This is the accessible text file for GAO report number GAO-05-78 entitled 'Nursing Home Deaths: Arkansas Coroner Referrals Confirm Weaknesses in State and Federal Oversight of Quality of Care' which was released on November 17, 2004. This text file was formatted by the U.S. Government Accountability Office (GAO) to be accessible to users with visual impairments, as part of a longer term project to improve GAO products' accessibility. Every attempt has been made to maintain the structural and data integrity of the original printed product. Accessibility features, such as text descriptions of tables, consecutively numbered footnotes placed at the end of the file, and the text of agency comment letters, are provided but may not exactly duplicate the presentation or format of the printed version. The portable document format (PDF) file is an exact electronic replica of the printed version. We welcome your feedback. Please E-mail your comments regarding the contents or accessibility features of this document to Webmaster@gao.gov. This is a work of the U.S. government and is not subject to copyright protection in the United States. It may be reproduced and distributed in its entirety without further permission from GAO. Because this work may contain copyrighted images or other material, permission from the copyright holder may be necessary if you wish to reproduce this material separately. Report to Congressional Requesters: United States Government Accountability Office: GAO: November 2004: Nursing Home Deaths: Arkansas Coroner Referrals Confirm Weaknesses in State and Federal Oversight of Quality of Care: GAO-05-78: GAO Highlights: Highlights of GAO-05-78, a report to congressional requesters: Why GAO Did This Study: GAO was asked to assess the effectiveness of nursing home oversight by considering the effect of a unique Arkansas law that requires county coroners to investigate all nursing home deaths. Coroners refer cases of suspected neglect to the state survey agency and law enforcement entities such as the state Medicaid Fraud Control Unit (MFCU). The Centers for Medicare & Medicaid Services (CMS) contracts with survey agencies in every state to periodically inspect nursing homes and investigate allegations of poor care or neglect. MFCUs are charged with investigating and prosecuting resident neglect. GAO examined (1) the results of Arkansas coroner investigations, (2) the state survey agency’s experience in investigating coroner referrals, and (3) whether weaknesses in state and federal nursing home oversight identified in prior GAO reports were evident in the survey agency’s investigation of coroner referrals. What GAO Found: According to the Pulaski County coroner, he referred 86 cases of suspected resident neglect to the state survey agency for the period July 1999, when the Arkansas law took effect, through December 2003. Agency officials said that other state coroners referred four cases during this time period. Importantly, these 86 referrals constituted just 2.2 percent of all nursing home deaths the coroner investigated. However, the referrals included disturbing photos and descriptions of the decedents, suggesting serious, avoidable care problems; more than two-thirds of the 86 referrals listed pressure sores as the primary indicator of neglect. Some photos of decedents’ pressure sores depicted skin conditions so deteriorated that bone or ligament was visible, as were signs of infection and dead tissue. The referrals involved 27 homes, over half of which had at least 3 referrals. Arkansas state survey agency officials told GAO that they received 36 (fewer than half) of the Pulaski County coroner’s referrals. The 50 referrals not received described decedents’ conditions similar to those the survey agency did receive. Of the 36 referrals for alleged neglect that it received, the survey agency complaint investigations substantiated 22 and eventually it closed the home with the largest number of referrals. However, the agency’s investigations often understated serious care problems—both when neglect was substantiated and when it was not. For 11 of the 22 substantiated referrals, the state survey agency either cited no deficiency for the decedent or cited a deficiency at a level lower than actual harm for the predominant care problem identified by the coroner. In contrast, MFCU investigations of many of the 11 referrals found the homes negligent in caring for decedents, and the MFCU reached settlements with the owners of several homes. In half of the 14 referrals not substantiated, the MFCU or an independent expert in long-term care either found neglect or questioned the “not substantiated” finding. Moreover, they found gaps and contradictions in the medical records for some decedents, raising a question about the survey agency’s conclusions that the same records indicated appropriate care had been provided. GAO’s prior work on nursing home quality of care found that weaknesses in federal and state oversight nationwide contributed to serious, undetected care problems indicative of resident neglect. GAO’s review of the Arkansas survey agency’s investigations of coroner referrals confirmed that serious, systemic weaknesses remain. Oversight weaknesses GAO previously identified nationwide and those it found in Arkansas included (1) complaint investigations that understated the seriousness of allegations and were not timely; (2) predictable timing of annual state surveys that could enable nursing homes so inclined to cover up deficiencies; (3) survey methodology weaknesses, coupled with surveyor reliance on misleading medical records, that resulted in missed care problems; and (4) a policy that did not always hold homes accountable for neglect associated with a resident’s death. What GAO Recommends: GAO recommends that the CMS Administrator revise CMS’s policy on citing deficiencies to better ensure that nursing homes are held accountable for care problems identified after a resident’s death. CMS concurred with GAO’s recommendations and listed numerous initiatives it plans in response to the report’s findings. www.gao.gov/cgi-bin/getrpt?GAO-05-78. To view the full product, including the scope and methodology, click on the link above. For more information, contact Kathryn G. Allen at (202) 512-7118. [End of section] Contents: Letter: Results in Brief: Background: Coroner Referrals of Suspected Neglect, While Few in Number, Indicated Serious Care Problems: The State Survey Agency's Investigation of Coroner Referrals Often Understated Neglect of Residents: Resident Neglect May Go Undetected Because of Well-Documented Oversight Weaknesses: Conclusions: Recommendations for Executive Action: Agency and State Comments and Our Evaluation: Appendix I: Coroner Referrals for Pressure Sores and the Seriousness of Deficiencies Cited on Standard Surveys: Appendix II: Coroner Referrals That the State Survey Agency Reported as Not Received, Substantiated, or Not Substantiated: Appendix III: Comments from the Centers for Medicare & Medicaid Services: Appendix IV: Comments from the Arkansas Department of Human Services: Related GAO Products: Tables: Table 1: Possible Outcomes of State Survey Agency Complaint Investigations: Table 2: Scope and Severity of Deficiencies Identified during Nursing Home Surveys: Table 3: Description of Pressure Sore Stages: Table 4: Pulaski County Coroner Referrals Received by State Survey Agency and MFCU, July 1999 through December 2003: Table 5: Extent to which the State Survey Agency Cited Serious Deficiencies for Substantiated Referrals from the Pulaski County Coroner: Table 6: Six Coroner Referrals Where the MFCU Found Negligence by the Nursing Home but the State Survey Agency either Cited No Deficiency or a Deficiency at Less than Actual Harm for the Decedent: Figures: Figure 1: Predominant Care Problems Identified in Pulaski County Coroner Referrals to State Survey Agency and the MFCU, July 1999 through December 2003: Figure 2: Number of Pulaski County Coroner Referrals of Suspected Neglect, by Nursing Home, July 1999 through December 2003: Figure 3: Elapsed Working and Calendar Days between Receipt of Coroner's Referral and Start of Investigation by Arkansas State Survey Agency: Abbreviations: CMS: Centers For Medicare & Medicaid Services: MFCU: Medicaid Fraud Control Unit: OSCAR: On-Line Survey, Certification, and Reporting system: United States Government Accountability Office: Washington, DC 20548: November 12, 2004: The Honorable Charles E. Grassley: Chairman: Committee on Finance: United States Senate: The Honorable Christopher S. Bond: United States Senate: An October 2002 series in the St. Louis Post Dispatch concluded that avoidable deaths of vulnerable nursing home residents was a widespread but rarely investigated problem. The series spotlighted an Arkansas law requiring investigations by county officials, such as coroners, of all nursing home deaths.[Footnote 1] Under this law, deaths associated with suspected resident neglect, including poor quality care, are referred to the state survey agency and to law enforcement entities. The Centers for Medicare & Medicaid Services (CMS), the federal agency responsible for managing Medicare and Medicaid, contracts with survey agencies in every state to oversee the quality of nursing home care. In 1998, we reviewed allegations that thousands of California nursing home residents died because of poor care. We found oversight weaknesses that were systemic and not limited to California. Despite federal and state oversight, over half of the decedents in our sample had received unacceptable care that sometimes endangered their health and safety.[Footnote 2] We also found that state surveyors sometimes classified deficiencies at homes where residents had died as less serious than warranted. Our subsequent reports on nursing home quality continued to demonstrate that (1) an unacceptably large proportion of nursing homes--one-fifth as of early 2002--harmed residents and (2) states' periodic inspections of nursing homes failed to identify all serious deficiencies, such as preventable weight loss and pressure sores.[Footnote 3] Our preliminary work on this report found that the 1999 Arkansas law was the only such law nationwide.[Footnote 4] You asked us to consider Arkansas's experience with required coroner investigations to assess the effectiveness of nursing home oversight by the Arkansas state survey agency and by CMS. Specifically, we examined (1) the results of Arkansas coroner investigations of nursing home resident deaths, (2) the experience of the Arkansas state survey agency in investigating suspected cases of resident neglect referred by county coroners, and (3) whether systemic weaknesses in state and federal nursing home oversight identified in our prior reports were evident in the survey agency's investigations of coroner referrals.[Footnote 5] To identify the results of nursing home death investigations by Arkansas's 75 coroners, we asked the Arkansas Office of Long Term Care, the state survey agency, to identify referrals from each county coroner since the law's effective date.[Footnote 6] Because the agency told us that all but four of the referrals were made by the Pulaski County coroner, where the state capital Little Rock is located, we focused on that county's referrals. We obtained and reviewed copies of the coroner's referrals, including the investigative reports, autopsy reports (if one was conducted), and photos of decedents that documented suspected care problems. We interviewed the Pulaski County coroner to determine how reported deaths were investigated, the basis for determining when referrals were warranted, and the process for transmitting referrals to the state survey agency and law enforcement entities. To evaluate state survey agency investigations of coroner referrals of suspected nursing home neglect, we asked the Arkansas survey agency to provide documentation on the results of its investigations. Since the agency treats such referrals as complaints, we reviewed the agency's guidance to surveyors on complaint investigations and discussed the procedures with agency officials. We followed up with agency staff to clarify facts regarding specific investigations of coroner referrals, as needed. To assess the overall quality of care provided at homes with coroner referrals, we obtained data from the survey agency on other complaints against these homes and analyzed data in CMS's On-line Survey, Certification, and Reporting system (OSCAR). CMS officials generally recognize OSCAR data to be reliable, and we judged it to be appropriate for our work. Since the Pulaski County coroner referrals were also sent to the Arkansas Medicaid Fraud Control Unit (MFCU), we obtained copies of its investigative files. MFCUs are charged with investigating and prosecuting Medicaid provider fraud and incidents of patient abuse and neglect. In Arkansas, the MFCU is located within the office of the state attorney general. We compared the results of the state survey agency and MFCU investigations to identify similarities and differences in their findings. For some coroner referrals of suspected resident neglect for which we questioned the state survey agency's decision to not substantiate the existence of serious care problems, we asked a professor of nursing with expertise in long-term care to assess the consistency between the findings from the agency's investigations and the decedents' conditions as documented by the coroner. The expert's assessment was based on a review of the various investigative reports, medical records we obtained, and photos of decedents taken by the coroner. We also discussed our evaluation of the investigations with officials from the Arkansas state survey agency, the MFCU, and CMS. To identify whether systemic weaknesses in state and federal nursing home oversight were evident in the survey agency's investigations of coroner referrals, we reviewed our findings regarding the Arkansas state survey agency's investigations in the context of our prior work on nursing home quality. We conducted our work from August 2003 through October 2004 in accordance with generally accepted government auditing standards. Results in Brief: According to the Pulaski County coroner, he made 86 referrals to the state survey agency of nursing home deaths where neglect was suspected from July 1999, when the Arkansas law took effect, through December 2003. The 86 referrals, constituting 2.2 percent of the approximately 4,000 nursing home deaths the Pulaski County coroner investigated in the 4.5-year period, included disturbing photos and descriptions of the decedents that suggested the existence of serious, avoidable care problems. In over two-thirds of the coroner referrals, pressure sores were the predominant indication of suspected neglect identified during the physical examinations of the decedents, and for some decedents these were at the stage described as life-threatening. For example, the photos of some decedents' pressure sores depicted individuals with skin conditions so deteriorated that bone or ligament was visible, as were signs of infection and dead tissue. The coroner also cited injuries such as falls and broken bones in about 6 percent of the 86 cases. The referrals involved a total of 27 homes, over half of which had three or more referrals during the 4.5-year period. The Arkansas state survey agency informed us it received 36 coroner referrals--fewer than half of those the coroner said he referred--and the MFCU reported it received 51, almost three-fifths. According to the coroner, the referrals were hand delivered to ensure that none were lost and in March 2004, the coroner began requesting signed receipts. Of the 36 referrals that it investigated, the survey agency substantiated 22 and eventually closed the home with the largest number of referrals. However, the survey agency's investigations often understated serious care problems--for both substantiated and unsubstantiated referrals. For 11 of the 22 substantiated referrals, the state survey agency found other care problems but either cited no deficiency or cited a deficiency at a level lower than actual harm for the predominant care problem identified by the coroner. The MFCU's investigations of 6 of these 11 referrals, however, found the nursing homes negligent in providing care, in effect substantiating the existence of serious care problems. Moreover, the MFCU reached settlements with owners of several of the nursing homes. Although we did not examine each of the 14 unsubstantiated referrals in detail, the state survey agency's findings for seven decedents were questioned by the MFCU's investigation, which identified neglect, or by our expert consultant, who questioned the basis for the not-substantiated finding. Examples of neglect they identified included the development of avoidable pressure sores and the lack of a treatment plan. The MFCU and our expert consultant also found omissions and contradictions in the medical records for 4 of the 14 referrals, raising a question about the state survey agency's conclusions that the same records indicated appropriate care had been provided. We found the same serious, systemic survey and oversight weaknesses in the Arkansas state survey agency's investigation of coroner referrals that our prior work on nursing home quality of care identified nationwide. These weaknesses included (1) understatement of the seriousness of complaints and a failure to investigate serious complaints promptly; (2) predictable timing of state surveys, which could enable a home so inclined to cover up deficiencies; (3) survey methodology weaknesses that result in overlooked care problems; and (4) not holding nursing homes accountable for neglect associated with a resident's death. CMS discourages surveyors from citing a deficiency for a care problem involving a deceased resident unless the problem was so serious that it contributed to or caused a resident's death or unless the same problem can be identified for individuals still residing at the nursing home. If a similar problem is not identified during a complaint investigation that assesses care provided to current residents, it