to notify the APA that a crosswalk between DSM-4 and ICD-9 was compatible with the HIPAA standard. Ms. Humphreys concurred with assuring APA that mental health professionals could still use DSM. Ms. Pickett advised that, given information provided in that morning's testimony, they needed to start over on the white paper initiated since the last meeting that was to distinguish between coding guidelines and diagnostic criteria. She noted there were two issues concerning mental health disorders treated in a hospital. ICD-9-CM was the coding classification used in hospitals, but mental health providers billed on a professional claim using DSM-4. APA tried to harmonize them but, as Dr.Mirin mentioned that morning, some descriptors were different. Ms. Pickett said all that would be covered in the white paper.
Members discussed the other issue AHA presented regarding diagnostic criteria and clinical descriptions linked to DSM's numerical equivalent of what resembled ICD-9-CM code. There were subtle distinctions between descriptors in DSM-4 and ICD-9-CM. And descriptors didn't go in the HIPAA procedures, so payers could misinterpret the DSM CM codes. Any differences in the code descriptors had been harmonized in ICD-10-CM, but there would be at least a two-or-three-year hiatus when this situation would exist between DSM-4 and ICD-9-CM. Ms. Trudel remarked that, rather than put this interim issue in these recommendations, the Subcommittee might want to wait for the white paper and a sense of all the issues.
Dr. Zubeldia noted a request from the APA that, if DSM was adopted as the HIPAA code set, the association be designated as a code set maintainer. Dr. Zubeldia said that after hearing the next day's testimony they could draft a letter to the Secretary to bring to the June meeting. Dr. Cohn added that that the Subcommittee also had to develop a letter on the NPRMs. Ms. Humphreys said the staff would draft that letter for the Subcommittee to edit.
Dr. Cohn noted he was a member of the CPT editorial panel. As CPT was a HIPAA named medical code set, he recused himself from any votes on procedural coding issues. Dr. Zubeldia led the session.
Panel 3: Procedural Code Set Follow-up Reports -- HCPCS and AMA Update Processes
Mr. Beebe provided further information requested by the Subcommittee, based on his February 6thpresentation. Noting the Subcommittee had expressed concern about the proposed semi-annual early release of CPT Category III codes for new and emerging technology and the lack of coordination between CPT and the HCPCS's newly adopted quarterly update schedule. Since the February 6 meeting where this problem of the semi-annual updates was discussed, the CPT editorial panel discussed the issue of the Category III code release and implementation dates with several payer groups on the panel: Healthcare Insurance Association of America (HIAA), CMS, AND BCBSA. The panel and payer groups agreed on a July 1 release of the new Category III codes for January 1 implementation, giving payers a six-month window to incorporate those sets.
The CPT editorial panel chose to develop the Category III codes to address concerns of the healthcare device and pharmaceutical industries. The editorial panel code proposal criteria required that services receive full FDA approval for any devices or drugs used in the performance of the service. Candidates were also required to demonstrate that many physicians or non-physician healthcare professionals throughout the country performed the service or procedure. Mr. Beebe noted these criteria often delayed development of CPT codes for new and emerging technologies. To respond to pharmaceutical and device manufacturers' concerns about this delay, the panel developed Category III codes to provide standard descriptor language and codes to facilitate tracking and research on new and emerging technologies.
Mr. Beebe said the editorial panel was able to offer early release on Category III codes because the services weren't part of the Medicare fee schedule and didn't go to the AMA/Specialty Society RVS Update Committee (RUC) for relative value recommendations. CPT was used in connection with the Medicare Physician Fee Schedule and the panel had made every effort to accommodate the statutory requirements for annual updates to the schedule. Unless Congress acted to change this law, the panel would only consider early release for codes that didn't need to go to the RUC. AMA and the CPT editorial panel viewed the possibility of more frequent updates to medical code sets as potentially disruptive to the claims process. Because payers on the panel recommended a six-month implementation period for systems to accommodate Category III codes, AMA was concerned that more frequent updates could disrupt the claims processing system that already caused delayed payments to physicians.
On the issue of coordination between CPT and HCPCS code set release schedules, Mr. Beebe said AMA and the CPT editorial panel agreed on the need for greater communication and cooperation between developers. AMA believed coordination could only be achieved if the HCPCS alphanumeric editorial panel expanded its decision-making process. Mr. Beebe said the work of the alphanumeric editorial panel could benefit from the input of physician groups and he recommended that the CPT editorial panel be provided a seat and an opportunity to serve on the HCPCS editorial panel. Dr. Tracy Gordy, the AMA editorial panel chair, recently wrote to Tom Grissom, Director of Center for Medicare Management, voicing similar concerns and asking for greater physician and non-physician participation in the HCPCS panel.
Mr. Beebe noted the Subcommittee had indicated that the pricing of CPT was a concern, particularly related to the cost of acquiring information on CPT from the Internet and the cost of CPT Assistant. Concern was also expressed regarding a rumor that the AMA had new licensing prices and policies for use with CPT. Mr. Beebe said AMA had no intention of changing its licensing prices and policies. CPT electronic data files were licensed directly to end users (internal use license) and to product developers who wished to incorporate CPT in their products (distribution license). Royalties for internal use and distribution licenses were $10 per user per product (plus the initial cost of the data file, roughly $55). Volume discounts to the $10 per user fee were available.
Historically, CPT pricing was developed and based on individual users. Some software programs were now priced based on the number of concurrent users. Mr. Beebe said the individual-user price structure, which didn't require a computerized user counter function to monitor the number of concurrent users, was simpler to implement and administer. The modest CPT royalty was, in part, reflective of the broader user base of individuals, rather than concurrent users.
Mr. Beebe said AMA encouraged fair use (limited use that met specified requirements and didn't require license) under U.S. copyright law. AMA didn't require CPT license for CPT codes used in connection with super bills, patient bills or incidental viewing of CPT codes. Mr. Beebe said a hospital that provided patient access to their medical records would likely fall under fair use and the patient wouldn't pay. The hospital would pay for CPT incorporated into the software that facilitated development of the patient access tool or development of the bill. Mr. Beebe said Internet pricing and policy were the same as the distribution licensing policy with security features to protect copyright. AMA signed two amendments with CMS allowing free use of CPT on the CMS and their carriers' Web sites to communicate Medicare policies. Users could also look up CPT codes and Medicare fee schedule information on the Internet. Users could search by CPT code or key words in the long descriptor; the search returned long descriptors of the CPT service and Medicare fee information by geographic area. Subscriptions to CPT Assistant (a monthly newsletter that provided accurate, up-to-date information on the application and interpretation of CPT codes) were $169 (AMA members $115) per 12-issue year.
CPT Editorial Panel meetings were open to participating specialty societies and the Health Care Professionals Advisory Committee. Any member of the advisory committee and staff could attend panel meetings throughout deliberation and voting on coding changes, including development of code language and guidelines. Individuals and organizations requesting coding changes were welcome to present their views to the panel and remain throughout deliberations and voting. Panel voting was conducted electronically to maintain the confidentiality AMA believed was necessary to preserve the integrity of this editorial process where reconsideration and refinement were inherent steps in developing a quality product.
CMS, BCBSA, HIAA, AHA, AHIMA, representatives from the managed care industry and from HCPAC with rotating representatives from medical specialty societies participated in the CPT editorial process. Panel decision making was conducted in open by secret ballot; payers, physicians and non-physicians were permitted to vote. Although panel meetings were closed (unlike ICD-9-CM meetings), Mr. Beebe said panel decisions were made in the open.
Although the ICD-9-CM maintenance process was identified as the most workable approach in the preamble to the final rule for transactions and codes sets, AMA considered the CPT editorial process another desirable alternative. The ICD-9 Coordination and Maintenance Committee conducted open meetings; anyone was welcome to speak on an issue, and committee agendas were posted in advance on the CMS Web site. However, decisions weren't made at the meetings and the group making decisions wasn't specified and contained no representation outside the government. CPT editorial panel meetings were open only to participating physician and non-physician healthcare professional organizations, not the public. A specified group that contained representation from physicians, hospitals, payers and non-physicians made decisions at panel meetings.
AMA viewed the two processes as different approaches that achieved similar results in terms of openness and opportunity for inclusion. AMA believed CPT met the intent of HIPAA for an open update process by providing access to the editorial process to all interested parties, giving stakeholders a voice in deliberations, and broadening decision making to include a diverse group of stakeholders on a rotating basis, as practical. Acknowledging that the demands of a rigorous and thorough editorial process limited how many individuals could be involved in making decisions, Mr. Beebe said AMA was exploring ways of using the Internet to post agenda material and allow public comment. He said policies surrounding the release of CPT codes and licensing of CPT and the panel process were always developed to maximize the public interest.