is assumed to have been recognized by the home and corrected. However, our prior work demonstrated, and our work in Arkansas confirmed, that (1) nursing home records can contain misleading information or omit important data, making it difficult for surveyors to identify care deficiencies during their on-site reviews; and (2) states' surveys of nursing homes do not identify all serious deficiencies, such as preventable weight loss and pressure sores. Given the results of our prior work, we believe that the serious, undetected care problems identified by the Pulaski County coroner are likely a national problem not limited to Arkansas. We are recommending that the Administrator of CMS revise CMS's policy on citing deficiencies to better ensure that nursing homes are held accountable for care problems identified after a nursing home resident's death. CMS concurred with our recommendations to revise its policy on citing deficiencies for past noncompliance and also identified more than a dozen additional initiatives it plans to take to address shortcomings in the nursing home survey process. CMS commented that the focus of its initiatives, such as additional guidance on the scope and severity of deficiencies, would be broad, a recognition that the shortcomings we identified were systemic and not limited to Arkansas. Both CMS and the state survey agency raised concerns about the discrepancy we reported between the number of referrals the coroner said he made (86) and the number the survey agency said it received (36). In addition, the state survey agency commented that we had understated the number of investigations it actually conducted. We revised the report to address these concerns. In oral comments, the Pulaski County coroner indicated that he believes the law has had a significant, positive impact on the quality of care provided to nursing home residents in Pulaski County. The MFCU did not provide comments. We incorporated technical comments from CMS, the state survey agency, and the Pulaski County coroner, as appropriate. Background: Combined Medicare and Medicaid payments to nursing homes for care provided to vulnerable elderly and disabled beneficiaries totaled about $64 billion in 2002, with total federal payments of approximately $45.5 billion. Oversight of nursing home quality is a shared federal-state responsibility. On the basis of statutory requirements, CMS defines standards that nursing homes must meet to participate in the Medicare and Medicaid programs, and contracts with states to assess, through annual surveys and complaint investigations, whether homes meet these standards. CMS is also responsible for monitoring the adequacy of state survey activities. Arkansas's unique 1999 law requires investigations by county officials, such as coroners, of nursing home residents' deaths and referral of any cases of suspected neglect to the state survey agency and the MFCU. Standard Surveys: Every nursing home receiving Medicare or Medicaid payments must undergo an unannounced standard survey not less than once every 15 months, and the statewide average interval for these surveys must not exceed 12 months.[Footnote 7] A standard survey entails a team of state surveyors, including registered nurses, spending several days in the nursing home to assess compliance with federal long-term care facility requirements, particularly whether care and services provided meet the assessed needs of the residents and whether the home is providing adequate quality of care, such as preventing avoidable pressure sores, weight loss, or accidents. State surveyors assess the quality of care provided to a sample of residents during the standard survey, which is the basis for evaluating nursing homes' compliance with federal requirements. CMS establishes specific investigative protocols for state surveyors to use in conducting these comprehensive surveys. These procedural instructions are intended to make the on-site surveys thorough and consistent across states. When a deficiency is identified, the nursing home is required to prepare a plan of correction that must be approved by the state survey agency. Our earlier work indicated that facilities could mask certain deficiencies, such as routinely having too few staff to care for residents, if they could predict the survey timing; CMS therefore directed states, effective in 1999, to (1) avoid scheduling a home's survey for the same month of the year as the home's previous standard survey and (2) begin at least 10 percent of standard surveys outside the normal workday (either on weekends, early in the morning, or late in the evening). Complaint Investigations: Complaint investigations provide an opportunity for state surveyors to intervene promptly if quality-of-care problems arise between standard surveys. A nursing home resident, family member, friend, nursing home employee, or others may file complaints. CMS requires the investigation of complaints that represent immediate jeopardy to resident health and safety within 2 working days and considers such complaints to be those where one or more of the conditions alleged in the complaint, if true, may have caused or is likely to cause serious injury, harm, impairment, or death to a resident. Beginning in 1999, CMS required investigation of complaints that allege harm to a resident (but which do not rise to the level of immediate jeopardy) within 10 working days, but did not provide detailed guidance to the states about what constitutes harm until November 2003. In November 2003 guidance, CMS generally defined two categories of complaints representing harm: (1) those that, if true, would impair the resident's mental, physical, and/or psychosocial status, which must be investigated within 10 working days, and (2) those that would not significantly impair the resident's mental, physical, and/or psychosocial status, which must be investigated within 45 calendar days. Other complaints that do not rise to the level of either immediate jeopardy or harm do not have to be investigated until the home's next survey, or in some cases, not at all if the state survey agency can determine with certainty that no investigation, analysis, or action is necessary. The requirements identified in the November 2003 guidance became effective on January 1, 2004. Generally, nurse surveyors investigate complaints onsite at the nursing home by reviewing medical records and interviewing staff and residents. The investigations typically include a sample of residents in addition to the resident who is the subject of the complaint to help determine if the problems are systemic. Depending on the volume of complaints against a particular home, several complaints for different residents may be investigated concurrently. Each complaint may contain one or more allegations that a facility is violating federal quality-of-care standards. For example, a single complaint could allege problems with resident abuse, treatment of pressure sores, and proper feeding and hydration. In the course of complaint investigations, the state survey agency can either substantiate or not substantiate the specific allegations or discover other, unreported violations of federal standards (see table 1). A substantiated complaint, however, does not necessarily mean that the state survey agency found neglect of the resident who was the subject of the complaint but rather may indicate other, unrelated care problems. If the state survey agency finds a current violation of a federal standard during a complaint investigation--even if the violation does not relate to the specific allegations being investigated or the residents who are the subject of the complaint--it is required to cite a deficiency against the home. If a complaint investigation reveals no current violation of federal standards but determines that an egregious violation of federal standards occurred in the past that was not identified during earlier surveys, a deficiency known as past noncompliance should be cited and a civil monetary penalty imposed. CMS does not define egregious but indicates that it includes noncompliance related to a resident's death. Table 1: Possible Outcomes of State Survey Agency Complaint Investigations: Complaint outcome: Substantiated: Deficiency; Basis of outcome: The investigation revealed a current violation of federal standards and resulted in the citation of one or more deficiencies.[A] The deficient practice had not been identified and corrected by the home prior to the investigation. Complaint outcome: Substantiated: Past noncompliance; Basis of outcome: If the investigation revealed a past egregious violation of federal standards, such as causing the death of a resident, but identified no current violation, the home should be cited for past noncompliance and assessed a civil monetary penalty.[B]. Complaint outcome: Substantiated: No deficiency; Basis of outcome: The investigation revealed a nonegregious past violation of federal standards but the home had a quality assurance program in place that identified the deficient practice, took appropriate corrective action prior to the investigation, and implemented measures that prevented a recurrence. Complaint outcome: Not substantiated: No deficiency; Basis of outcome: The investigation identified no violation of federal standards. Source: CMS. [A] When a home does not participate in Medicare or Medicaid, the state may cite deficiencies under its state licensing regulations. [B] CMS does not define egregious but notes that it includes situations that caused the death of a resident. [End of table] Deficiency Reporting: Quality-of-care deficiencies identified during either standard surveys or complaint investigations are classified in 1 of 12 categories according to their scope (i.e., the number of residents potentially or actually affected) and their severity. An A-level deficiency is the least serious and is isolated in scope, while an L-level deficiency is the most serious and is considered to be widespread in the nursing home (see table 2). States are required to enter information about surveys and complaint investigations, including the scope and severity of deficiencies identified, in CMS's OSCAR database. Since 1998, such information has been available to the public through CMS's Nursing Home Compare Web site. Table 2: Scope and Severity of Deficiencies Identified during Nursing Home Surveys: Severity: Immediate jeopardy[B]; Scope[A]: Isolated: J; Scope[A]: Pattern: K; Scope[A]: Widespread: L. Severity: Actual harm; Scope[A]: Isolated: G; Scope[A]: Pattern: H; Scope[A]: Widespread: I. Severity: Potential for more than minimal harm; Scope[A]: Isolated: D; Scope[A]: Pattern: E; Scope[A]: Widespread: F. Severity: Potential for minimal harm[C]; Scope[A]: Isolated: A; Scope[A]: Pattern: B; Scope[A]: Widespread: C. Source: CMS. [A] CMS defines the scope levels as follows: isolated--affecting a single or a very limited number of residents; pattern--affecting more than a very limited number of residents; and widespread--affecting a large portion of or all residents. [B] Actual or potential for death/serious injury. [C] Nursing home is considered to be in "substantial compliance." [End of table] CMS Oversight: CMS is responsible for overseeing each state survey agency's performance in ensuring nursing homes' compliance with federal standards for quality of care. Its primary oversight tools are statutorily required federal monitoring surveys conducted annually in at least 5 percent of Medicare and Medicaid nursing homes surveyed by each state, on-site annual state performance reviews instituted during fiscal year 2001, and analysis of periodic oversight reports that have been produced since 2000. Federal monitoring surveys can be either comparative or observational. A comparative survey involves a federal survey team conducting a complete, independent survey of a home within 2 months of the completion of a state's survey in order to compare and contrast the findings. In an observational survey, one or more federal surveyors accompany a state survey team to a nursing home to observe the team's performance. Roughly 81 percent of federal surveys conducted in fiscal year 2003 were observational. State performance reviews, implemented in October 2000, measure state performance against seven standards, including statutory requirements on survey frequency, requirements for documenting deficiencies, timeliness of complaint investigations, and timely and accurate entry of deficiencies into OSCAR. These reviews replaced state self-reporting of their compliance with federal requirements. In October 2000, CMS also began to produce 19 periodic reports to monitor both state and regional office performance. The reports are based on OSCAR and other CMS databases. Examples of reports that track state activities include pending nursing home terminations (weekly); data entry timeliness (quarterly); tallies of state surveys that find homes deficiency-free (semiannually); and analyses, by state, of the most frequently cited deficiencies (annually). These reports, in a standard format, enable comparisons within and across states and regions and are intended to help identify problems and the need for intervention. Certain reports--such as the timeliness of state survey activities--are used to monitor compliance with state performance standards. The Arkansas Law: In July 1999, Arkansas enacted a law requiring nursing homes to immediately report the deaths of residents to the local coroner, regardless of the cause of death.[Footnote 8] The law included a similar reporting requirement for a hospital when a resident died within 5 days after transferring from a nursing home. Coroners who find reasonable cause to suspect that the death is due to maltreatment are directed to report their findings to the state Department of Human Services and to law enforcement and the appropriate prosecuting attorney.[Footnote 9] The statute leaves the scope of the investigation up to each coroner. Like most states, Arkansas already required unnatural deaths to be reported to the coroner for investigation before enactment of the 1999 law.[Footnote 10] According to a coroner who was instrumental in demonstrating the need for the legislation, nursing home administrators chose to release decedents to funeral homes despite the existing requirement for a coroner investigation of deaths that occurred under suspicious circumstances. From 1994 to 1998, this coroner's office conducted six exhumations of nursing home residents and, after full postmortem examinations, all six were determined to have died unnatural deaths. Two cases were ruled medication errors and four were deaths caused by suffocation. For example, one resident was found to have suffocated while tied to his nursing home bed, but the home never reported the death to the coroner. Coroner Referrals of Suspected Resident Neglect: The Arkansas state survey agency, an entity within the Department of Human Services, and the MFCU, an organization within Arkansas's Office of the Attorney General, receive and investigate coroner referrals. Referrals also may be sent to a local city or county prosecutor. The Arkansas state survey agency treats referrals of suspected neglect of nursing home residents as complaints. As with other complaints, they are prioritized for investigation on the basis of the seriousness of the allegations. Arkansas, like other states, has additional categories with longer investigation time frames (45 days and next survey) for complaints judged to be less serious than immediate jeopardy (2 working days) and actual harm (10 working days). Complaint allegations are entered on an intake form that also includes the source of the complaint and eventually the outcome of the investigation. To document their actions, Arkansas surveyors generally prepare a one-to two-page summary specifically describing how the complaint was investigated and which specific allegations were or were not substantiated. Typically, the individual who filed the complaint is informed about the results of the complaint investigation. The Arkansas state survey agency uses a computerized system to track the status of complaint investigations. In Arkansas, the MFCU's authority to investigate resident abuse and neglect is limited to nursing homes that receive Medicaid reimbursement; therefore, it cannot investigate such allegations in a nursing home that only participates in Medicare or that only accepts private pay patients. Generally, MFCUs have concurrent jurisdiction with local investigative and prosecutorial authorities and can both investigate and prosecute such cases statewide.[Footnote 11] On the basis of an investigation, a MFCU can initiate criminal actions in state court but must first obtain permission from the local prosecutor. In such cases, the focus is not on whether a home is providing appropriate care but rather on whether the MFCU can substantiate in court that an act of neglect occurred. These cases may be settled out of court by a payment to the state's Medicaid program without an admission of guilt. Coroner Referrals of Suspected Neglect, While Few in Number, Indicated Serious Care Problems: Of the approximately 4,000 nursing home deaths investigated by the Pulaski County coroner from July 1999 through December 2003, the coroner informed us that he identified and referred 86 cases (2.2 percent) of suspected resident neglect to the state survey agency and the MFCU.[Footnote 12] Even when measured against the number of complaints filed against nursing homes and abuse and neglect case referrals to the MFCU, the number of coroner referrals was very small. However, the coroner's referrals, many accompanied by photos, often depicted signs of serious, avoidable care problems. According to the Pulaski County coroner, his staff generally arrives at the nursing home or hospital within 15 to 20 minutes after the notification, which is expected to be immediate, of a resident's death.[Footnote 13] Facilities have been instructed not to disturb the resident's body. The initial on-site investigation consists of (1) a physical examination of the body, which is photographed; (2) interviews with the treating physician, staff, and perhaps family members; and (3) a review of the decedent's medical records, including a comparison of doctors' prescriptions and nurses' notes to ensure that medications were properly administered. During the investigation, the coroner's staff looks for several key indicators of whether a decedent may have received poor care, including significant weight loss; dehydration; pressure sores; undocumented injuries, such as bruises or skin tears; and interviews with family members. Many of these care indicators are similar to those examined during the state survey agency's annual inspection of every nursing home. Before releasing the body to a funeral home, the coroner may order a toxicology report or ask the state medical examiner to conduct an autopsy to determine whether care problems, such as a medication error or blood poisoning (sepsis) from infected pressure sores, contributed to the resident's death. Of the 86 residents referred by the coroner to the state survey agency and the MFCU, 14 had autopsies completed. Pressure sores, typically serious and often numerous, were the predominant indication of care problems identified in 67 percent of the coroner's referrals (see fig. 1).[Footnote 14] Pressure sores are caused by unrelieved pressure on the skin that squeezes the tiny blood vessels supplying the skin with nutrients and oxygen, causing the skin and ultimately, underlying tissue to die. Most pressure sores can be prevented with adequate nutrition, sanitation and frequent repositioning of the resident.[Footnote 15] In some of the coroner's photos, bone or ligament was visible, as were signs of infection or dead tissue, indicating a serious stage IV pressure sore (see table 3). Figure 1: Predominant Care Problems Identified in Pulaski County Coroner Referrals to State Survey Agency and the MFCU, July 1999 through December 2003: [See PDF for image] Note: Although the referrals sometimes identified multiple care problems, we attempted to identify the primary cause for each of the coroner's 86 referrals. [A] Skin tears and multiple bruises are serious and painful injuries for older individuals and should not be considered in the same context as cuts and bruises sustained by healthy and younger adults. A skin tear is a traumatic wound occurring principally on the extremities of older adults as a result of friction alone or shearing and friction forces that separate the top layer of skin from the underlying layer or both layers from the underlying structure. A skin tear is a painful but preventable injury. See Sharon Baronski, "Skin Tears: Staying on Guard Against the Enemy of Frail Skin," Nursing 2000, vol. 30, no. 9 (2000). [B] Care problems categorized as "other" included possible medication errors (3 decedents), a catheter problem (1 decedent), a resident with poor oral hygiene (1 decedent), a resident setting himself on fire (1 decedent), a home's failure to resuscitate a resident (1 decedent), a resident choking on food (1 decedent), a home's staff taking actions not approved by a physician (1 decedent), malnourishment (1 decedent), a family telling the coroner of poor care (1 decedent), a resident having difficulty breathing (1 decedent), and a resident suffering from a gangrenous colon (1 decedent). [End of figure] Table 3: Description of Pressure Sore Stages: Stage I; Description: Skin is not broken but is red or discolored and does not return to normal within 30 minutes after pressure is removed. Stage II; Description: The topmost layer of the skin is broken, creating a shallow open sore; there may be drainage. Stage III; Description: The break in the skin extends through the second skin layer into the tissue below the skin. The wound is deeper than in stage II. Stage IV; Description: The tissue breakdown extends into the muscle and can extend as far as the bone. Typically, there is considerable dead tissue and drainage. Stage IV may be life-threatening. Source: University of Washington, Spinal Cord Injury Pamphlet, "Taking Care of Pressure Sores." [End of table] Other indications of care problems identified by the coroner included bruises, abrasions, and skin tears (12 percent) and falls or broken bones (6 percent). For one referral, the bruise covered the decedent's entire upper chest and for another the arm from the elbow to the shoulder. In about 15 percent of referrals, the indications of care problems identified by the coroner were difficult to categorize, such as a decedent with a catheter whose penis was bloody and irritated, a resident who died when he attempted to burn off his restraints with a cigarette lighter, and a resident who was taken to the hospital with breathing problems. An autopsy of the last resident revealed the presence of toxic or excessive levels of drugs that likely caused the respiratory problems and contributed to the development of pneumonia and to death. For some referrals, the coroner found evidence of multiple care problems. For example, a 1999 referral involved a decedent with a 9- square inch pressure sore on her lower back, a gangrenous foot, and ants on her feeding tube and wounds. According to the resident's daughter, the odor in her mother's room at the nursing home was so great that she had to leave. The autopsy attributed the gangrene to arteriosclerosis that restricted the blood supply to her legs but also found that the resident suffocated when dried mucus that had accumulated in her mouth broke off and blocked her breathing passage. According to the MFCU, her wounds and oral care appeared to have been neglected for some time. The 86 cases of suspected resident neglect occurred in 27 nursing homes.[Footnote 16] Although it is difficult to precisely identify the proportion of Pulaski County nursing homes that had referrals because facilities closed and opened during the time period we examined, over half of the 27 homes had three or more referrals (see fig. 2). Fourteen homes accounted for almost 80 percent of the referrals. Some homes had a pattern of referrals spanning several years. For example, one home had seven referrals--one in 1999, two in 2000, two in 2001, and another two in 2002. Three of these seven referrals involved stage IV pressure sores, some of which were blackened with dead tissue, and one referral involved a resident who died because of an overdose of drugs administered by the nursing home. Nineteen of the 27 nursing homes were referred by the Pulaski County coroner, many of them more than once, because the deceased residents had pressure sores (see app. I). Eleven of the 12 referrals for one home involved pressure sores.[Footnote 17] The standard surveys of these homes, however, infrequently raised concerns about the care provided to prevent and treat pressure sores. As of November 2003, 15 of the 19 homes had not been cited on any of the previous four standard surveys for a pressure sore deficiency at the actual harm level or higher, while 3 homes each had one such deficiency.[Footnote 18] Figure 2: Number of Pulaski County Coroner Referrals of Suspected Neglect, by Nursing Home, July 1999 through December 2003: [See PDF for image] [End of figure] The State Survey Agency's Investigation of Coroner Referrals Often Understated Neglect of Residents: According to Arkansas state survey agency officials, the agency received 36 coroner referrals of suspected resident neglect, less than half of the 86 referrals the coroner said he made. The agency's investigations of these coroner referrals often understated serious care problems--both when neglect was substantiated and not substantiated (see app. II). Even in the majority of substantiated referrals, the state survey agency failed to cite serious deficiencies involving care problems for the decedents who were the subject of the referrals, in effect not confirming the predominant care problems identified by the coroner. The MFCU's investigations of many of these same referrals, however, frequently found that facilities had been negligent in caring for the decedents by identifying serious lapses in care. In half of the referrals not substantiated by the state survey agency, either the MFCU investigation found neglect or we questioned the basis for the "not substantiated" findings, and our concerns were confirmed by a professor of nursing with expertise in long-term care. Moreover, the MFCU found inconsistencies in the medical records for some decedents, raising a question about the state survey agency's conclusion that the same records indicated care had been provided. Fewer than Half of the Coroner Referrals Were Received by the State Survey Agency: Although the Pulaski County coroner told us that he had referred 86 cases of suspected resident neglect from July 1999 through December 2003, Arkansas state survey agency officials said that they received fewer than half (see table 4) and investigated all but one of the referrals they received.[Footnote 19] MFCU officials, however, indicated that they received almost three-fifths of the 86 referrals.[Footnote 20] The MFCU received all but three of the referrals received by the state survey agency. Overall, 32 coroner referrals were not investigated by either agency.[Footnote 21] Table 4: Pulaski County Coroner Referrals Received by State Survey Agency and MFCU, July 1999 through December 2003: Year of resident's death: 1999[A]; Number of referrals: 20; Received by state survey agency: 4; Received by MFCU: 2. Year of resident's death: 2000; Number of referrals: 24; Received by state survey agency: 17[B]; Received by MFCU: 22. Year of resident's death: 2001; Number of referrals: 23; Received by state survey agency: 11; Received by MFCU: 17. Year of resident's death: 2002; Number of referrals: 18; Received by state survey agency: 3; Received by MFCU: 9. Year of resident's death: 2003[C]; Number of referrals: 1; Received by state survey agency: 1; Received by MFCU: 1. Year of resident's death: Total; Number of referrals: 86; Received by state survey agency: 36[B]; Received by MFCU: 51[D]. Source: Coroner's office, Pulaski County; Arkansas state survey agency; and the MFCU. [A] The Arkansas law became effective in July 1999 and the state survey agency received its first referral on September 27, 1999. [B] Although the state survey agency lacked routine documentation describing its investigation of two coroner referrals, we included these referrals in our analysis because agency officials were able to tell us the outcome of the investigations. However, we excluded three other coroner referrals that survey agency officials told us they had received but for which they could neither document their investigations nor tell us the outcomes. [C] The coroner eventually referred six 2003 resident deaths to the state survey agency and the MFCU. We excluded five of the six because they were not actually referred until early 2004. [D] The MFCU received all but 3 of the 36 referrals received by the state survey agency. [End of table] According to the coroner, all the referrals were hand delivered rather than mailed to ensure that none were lost, but officials at the state survey agency and the MFCU told us that they did not know how referrals were delivered.[Footnote 22] We found inconsistencies in agency and MFCU recordkeeping. For example, the state survey agency told us that it had received five referrals on the coroner's list but could not provide a copy of any complaint intake forms for them or the results of its investigations for three of the five referrals. While a MFCU official told us that three other referrals were forwarded to it by the state survey agency, not the coroner, the state survey agency had no record of these referrals. The 50 coroner referrals not received by the state survey agency were similar to those received. For example, one decedent had large, unexplained bruises on her chest, upper right arm, and back, including a mass of more than nine square inches that likely consisted of clotted blood from a broken blood vessel. A second decedent had five pressure sores--lower leg, heel, lower back, and both hips; according to the coroner's report, one of the pressure sores was "draining a dark- colored, pus-filled, and foul-smelling fluid." The decedent's medical records indicated admission to the nursing home 6 months before death without any pressure sores. A third decedent had 10 pressure sores with dead tissue on one heel. A fourth decedent had a large tear on the upper arm, a pressure sore on one foot with dead tissue extending to mid-calf, and a stage IV pressure sore on one buttock. Three coroner referrals not received by the state survey agency but investigated by the MFCU found negligent care that resulted in settlements and payments by the facilities. Serious Deficiencies Seldom Cited for Care Problems Involving Decedents, Even Though Referrals Were Often Substantiated: With the exception of one home, we found that state survey agency complaint investigations of coroner referrals often failed to cite serious deficiencies for the decedents being investigated, even though over half of the referrals investigated were substantiated. Overall, the state survey agency substantiated 22 of the 36 coroner referrals it investigated at 12 nursing homes.[Footnote 23] However, the state survey agency cited actual harm or higher-level deficiencies in quality of care, abuse/neglect, or both for only 11 of these 22 substantiated referrals (see table 5). Table 5: Extent to which the State Survey Agency Cited Serious Deficiencies for Substantiated Referrals from the Pulaski County Coroner: Nursing home: A; Number of referrals substantiated: 7; Deficiencies cited for coroner referred decedents at actual harm or higher level in quality of care and/or abuse/neglect: Deficiency cited: 6 decedents; Deficiencies cited for coroner referred decedents at actual harm or higher level in quality of care and/or abuse/neglect: No deficiency cited: 1 decedent. Nursing home: C; Number of referrals substantiated: 2; Deficiencies cited for coroner referred decedents at actual harm or higher level in quality of care and/or abuse/neglect: No deficiency cited: 2 decedents. Nursing home: E; Number of referrals substantiated: 2[A]; Deficiencies cited for coroner referred decedents at actual harm or higher level in quality of care and/or abuse/neglect: Deficiency cited: 1 decedent; Deficiencies cited for coroner referred decedents at actual harm or higher level in quality of care and/or abuse/neglect: No deficiency cited: 1 decedent. Nursing home: I; Number of referrals substantiated: 2[A]; Deficiencies cited for coroner referred decedents at actual harm or higher level in quality of care and/or abuse/neglect: Deficiency cited: 1 decedent; Deficiencies cited for coroner referred decedents at actual harm or higher level in quality of care and/or abuse/neglect: No deficiency cited: 1 decedent. Nursing home: L; Number of referrals substantiated: 2; Deficiencies cited for coroner referred decedents at actual harm or higher level in quality of care and/or abuse/neglect: No deficiency cited: 2 decedents. Nursing home: B; Number of referrals substantiated: 1; Deficiencies cited for coroner referred decedents at actual harm or higher level in quality of care and/or abuse/neglect: No deficiency cited: 1 decedent. Nursing home: D; Number of referrals substantiated: 1[A]; Deficiencies cited for coroner referred decedents at actual harm or higher level in quality of care and/or abuse/neglect: No deficiency cited: 1 decedent. Nursing home: N; Number of referrals substantiated: 1; Deficiencies cited for coroner referred decedents at actual harm or higher level in quality of care and/or abuse/neglect: Deficiency cited: 1 decedent. Nursing home: Q; Number of referrals substantiated: 1; Deficiencies cited for coroner referred decedents at actual harm or higher level in quality of care and/or abuse/neglect: Deficiency cited: 1 decedent[B]. Nursing home: T; Number of referrals substantiated: 1; Deficiencies cited for coroner referred decedents at actual harm or higher level in quality of care and/or abuse/neglect: No deficiency cited: 1 decedent. Nursing home: X; Number of referrals substantiated: 1; Deficiencies cited for coroner referred decedents at actual harm or higher level in quality of care and/or abuse/neglect: No deficiency cited: 1 decedent. Nursing home: AA; Number of referrals substantiated: 1; Deficiencies cited for coroner referred decedents at actual harm or higher level in quality of care and/or abuse/neglect: Deficiency cited: 1 decedent. Nursing home: Total; Number of referrals substantiated: 22; Deficiencies cited for coroner referred decedents at actual harm or higher level in quality of care and/or abuse/neglect: Deficiency cited: 11; Deficiencies cited for coroner referred decedents at actual harm or higher level in quality of care and/or abuse/neglect: No deficiency cited: 11. Source: Arkansas state survey agency complaint investigation reports. Note: Of the 22 substantiated referrals for residents who died at these homes, 18 were referred for pressure sores, two for bruising, one for a fall, and one for catheter problems. [A] One referral was substantiated without any deficiencies. Even though the investigation revealed a past violation of federal standards, no deficiencies were cited because the home had a quality assurance program in place that identified the deficient practice, took appropriate corrective action prior to the investigation, and implemented measures that prevented a recurrence. [B] Past noncompliance was cited for pressure sores at the immediate jeopardy level. Past noncompliance may be cited when no current violation of federal standards is found but the past violation was so egregious that the home should be cited for a deficiency and a civil monetary penalty imposed. [End of table] Nursing home A accounted for 6 of 11 citations for neglect of decedents at the actual harm or higher level (see table 5). The neglect involved inadequate care to prevent and treat pressure sores. The home was terminated from participation in Medicare and Medicaid in November 2000, about 5 months after the first of a series of state survey agency complaint investigations initiated as a result of coroner referrals.