Panel 3: Procedural Code Set Follow-up Reports -- HCPCS and AMA Update Processes
Mr. Gustafson noted the initials that now stood for Healthcare Common Procedure Coding System (HCPCS) had represented the Healthcare Financing Administration before the administration changed its name. Three levels of codes were used for ambulatory procedures. Level I codes were the CPT 4 codes the AMA editorial panel maintained. Another level of codes were alphanumeric. Level III were local codes (used largely by state Medicaid agencies and some local carriers) and were being phased out. HCPCS usually referred to Level II codes (e.g., codes for drugs, durable medical equipment, outpatient pass-through items, medical and surgical supplies,).
The National Panel consists of three organizations: the CMS, BCBSA, and HIAA. Consultants (which include AMA, Department of Veterans' Affairs, a durable medical equipment contractor and two private insurers) received full information on the panel agenda and provided comments. The CMS HCPCS workgroup, which had representatives from Medicare and Medicaid as well as State Medicaid, made recommendations to the national panel and provided a channel for development of activities. The panel received applications annually in April. New codes arose in January. Mr. Gustafson said members meet frequently with interested parties and established codes on their own motion. The panel convened three times a year to determine codes for the national cycle.
Mr. Gustafson said payer systems permitted use of unlisted codes when a specific code wasn't available. Claims involving that code kicked out for review on a by-hand basis at the carriers. Almost all of the billion claims Medicare dealt with each year were automatically processed. But Mr. Gustafson said it was appropriate to hand process small numbers of claims, rather than set up a special code.
Mr. Gustafson noted provisions for quarterly updates of C codes used in Medicare's hospital outpatient perspective payment system to identify newly arrived drugs and devices early in their life cycle. He commented on the trade between the speed one could move and the administrative burden on stakeholders. More frequent updates could cloud the “data picture” and analysis, particularly in terms of research and epidemiological concerns. But Mr. Gustafson said, as a payer, the most significant administrative burden was on their and their providers' systems. He concurred with what they'd heard about providers favoring longer cycles, noting a preference for annual cycles. However he noted suppliers and device manufacturers preferred to move into a more explicit payment status.
Mr. Gustafson remarked that Medicare established G codes for professional services used within its system. One might need to quickly add a code because Congress declared a benefit would begin on a certain date; a large volume of claims was anticipated before the next coding cycle; for programmatic, policy purposes; or to maintain codes distinct from those in the CPT series (e.g., the CPT code for vena puncture included with finger and ear sticks, which Medicare didn't cover; a more specific code was needed for them).
Addressing a concern that there was no general implementation guide for HCPCS, Mr. Gustafson noted this information was contained within the operating instructions from payers. Pointing out that Medicare was driven by statute and had to code appropriately, he said CMS provided extensive manuals and program memoranda (for carriers, intermediaries and providers) and, as payers, insisted on maintaining control over payment policy.
Mr. Gustafson said much of the HCPCS activity was in durable medical equipment. A schedule for public hearings providing three opportunities annually for manufacturers to make formal presentations to the CMS workgroup was announced last November in The Federal Register. Agendas and proposed recommendations were announced in advance and CMS could revise its recommendation, based on that discussion, before forwarding it for the HCPCS panel's consideration. Panel meetings weren't open to the public, but agendas were posted on the Web site and anyone could make comments that were reviewed by the panel.
Discussion
Asked about a process for making temporary codes permanent, Dr. Lori Feinberg, who represented CMS on the CPT editorial panel, explained that Medicare temporary codes were sometimes made permanent. (Medicaid also had that capacity and might do more in the future.) Proposals went to the committee and were put on the agenda. After public comment, they were voted on and adopted during the update cycle. Noting several home-health home infusion codes had become part of the CPT code set, she said more would be brought into the permanent code set. Ms. Humphreys noted the home infusion people had said the S codes in HCPCS were exactly what they needed and that users found the rubric “temporary” unsettling. Dr. Feinberg said the panel would consider requests asking that a temporary code be made permanent.
Dr. Cohn remarked that, as they moved to get rid of local codes, it was important that the HCPCS national panel not be impeded. However, he noted that over the year the Committee heard testimony from HCPCS and others concerned about what they perceived as a lack of coordination between HCPCS alphanumeric and CPT HCPCS Level I codes. He asked Mr. Gustafson and Mr. Beebe to describe their plans for improved coordination. Dr. Feinberg remarked that, in the past, agenda materials for every HCPCS had gone to CPT. But, with the move towards coordination, materials went to Mr. Beebe. He proposed that, rather than implement a new vehicle, they make the task more important. Mr. Beebe observed that those making decisions at HCPCS were all payers, and suggested that the process would be fairer and coordination would improve if providers were also involved.
Recalling that several years ago the Committee recommended that the HCPCS board be expanded for greater representation from Medicaid agencies and that it consider how the criterion that any new code had to have volume usage might stifle adoption of new procedures, Mr. Blair asked if the recommendations were still being considered. Mr. Gustafson expressed CMS's belief that the agency, which sponsored both programs, adequately represented both at the national level. He noted there was Medicaid representation at the workgroup levels, as well. He added that the question of who ought to be on the national panel was one the panel determined. Mr. Gustafson added that CMS didn't have a hard-and-fast rule about volume but made judgment calls based on what people said about items or services that were likely to be important and widespread. He emphasized that whether an item was paid for wasn't dependent upon it having a code: small-volume items could be paid for utilizing unlisted codes. Dr. Feinberg agreed. Sometimes a code was split out to discriminate between services and identify types to be paid. But the beneficiary could get products for which there was no code. Hundreds of HCPCS codes were added, changed or modified every year.
Following-up on the issue of guidelines, Ms. Humphreys said she understood the distinction between payment and coding guidelines. But noting that Medicare and Medicare were big payers, she asked if CMS had official coding guidelines for HCPCS about what code could or could not be used for a particular service. They'd heard from the field that different payers took the same service or procedure and had slightly different ways that they wanted it coded. She pointed out that this was what administrative simplification was supposed to eliminate. Dr. Feinberg gave an example of a code called Care Plan Oversight that Medicare used to cover that was worded to match their payment policy. A professional organization said the code would be better if they added another piece. Medicare said they could do that, but clarified that the change wouldn't qualify for payment. The CPT panel adopted the code and Medicare marked the code non-covered. The status was published annually in CMS's Federal Register notice.
Dr. Zubeldia noted that when CPT and HCPCS initially came out many years ago the instruction was to use the code in the CPT and, if there wasn't one, to look for a HCPCS code. That migrated to use the HCPCS code in preference to the CPT code, but wasn't consistent from payer-to-payer or-code-to code. He asked if there were guidelines on which code to use. Dr. Feinberg said that although they could publish the rule saying which codes must be used, they hadn't solved the problem of providers and provider organizations having to maintain a whole set of independent rules until they achieved consistency. Mr. Beebe said this was a big problem for AMA. AMA spent a lot of time helping people understand how to code and drew a hard, fast line between payment and coding policy due to anti-trust. AMA didn't get involved in payment policy, but was very involved in developing coding guidelines and rules. CPT had several levels of coding rules to help people understand and uniformly apply CPT codes, conventions and guidelines. AMA believed that without adoption of those rules (e.g., What was an add-on code in CPT? How were add-on codes to be applied and used with respect to other codes in the code book?) conventions were meaningless and didn't create an administrative simplification plan.
Mr. Gustafson supported the payers' perspective that they could only pay for what they were supposed to pay for. He said this was an issue because creating a uniform set of coding guidelines for CPT that applied to all payers could lead to situations where Medicare payment policy was directed not by coding, but by coding guidelines created by another organization. Ms. Humphreys said she easily believed this could be the case, but she also believed uniform coding guidelines were possible. She noted they'd heard from the field that these guidelines would be valuable. Mr. Beebe remarked that CPT modifiers allowed the professional to modify the circumstances under which the service or procedure was performed. It didn't change what was actually done, only the circumstances surrounding it. Modifiers were accepted under HIPAA, and all payers had to be able to accept them. Payers didn't have to change payment policy based on CPT modifiers. But everyone had to be able to accept the add-on code.
Dr. Feinberg remarked that, in a sense, HCPCS was a transaction system. A code that had been turned off triggered a response that it didn't describe a covered service. She noted that there was one code for PET scans in the CPT book. There were prescriptive limits on what was paid for PET and they had created G codes for PET for solitary lung nodules or for breast cancer staging. But they couldn't accept a claim for PET alone, because that didn't give enough information about whether they could pay the claim. The claim was accepted and instructions were sent back to code it in a different way so that it could be processed. Dr. Feinberg said that information was available in public pronouncements and on the Web site. She said a compilation of the program memorandum on codes could be made more accessible.