[Footnote 24] Although the agency found that six of the coroner-referred decedents had been neglected by home A, the results of this home's March 2000 standard survey and the timing and results of some complaint investigations prior to its closure were inconsistent with those findings. We identified the following inconsistencies in surveys of this home: * The home's March 3, 2000, standard survey found no deficiencies other than a C-level deficiency (potential for minimal harm) for inadequate housekeeping and maintenance, including a water-damaged ceiling tile, soiled carpeting, and worn upholstery on a sofa. The survey's resident sample, however, included a resident who died in mid-April, less than 6 weeks after the standard survey, with five stage IV pressure sores. * Even though the photos accompanying coroner referrals for four decedents suggested serious, systemic care problems, the state survey agency did not initiate a complaint investigation until May 16, 2000, about 3 weeks after receiving the referrals, which were all sent at the same time.[Footnote 25] CMS guidance requires that such complaints be investigated within 2 to 10 days, but state survey agency officials told us that they often gave a higher priority to investigating serious complaints for living residents. The state survey agency cited actual harm deficiencies for quality of care for three of the four decedents because similar care problems were found for current residents at the facility. * The May 16 investigation, however, included March 27 and April 3 complaints from family members of one resident alleging that he (1) had deteriorating, unbandaged pressure sores and (2) was left wet and soiled for long periods, a situation that could have contributed to worsening pressure sores.[Footnote 26] These allegations went uninvestigated for almost 2 months until they were confirmed in May. Investigation of a subsequent July complaint for this resident documented further deterioration of the pressure sores that began on his buttocks and extended all the way up his back. * Although this same resident was included in the sample of a subsequent September 2000 complaint investigation, his continuing pressure sores were not cited during that investigation. A final complaint investigation at the home about 6 weeks later--following the resident's death--found that he had 28 pressure sores when he died; 7 of the pressures sores, 2 of which were stage IV, did not have a physician's order for treatment. Only five of the referrals for decedents at other homes resulted in the citation of a deficiency at the actual harm or higher level for the decedent in question (see table 5). The deficiencies cited involved quality of care or abuse/neglect for four of the five decedents. For one of the five decedents, who had numerous, serious pressure sores, no current violations of federal standards were identified during the investigation of the coroner's referral. Under CMS guidance, surveyors would need to identify a current resident with inadequate treatment to prevent and heal pressure sores in order to cite a pressure sore deficiency at the actual harm level. However, the surveyor determined that an egregious past violation of federal standards involving this decedent warranted citing a deficiency known as past noncompliance and imposition of a civil monetary penalty.[Footnote 27] Because the deficiency occurred in the past and was assumed to have been corrected by the facility, a plan of correction was not required and no deficiency could be cited for the underlying care issue--inadequate treatment to prevent and heal pressure sores.[Footnote 28] Although Arkansas state survey agency officials told us that they frequently cite past noncompliance, we found that it was cited for only one coroner referral.[Footnote 29] For the remaining 11 substantiated coroner referrals, the state survey agency cited either no deficiency for the decedent or cited a deficiency at a level lower than actual harm for the predominant care problem identified by the coroner, even though the MFCU's investigations found neglect for six of the decedents, in effect substantiating the existence of serious care problems in these cases (see table 6). The MFCU's findings raise a question about the thoroughness of state survey agency complaint surveys. Because the nature of the problems identified by the coroner in these 11 referrals did not appear to differ significantly from referrals for home A that were substantiated at the actual harm or higher level (see table 5), we asked the state survey agency to review the 11 referrals to determine why no serious deficiencies were cited and if past noncompliance should have been cited. Noting their current heavy workload, state survey agency officials agreed to review 2 of the 11 cases. They told us that they could not cite an actual harm pressure sore deficiency for either decedent because the decedents were not in the facility at the time of the complaint investigations and under CMS guidance, surveyors would need to identify a current resident with inadequate treatment to prevent and heal pressure sores in order to cite a pressure sore deficiency at the actual harm level. In one of these cases, however, agency officials told us that they should have cited past noncompliance because of the serious nature of the decedent's condition. Table 6: Six Coroner Referrals Where the MFCU Found Negligence by the Nursing Home but the State Survey Agency either Cited No Deficiency or a Deficiency at Less than Actual Harm for the Decedent: Home: B; Resident: Resident 59; Problems identified by coroner: Numerous pressure sores; ulcers on the roof of decedent's mouth; leaking feeding tube; Results of investigation: State survey agency: No deficiency was cited for this decedent, but a deficiency for pressure sores was cited at the D level for another resident; Results of investigation: MFCU: Negligence found and fraud case is pending. Home: E; Resident: Resident 5; Problems identified by coroner: Numerous pressure sores; dirty unchanged bandages; ulcer on the roof of decedent's mouth; resident and medical equipment covered with live ants; foot and ankle in advanced stages of decomposition; Results of investigation: State survey agency: No deficiency was cited for this decedent; Results of investigation: MFCU: Inadequate care found, leading to a $30,000 settlement agreement with the home.[A]. Home: I; Resident: Resident 40; Problems identified by coroner: Numerous pressure sores; ulcers on the roof of decedent's mouth; Results of investigation: State survey agency: No deficiencies cited for this decedent; Results of investigation: MFCU: "Absence of care" found and fraud case is pending. Home: L; Resident: Resident 25; Problems identified by coroner: Pressure sores and skin discoloration; Results of investigation: State survey agency: Cited the home for a B- level deficiency for this resident due to incomplete records. (It also cited pressure sores at the immediate jeopardy level but not for this decedent); Results of investigation: MFCU: Included among 42 residents of a chain of nursing homes whose care the MFCU found negligent, leading to a $1.5 million settlement with the owners.[A]. Home: L; Resident: Resident 52; Problems identified by coroner: Numerous pressure sores and skin tears; Results of investigation: State survey agency: Cited the home for two B-level deficiencies for this decedent, both related to the home's recordkeeping.[B]; Results of investigation: MFCU: Included among 42 residents of a chain of nursing homes whose care the MFCU found negligent, leading to a $1.5 million settlement with the owners.[A]. Home: T; Resident: Resident 48; Problems identified by coroner: Bruises on face and head, possibly due to falls; family told coroner that the home did not monitor resident properly to avoid falls; Results of investigation: State survey agency: Cited the home for two E-level violations for this decedent--one for improper use of restraints and one for accident prevention; Results of investigation: MFCU: Found evidence of neglect, but MFCU cited insufficient resources as the reason for not pursuing the case. Source: GAO analysis of Pulaski County coroner referrals and Arkansas state survey agency and MFCU investigative reports. [A] As of January 2004, 12 coroner referrals were included in MFCU settlements totaling $1,767,000 with five nursing homes. Some of the settlements, however, involved residents who were not referred by the Pulaski County coroner. For example, the largest settlement for $1.5 million involved 42 residents, 2 of whom were referred by the coroner. [B] The state survey agency noted that this home had been cited for immediate jeopardy for pressure sores during a survey conducted about 5 weeks before this decedent's death. Although the decedent was a resident of the home during the earlier survey, she was not included in the sample of residents reviewed at that time. [End of table] State Survey Agency Decision Not to Substantiate Some Coroner Referrals Was Questionable: On the basis of the MFCU's investigations and our own review, we question the state survey agency's decision not to substantiate more of the coroner's referrals or forward them to another agency for further investigation. Overall, the state survey agency did not substantiate 14 of the 36 coroner referrals that it investigated.[Footnote 30] Although we did not assess each of the 14 unsubstantiated referrals in detail, the state survey agency's findings for 7 decedents were challenged either by the results of the MFCU's investigations or by an expert review conducted at our request. Both the MFCU and our expert noted omissions and contradictions in the medical records of some of the 14 decedents, raising a question about the state survey agency's conclusions that the same records indicated care had been provided. The MFCU's investigations identified neglect of two decedents that the state survey agency failed to substantiate.[Footnote 31] In one of the cases, the MFCU found that the nursing home failed to (1) accurately assess changes in the resident's status, allowing the resident to develop stage II pressure sores before the staff was even aware that he had a skin problem; (2) track the resident's ability to perform certain basic activities of daily living; (3) routinely monitor his weight despite continued weight loss; and (4) follow physician orders, sometimes delaying prescribed treatment. In the other case, the MFCU found that the nursing home failed to provide necessary treatment, rehabilitation, care, food, and medical services. In particular, the resident had no skin breakdown upon admission to the facility. But 7.5 months later, she had six pressure sores, including one on her right hip that was almost 4 inches across and had progressed to stage IV and two others that had progressed to stage III. There was no comprehensive care plan to address the resident's pressure sores. Other care was also found negligent. For example, during a hospital stay about 2 months before the resident's death, the hospital found a large area on the back of her tongue with a thick buildup of saliva that had not been properly cleaned at the nursing home for up to 7 days. For five other coroner referrals not substantiated by the state survey agency, the expert agreed that we had a basis to question the state survey agency's findings.[Footnote 32] For example, the expert found that (1) some facilities were not removing the dead tissue around pressure sores; (2) the color of one decedent's skin suggested it was urine stained, a situation that contributes to skin breakdown and infection; and (3) two decedents were not receiving oral care, the lack of which the expert characterized as "profound" for one decedent. For three of the five cases, the expert found evidence that neglect contributed to the residents' physical condition as documented in the coroner's referrals. In general, the expert found the degree of skin damage and pressures sores in the reviewed cases to be "very suspicious" and concluded that preventive measures, such as special mattresses, would have precluded the development of such severe pressure sores, despite the decedents' health status. The expert also found the scarce and inconsistent mention of pain assessment and management to be suspicious enough to warrant concern about abuse.[Footnote 33] Although three of the five deceased residents were receiving hospice care at the nursing home, our expert questioned the apparent lack of care for these residents. Ideally, hospice care provides consistent pain assessment and intervention, measures to prevent further skin breakdown and the associated discomfort, and local treatment to minimize odor. These standards are inconsistent with not changing pressure sore dressings, even if a family member asks not to have them changed. Finally, our expert questioned if some of the facilities had a quality assurance process in place to identify systemic problems, such as the incidence of pressure sores. We found that the state survey agency had cited the facility where two of the five decedents had resided for immediate jeopardy regarding the federal requirements to maintain a quality assurance committee that meets regularly. This deficiency was cited about 9 months before and 9 months after the residents' deaths. In two of the five cases, the state survey agency had concluded that serious pressure sores were acquired during hospitalizations but did not identify other care problems noted by our expert consultant. For example, one of the nursing homes failed to remove dead tissue around the pressure sores, an indication of poor care. In addition, the expert noted the lack of oral care for one of these decedents, again raising questions about the quality of care provided by the home. Even if the state survey agency had justifiably concluded that the decedents' serious pressure sores were acquired during hospitalizations rather than in the nursing homes where the residents died, neither case was referred to Arkansas's Division of Health Facility Services, the entity responsible for oversight of hospitals that serve Medicare and Medicaid beneficiaries. State survey agency officials agreed that it might have been appropriate to refer such cases to this division. CMS's 1999 guidelines for complaint investigations instruct state survey agencies to refer cases to another agency when it lacks jurisdiction. Omissions and contradictions in the medical records for four other decedents whose referrals were not substantiated raise a question about the state survey agency's conclusions that these same records indicated care had been provided. For example, in two cases, the MFCU found numerous omissions in the facility's care and treatment records, such as missing entries on the medication records and nurse assistant flow sheets, as well as a discrepancy as to when a pressure sore was first noted. In another case, the MFCU concluded that there were so many documentation problems that it was difficult to follow the course of one decedent's care, including late entries that were "questionable and too many." In addition, in another case, our expert consultant found that the seriousness of a pressure sore was understated by the home. Federal surveyors also found evidence that state surveyors missed or failed to cite deficiencies, including some that harmed residents. A March 2000 federal comparative survey of an Arkansas nursing home, some of whose residents were the subject of coroner referrals, found care issues that had not been identified by the state survey agency.[Footnote 34] A comparative survey is conducted within 2 months of a state survey to independently verify its accuracy. Overall, federal surveyors cited 19 health-related deficiencies that state surveyors did not, including failure of the nursing home to develop and implement effective procedures to prevent neglect and abuse of residents. Three of the 19 deficiencies that state surveyors did not identify were cited by federal surveyors at the actual harm level: failure to provide (1) necessary care and services to maintain a resident's highest well being; (2) good nutrition, grooming, and personal and oral hygiene; and (3) treatment and services to increase and prevent further degradation in a resident's range of motion. Federal surveyors also cited a widespread failure in infection control procedures at the potential for more than minimum harm level. One of the coroner-referred deaths at this facility occurred within 6 weeks of both the state and federal surveys that were about 1 month apart. The decedent arrived in the hospital emergency room with a fever of 104°, an indication of infection, as well as ragged tears on his right knee and shin and serious pressure sores on both buttocks. Though documentation was not available, a state survey agency official told us that this complaint was unsubstantiated. Resident Neglect May Go Undetected Because of Well-Documented Oversight Weaknesses: Because of oversight weaknesses that are well-documented nationwide, neglect of nursing home residents may often go undetected. We found the same systemic oversight weaknesses in the Arkansas state survey agency's investigation of coroner referrals that our prior work on nursing home quality of care identified nationwide. These oversight weaknesses include (1) complaint investigations that understated the seriousness of the allegations and were not conducted promptly; (2) annual standard survey schedules that allowed nursing homes to predict when the next survey would occur; (3) survey methodology weaknesses, coupled with surveyor reliance on misleading medical records, that resulted in overlooked care problems; and (4) a policy that did not always hold nursing homes accountable for care problems identified after a resident's death. Serious Complaints Were Inappropriately Prioritized and Not Promptly Investigated: In 1999, we reported that many survey agencies in the 14 states we examined often assigned inappropriately low investigation priorities to complaints and failed to investigate serious complaints promptly.