Ms. Humphreys suggested that, while everyone sought uniform payment policy, in its absence they'd like uniform coding approaches. The payment policy could be added on top of a uniform approach. She said that her understanding was that HCPCS wasn't designed just to meet CMS's needs. It was a national coding system for particular services and it didn't matter if the claim came to you or a third party, if that service was covered by HCPCS, there had to be a HCPCS code and a description of when to use it. That was the whole purpose of HIPAA.
Dr. Feinberg said the CPT panel could make an effort to organize its program memoranda to deal better, at least from a Medicare perspective, with the HCPCS coding, and post it on the HCPCS Web site. She said the coding guidelines might be specific: e.g., although CPT was willing to have S codes be permanent codes, their payment policy was based on a completely different theory for paying for those services and S codes would be denied. Ms. Humphreys commented that people might not want to send them, but only to use them in the same way and have them listed in an official HIPAA code set as permanently available. Dr. Zubeldia said it sounded as though they were trying to enforce payment policies through the code set. Dr. Feinberg said that was the only information they had about the service. Dr. Zubeldia suggested it would be easy to tell the provider not to bill for vena puncture done through an ear prick. Mr. Gustafson pointed out that the CPT manual would contradict payment policy. Dr. Feinberg concurred; the code needed to describe the service.
Dr. Yasnoff restated Ms. Humphreys' perspective. He suggested that the purpose of codes was to represent what the provider did in sufficient detail that all the decisions down line could be made. If the purpose of the coding system was only to facilitate payment, rather than represent reality, one ended up with a coding system without uniformity. He posed the scenario of a set of medical providers who painted patients' toes red and he said he'd argue that there should be a code for painting patients' toes red. Probably, no payers would cover that service, but it would represent reality. In order to make payment decisions, the codes needed to represent what providers did. He said a code for vena puncture that included services one didn't pay for wasn't a payment problem, but a problem with the code representing reality. He contended that representing reality was important because presumably there was more potential for agreement on representing reality than payment policy. The plurality of payment policies would only change over time. Dr. Feinberg asked if there would then be another code for painting patients' toenails blue? A distinction had to be made about where one stopped describing the service in detail. She noted there was always tension in a coding system. When somebody paid for a service and somebody didn't, that had to be distinguished at the code level. Dr. Yasnoff replied that, if there was a clinical reason to paint someone's toes another color, they would want different codes. And a 100,000 claims a year for painted toes would probably be treated differently than one claim in a decade.
Ms. Humphreys remarked that if the principle was describing procedures to be done, then, when a HIPAA transaction wasn't being sent, no one had to worry. But a HIPAA transaction had to be covered. The factors were what was sent and who was paying for it? Dr. Cohn commented on the difficulty of these issues where the world of coding hit the real world of billing. The process was imperfect. He said what he was asking for was better coordination. He was concerned that, as they moved towards HCPCS II becoming a national code used by everybody, that the panelists look at overlaps, redundancies, and areas of confusion. He said he anticipated that areas like home infusion would be able to build, either through CPT or S codes, sending codes either way and communicating on a bi-weekly basis concerns of various groups. He said they needed to bird dog and follow this process in order to coordinate better.
Dr. Feinberg suggested that knowing the situations in which coders had difficulty with the HCPCS codes would help in understanding the guidelines needed. She asked people to send input to the HCPCS Web site and said they'd make the problematic coding information (which could be a model for other payers) as accessible as possible. Aetna might have a different code than Cigna but, knowing the kind of information considered helpful, they could potentially produce a guide helpful to coders. Ms. Humphreys questioned why Aetna would tell somebody to code the same service differently than Medicare. She asserted that the same service should have the same coding; one guideline ought to fit both in terms of the description of service. Dr. Feinberg gave an example of a code for a shoulder replacement. A redo of that replacement was harder to do than the initial surgery and could be coded with the shoulder replacement code and a modifier indicating it was a complex case. Some surgeons wanted the codes broken out to initial shoulder replacement and replacement shoulder replacement. But the bulk of surgeons liked it the way it was because, presumably, that would change the price of both services. Momentum propelled some codes when provider interests were involved. Ms. Humphreys pointed out that with HIPAA that would no longer be possible.
Dr. Feinberg suggested that when an S code was created, promulgated and publicized it provided a distinct enough service. Mr. Blair asked how that differed from the process already in place. The comments the Subcommittee received before they made their recommendations several years ago conveyed widespread frustration that new codes weren't being dealt with in a quick, responsive way and that the perspective was focused on payment, rather than coding. Mr. Blair said he believed that in the HIPAA world the process for determining which new codes should be added had to reflect a broader responsibility. The scope had increased and they were hearing that the processes for adding new codes and modifying and updating existing codes had to be viewed in terms of this broader role.
Ms. Greenberg reinforced what everybody was saying with an example. She observed that, unlike the pre-HIPAA era, the post-HIPAA world wasn't payer-specific when it came to data. That didn't mean payers didn't have separate policies. But, from a public health or outcomes point of view, uniform data was essential across all payers and, hopefully, settings. That was what HIPAA was all about. Noting she represented the Public Health Data Standards Consortium on the NUBC, she remarked that UB-92 recognized that it had to move away from being so payer-specific. Ms. Greenberg explained that if Medicare needed a code for a legislative reason they put it in and, if BCBSA needed a code for the same purpose, then they had to harmonize the same code. She noted that Medicaid had taken thousands of state Medicaid codes for abortion and created a set of national codes. Ms. Greenberg said the pre-HIPAA model, with CPT having non-specific codes and HCPCS specific codes, no longer worked. She agreed that classifications didn't work without guidelines and she emphasized the need to collect uniform data and let people determine their own payment policies. She reiterated the need to get out of the current framework of payer-specific datasets and classifications. One criterion the Committee had articulated for classification was that it should be hierarchical. When there was both a specific and non-specific code, one was supposed to use the more specific code, even if the payer was willing to pay on the non-specific code. Using more specific codes was the essence of a coding guideline and how ICD worked. If they wanted to look at outcomes and resource utilization, then they needed the most specific codes.
Dr. Cohn commented that in addition to recognizing Levels I and II, it would be good to have recognition between alphanumeric and HCPCS Level I and enough coordination among various parts of HCPCS to produce a single set of guidelines. Recalling that Mr. Gustafson said Medicare processed nearly a billion claims, most never seen by human eyes, Dr. Fitzmaurice noted they'd seen examples of codes lumped or split and services, devices, or supplies added or subtracted from the definition of code retirement. He said he inferred from what Dr. Feinberg said that she could write a compendium history of any code change, and he asked if HCPCS and CPT-4 had considered adding a unique code to the existing code structure so that that one code that never changed referred to one definition and was replaced whenever that meaning changed. Over time, there might be four or five definition changes, and each unique number would indicate what that code definition intended at the time it was submitted or paid or appeared on another record. It could reduce miscoding, coding errors and misinterpretations of codes and was important, not only for payment and research, but for drug alerts, device and product recalls, reverse genomic studies, quality improvement, patient safety, and medical areas.
Dr. Feinberg said CPT felt somewhat constrained by the five-field character structure implemented in probably, well over a million outpatient places that transmit Medicare claims electronically. But, she said the rule at the workgroup level was that when any substantive change was made to an existing code its number was also changed. The version of HCPCS used internally had a stop-and-start date for every code. Deleted codes stayed in the historical data book. The code book was loadable from the Web site, but it was large and buying a printed copy could be easier. Dr. Zubeldia noted that there were no start-and-stop dates and only HCPCS codes from the last three years could be downloaded. Dr. Fitzmaurice said that, until there was a unique number for each definition, people wouldn't know a definition had changed or which was intended on a claim. He asked the code maintainers to consider putting changes in the HIPAA record structure into their coding system. If people wanted to incur the cost of expending fields in their computerized patient and billing records, they could. The cost of not doing it was all the possible errors. People had to make the benefit cost analysis.
Mr. Gustafson commented that the discussion had been interesting and fruitful. CMS was very conscious of the frustrations reflected by the panel and attempted to understand and be receptive to solving them to the extent it could. CMS believed coherent coding policy was a good thing and that the agency should attempt to work more enthusiastically and creatively than perhaps it had in the past with the coordination of CPT versus HCPCS. Undoubtedly there were ambiguities and overlaps in the coding systems that weren't necessary and could be pruned to everyone's advantage. He said CMS would discuss with Mr. Beebe ways they might do that more fruitfully and that he was confidant others could point out circumstances to them. But Mr. Gustafson cautioned the Subcommittee that, while a uniform coding system everybody could use for all purposes was a desirable goal, it was also a substantial achievement to aspire to all at once. As CMS understood it, HIPAA created a situation where electronic transactions had to be communicable across networks between providers and payers in a uniform, understandable fashion so everyone knew what a code was and could respond appropriately, including sending it back, saying one didn't pay for that code. But Mr. Gustafson said that, by itself, was a comparatively limited achievement. Aspirations he'd heard that day related to the next level. Agreeing on what the coding rules were depended on how granular everyone wanted to make the payment system and how speedily it could be updated to reflect those needs. How far they could go wasn't obvious, but he suggested there was a need to acknowledge that goal, and that it might take them a long while to move toward it.