[Footnote 35] Such practices may delay the identification and correction of care problems that may involve other residents of a nursing home in addition to the resident who is the subject of the complaint. Based on our draft report, CMS reviewed the Arkansas state survey agency's prioritization of the 36 coroner referrals the agency said it received. CMS concluded that about 31 percent of the referrals should have been prioritized for more prompt investigation.[Footnote 36] Furthermore, CMS found that 5 referrals prioritized by the state as requiring an investigation within 10 working days suggested the potential for immediate jeopardy and should have been prioritized for investigation within 2 working days.[Footnote 37] The state survey agency prioritized 6 other referrals as not requiring investigation for up to 45 days, but CMS indicated that 1 of these referrals should have been prioritized for investigation within 2 days, and the remaining referrals within 10 working days (actual harm).[Footnote 38] Although the state survey agency classified most of the 36 referrals as requiring investigation within 10 working days, we found a significant disparity between the prioritization it assigned and the time it actually took to conduct the investigations. As shown in figure 3, 16 referrals were investigated in 10 working days or less and 19 referrals took between 11 and 290 working days to investigate.[Footnote 39] Identifying time frames in terms of working days, as CMS's guidance requires, however, understates the actual elapsed time between receipt and investigation of referrals. The average elapsed time from the date the survey agency received a referral until it initiated its investigation was 46 calendar days. Seven referrals were not investigated for between 91 and 425 calendar days and the investigation of an additional 11 referrals took between 21 and 90 calendar days (see fig. 3). State survey agency officials told us that because of surveyor turnover and the number of complaints received from all sources, the agency could not investigate all coroner complaints quickly; CMS has identified untimely complaint investigations in many other states. Moreover, Arkansas state survey agency officials told us that they gave priority to allegations involving residents who were still living in a facility over comparable allegations involving deceased residents, even though the coroner's referrals were accompanied by photos that suggested the possibility of systemic care problems. Figure 3: Elapsed Working and Calendar Days between Receipt of Coroner's Referral and Start of Investigation by Arkansas State Survey Agency: [See PDF for image] Note: One of the 36 referrals is excluded from this figure because the state survey agency was unable to identify the date the referral was received from the coroner. [End of figure] Predictable Surveys Allow Nursing Homes to Conceal Care Problems: In 1998 and subsequent work, we found that nursing homes could conceal care problems if they chose to do so because annual state surveys were often predictable.[Footnote 40] For example, a home could (1) significantly change its level of care, food, and cleanliness by temporarily augmenting its staff just prior to or during the period of the survey and (2) adjust resident records to improve the overall impression of the home's care. We believe that the striking disparity between annual survey findings that failed to identify serious problems in preventing and treating pressure sores and the numerous instances of serious pressure sores identified by the coroner is partly the result of the predictability of annual surveys. In July 2003, we reported that standard surveys in Arkansas, as well as those nationwide, continued to be highly predictable. In 2003, we reported that the timing of 36 percent of Arkansas's most recent surveys (34 percent nationwide) could have been predicted by nursing homes.[Footnote 41] We considered nursing home surveys predictable if homes were surveyed within (1) 15 days of the 1-year anniversary of their prior survey (28 percent for Arkansas) or (2) 1 month of the maximum 15-month interval between standard surveys (8 percent for Arkansas).[Footnote 42] The director of the Arkansas state survey agency acknowledged that the predictability of the state's standard surveys allowed homes to mask problems by having more staff on hand during surveys. On the basis of the finding in our 2003 report, she told us she has tried to reduce survey predictability, in part by using computer programs to vary the timing of homes' surveys. For 168 of Arkansas's approximately 236 nursing homes surveyed since our last report (August 1, 2003, through June 22, 2004), 22.6 percent of the surveys were predictable. In 1998, we recommended that CMS segment the standard survey into more than one review throughout the year, simultaneously increasing state surveyor presence in nursing homes and decreasing survey predictability.[Footnote 43] Although CMS disagreed with segmenting the survey, it did recognize the need to reduce predictability. CMS directed states in 1999 to (1) begin at least 10 percent of standard surveys outside the normal workday (either on weekends, early in the morning, or late in the evening) and (2) avoid scheduling, if possible, a home's survey for the same month of the year as the home's previous standard survey. We reported previously that CMS's focus on so-called staggered surveys, though beneficial, was too limited to reduce survey predictability.[Footnote 44] Survey Methodology Weaknesses and Misleading Medical Records Contribute to Undetected Care Problems: Our 1998 work on California nursing homes revealed that surveyors may overlook significant care problems because (1) the federal survey protocol they follow does not rely on an adequate sample for detecting potential problems and their prevalence and (2) some resident medical records omit or contain misleading information.[Footnote 45] Because CMS has not yet completed the redesign of the survey methodology, nearly 7 years later Arkansas surveyors, as well as those in other states, still rely on a flawed survey methodology to detect resident care problems. As noted earlier, omissions and contradictions in the decedents' medical records, as well as the coroner's photos, sometimes raised questions about whether appropriate care had been provided in cases the state survey agency did not substantiate. Our 1998 report recommended that CMS revise federal survey procedures by using a stratified random sample of resident cases and reviewing sufficient numbers and types of resident cases. Under development since 1998, CMS's redesigned survey methodology is intended to more systematically target potential problems at a home and give surveyors new tools to better document care outcomes and conduct on-site investigations. Use of the new methodology could result in survey findings that more accurately portray the quality of care provided by a nursing home to all residents. CMS officials told us that the new methodology would be piloted in 2005 in conjunction with an evaluation that compares its effectiveness with that of the current survey methodology. Our work in Arkansas suggested the existence of sampling problems, underscoring the importance of implementing the revised survey methodology. For example, three residents with serious pressure sores who died on March 7, March 29, and April 3, 2000, and were the subject of coroner referrals were not included in the resident sample for one home's March 3, 2000, annual standard survey. The survey failed to identify any pressure sore or other quality of care deficiencies. It is difficult to understand how residents with such serious care problems could have been omitted from the survey. In addition, the extent of the physical deterioration of some decedents where the MFCU identified neglect but the state survey agency did not find similar problems for current residents also raises a question about state survey agency sampling methodology because the seriousness of decedents' conditions suggested that care problems were systemic. In some coroner referrals that the state survey agency did not substantiate, surveyors noted that the medical records indicated that care had been provided. However, the MFCU found omissions and contradictions in decedents' medical records, including missing entries and late entries that were "too many and questionable." The medical record for one decedent showed the resident's height as 10 inches different from the height in her nutritional assessment (height is an important factor in determining a resident's appropriate weight). Since surveyors screen residents' medical records for indicators of improper care, misleading or inaccurate data may result in care deficiencies being overlooked. We also found evidence that Arkansas surveyors took medical records at face value even when these records were contradicted by color photos that documented decedents' physical conditions. For example, our expert consultant found that the coroner's photos of one decedent clearly showed that dead tissue around pressure sores had not been removed even though the state surveyor cited medical records indicating such care was provided just 11 days before the resident's death. The coloration of the same decedent's skin also suggested that she was left in her own waste for extended periods. However, the surveyor noted that the family's concern about staff's unresponsiveness to resident call lights was not substantiated because residents who were interviewed said that staff response was prompt. Under CMS Policy, Nursing Homes Not Always Held Accountable for Past Noncompliance: In our current work, we found that many Arkansas nursing homes with coroner referrals escaped accountability for providing poor care when the state survey agency investigated the neglect of nursing home residents after their deaths. We believe that CMS's vague policy on past noncompliance is partly responsible for this situation. First, the Arkansas state survey agency did not always cite past noncompliance when warranted. For example, the MFCU found that nursing homes had neglected eight decedents referred by the coroner but the state survey agency either cited no deficiency for the decedents, cited a deficiency at a level lower than actual harm for the predominant care problems identified by the coroner, or found the referrals to be unsubstantiated. According to state survey agency officials, care problems similar to those of the decedents were not identified in a sample of current residents and, under CMS policy, the decedents' care problems were assumed to have been identified and corrected by the home. Second, for the one coroner referral that the Arkansas state survey agency did cite for past noncompliance, the home was not required to prepare a plan of correction because no current deficiency was identified. When past noncompliance is identified, it is recorded in OSCAR and on CMS's Nursing Home Compare Web site simply as past noncompliance without additional information on the specific deficient practice(s), such as failure to prevent and treat pressure sores. Moreover, CMS policy discourages citing past noncompliance unless the violation is egregious. Although CMS officials indicate that "egregious" includes noncompliance related to a resident's death, the term is undefined and is not used in CMS's scope and severity grid, which defines serious deficiencies as actual harm or immediate jeopardy.[Footnote 46] According to CMS officials, the objective of its survey policy is to focus surveys on current residents and care problems rather than on poor care provided in the past. We question CMS's assumption that if a decedent's care problem is not found to affect other residents at the time of a complaint investigation, it was identified earlier by the home and corrected.[Footnote 47] On the basis of our past work, it is also possible that the state survey agency's complaint investigation missed serious care issues. CMS and Arkansas state survey agency officials agreed that the poor physical condition of the decedents referred by the coroner suggested the existence of systemic care problems. Conclusions: The Arkansas law requiring coroner investigations of nursing home residents' deaths has helped to demonstrate that a small number of residents died in deplorable physical condition. The Arkansas law also confirmed the systemic weaknesses in state and federal oversight of nursing home quality of care that we identified in prior reports. On the basis of our prior work, we believe it is likely that serious care problems similar to those identified by the Pulaski County coroner exist in other Arkansas counties and in other states. Despite Arkansas's annual standard surveys and intervening complaint investigations, the negligent care provided to some residents before they died was never detected. In addition, complaint investigations initiated by the state survey agency in response to coroner referrals often failed to cite deficiencies for serious care problems that, according to the MFCU's investigations and our expert consultant, constituted or suggested neglect. Even when the Arkansas state survey agency found the neglect to be egregious, it did not hold the nursing home accountable by citing a little used deficiency known as past noncompliance. We believe that CMS's policy on past noncompliance is flawed for three reasons. First, the policy involves considerable ambiguity. CMS does not define what constitutes an egregious violation yet implies that one exists where care problems relate to a resident's death, which is often difficult to demonstrate without an autopsy. Moreover, the term egregious is not clearly related to CMS's scope and severity grid, which defines serious deficiencies as actual harm or immediate jeopardy. Second, CMS's policy on past noncompliance does not hold homes accountable for negligence associated with a resident's death unless similar care problems are identified for current residents. CMS assumes that (1) similar care problems were not found because they have already been identified and corrected by the home and (2) the state survey agency did not miss the deficiency for current residents. However, our prior work demonstrated, and our work in Arkansas confirmed, that (1) nursing home records can contain misleading information or omit important data, making it difficult for surveyors to identify care deficiencies during their on-site reviews and (2) states' surveys of nursing homes do not identify all serious deficiencies, such as preventable weight loss and pressure sores. Third, the policy obscures the nature of the specific care problem, such as avoidable pressure sores, because the only deficiency reported in OSCAR and to the public on CMS's Nursing Home Compare Web site is "past noncompliance." We believe that the goal of preventing resident neglect by requiring nursing homes to comply with federal quality standards is inconsistent with a policy that discourages citing deficiencies because the harm was simply not egregious enough or was potentially missed for current residents. Recommendations for Executive Action: We recommend that the Administrator of CMS revise the agency's current policy on citing deficiencies for past noncompliance with federal quality standards by taking the following two actions: * hold homes accountable for all past noncompliance resulting in harm to residents, not just care problems deemed to be egregious, and: * develop an approach for citing such past noncompliance in a manner that clearly identifies the specific nature of the care problem both in the OSCAR database and on CMS's Nursing Home Compare Web site. Agency and State Comments and Our Evaluation: We provided a draft of this report to CMS; the Arkansas Department of Human Services, Office of Long Term Care (the state survey agency); the Arkansas MFCU; and the Pulaski County coroner. We received written comments from CMS and the survey agency, and oral comments from the coroner. The MFCU stated that it did not have comments. CMS concurred with our recommendations to revise its policy on citing deficiencies for past noncompliance and also identified more than a dozen additional initiatives it plans to take to address shortcomings in the nursing home survey process. CMS commented that the focus of its initiatives, such as additional guidance on the scope and severity of deficiencies, would be broad, in effect supporting our conclusion that the shortcomings we identified were systemic and not limited to Arkansas. CMS and the state survey agency raised concerns about (1) the discrepancy we reported between the number of referrals the coroner said he made (86) and the number the survey agency said it received (36) and (2) the relevance of survey predictability to complaint investigations based on coroner referrals. In addition, the state survey agency commented that we had understated the number of investigations it actually conducted. (CMS's comments are reproduced in app. III,[Footnote 48] and the state survey agency's comments are reproduced in app. IV.) Our evaluation of CMS, survey agency, and coroner comments covers the following six areas: CMS's past noncompliance policy, shortcomings in state survey agency investigations, lessons from implementing the Arkansas law, the number of coroner referrals and survey agency investigations, survey predictability and methodology redesign, and the impact of the Arkansas law. CMS Policy on Past Noncompliance: CMS agreed with our recommendations to revise its past noncompliance policy. We found that some nursing homes were not held accountable for serious deficiencies, even though some coroner referrals were substantiated, because of flaws in CMS's policy governing past noncompliance. Following a planned review of the policy, CMS indicated that it would (1) clarify expectations for the manner in which state survey agencies should address past deficiencies that have only recently come to light, (2) further define important terms, particularly egregious, (3) ensure that the specific nature of the care problems was identified in OSCAR, and (4) strengthen criteria for determining whether a nursing home had actually taken steps to address deficiencies that contributed to past noncompliance. CMS did not indicate whether it also planned to identify the specific nature of deficiencies associated with past noncompliance on its Nursing Home Compare Web site, but we continue to believe that posting such information would provide valuable assistance to consumers. Shortcomings in State Survey Agency Investigations Nationwide: Because of the seriousness of the shortcomings identified in our report, CMS sent a clinical fact-finding team to Arkansas for 3 days after receiving a draft of our report. The CMS clinical team found that some, but not all, of the referrals for which lower-level deficiencies were cited should have received a higher-level severity rating. In addition, from among six coroner referrals that were not substantiated by the survey agency, the team believed two should have been substantiated, a higher disparity rate than CMS said it has typically found for Arkansas surveys in general. As a result of its team's visit, CMS concluded that additional training and clarification of its guidance were warranted, including (1) increased training for state surveyors in determining the appropriate scope and severity of deficiencies as well as the development of additional CMS guidance and analysis of patterns in state deficiency citations and (2) the development of an advanced course in complaint investigations to be piloted in Arkansas and evaluated for potential expansion and replication nationwide. CMS noted that these initiatives would be applied broadly, a recognition that the shortcomings we identified were systemic and not limited to Arkansas. While we fully support CMS's new initiatives, timely and sustained follow-up to ensure effective implementation is critical; earlier CMS initiatives to address these same problems were not timely or were ineffective. We reported in July 2003 that CMS began a complaint improvement project in 1999 but did not provide more detailed guidance to states until almost 5 years later.