Dr. Zubeldia noted issues had been identified: a need for coordinated coding guidelines, a way to make temporary codes permanent (e.g. removing the temporary label), addressing the distinction between specificity of codes and enforcement of payment policy. He suggested that, even though they'd heard that Medicaid was adequately represented, the proliferation of local codes by state Medicaids might be a reflection of their feeling that they weren't properly represented. And he noted the Subcommittee needed to look at the structural limitations of the HCPCS codes and whether that was an appropriate technical structure or a straightjacket that prevented further granularity of the codes.
Dr. Cohn reiterated that his biggest concern was the lack of coordination. Rather than solving the entire problem with HCPCS by August, he said he'd personally be satisfied with a report on a plan for coordination with some actual achievements. Dr. Zubeldia remarked that the overall framework needed to be the HIPAA structure where the HCPCS codes were no longer private and proprietary to a payer or a series of islands that constituted a multitude of payers with their own local codes. The HIPAA requirement that there be no more local codes encompassed the entire healthcare industry. Mr. Gustafson said CMS would look at these issues. He said he would confer with his colleagues and report back to the Subcommittee about continuing this discussion at the August or following meeting.
Follow-up Report from Alternative Link and AMA
Mr. Beebe explained that at the Subcommittee's April 10 meeting, several testifiers demonstrated gaps in the nationally designated code sets. Among them were codes to describe services provided by alternative therapy providers. The Subcommittee had recommended that AMA and Alternative Link work together through the CPT editorial process to develop codes to facilitate the reporting of alternative therapy services under HIPAA.
The CPT code set was maintained by AMA through the CPT editorial panel. Mr. Beebe said the panel's process was well-established, publicized and documented. Anyone could submit proposals for a new or revised CPT code. Instructions and an application were available on the AMA CPT Web site (www.ama-assn.org/cpt) along with a schedule of meetings and deadlines. A well-established CPT editorial process had to be followed. Requiring everyone to follow the same process, AMA and the panel could consistently apply the same criteria to code proposals; maintain clinical accuracy, relevance and consistency with generally accepted healthcare practices; and develop uniform codes that were internally consistent, mutually exclusive, and functionally integrated into the overall CPT code structure and conventions.
The CPT editorial process was an essential component of the development of the CPT code set and couldn't be bypassed or shortcut. Mr. Beebe said circumventing the process would lead to internal inconsistencies, structural defects and inferior descriptions and wouldn't be fair to others. Adherence to the editorial process was an important part of CPT's integrity as a HIPAA-designated code set, as well as a definitive characteristic of CPT codes.
A 112 member advisory committee supported the CPT editorial panel. National medical specialty societies and the Health Care Professionals Advisory Committee (HCPAC), AMA's non-physician healthcare professional group, nominated committee members. Due to these organizations efforts, CPT contained creative descriptions of services and procedures performed by non-physician healthcare professionals, including some services that might be considered alternative therapy services, including therapeutic massage, acupuncture and chiropractic manipulation.
The AMA and the editorial panel recognized that CPT's position as a designated national HIPAA code set might require the development of additional codes to describe other services provided by alternative therapy providers. It might also be necessary to expand the membership of the HCPCS to include professional organizations representing additional alternative therapy providers. Mr. Beebe said AMA was committed to making these changes, but any activity designed to expand or change the CPT code set must follow the CPT editorial process.
Mr. Beebe said the editorial panel was working on a process to include more codes for alternative therapy services. Based on its experiences with the HIEC and home infusion coding, the editorial panel had adopted a workgroup approach to the evaluation and development of descriptions for alternative therapy services for possible inclusion in CPT. The chair of the editorial panel appointed an Alternative Therapy Workgroup consisting of national organizations representing providers of alternative therapy, non-physician healthcare professionals from HCPAC, and members of the panel and the payer community as well as Alternative Link. The workgroup was charged with evaluating the reporting needs and existing claims requirements of alternative therapy payers and providers; development of a framework, including criteria, for review of descriptions of alternative therapy services by the CPT/HCPAC Advisory Committee and the panel; the selection of discreet fields of alternative therapy to begin development of code proposals; and development of CPT code change proposals for alternative therapy. By bringing relevant stakeholders together under its leadership, the panel could resolve issues and problems more quickly and with better results than by following a more independent code development process. Using a multi-disciplinary workgroup also allowed the panel and advisory committee to have access to the most current information on alternative therapy and served to internally scrutinize coding proposals.
The AMA CPT staff had begun to assemble the membership of the Alternative Therapy Workgroup and was working on scheduling meetings. Mr. Beebe said AMA made serious, deliberate attempts to work with Alternative Link. Since the April meeting, they'd met once, held a lengthy conference call, and exchanged letters and e-mails. AMA had expressed willingness to collaborate through the CPT editorial process to develop code change proposals and had invited the company to participate in the Alternative Therapy Workgroup. Mr. Beebe said the meetings had increased an understanding of their positions, but hadn't produced agreement on how to develop CPT codes for alternative therapy.
Mr. Beebe noted that the Subcommittee had recommended collaboration through the CPT process to develop codes to facilitate reporting of alternative therapy services under HIPAA, but hadn't asked AMA and Alternative Link to explore options for incorporating ABC codes into CPT. He emphasized the difference between creating codes for a service and importing codes for that service developed by another entity was significant because the editorial panel process guaranteed a critical level of consistency in procedure coding, code application and clinical decision making as to what constituted appropriate descriptions of healthcare services. These functions would be eroded by the wholesale inclusion of another group of codes.
Alternative Link had argued that the editorial panel, by developing codes for alternative therapies, was appropriating Alternative Link's intellectual property without remuneration. Mr. Beebe contended that wasn't the case. AMA didn't intend to utilize Alternative Link's pre-existing codes, but to create new codes that met CPT editorial panel standards and were an integral part of a coherent work of coding.
Alternative Link had developed a coding work and database for alternative therapy that linked providers with state licensing requirements. However, Mr. Beebe pointed out that Alternative Link wasn't a professional organization and didn't represent any element of the alternative therapy provider community. Alternative Link, not providers of alternative therapy services, had voiced concern that their services wouldn‘t be recognized under HIPAA and that their business needs wouldn't be met. Mr. Beebe suggested that the company might have developed a product that was jeopardized by the HIPAA transaction standards.
Once the Alternative Therapy Workgroup established a functional framework and alternative therapy codes were developed by AMA through the editorial panel process, AMA offered to collaborate on development of a crosswalk from CPT alternative therapy codes to alternative therapy codes in Alternative Link's ABC coding system. Such a crosswalk would enable Alternative Link to utilize its databases to link services with providers and state licensing information, thereby providing the claims processing functions that were an important part of the business need of Alternative Link.
Mr. Beebe said AMA took its responsibility to maintain CPT very seriously. AMA and the editorial panel recognized CPT's designation as a national HIPAA code set for electronic healthcare imposed an obligation to be responsive to the needs of healthcare payers and providers. However, Mr. Beebe requested that those seeking changes to CPT be required to follow the established process. CPT might have been selected as a HIPAA-designated standard due to its general acceptance, but AMA believed it was generally accepted due to respect for the CPT panel, its editorial process, and the substance of the CPT code work.
Follow-up Report from Alternative Link and AMA
Ms. Gianinni said Alternative Link's understanding from the April hearings and subsequent meeting with AMA and NCVHS and CMS staff was that AMA was advised to review the functionality and value of the 4,000 market-ready ABC codes. Depending on the significance of these codes to public health, health insurance business needs, and congressional intent, AMA was to come to an agreement about how to best make use of the codes and establish an appropriate relationship, including an acceptable approach to any relevant intellectual property issues. Based on that agreement, AMA was to collaborate to integrate the functionality of ABC and CPT codes in a way that supported public health, health industry business needs, and congressional intent under HIPAA.
Ms. Gianinni said that during those interactions facilitated by a CMS staff member, AMA was unreliable, self-serving and predatory in its approach to the collaboration. She said AMA's newly proposed alternative therapy CPT coding initiative established AMA as a direct competitor, rather than collaborator. And she said the role AMA proposed for itself would position it to: delay code set availability and keep alternative medicine, nursing and other non-allopathic physician practitioners disenfranchised; manipulate code assignment to make head-to-head scientific comparisons of conventional physician and non-conventional, non-physician interventions difficult; reverse engineer Alternative Link's intellectual property; and enhance the political strength of allopathic physicians by dominating a critical element of the healthcare infrastructure. Ms. Gianinni said AMA's stance was particularly disturbing because the DSMO and Subcommittee assured Alternative Link that AMA had been counseled to be cooperative, transparent, impartial and socially responsible (duties of a designated standard) and that HHS believed AMA would be more collaborative than it had in the past.
Ms. Gianinni said AMA had exhibited aggressive, competitive behavior: withholding a non-disclosure agreement, proposing a crosswalk from ABC codes to non-existent CPT codes, and asking Alternative Link to refocus its business and claims edits scrub application. She said AMA showed virtually no interest in examining the ABC codes. AMA cancelled two meetings and announced it would establish an alternative therapies workgroup and develop competitive CPT codes with or without Alternative Link's involvement.
Ms. Gianinni noted that in a presentation to the AMA Board of Trustees on December 4, 2000, the Chair stated, “You might have never realized the huge investment physicians have in coding, not in terms of dollars, but an investment in physician autonomy. AMA administration of a physician-driven process, coding, has given us a cohesive voice in healthcare delivery. A voice that makes us stronger and louder than we were as individuals.” Ms. Gianinni said AMA was a politically astute organization. Its members were politically active and strategically placed. The AMA was well represented on the DSMOS, permanently chaired the National Uniform Claims Committee, and held a seat on the NUBC. Conventional allopathic physicians served in HHS and on the NCVHS. Allopathic physicians also held leadership positions in HIAA and other associations and organizations with significant influence on code development.
With this level of advocacy, Ms. Gianinni said AMA maintained control over this critical component of the healthcare infrastructure, despite well known weaknesses in the CPT code set and editorial process. As early as 1993, NCVHS reported that AMA's code development process was flawed and widely viewed as favoring AMA membership. Still, the AMA CPT code set was named a designated standard under HIPAA. Since then, eight codes were included in CPT for alternative medicine. AMA's CPT editorial panel was comprised almost exclusively of conventional medical doctors, as was the CPT advisory committee. The Health Care Professional Advisory Committee and the newly formed Alternative Therapies Workgroup included some other practitioners, but both failed to include over 90 specialty associations, not including over 50 specialties of nursing. Despite the advisory bodies within AMA, final decisions were still made by a decision making body that was almost exclusively allopathic physicians.
Ms. Gianinni said Alternative Link's intent wasn't to vilify AMA, but to convey that, even under the best circumstances, a political organization with an allopathic physician agenda would be hard pressed to divorce itself from self-interest and demonstrate cooperative, transparent, impartial and socially responsible behavior in code development for alternative medicine, nursing and other non-allopathic physician interventions.
Ms. Gianinni described an urgent need for these codes. She noted the Journal of the American Medical Association reported that nearly a quarter of a million individuals died in the U.S. each year as a direct result of hospital and physician errors and negative effects of appropriate conventional treatments. Ms. Gianinni said Americans increasingly turned to alternative medicine, nursing and other non-allopathic physician interventions as a primary source of care. Complementary and alternative medicine had a compounded annual growth rate of 38 percent. More visits were made to them than to primary care practitioners.
Ms. Gianinni cautioned that public health might suffer if AMA was granted coding authority over alternative medicine, nursing and other non-allopathic physician interventions. She described a scenario where AMA was politically motivated in the development of codes for non-conventional medicine and she suggested the codes might be defined with insufficient granularity or that interventions could intentionally be bundled and coded in a way that precluded the demonstration of superior outcomes for alternative medicine, nursing and other non-physician interventions. She emphasized that the functionality of ABC codes was needed immediately to support cost effective, evidence-based and integrative healthcare. And she described the cost of delays because of unnecessary administrative insufficiencies in terms of potential loss of human lives and dollars wasted.
Discussion
Mr. Beebe said the editorial panel had codes for alternative medicine and that he would explain why it didn't have more. He said AMA recognized its obligation to work with providers of alternative therapy services to see what their needs and requirements were under HIPAA, expand the codes CPT currently had, and explore current needs for reporting under HIPAA. Mr. Beebe noted that was what AMA formed the Alternative Therapy Workgroup to do.
Synthia Molina, Alternative Link's CEO, said their desire from the beginning was to work with AMA at the highest levels to try and re-think the situation. Alternative Link believed there was a fundamental shift, not only in consumer demand around healthcare but also in code set needs under HIPAA. She noted the ABC codes were developed for the sole purpose of meeting congressional intent under HIPAA. They had a hierarchical structure and many features CPT codes lacked. And they were designed to create a minimal burden on health industry stakeholders. They fit into existing fields used for CPT codes and could fit into the information systems of payers, TPAs, and provider organizations. Alternative Link wanted to raise the level of discussion from the CPT editorial process and discuss this as business process re-engineering to assist the industry in getting the code sets it needed.
Ms. Molina said Alternative Link would like to be a designated standard. She emphasized that the need for these codes was immediate. AMA had emphasized how much effort had gone into separating code assignment from payment. Ms. Molina said everyone familiar with coding understood that not assigning codes equated almost directly to no payment. Through non-coding, it was possible to violate anti-trust, as well. Alternative Link sought to ensure that possibility didn't exist. She said Alternative Link admired many things AMA had done. But she said the fact that of the 4,000 codes for complementary and alternative medicine developed over the last half decade, only eight were developed under the CPT editorial process indicated a problem with the process itself.
Dr. Zubeldia reviewed several options discussed in the April meeting. One option was for the ABC codes to be a separate HIPAA standard. They'd noted that a qualifier couldn't be added to the transactions until version 4050, which meant the codes couldn't be used for at least four or five years. They'd also discussed that they couldn't just merge two different coding structures into one without great disruption. AMA and Alternative Link had said they would work together to see what alternative codes could be included in CPT to resolve the immediate needs of the industry and look for a longer-term solution. He said he wasn't sure that they were going in that direction.
Ms. Gianinni said one thing that stalled their conversations was AMA's stance that the CPT editorial process was the only process for creating terminology. Alternative Link doubted that process would be open enough now to adopt the need today for these codes. She suggested the best solution was to put the codes into HCPCS, where there weren't all the problems and intellectual property issues that existed within AMA. Ms. Gianinni noted millions of dollars were invested in this work. Not having its value recognized negated work done over the last seven years within the guidelines of HIPAA, the SDOs and in line with everything HIPAA had sought to accomplish since 1996.
Ms. Gianinni clarified that both code sets had a five-character structure. HCPCS were part alpha, part numeric. ABC codes were all alpha. She said there probably were a few crosswalks with HCPCS and if duplication existed it was minimal. The codes didn't overlap in their numerical designation.
Ms. Greenberg asked how the codes would be maintained and how Alternative Link's concern about intellectual property would be addressed. Ms. Gianinni compared the issue to a licensing agreement built with CMS for special use of dental codes maintained by the American Dental Association. Not Alternative Link, but the non-profit organization that grew out of Alternative Link when the ABC codes were disposed of would be the maintenance organization. Alternative Link would take responsibility until that could be accomplished. She confirmed Mr. Blair's understanding that Alternative Link was proposing to make a significant financial sacrifice in order to try to get these code sets out where they could be used.
Dr. Zubeldia asked if all the ABC codes began with a letter that currently wasn't used in HCPCS. Ms. Gianinni said there was repetition; both had a code that began with A. Alternative Link thought that was a database issue that was relatively easily solved. Dr. Zubeldia asked if the alternative medicine codes wouldn't be sprayed through multiple HCPCS categories because of the first letter. Ms. Gianinni replied that they didn't know yet, but thought that bore exploration to see how best to effectively preserve both structures.
Dr. Cohn observed that what HCPCS decided to do with its code set was an issue for the HCPCS national panel, not something the Committee dictated. Ms. Trudel said HCPCS would want to check with its general counsel as to whether this constituted code set maintenance or pushed into new-version territory. They were talking about 11 million codes. Ms. Gianinni said an all alpha code structure offered 11.8 million code possibilities. Expanding out to the possibility of numeric and alpha in a five-character space resulted in 45,000,000 possibilities. Ms. Gianinni contended that with that number of coding possibilities there was a solution for everyone. Ms. Trudel responded that her point was that that was far more than the 4,000 codes in HCPCS by an order of magnitude.