[Footnote 49] Similarly, we reported that CMS began developing more structured guidance for surveyors in October 2000 to address inconsistencies in how the scope and severity of deficiencies are cited across states, but the first installment on pressure sores had not yet been released as of September 2004.[Footnote 50] Our 2003 report also noted that CMS began annual reviews of a sample of deficiency citations from each state in October 2000 to identify shortcomings and the need for additional training, but CMS's recognition that additional guidance and training are required raises a question about the sufficiency and effectiveness of these reviews. Furthermore, we believe that other factors may be contributing to survey shortcomings. Our 2003 report noted that some state officials cited inexperienced surveyors, the result of a high turnover rate, as a factor contributing to the understatement of serious quality of care deficiencies. CMS commented that the photos conveyed from the coroner's office were graphic, serious, and require careful investigation. The CMS clinical team found that the photos were very helpful in a number of investigations. We agree with CMS's view that the photos alone do not represent sufficient evidence to render a conclusion that there was poor care, neglect, or avoidable outcomes, or that the nursing home caused the death. On the basis of its visit to Arkansas, the CMS clinical team concluded that not all referred cases could be substantiated with the photos, medical records, and other information available to it; as we noted in the report, our expert consultant reached the same conclusion on two of the seven cases she reviewed. We nevertheless continue to believe that the state survey agency at times appeared to dismiss photographic evidence of potential neglect and to rely instead on observations of and interviews with current residents. In response to our findings, CMS said it would study the issues involved in the use of photos and would issue additional guidance for use by state survey agencies. Lessons from Implementing Arkansas's Law on Nursing Home Deaths: CMS made a number of observations about lessons from the Arkansas experience that would improve the effectiveness of mandatory reporting systems, such as the coroner referrals required by the Arkansas law. These lessons related to the implementation of the Arkansas law by local coroners and the quality and timeliness of referrals made by the Pulaski County coroner. We agree that these factors are important to the ability of state survey agencies to promptly and effectively complete their own investigations based on coroner referrals of potential neglect. However, because we lack the authority to evaluate the implementation of state laws, we excluded such an analysis from the scope of our work. We do have the authority to evaluate the performance of federally funded entities--such as the state survey agency and the MFCU--that are responsible for ensuring that Medicare and Medicaid nursing home residents receive quality care, and we therefore focused our work on how these entities responded to the cases referred to them. In particular, CMS highlighted the lack of referrals from most Arkansas coroners and the processes followed by coroners, primarily the Pulaski County coroner, in making referrals to the state survey agency. During our interviews, the Pulaski County coroner and MFCU officials demonstrated their awareness of the absence of an enforcement mechanism in the state law to ensure that nursing homes and coroners comply with the law; the Pulaski County coroner told us that he intends to pursue this issue with the state legislature. According to CMS, the quality of the documentation provided by coroners did not conform to key principles of forensic science, such as embedded photo dating and subject identification, photo scale metrics and color charting, and interviews with residents' physicians. While the coroner referrals may have lacked these features, the referral packages we examined clearly identified the decedents, the time the coroner's office was notified of the deaths, and the time the coroner's staff arrived at the homes. It is also clear from the documentation that the photos were taken shortly after death. Requiring such a level of forensic evidence from the coroner substantially exceeds the burden of proof the state survey agency requires for other complaints filed, which is how the coroner referrals are treated. The coroner referrals are intended to be the starting point for the state's investigation, not a substitute for its own thorough investigation. Both CMS and the state survey agency expressed concern about the elapsed time between the dates of death and the receipt of coroner referrals by the survey agency. In particular, they noted that our analysis excluded five referrals the coroner made in 2004 that related to deaths in 2003, with the elapsed times from the deaths to receipt of the referrals ranging from 222 to 400 days.[Footnote 51] We excluded these five referrals because they had not yet been referred when we completed our data collection for this report, which covered referrals for the period July 1999 through December 2003.[Footnote 52] In principle, we agree with CMS's view that the value of a timely investigation by the state survey agency can be influenced by the length of time associated with referrals, even though we found that the coroner's referral of several cases up to 4 months after the residents' deaths did not appear to have handicapped the investigations. For example, the state survey agency substantiated three coroner referrals with deficiencies at the actual harm and immediate jeopardy level even though the referrals were not received for between 65 and 106 calendar days after residents' deaths. Although the survey agency did not substantiate one coroner referral that was not received until 102 days after the resident's death, the MFCU found neglect. For the 36 referrals the survey agency said it received from the coroner for the period we analyzed, the average elapsed time from the date of death until the coroner made his referral was 38 days (ranging from zero to 180 days), whereas the average elapsed time from the date the survey agency received the referral until it initiated its investigation was 46 days (ranging from zero to 425 days).[Footnote 53] Notwithstanding these elapsed times for coroner referrals and state investigations, CMS commented that it would study its priority criteria for complaint triage and refine its policy with regard to the treatment of and response to complaints. Number of Coroner Referrals and State Survey Agency Investigations: Both CMS and the state survey agency questioned the validity of the number of Pulaski County coroner referrals, commenting that we lacked independent verification of the number actually referred; they also believed that the report's language suggested referrals had been received but not investigated. We revised the report to make it clear that the coroner told us he had referred 86 cases of suspected neglect of deceased nursing home residents to the state survey agency and the MFCU for investigation (and, as noted below, we reviewed the related case documentation for each of the 86 referrals). We also revised the report to clarify that the state survey agency investigated the 36 coroner referrals that it told us it had received.[Footnote 54] CMS asserted that the coroner was unable to provide its clinical team with a list of his referrals; however, CMS's comments do not reflect that the coroner's case files were not automated. We compiled a list of the 86 referrals ourselves. Our list was based on documentation provided by the coroner for each of the cases he told us he referred, including a narrative summary describing the suspected neglect, copies of decedents' medical records, autopsy reports, and photos documenting the decedents' conditions. Although the state survey agency and the MFCU told us that they did not receive all 86 coroner referrals, we believe that the MFCU's receipt of almost three fifths of the coroner's referrals (compared with the state survey agency's receipt of fewer than half) provides independent corroboration that the Pulaski County coroner made more than 36 referrals during the 4.5-year period we examined. As noted in the report, the coroner was instrumental in securing passage of the law, a fact that is inconsistent with the suggestion that the coroner withheld referrals. To address the disparity in the number of referrals the coroner told us he made and the number the state survey agency and the MFCU told us they received, the coroner began requiring signed receipts in March 2004, a practice reflected in our draft report. The state survey agency commented that we had understated the number of investigations of nursing home deaths it had conducted. The agency identified 22 investigations that, in most cases, were based on the receipt of a complaint from individuals other than the coroner. * We excluded 9 of these 22 investigations because they were conducted prior to the residents' deaths. For example, one complaint of alleged rape of a 91-year-old resident was filed by a hospital that found the resident had a sexually transmitted disease. The complaint was not substantiated. The coroner's investigation of the resident's death 5 months later resulted in a referral based on seven serious pressure sores on the decedent's feet, lower back, and hips, a problem that was not noted during the hospitalization. * We revised our analysis to include 1 of the 22 cases because the coroner confirmed that he had indeed made the referral. Thus, we adjusted the number of coroner referrals from 85 in the draft report to 86 in the final report. We also revised the number of referrals the state survey agency said it received from 35 to 36. We confirmed that this additional referral was not received or investigated by the MFCU. * For 7 cases, we determined that the allegations in the non-coroner complaints were similar to the concerns raised by the coroner's investigations and have added footnotes in the appropriate sections of the report, depending on whether the investigations substantiated (2 complaints) or did not substantiate (5 complaints) the complainants' allegations. * For the remaining 5 cases, we made no changes in the report.[Footnote 55] In one case, the survey agency's complaint investigation focused on an issue different from the suspected neglect identified by the coroner. In four other cases, the agency included the decedents' records in its resident samples during standard surveys. The decedents were not included in any deficiencies cited during these surveys and, importantly, the surveyors lacked the coroner's photos of pressure sores, which would have been particularly useful in raising questions about the care provided as documented in the decedents' medical records. Survey Predictability and Methodology: Both CMS and the state survey agency questioned the relevance of survey predictability to complaint investigations resulting from coroner referrals and suggested we delete this analysis from the final report. Neither organization commented on our assessment of the impact of survey methodology weaknesses and misleading medical records on detecting quality-of-care problems. We retained this analysis in the final report because we believe the issues of survey predictability and methodology are relevant to state survey complaint investigations of coroner referrals. Our 1998 and subsequent work found that predictable surveys allowed homes so inclined to (1) significantly change the level of care, food, and cleanliness by temporarily augmenting staff just prior to or during a survey, and (2) adjust resident records to improve the overall impression of the home's care.[Footnote 56] We also reported in 1998 that surveyors may overlook significant care problems during annual surveys because of survey methodology weaknesses and omissions or misleading information in resident medical records. Although the predominant care problem identified in 67 percent of the coroner's referrals involved serious pressure sores, most of the nursing homes referred had not been cited for a pressure sore deficiency at the actual harm level or higher on any of their previous four standard surveys. We believe that the striking disparity between annual survey findings and the predominant care problems identified by the coroner relates to the predictability of annual surveys, weaknesses in survey methodology, and misleading medical records--all of which contribute to the phenomenon of undetected care problems. Our work in Arkansas suggested the existence of sampling problems in a home whose annual survey failed to detect any quality-of-care problems, even though three residents, all with serious pressure sores, died within 1 month. The fact that none of these residents was included in the nursing home's annual standard survey underscores the importance of implementing a revised survey methodology that CMS has had under development for 7 years. Our report also provides several examples where misleading medical records contributed to the failure of the Arkansas state survey agency to detect care problems that the MFCU or our expert consultant identified and were obvious in some of the coroner's photos of decedents. CMS further commented that our analysis of survey predictability resurrected prior reports and recommendations to which CMS has previously responded and that we failed to acknowledge CMS and state survey agency progress in reducing survey predictability. We believe that CMS's comments are inaccurate. In our 1998 report, we recommended segmenting the survey into more than one review throughout the year to reduce survey predictability. CMS responded to this recommendation by requiring that 10 percent of state annual surveys be conducted on weekends, at night, or early in the morning. Despite CMS's introduction of "off hour" surveys, we reported in 2003 that about one-third of state surveys remained predictable (36 percent in Arkansas). Contrary to CMS's comments, the draft report did acknowledge that Arkansas appeared to be making progress in reducing survey predictability through the use of computer programs to vary the timing of homes' surveys. Impact of the Arkansas Law: In oral comments, the Pulaski County coroner indicated that our report was fair and accurate. He also told us that he believes the law has had a significant, positive impact on the quality of care provided to nursing home residents in Pulaski County. In particular, he rarely finds decedents with serious pressure sores and the pressure sores he does find are not as serious as those in earlier referrals. He also cited the declining number of referrals--only six 2003 resident deaths were referred compared to 18 in 2002. He also provided technical comments that we incorporated as appropriate. As arranged with your offices, unless you publicly announce its contents earlier, we plan no further distribution of this report until 30 days after its issue date. At that time, we will send copies of this report to the Administrator of the Centers for Medicare & Medicaid Services and appropriate congressional committees. We also will make copies available to others upon request. In addition, the report will be available at no charge on the GAO Web site at http://www.gao.gov. Please contact me at (202) 512-7118 or Walter Ochinko, Assistant Director, at (202) 512-7157 if you or your staffs have any questions. GAO staff who made key contributions to this report include Jack Brennan, Lisanne Bradley, Patricia A. Jones, and Elizabeth T. Morrison. Signed by: Kathryn G. Allen: Director, Health Care--Medicaid and Private Health Insurance Issues: [End of section] Appendix I: Coroner Referrals for Pressure Sores and the Seriousness of Deficiencies Cited on Standard Surveys: Nursing home: A; Number of coroner referrals for pressure sores: 11; Number of deficiencies cited for pressure sores on homes’ standard surveys[A]: Below