Ms. Molina said, obviously, Alternative Link had an interest in return on the investment investors made and business models to accomplish that. But she said they were also so adamant about the immediate need for these codes for health services, research, and clinical trials that Alternative Link was willing to integrate the functionality of the code set into existing designated standards or be named a designated standard. Ms. Molina said a good follow-on action would convey that serious consideration would be given to the code set functionality, the development and maintenance process, the real need out in the industry, and an opportunity for Alternative Link to bring that information back to the table and suggest specific solutions. Remarking that one thing that put organizations out of business was an inability to recognize a fundamental shift in the marketplace, Ms. Molina said Alternative Link recognized such a shift and that, because of the timing of the ABC code development, brilliant solutions were embedded in the code set that could solve some of the problems discussed earlier. Alternative Link had some challenges in trying to make sure that when HIPAA was fully implemented the industry could function. Enormous areas of healthcare didn't have the codes and, with a mandate that only approved codes could be used, the industry would be at a standstill. Ms. Molina said Alternative Link had part of the solution and would like to be seriously considered.
Mr. Beebe addressed the reason CPT had only eight codes for alternative medicine. The CPT process was predicated on the idea that physicians, non-physicians, and payers came to the panel for services they needed to describe for financial and administrative transactions. CPT had eight codes because the American Chiropractic Association had said they needed five codes for chiropractic manipulation in CPT. The American Medical Acupuncture Association said they needed acupuncture codes and the panel also worked with them to again meet the demands of providers for claims. Mr. Beebe said the fact that CPT only had eight codes indicated to him that, at this point, there wasn't a business need on the part of healthcare providers to submit claims. There might be a sea change and need in the future. When there was, alternative medicine providers would come to the panel and ask for additional codes.
Dr. Zubeldia said he'd heard that Alternative Link wanted to introduce the functionality of the coding structure into HCPCS now, not necessarily the codes themselves. Ms. Molina confirmed Alternative Link's interest, so long as there was a viable business model. Dr. Zubeldia said perhaps HCPCS could look at dedicating one letter and might even change all the codes to introduce new functionality that represented the needs of other providers. Ms. Gianinni noted there was a lot of structural intelligence in this code set that Alternative Link could offer as a solution to the growth problems. Dr. Zubeldia noted that HCPCS's structure today was one letter and four numbers. In order to introduce functionality, all the descriptors might have to be renumbered. Ms. Molina reiterated that if there was a viable business model and Alternative Link could integrate and implement the code set promptly, they'd be interested
Dr. Zubeldia said the Subcommittee needed to hear from HCPCS on this, but not that day. They also had to hear from providers and payers. Mr. Blair questioned that there was any reason they needed to render a judgment or verdict. If the codes were acceptable to HCPCS and they worked out an agreement, they wouldn't have to wait for approval. Dr. Zubeldia noted it would be a major change. Dr. Cohn said the Committee's role had to do with updating HIPAA mandated standards. The question was would this be enough of a sea change to need a proposed rule. Ms. Humphreys agreed with Dr. Cohn's assessment and found Ms. Trudel's concern reasonable. But she said that if there was a major gap and working with a code set maintainer became a Catch-22, and inclusion of all the codes and fields that were important was too big a change for an existing code, then either another standard had to be designated or major expansion of the existing code set had to be acceptable. Dr. Cohn said they should talk with the national panel about coordination in the HCPCS area. At the end of the day, they wanted to be sure there wasn't a lot of redundancy and overlap with whatever they did. Dr. Yasnoff said that this got back to the issue of codes and the representation of reality versus codes and payment of claims. The reality of what was happening had to be represented in order to know what should be paid for. Without a representation of reality, there was no way to evaluate whether there were any positive outcomes from what was occurring.
Mr. Blair said that if Alternative Link worked with HCPCS, there wasn't any need to ask them to wait. Maybe by the next meeting, both could report whether there were any difficulties and if they needed help or assistance. Dr. Zubeldia agreed that both organizations needed to explore a possible solution; then it would be up to the Subcommittee to recommend whether that possible solution was moved under HIPAA. Mr. Blair said HCPCS was able to wind up saying whether it could accommodate part or all of this and whether or not it was a major change. Ms. Greenberg said she didn't disagree that if the HCPCS editorial panel and Alternative Link could have productive conversations, they should have them. But she cautioned against the impression that in the eyes of the code set developer and maintainer this represented something that required a rule. If that was up to people who maintained 9-ICM, they'd probably say 10-CM was just an update. She said it wasn't solely the role of the maintainer of a code set to decide whether something was just updating or required regulation under HIPAA. Noting that HCPCS was probably better able than the Committee to determine whether additional codes could fit into their structure and assess the impact, Mr. Blair asked again why HCPCS should be asked to wait.
Dr. Zubeldia proposed that the HCPCS national panel present a report at the next meeting. He said they should also hear providers' and payers' thoughts about a proper home for this coding structure. Ms. Molina encouraged the panel to invite academics and NCAM to the next meeting, so that the need to measure outcomes was represented as well as payment. Dr. Cohn agreed with Dr. Zubeldia. But he said that pending the outcome of whatever discussion or recommendations the national panel had, they couldn't determine the need to hear from the constituency.
Impact of Elimination of Local Codes on Claims Processing for Nursing Services:
Ms. Bickford said ANA was the only full service association representing the nation's registered nurses through its 54 constituent member associations. It's 2.7 million members included registered nurses working and teaching in every healthcare sector across the nation. Nurses formed the largest healthcare workforce in the nation. From nurse midwives attending delivery to geriatric nurses practitioners and from clinical specialists managing end-of-life care to staff nurses tending during times of acute injury or illness, nurses were integral to healthcare across the life span. Always seeking better outcomes, nurses touched patients and managed teams of healthcare professionals in hospitals, clinics, community health centers, offices, nursing homes and patients' homes. Nurses were the ones who most often cared for patients for extended periods and managed the technologies incorporated into their healthcare experiences.
Ms. Bickford said ANA didn't considered the removal of local codes an issue significant to nursing, because nursing hadn't used them, but she noted substantive issues remained that she discussed in light of both days' ongoing discussion in which, she noted, nurses had been partners. Ms. Bickford emphasized that nursing came to healthcare delivery with its own view. Nursing encompassed the prevention of illness; alleviation of suffering; and the protection, promotion, and restoration of health in the care of individuals, families, groups and communities. She noted this was a very different practice model from the existing pathology-based, diagnostic and procedure HIPAA code sets. Nurse's holistic healthcare considered the patient and his or her environment in context. It involved underserved, indigent and vulnerable populations. Nurses looked at problem solving and patient education, demonstrated advocacy for the patient and their informed decision making, and often incorporated complementary and alternative care modalities. Identification and ongoing evaluation of patient outcomes closed the loop. With these considerations, Ms. Bickford said standardized code sets were developed that permitted linkage of the diagnosis or problem, expected outcomes, interventions and actual outcomes for evidence-based practice.
Ms. Bickford reported that nursing languages were incorporated into the UMLS. SNOMED CT First Release included NANDA Taxonomy II and the nursing diagnoses/problems from HHCC, Omaha System, and PNDS. SNOMED CT Second Release, due in July, would include interventions from PNDS. A third release in December would include more interventions from HHCC, Omaha System, and Nursing Interventions Classification. In 2003, SNOMED CT would include outcomes from the HHCC, Omaha System, and the Nursing Outcomes Classification. Ms. Bickford explained that ANA strove to integrate nursing languages and code sets into information systems environments that supported their description of clinical practice. ANA hadn't been concerned about reimbursement, but focused on the patient and describing reality, a point they'd heard discussed often in the last two days. Ms. Bickford noted ANA had been a long-time participant in standards and code set development with ANSI, ASTM, HL7 and at the international level. ANA had been working on ICD-10-CM and had participated in CPT-4 and the CPT-5 discussion for the proposed revision.
Ms. Bickford identified substantive issues confronting the nursing specialties. Codes didn't capture the essence of the care delivered, accurately reflect time spent with a patient, or allow identification and differentiation of practitioners and practices. Codes were inconsistently applied throughout the states and missed all the teaching components, making a significant difference in outcomes. Ms. Bickford noted current codes only reflected disease and that disease and illness weren't the same thing. Prevention, coordination and case management, palliative and transitional care, rehabilitation, pain relief, complementary and alternative modalities were missing in code sets. “Incident to” for Medicare B didn't permit correct recording and tracking of who completed an activity or intervention. High risk, high need patients fell outside the “average” code and CMS couldn't track where money was spent. Innovative interventions and care delivery solutions couldn't be tracked--areas where nurses demonstrated significant influence in reducing costs. Multiple reporting requirements by differing payers required submitting claims as HCPCS, for some, and CPT for others, an undue information management burden. Excessive reporting requirements from multiple entities were extremely burdensome when one designated core data set could suffice, embracing the “write once, read many times” principle.
ANA recommended: implementing ICD-10-CM for diagnoses as soon as possible, reflecting nurses and other non-physician provider's work in codes and systems, including attachments to HL7 messages as an automated solution to tease out clinical details that better inform decision-making about quality and cost, and designating permanent nurse seats on the Committee and CPT editorial panel. Ms. Bickford reiterated that removal of local codes had little impact on nursing because these services didn't exist in the existing code sets.