actual harm: 1. Nursing home: B; Number of coroner referrals for pressure sores: 5; Number of deficiencies cited for pressure sores on homes’ standard surveys[A]: Actual harm or higher: 1; Number of deficiencies cited for pressure sores on homes’ standard surveys[A]: Below actual harm: 2. Nursing home: C; Number of coroner referrals for pressure sores: 5; Number of deficiencies cited for pressure sores on homes’ standard surveys[A]: Below actual harm: 1. Nursing home: D; Number of coroner referrals for pressure sores: 4; Number of deficiencies cited for pressure sores on homes’ standard surveys[A]: Actual harm or higher: 1; Number of deficiencies cited for pressure sores on homes’ standard surveys[A]: Below actual harm: 1. Nursing home: E; Number of coroner referrals for pressure sores: 3. Nursing home: F; Number of coroner referrals for pressure sores: 3; Number of deficiencies cited for pressure sores on homes’ standard surveys[A]: Below actual harm: 2. Nursing home: G; Number of coroner referrals for pressure sores: 3. Nursing home: H; Number of coroner referrals for pressure sores: 3; Number of deficiencies cited for pressure sores on homes’ standard surveys[A]: Below actual harm: 2. Nursing home: I; Number of coroner referrals for pressure sores: 3. Nursing home: J; Number of coroner referrals for pressure sores: 3. Nursing home: K; Number of coroner referrals for pressure sores: 4. Nursing home: L; Number of coroner referrals for pressure sores: 2; Number of deficiencies cited for pressure sores on homes’ standard surveys[A]: Actual harm or higher: 1. Nursing home: M; Number of coroner referrals for pressure sores: 2. Nursing home: N; Number of coroner referrals for pressure sores: 1. Nursing home: O; Number of coroner referrals for pressure sores: 2. Nursing home: P; Number of coroner referrals for pressure sores: 1; Number of deficiencies cited for pressure sores on homes’ standard surveys[A]: Below actual harm: 2. Nursing home: Q; Number of coroner referrals for pressure sores: 1; Number of deficiencies cited for pressure sores on homes’ standard surveys[A]: Below actual harm: 1. Nursing home: R; Number of coroner referrals for pressure sores: 0; Number of deficiencies cited for pressure sores on homes’ standard surveys[A]: Below actual harm: 2. Nursing home: S[B]; Number of coroner referrals for pressure sores: 1. Nursing home: T; Number of coroner referrals for pressure sores: 0. Nursing home: U; Number of coroner referrals for pressure sores: 0; Number of deficiencies cited for pressure sores on homes’ standard surveys[A]: Actual harm or higher: 2; Number of deficiencies cited for pressure sores on homes’ standard surveys[A]: Below actual harm: 1. Nursing home: V; Number of coroner referrals for pressure sores: 0; Number of deficiencies cited for pressure sores on homes’ standard surveys[A]: Below actual harm: 1. Nursing home: W; Number of coroner referrals for pressure sores: 0; Number of deficiencies cited for pressure sores on homes’ standard surveys[A]: Below actual harm: 1. Nursing home: X; Number of coroner referrals for pressure sores: 0. Nursing home: Y[B]; Number of coroner referrals for pressure sores: 0. Nursing home: Z; Number of coroner referrals for pressure sores: 1. Nursing home: AA[B]; Number of coroner referrals for pressure sores: 0. Nursing home: Total; Number of coroner referrals for pressure sores: 58; Number of deficiencies cited for pressure sores on homes’ standard surveys[A]: Actual harm or higher: 5; Number of deficiencies cited for pressure sores on homes’ standard surveys[A]: Below actual harm: 17. Source: GAO analysis of coroner referrals and OSCAR data. [A] Includes last four state surveys for each home as of October 24, 2003,with the exception of homes Q and Z, which include the last four surveys as of July 30, 2004. [B] The state survey agency is not required to survey these facilities under federal law. [End of table] [End of section] Appendix II: Coroner Referrals That the State Survey Agency Reported as Not Received, Substantiated, or Not Substantiated: Nursing home: A; Number of referrals: 12; Number not received: 5; Number substantiated: 7. Nursing home: B; Number of referrals: 7; Number not received: 3; Number substantiated: 1; Number not substantiated: 3. Nursing home: K; Number of referrals: 7; Number not received: 5; Number not substantiated: 2. Nursing home: C; Number of referrals: 6; Number not received: 4; Number substantiated: 2. Nursing home: F; Number of referrals: 5; Number not received: 3; Number not substantiated: 2. Nursing home: H; Number of referrals: 5; Number not received: 5. Nursing home: D; Number of referrals: 4; Number not received: 2; Number substantiated: 1[A]; Number not substantiated: 1. Nursing home: G; Number of referrals: 4; Number not received: 1; Number not substantiated: 3. Nursing home: M; Number of referrals: 4; Number not received: 2; Number not substantiated: 2. Nursing home: E; Number of referrals: 3; Number not received: 1; Number substantiated: 2[A]. Nursing home: I; Number of referrals: 3; Number not received: 1; Number substantiated: 2[A]. Nursing home: J; Number of referrals: 3; Number not received: 2; Number not substantiated: 1. Nursing home: L; Number of referrals: 3; Number not received: 1; Number substantiated: 2. Nursing home: R; Number of referrals: 3; Number not received: 3. Nursing home: O; Number of referrals: 2; Number not received: 2. Nursing home: P; Number of referrals: 2; Number not received: 2. Nursing home: S; Number of referrals: 2; Number not received: 2. Nursing home: T; Number of referrals: 2; Number not received: 1; Number substantiated: 1. Nursing home: N; Number of referrals: 1; Number not received: 0; Number substantiated: 1. Nursing home: Q; Number of referrals: 1; Number not received: 0; Number substantiated: 1. Nursing home: U; Number of referrals: 1; Number not received: 1. Nursing home: V; Number of referrals: 1; Number not received: 1. Nursing home: W; Number of referrals: 1; Number not received: 1. Nursing home: X; Number of referrals: 1; Number not received: 0; Number substantiated: 1. Nursing home: Y; Number of referrals: 1; Number not received: 1. Nursing home: Z; Number of referrals: 1; Number not received: 1. Nursing home: AA; Number of referrals: 1; Number substantiated: 1. Nursing home: Total; Number of referrals: 86; Number not received: 50; Number substantiated: 22; Number not substantiated: 14. Source: Arkansas state survey agency. Note: Data on referrals made from July 1999 through December 2003 are based on information provided by the Pulaski County coroner. [A] One referral was substantiated without any deficiencies. [End of table] [End of section] Appendix III: Comments from the Centers for Medicare & Medicaid Services: A portion of CMS's response was based on tables presented in attachment 1 to its comments. Because the tables did not accurately reflect the coroner cases discussed in our report, CMS submitted an amended attachment 1, which we have substituted for the original attachment 1. CMS, however, did not make corresponding changes on pages 6 and 7 of its letter. We have marked the text on those pages in the letter where the information in the amended attachment 1 supercedes data presented in the letter. DEPARTMENT OF HEALTH & HUMAN SERVICES: Centers for Medicare & Medicaid Services: Administrator: Washington, DC 20201: DATE: SEP 15 2004: TO: Kathryn G. Allen: Director, Health Care-Medicaid and Private Health Insurance Issues: Government Accountability Office: Signed by: FROM: Mark E. McClellan, M.D., Ph.D.: Administrator: SUBJECT: Government Accountability Office (GAO) Proposed Report: Nursing Home Deaths: Arkansas Coroner Referrals Confirm Weaknesses in State & Federal Oversight of Quality of Care (GAO-04-980): Thank you for the opportunity to comment on the above-referenced, draft report. The issues raised in the report are important ones. For that reason, and because of significant gaps in the information available, we sent a clinical team to Arkansas for three days. The Centers for Medicare & Medicaid Services (CMS) fact-finding team reviewed a number of the cases referred to the state Survey Agency (SA) over the course of the 4.5 years covered by the GAO study. The team reviewed evidence associated with the coroner case referrals, including photos. The team reviewed complaint investigation files and complaint systems related to these referrals in the state SA. The team conducted numerous interviews among officials of the Pulaski county coroner's office, the state SA, and the Medicaid Fraud Control Unit (MFCU) located in the state's Attorney General's office. In the absence of a common descriptive label in the law itself, we will simply refer to it as the "Coroner Referral Law" for ease of reference. The Arkansas "Coroner Referral Law," unique in the nation, represents an important case study. There are valuable lessons that can be gained from the Arkansas experience and from your report. 1n the comments below we highlight some of those lessons, affirm in principle the GAO recommendation to CMS, and itemize more than a dozen additional action steps that CMS is undertaking in order to make the most effective use of insights obtained from the Arkansas experience. The Pulaski county coroner reports that he referred 85 cases over 4.5 years to both the SA and MFCU, about 2.1 % of the total number of nursing home deaths during that period. The SA reports, however, that it received fewer cases than reported by the coroner, and then investigated 100% of those that it did receive. The MFCU also reported that it actually received fewer cases than the number stated by the coroner. Although the number of allegations of abuse or neglect referred by the coroners and substantiated by the survey agency represented only about 0.5% of all nursing home deaths (21 of more than 4000 reported to the coroners), the conditions of the affected individuals were extremely serious and merit close review. Lessons for Mandatory Reporting Systems: CMS, and any state that contemplates a mandatory reporting and investigation law similar to the law in Arkansas, can benefit greatly from examining the experience under the Arkansas Coroner Referral Law since 1999. Only 1 of 75 coroners seems to have had substantial participation in the referral and investigation system (about 95% of reported referrals came from just one coroner). This experience points to the need for enforcement provisions in any such mandatory referral system. The two agencies receiving referrals from the Pulaski county coroner state Medicaid Fraud Control Unit (MFCU) and SA report that they did not receive 40-59% of the referrals (respectively) that the Pulaski county coroner states he made. This phenomenon points to the importance of having crystal clear referral and documentation systems in place from the referring agencies (in this case, the coroners'offices). The fact that referrals from coroners' offices to the state survey agency ranged from 2 days to 415 days (after receipt of the notice from the nursing home) suggests the need for clear timelines for all participating agencies. The CMS clinical team also found that a considerable amount of the documentary evidence conveyed from coroners' offices did not conform to key principles of forensic science. Examples include a lack of embedded photo dating, lack of embedded subject identification, lack of scale metrics and color charting in the photos, lack of adequate records for interview sources, lack of interviews with residents' physicians. These experiences suggest the need for clear investigatory protocols and a method to ensure that the protocols are scrupulously followed by coroners' offices. The fact that the qualifications of the coroners varies so widely, that the coroners are rarely physicians (about 80% are reported to be funeral home directors) suggests the need for either additional training to handle the more sophisticated responsibilities dictated by the Arkansas law, or qualification standards. The point of these observations is not to criticize the Arkansas coroners, for they work within a larger system structured by the law itself. The point is that an effective referral and investigation system depends on the action of many participants whose individual contributions must integrate. Adequacy, timeliness, and quality of the referrals are important because they: * Aid or impede the survey agency's subsequent investigation. * Affect the SA's judgement about how urgently the investigation ought to be done compared to other newly-arriving complaints. * Affect the freshness of evidence or witness testimony if referrals are delayed. Affect CMS' ability to hold survey agencies accountable since no agency can fully prove a negative (e.g. "prove that you did not receive the number of referrals that the coroner thinks was sent"). CMS Policy on Past Non-Compliance: The CMS policy on "past non- compliance," incorporated into regulation and operational manuals, is intended to avoid penalizing nursing homes for problems for which they have taken clear remedial action and that occurred prior to more recent surveys of the nursing homes but only recently came to the attention of the SA. We fully agree in principle with the GAO recommendation that the CMS policy for past non-compliance merits additional work in light of the Arkansas experience. Following our review we will: * Clarify Expectations for Past Deficiency Findings for Nursing Homes: CMS will clarify expectations for the manner in which state survey agencies should address past deficiencies that have only recently come to light. * Further Define Terms for Past Non-Compliance: Further define important terms, particularly the term "egregious." * Enhance CMS Information System: Add the capacity to record the fact and the nature of past noncompliance in the CMS information system ("OSCAR"). * Strengthen Criteria for NH Correction of Deficiencies: Strengthen the criteria for determining whether past noncompliance has actually been corrected by nursing homes (NHs). This action moves beyond the GAO recommendation because we believe that the degree of systemic remedy actually put in place by the nursing home ought to be the most important determinant for present-day enforcement of past noncompliance that only more recently came to the attention of the survey agency. Referrals from the Coroner's Offices. With regard to referrals from the Pulaski county coroner, the GAO report states that the SA investigated "fewer than half of those the coroner referred..." (p.4). This statement is unconfirmed and ought to be removed. The only confirmed facts are: * The full number of referrals that the Pulaski coroner states that he made were not confirmed as having been delivered to either the SA or the Medicaid Fraud Control Unit of the Arkansas Attorney General's Office. * The SA investigated all referrals that they confirm as having been received. The draft GAO report also criticized both the SA and MFCU for record- keeping practices, making no mention of the coroners' offices. In contrast, we found that the state SA had a competent system for managing complaints. A second state agency (the MFCU) similarly reported receiving far fewer than the 85 referrals that the coroner's office suggested (51 rather than 85, according to the draft GAO report). The Pulaski county coroner's office was not able to provide CMS with a listing of the 85 referrals it reportedly made to both the SA and MFCU. The CMS on-site review of the SA's system, the inability of the coroner's office to provide CMS with either a list of the 85 referrals or confirmation that they were delivered, and the fact that two different state agencies both report that they did not receive the full number of referrals claimed by the coroner, are facts that lead us to conclude that the state survey agency received fewer referrals than 85. Of these, the SA investigated 100%. We suggest that those sections of the GAO report that baldly accept the coroner's number of 85 referrals (e.g. top of p. 4) be adjusted to reflect the fact that evidence is lacking to confirm that 85 referrals were actually delivered. And to state on page 19 that "Overall, 32 coroner referrals were not investigated by either agency" is simply inaccurate in light of the fact that receipt of the referrals is not confirmed by either of the two agencies (SA or MFCU). Insofar as CMS jurisdiction applies only to the SA and MFCU, we will undertake the following action to augment the national automated complaint tracking system that CMS and states implemented in 2004: * Design Feedback Systems for Agency Referrals: We will design, with Arkansas and other states, a model feedback report for coroners, law enforcement and other agencies that have a mandatory reporting obligation. The feedback report will contain pertinent information on all referrals received by the SA from the relevant source. The report will provide a structured means by which the referral source may check its own records and identify any problems it perceives. Timeliness of Investigations: The Arkansas "Coroner Referral Law" presents special challenges for public policy. We will: * Study the Priority Criteria for Complaint Triage: Study the Arkansas experience in terms of its implications for CMS requirements for the speed and triage priorities for responding to complaint referrals made on behalf of individuals who are deceased. * Refine CMS Policy on Responding to Complaints: Refine CMS policy with regard to the treatment of complaints. Among the many questions worth considering are the following: What priority should be accorded to referrals that involve persons who have been deceased for a considerable period of time prior to receipt by the SA, compared to complaints involving current residents? How can patterns of potential abuse or neglect best be identified to establish context for any referral, particularly patterns that ought to move a referral to top priority regardless of how long a person has been deceased prior to receipt by the SA? In the Arkansas experience, the CMS clinical team found that many referrals from the coroner's office were significantly delayed. Attachment 1 to this