Discussion
Asked about the use of the ABC codes, Ms. Bickford said in its effort to articulate what nursing did and provide a reporting mechanism, ANA strongly supported development of standardized languages code sets. The ANA recognition program looked for innovations in how nurses might begin to describe their reality. NANDA was the nursing diagnosis entity and there was the home healthcare classification. ANA also recognized the ABC codes as a language that supported nursing because its structure integrated the nursing interventions classifications. ABC codes included alternative therapies, which had become part of nursing practice in many cases. The code set reduced redundancy because a mechanism to identify which provider performed the activity was embedded in the structure: one didn't have to use a nursing intervention code. And she noted the costing piece was critical as the throes of the nursing shortage were compounded by an aging population with increasing healthcare consumption.
Responding to Ms. Bickford's recommendation for a nursing seat on the Committee, Ms. Greenberg clarified that the membership of the Committee (unlike the CPT editorial board which represented particular constituencies) was selected for their expertise in a variety of areas. While the Committee had had some nurses over the 50 years, Ms. Greenberg said she would be the first to admit they'd had a lot more doctors than nurses. But the Committee didn't have slots for different professions, and when anyone served they didn't represent a constituency. Dr. Zubeldia noted that the Subcommittee had a nurse--Ms. Bickford.
Observing that they all struggled with the boundaries delineating HIPAA-mandated code sets and transactions versus clinical terminologies and transactions, Dr. Cohn said the nurses' terminologies, actions, and diagnoses were part of a gray zone between the two. So far, in dealing with the HIPAA transactions, the Subcommittee had focused on things often transported in 837 transactions, which were payment-based claims. As Ms. Bickford commented, there wasn't much here that was part of that, though nurses often were paid specifically for the work done, which was very important. Later in the year, as the Committee moved into terminology discussions, Dr. Cohn said he was sure the Subcommittee would pay more attention to this area and how best to handle it.
Ms. Bickford conveyed the increasing concern of practitioners in nurse-managed call and home health centers that there wasn't a mechanism to reflect their care delivery. Structuring currently in place resulted in alternative care delivery processes that weren't recognized and couldn't be counted or reimbursed. She pointed out that nursing had done significant work in the outcomes arena. She said she didn't know of anyone else working in that arena. She asked the Subcommittee to consider that as they looked at evidence-based practice and informed decision making on resource consumption and allocations and innovative solutions.
Mr. Blair said he anticipated that as they investigated medical terminologies within the domain of recommendations for PMRI standards and code sets that the nursing terminologies would fit in. Dr. Cohn said he didn't know exactly where this fit, but he said it certainly was an area the Subcommittee would look at fairly hard. Although it wasn't within the HIPAA financial and administrative transactions, Mr. Blair suggested it was within the HIPAA PMRI standards. Ms. Bickford said that was why ANA had participated in the standards work all along, trying to assure this was included. Ms. Bebee noted the PMRI report from July 2000 contained a figure depicting healthcare terminology domain areas. The nursing codes were a big part of it.
Ms. Bickford remarked that they'd just heard a significant discussion about incorporation of the alternative therapies into existing code structures. She said they'd have the same difficulty if they did that for nursing, because they, too, already had an existing structure. Similar to the ABC codes, they had their own intellectual work and were in the same predicament. Mr. Blair reflected that Alternative Link had brought up a sense of urgency first tied to the billing issues, but also linked to the need to have the code sets for research. He said his sense of the urgency expressed concerned it happened before they'd gone into a thoughtful, methodical process of evaluating the terminologies for PMRI standards.???
Subcommittee Discussion on Issues Presented During Morning Sessions
Dr. Cohn thanked Dr. Zubeldia for chairing the last day and a half and making progress on a number of difficult issues. Dr. Zubeldia suggested that bringing back yesterday's discussion on the recommendation and moving forward with ICD-10-CM and PCS, for the same uses ICD-9 served now would probably be the best conclusion for the meeting. From what they'd heard that second morning, the recommendation was still to move forward with ICD-10. Dr. Zubeldia clarified they were talking about diagnosis and procedures (ICD-9 procedure codes, which was institutional use). Mr. Beebe said he understood that the Committee was interested in collecting additional data on the impact of PCS in terms of the PCS issue.
Ms. Humphreys summarized what evolved yesterday. The current situation was quite broken in in-patient care, both on the diagnosis and procedure side. The Committee might recommend to the Secretary that they move forward with an NPRM related to ICD-10-CM and ICD-10-PCS, with a two year implementation process once whatever came out of those comments was finalized. One question was whether this was only on the diagnosis or procedure side--or both.
Ms. Trudel recapped the recommendations. ICD-10-CM and ICD-10-PCS were recommended for in-patient only. Implementation would be at the same time. There would be a two-year-plus implementation period after the final rule. The effective date would be no sooner than October 2005. And the effective date would be on-cycle (October or January) when coding updates were currently expected and managed. There was a need for a date certain, a recommendation for corresponding coding guidelines, and acknowledgement that CM and PCS were both broken. Ms. Humphreys noted that since 1993 the Committee had been on record that the existing code sets were broken in terms of the ability to collect data. They'd heard a lot in previous hearings about the procedure codes. Yesterday, they'd heard compelling examples about the problem on the diagnosis side.
Noting they'd heard a lot of people say they didn't want to wait until 2007 or 2008, Ms. Trudel asked if the Subcommittee wanted to put an end date on the recommendation that an effective date would be no sooner than October 2005. Dr. Zubeldia agreed there was a strong message that this shouldn't float forever. While he agreed that they'd heard people throwing out dates, based on hypotheticals, Dr. Cohn contended that the critical piece was having enough time from the effective date for a successful implementation. Dr. Cohn cautioned that it was dangerous to give end dates. They couldn't guarantee when a final rule would come out. He said they'd heard urgency in the last two sessions about ICD-9, vol. 1-3 being broken, but he contended he didn't hear there was an emergency. Ms. Humphreys noted the issue brought up yesterday was that, given these complex and controversial issues, a statement from the Committee to HHS proposing a rule followed by implementation two years later, without a strong statement about replacing these within some specific time frame, might just “put the monkey on somebody else's back.” Meanwhile, it was very difficult to describe what went on in hospitals, both in terms of what the patients suffered from and how everybody was affected in terms of the diagnosis portion. Dr. Cohn said there ought to be a way to wordsmith exactly what Ms. Humphreys said. They didn't want to give people a sense that there wasn't urgency. Ms. Humphreys said they could work on a draft for the Committee to review.
Members clarified that when they spoke of an effective date of October 2005 they meant a mandatory compliance date of October 2005. Most people they'd heard advocated that it wouldn't be possible to implement in an effective way before that date. Noting that they'd heard a lot about implementation planning, Dr. Cohn said he thought they needed to inform the Secretary that, as this moved forward, HHS carried a responsibility to develop a national implementation plan and assist with successful implementation. Members noted they'd heard people recommending getting WEDI-SNIP involved and that they had to include milestones in the implementation plan. A draft letter to the Secretary will be circulated to the Subcommittee and the full committee prior to the next Committee meeting, and time will be provided on the June breakout agenda for final wordsmithing.
Subcommittee Discussion on Additional Carryover Issues
Ms. Pickett asked to move the open issue about what the Subcommittee would recommend about DSM being a crosswalk to the June or subsequent agenda. Based on APA's testimony, additional discussions the Subcommittee had, and what they'd heard yesterday, she reported a number of issues probably had to be rolled into the white paper.
Ms. Pickett noted version control was an open issue raised by presenters at the February and April meetings. The new version of ICD-9-CM became effective annually on October 1; CPT became effective each January 1. But some payers and providers continued to use previous code sets after those dates. All this was a burden to providers who worried about which code set to send. And some state DRG programs used two-to-three year old groupers, which didn't allow for the acceptance of any new codes. Ms. Humphreys said the issue was whether everyone understood that, according to HIPAA, they legislatively still had the problem about October-January. They had to resolve what would happen. She noted that morning's discussions about whether everybody was quarterly, three times or twice a year a year, but not on the same date. As with the local codes, the question was did everybody really understand HIPAA? Did they actually know that, theoretically, their two-year-old grouper was going to be an item of the past when HIPAA was implemented? Ms. Pickett said that was an excellent question, because they didn't have the answer. Dr. Cohn remarked that issues around this (e.g., how long old codes were used versus new codes) had been brought up before. Ms. Humphreys recalled long conversations about the way HIPAA stated that submissions were done with the code set in place at the time the service was delivered. People couldn't renovate all of their health records. Therefore, there was a built-in, natural requirement to be able to process in the old way, because not all claims would be submitted by midnight on the date codes switched over. Ms. Humphreys said she didn't know what statutes of limitation were common in the various segments of this multifarious industry, but either AHA or AHIA could provide details.
Ms. Prophet clarified that today the proper protocol was that the codes were effective on the date service was rendered or, in the case of in-patient discharge, the date of discharge. She said she wasn't aware of Medicare having a different requirement for any particular claim. The problem, as Ms. Pickett had pointed out, was that some payers had grace periods in which one could continue to submit old codes for anywhere from 30-to-90 days after the effective date because either the provider or payer system hadn't been updated to reflect them.
Dr. Zubeldia remarked that when Medicare carriers updated their CPT tables, they had an expiration date on each code indicating when each code ceased to be effective. When they updated the ICD-9 tables October 1, that changed. As of that date, one couldn't file a claim with an old code, because there was no expiration data associated with codes in that table. Ms. Prophet said, technically, that wasn't correct. When the official addenda came out, it clearly announced that this was effective for services as of October 1. Dr. Zubeldia said he was concerned that this wasn't how vendors systems worked. He said it might be an education process. Dr. Cohn asked if the Subcommittee wanted vendors or payers to come in and identify any version control issues and try to straighten them out. Dr. Zubeldia said making it clear up front that this could be a change over what everyone did today might be enough. Dr. Cohn said it sounded like an issue they needed to follow.
Dr. Cohn said the other issue was the standard use of code set guidelines that they'd talked about that morning. Ms. Pickett noted this was an across-the-board issue, with various payers interpreting use of the code sets differently. Even if it was standardized by CMS and use of the alphanumeric HCPCS, others might have a different interpretation and reporting requirement for a given service. Since everyone used HCPCS, the issue loomed large. Ms. Pickett emphasized that was only part of the picture. She reiterated that AHA or AHIA could provide additional information on how, in the absence of definitive guidelines, the burden played out for a provider told to report a service with different combinations of codes for different payers. Ms. Humphreys suggested there were two issues. Some HIPAA code sets didn't have definitive guidelines for coding; others did. But individual payers had their own local procedures that they told people to follow, even though there was a national set of guidelines for diagnosis. And Ms. Humphreys said there might be conflicting guidelines that payers were used to providing. So, it was basically Ms. Greenberg's issue about whether payers actually grasped that, in the post-HIPAA world, the whole intent was to get rid of these local variations.
Dr. Cohn suggested they might need a letter for the September full Committee meeting stating concern about version control, guidelines, the standard use of codes, and a variety of other issues and recommendations the Secretary had to highlight and communicate so the industry was aware of them. This wasn't something the Committee could fix, but people had to know they needed to work on this as they implemented HIPAA. Ms. Guilfoy said the only problem with waiting until September was testing had to begin by April, assuming everybody filed for an extension. That was cutting this slim when systems were being changed to accommodate whatever the letter said.
Ms. Guilfoy also noted that the 837 institutional and professional formats and contents being required for the same service by different payers was a major issue for providers who'd thought they no longer would have to file different versions of the claim. She said getting payer-to-payer coordination of benefits would be impossible if one sent the institutional content to payer A, but payer B wanted the professional content. Payer A wouldn't have the professional content to send. Dr. Cohn said he didn't know that they could deal with that sooner, but it could be a topic for August. Ms. Guilfoy said a WEDI-SNIP workgroup was studying the current status of the issue. Noting they'd heard this issue before, Dr. Cohn said they should hear from WEDI and others about the problem and possible solutions. It would be useful if their letter offered some solutions. Dr. Zubeldia commented that the problem was that payers today used whatever they had; HIPAA didn't require them to change business practices. So, if their system needed NSF for home health, they continued requiring the 837 professional for home health. It was difficult to convince them that home health should be billed with the 837 institutional unless it was mandated. Ms. Guilfoy said part of the real problem was that the data elements for home health were spread across both the institutional and the professional claim. She agreed that it should be with the institutional; the problem was the data elements were in both forms.
Ms. Trudel cautioned that telegraphing that the Subcommittee was looking for a short-term solution by October 2003 was different than saying they realized a need to address this eventually. Giving health plans the impression that they had to use one or the other by October 2003 would complicate many people's implementation plans. Ms. Guilfoy remarked that payers in Minnesota were already used to this kind of processing; Minnesota had had a law since the mid-Nineties defining that institutional providers sent either the UB or EMC and professional providers sent either the 1500 or NSF. Ms. Guilfoy said she was checking if other states had similar legislation. She noted that, from the provider side, having to have the ability to generate two formats wasn't administrative simplification. Dr. Zubeldia responded that it wasn't the Committee's role to design standards. If a flaw or a hole in the design of the standards was identified, the standard's designers could close the hole and provide better instructions in the standards. Dr. Zubeldia said they needed to look at this soon. Ms. Guilfoy suggested they hear from WEDI, the DSMOs, UBC, and UCC.
Subcommittee Discussion of Future Agendas for HIPAA Code Set Issues
Dr. Cohn acknowledged receipt of letters from the National Association of Community Pharmacists and another group regarding having a single or dual formats for retail pharmacies related to professional supplies and services. Dr. Cohn reported this issue had been brought to the group and would probably be part of the discussion about the yearly process they had with the Subcommittee in October. He noted they had to talk with the DSMOs about their timeline and availability and whether there were other issues that they wanted to discuss in August.
Ms. Leon-Chisen suggested that the Subcommittee split the issues of version control and guidelines. Version control was problematic (especially with Medicaid claims where many payers had the current code sets, but the grouper for the Medicaid system might be several years old, as Ms. Pickett had mentioned), but Ms. Leon-Chisen said it could be taken care by alerting people about what HIPAA meant. The guidelines issue was more complex and would probably take longer to work out. The HIPAA code set only mentioned the official code guidelines under ICD-9-CM. They'd heard that day that there was no set of guidelines for HCPCS. She said AHA would work with CMS, AMA and others, providing examples and help in developing guidelines. Ms. Prophet said she'd been about to make the same point. She added that she didn't know what the letter on coding guidelines would say for code sets (e.g., CPT) that had rules, but HIPAA currently didn't require adherence. Dr. Cohn clarified that this issue had already been divided into two: situations without mandated guidelines and another where a guideline was required. But people didn't realize they needed to follow it. Ms. Humphreys suggested that CMS could involve itself more directly with the Medicaid group in terms of the old groupers. Dr. Cohn noted that the Subcommittee had to check with the DSMOs to see if they wanted to begin discussing some of these issues in August or October.
Ms. Trudel said the NPRM was posted on the CMS Web: www.cms.hhs.gov/hipaa. Dr. Cohn noted that, if the Committee sent a letter on this issue, after the thirty-day comment period the Subcommittee would be reviewing NPRMs. He said he anticipated that the letter wouldn't be lengthy and might be generally supportive, mirroring many of their recommendations and aspects they'd heard. He noted that there was just barely enough time to draft a letter for discussion at the June Subcommittee and full Committee meeting. They wouldn't be able to send it back for further revision after the full meeting; but they had that meeting in which to fix it up and get it approved.
Dr. Cohn said at the June meeting they'd have major discussions about the Administrative Simplification Compliance Act and the Subcommittee's responsibilities for reviewing a sample of the compliance delay request. Mr. Augustine was developing an approach and methodology for how they might evaluate it. Hopefully, members would receive a draft before the June meeting. Ms. Trudel said the statistics indicated requests were filed electronically seven days a week. Slightly over a thousand had been filed so far.
Dr. Cohn said the main focus of the August meeting hadn't been fully determined. Noting that some of the discussions around carryover issues were pertinent to HIPAA implementation, Dr. Cohn said he hoped they'd get back to that. He noted they also wanted to gain an understanding from the industry about other issues. If the DSMOs were prepared, they'd begin that discussion in August. In the second half of the year, the Subcommittee hoped to talk about lessons learned so far from the implementation and consider suggestions for how they might learn more and make recommendations to the Secretary on how to improve the process. Ms. Greenberg noted they'd also be reviewing the sample and recommending best practices and ways to overcome problems. And, getting back to PMRI, they'd begin looking at vocabulary standards. Dr. Cohn said he was trying to figure out how all the pieces fit in the puzzle. In June they might be able to plan the fall hearings. Mr. Blair remarked that it would be helpful to know the industry response to any letter HHS sent out concerning incentives for guidance to the industry, based on the Committee's initial recommendations on PMRI message format standards. Dr. Cohn noted the full Committee would meet on June 26-27. The Subcommittee would hold hearings August 28-29, October 22-23, and December 10-11. With that, Dr. Cohn adjourned the meeting.
I hereby certify that, to the best of my knowledge, the foregoing summary of minutes is accurate and complete.
/s/ 1/22/2003
_________________________________________________
Chair Date