letter contains a listing of the referral times and SA investigation times. The Pulaski county coroner stated that the coroner's referral process, from the date of the nursing home's call of a resident's death to case development and referral by the coroner to the SA, takes a maximum of 2 weeks. Yet, both the SSA and the MFCU received coroner cases that took significantly longer (see Attachment 1). The GAO report apparently omitted 5 of the following 6 cases that were reported by nursing homes to the coroner in 2003 but not referred from the coroner to the state survey agency until 2004. The CMS team reviewed the timeliness of referral to the SA of the 6 Pulaski County nursing home resident deaths in 2003 and found the following. Case #: 8966; Date of Coroner Report: 3/10/2003; Date of SA Receipt: 4/28/2004; Days: 415. Case #: 8965; Date of Coroner Report: 5/17/2003; Date of SA Receipt: 4/28/2004; Days: 347. Case #: 8974; Date of Coroner Report: 6/10/2003; Date of SA Receipt: 4/29/2004; Days: 324. Case #: 8970; Date of Coroner Report: 6/15/2003; Date of SA Receipt: 4/13/2004; Days: 303. Case #: 8973; Date of Coroner Report: 9/4/2003; Date of SA Receipt: 4/29/2004; Days: 238. Case #: 8779; Date of Coroner Report: 3/1/2004; Date of SA Receipt: 3/05/2004; Days: 4. [End of table] Overall, the CMS team found the following approximate average and median intervals of time for all cases (including those above that were omitted by GAO). Referral Interval: Median Time; Coroner – Receipt to Referral Time: 29 days; State Survey Agency Receipt to Investigation Time: 21 days. Referral Interval: Average Time; Coroner – Receipt to Referral Time: 75 days; State Survey Agency Receipt to Investigation Time: 37 days. Referral Interval: % Completed within 10 Days; Coroner – Receipt to Referral Time: 17%; State Survey Agency Receipt to Investigation Time: 33%. Referral Interval: Range; Coroner – Receipt to Referral Time: 2 - 415 days; State Survey Agency Receipt to Investigation Time: 0 - 164 days. [End of table] These data are included here not as criticism of the Arkansas coroner, but to establish context for evaluating actions of the state SA. The value of a timely investigation by the SA will be reduced by any significant interval of time required by referring entities before cases are referred to the SA. The CMS clinical team did find that the SA should have investigated a number of the cases in a more timely manner when evaluated strictly against CMS policy. In reviewing the "triage" decisions of the SA, the CMS team concurred with 81% and did not concur with 19%. It remains for us to evaluate that policy in light of the Arkansas experience, make needed adjustments or clarifications, communicate and train state agency staff on those policies, and then ensure that the timelines are met. We will do so. Cases Not Substantiated: The GAO report concluded that "In half of the referrals not substantiated by the state survey agency, either the MFCU investigation found neglect or we questioned the basis for the `not substantiated' finding and our concerns were confirmed by a professor of nursing with experience in long-term care." (p.18): The SA confirmed receiving 35 cases, investigated the 35, and substantiated 21 of the cases. Additional cases were investigated that GAO put on its list, but the SA's investigations were precipitated by complaints other than those reportedly made by the coroner and are not included here in our comments. While it is appropriate to question the 14 cases not substantiated, investigations by survey agencies are governed by rules of evidence and must be fact-based. This is appropriate, since allegations and convictions for abuse or neglect are serious, legal matters in which all parties have certain legal rights and responsibilities. The CMS clinical team reviewed a sample of 8 of the 14 cases that were not substantiated by the SA. For 2 of the 8 a final conclusion could not be reached without additional records from the nursing home that had been available to the SA investigators at the time the state officials conducted their investigation. Of the remaining 6 cases the CMS team concurred with 4 and disagreed with 2. The degree of disagreement (2 of 8) is higher than the typical disparity rate CMS has found for Arkansas surveys in general. The difference here could be an artifact of the small sample size (8 cases) or special challenges experienced by the state SA in investigating referrals of individuals who have been deceased for some time. We have reviewed these findings with the Arkansas SA and concluded that additional training is needed. To that end, CMS will: * Increase Complaint Investigation Training: Develop a curriculum for an advanced course in complaint investigations. * Pilot Advanced Programs for Complaint Investigation Training Work with state officials to sponsor such training and include other states in the region. * Evaluate and Expand Advanced Training for Complaint Investigations: Evaluate the training for potential expansion and replication in all regions. Understating the Seriousness of Deficiencies: The GAO report states that (a) some of the substantiated complaints did not result in citations for deficiency, and (b) some of those that were cited were done at lower levels of severity than warranted. Cases that were substantiated but did not result in a citation of deficiency primarily result from the policy governing past noncompliance. This policy is provided both in CMS operations manuals and in regulation at 42 CFR 488.430(b). As stated earlier, we will review this policy and take appropriate action pursuant to our review. The CMS clinical team agreed that some - but not all-of the other cases should have received a higher level of severity rating. We are discussing these findings with the state SA and have agreed with them to provide additional training and guidance for the future. The CMS monitoring and other available evidence indicates that Arkansas has not had a pattern of citing deficiencies at levels lower than expected during the 4.5 year period covered by the GAO study (they have generally been higher than national averages). We therefore take work of our clinical team and of GAO to mean that additional training and additional clarifications relative to scope and severity are warranted even more for other states than for Arkansas. We will do the following: * Increase Training in Classifying Deficiencies: We will arrange for state-specific or multi-state training to address issues related to the proper citation of deficiencies in terms of their scope and severity. * Issue Additional CMS Guidance on Deficiency Classification: We will issue additional guidance to states of the proper classification of identified deficiencies and the relationship of those citations to enforcement actions. * Increase Regional Office Follow-Up. * Increase CMS Analysis of Patterns in State Deficiency Citations: Following Through on Enforcement: Proper classification and citation of a deficiency is necessary, but not wholly sufficient. The ultimate goal must be (a) remedy of the problem, and (b) system changes that might be needed to prevent identical or similar problems in the future. To this end, CMS will: * Automate Enforcement Tracking: On October 1, 2004 CMS will implement, nationwide, a new electronic tracking and management system ("Aspen Enforcement Manager," or "AEM") for all types of actions that state survey agencies might take, in response to identified nursing home deficiencies, to promote the prompt and effective remedy of the problems. Predictability of Nursing Home Surveys: The GAO report summarizes previous GAO work related to the extent to which nursing homes might be able to predict the states' unannounced surveys. The frequency of the surveys is statutorily required every 15 months, with an average of 12 months. CMS requires both staggering of surveys (e.g. nursing homes should not have surveys around the same time of year each year) and off-hour surveys (e.g. weekends, evenings). The Arkansas SA has a track record of often exceeding the percentage of off-hour surveys required by CMS (10 percent). Substantial increases in the off-hour or staggered schedule requirements would have a fiscal impact on state budgets, or require that other important functions not be done. The GAO report would be more useful if it came to terms with these facts instead of resurrecting old GAO reports (to which CMS has responded) without acknowledging CMS and state improvements made since the original GAO reports, or the implications of even further advances. Since the issue of survey predictability is not especially germane to the analysis of the Arkansas Coroner Referral Law, we recommend that this portion of the GAO report be removed, or that the above issues be addressed. Use of Photographs: The photos conveyed from the coroner's office are graphic, serious, and fully require careful investigation. Photographs in any investigation can be (a) tremendously useful when accomplished with care, or (b) not useful, and sometimes misleading, when not accomplished according to generally-accepted rules of forensic science. The use of photographs in the survey process therefore poses important public policy challenges. To the extent that advances in technology, and laws such as the Arkansas Coroner Referral Law, become more common then we are well advised to study this area further. The CMS clinical team reviewed the photos and found that they were very helpful in a number of the investigations. The CMS team also reviewed the majority of cases that the SA did not substantiate. In a number of cases the team found that the surveyors' investigation did not result in a substantiated finding of abuse or neglect because: * The resident arrived to the NH with the pressure ulcers from other facilities, (the specific NH in which the resident resided could not have avoided the outcome). * Interventions by the primary care physician with resident families sought to address the challenges in meeting a resident's nutritional/ hydration needs without invasive lines/tubes (end of life choice decisions by individuals under hospice care, and a patient rights by regulation), * Some of the photos capture the presentation of peripheral vascular diseases, and skin conditions likely weakened by the use of certain medications (e.g. steroids) commonly used to treat arthritic, pulmonary, or cardiovascular diseases. In such cases, the meaning of the photographic information is not dispositive and can only be discerned through other investigation, including interviews with physicians. * In most cases the coroner's office did not interview physicians. In one case of graphic photos, interviews by surveyors determined that the physician had determined that the only option was bilateral amputation, a course that neither the family nor the hospice patient desired. Any one of the above scenarios can be further complicated by an individual resident's aged, demented, and/or debilitated body ravaged by several co-morbidities. The coroner's photos of significant pressure ulcers, or disease processes, are particularly useful for the coroner's purposes - determining whether there is sufficient evidence to warrant further investigation by the SA. The photos by themselves do not, however, represent sufficient evidence to render a conclusion of poor care, neglect, avoidable outcomes or that the NH caused the death. We therefore recommend that the GAO report acknowledge these facts (e.g. by inserting the phrase "potential neglect" on p. 4 when stating that photos of "...pressure sores were the predominant indication of neglect..." The investigation of complaints by the SA to substantiate (or not), and the work of the MFCU to build a case for civil or criminal prosecution, requires a variety of tools and information sources (e.g., interviews, medical record documentation) to best understand the circumstances of care or service. If, as the CMS team learned, those photos lack common requisites of forensic science then they are much less useful than could be the case. Examples of key problems include a lack of embedded photo dating, lack of embedded subject identification, lack of scale metrics and color charting in the photos lack of adequate records for interview sources, lack of interviews with residents' physicians. Some of the coroner's pictures were so close to the body part that we could not identify with certainty the body part captured. It is also unclear from the GAO draft report whether the professor of nursing had the benefit of the actual NH medical record, or the benefit of interviews conducted by the SA investigators who followed up on the referrals. In light of these issues CMS will: * Issue Guidance on Use of Photographic Evidence: CMS will undertake a study of the issues involved in the use of photographs and issue additional guidance for use by state survey agencies. We expect the guidance to the state SAs will also be useful to their partnering agencies (e.g. coroners). Pressure Ulcers: The GAO draft observes that, in about 2/3 of the coroner's referrals, pressure sores were the predominant indication of potential neglect. This points to the increasing importance of management, training and practice that can promote the effective prevention and treatment of pressure ulcers. CMS has already identified this area as an important focus for additional investment, and is undertaking the following actions: * Reducing Pressure Ulcers through Clear Goals: CMS has made the reduction in pressure ulcers in nursing homes as an important goal under the Government Performance and Results Act (GPRA): * Increasing Quality Improvement Collaborations: CMS, state Survey Agencies, and the Quality Improvement Organizations (QIOs) have embarked on collaborative efforts to increase and coordinate their respective efforts to work with nursing homes in the prevention and reduction of pressure ulcers. We appreciate the investment of time and energy that GAO devoted to reviewing the Arkansas experience. We hope in the future that GAO will be able to share its case information with us so that we may continue to investigate the many issues that this valuable case study raises. In the meantime, we are hope our comments are helpful in refining your report. Amended Attachment 1: Federal Review of AR OLTC Triage of 36 Pulaski Count Coroner's Referrals: [See PDF for image] [End of table] NOTE: Amending table #1 may slightly alter some final percentages in the text response. Any revisions do not change any of the CMS conclusions. CMS agrees with 24 of 35 (68.57%) priorities assigned by OLTC, and disagrees with 11 of 35 (31.43%). Column C - Complaint receipt dates obtained from the OLTC tracking system. Line 34 referral receipt date obtained from a note written by an OLTC employee. Line 36 referral receipt date handwritten on the coroner's report. CMS triaged all residents in participating facilities, including line 13 referral (complaint #4430) who was not in a certified bed. CMS did not triage the referral on line 36. The resident was living in a non-participating, licensed-only State Veterans Facility. Any care or service failure could not impact Medicare beneficiaries or Medicaid recipients. CMS has no jurisdiction over this facility nor State licensure actions regarding this facility. Federal Review of Five Pulaski Count Coroner's Referrals Received in 2004: [See PDF for image] [End of table] [End of section] Appendix IV: Comments from the Arkansas Department of Human Services: Arkansas Department of Human Services: Division of Medical Services: Office of Long Term Care: Mail Slot 5409: P.O. Box 8059: Little Rock, Arkansas 72203-8059: Telephone (501) 682-8487: TDD (501) 682-6789: Fax (501) 682-8551: Web Site: http://www.medicaid.state.ar.us/general/units/oltc: September 13, 2004: Walter Ochinko, Assistant Director: Medicaid and Private Health Insurance Issues: Health Care, Room 5A14: United States General Accounting Office: 441 G. Street N.W. Washington, DC 20548: RE: Arkansas Office of Long Term Care Comments to Draft GAO Report - NURSING HOME DEATHS: Arkansas Coroner Referrals Confirm Weaknesses in State and Federal Oversight of Quality of Care (GAO-04-980): Dear Mr. Ochinko: The Arkansas Department of Human Services believes that the Arkansas Coroner's Law reviewed in the report has the potential to help enhance quality of care to residents through appropriate oversight of nursing facilities. As the Arkansas Department of Human Services assisted the Pulaski County coroner in bringing the law into existence, the Department has a vested interest in its efficacy. We certainly recognize that the Pulaski County coroner has utilized the law to uncover deaths that required further examination on the part of the Office of Long Term Care. We therefore appreciate the opportunity to review the above referenced draft report and to offer comments. Such a cooperative review will no doubt ensure the accuracy of the report and help focus all parties on the issues. In that light, we tender the following observations and comments. Validity of Referral Numbers - Throughout the report the number of referrals from the Pulaski County Coroner is stated to be eighty-five (85). Further, this number is stated to be the number of referrals to both the State Survey Agency (SSA - the Office of Long Term Care) and to the State Medicaid Fraud Control Unit (MFCU) located in the Arkansas Attorney General's Office. The phrasing of the referral numbers is troublesome, and is discussed later in this letter. We offer the following examples, with page notations and suggested language that would accurately reflect the facts. The suggested language is italicized: Page: Highlight Page: Current Language: "From July 1999, when the Arkansas law took effect, through December 2003, the Pulaski County coroner referred 85 cases of suspected resident neglect to the state survey agency." Page: Highlight Page: Suggested Language: "From July 1999, when the Arkansas law took effect, though December 2003, the Pulaski County coroner claims to have referred 85 cases of suspected resident neglect to the state survey agency. The state survey agency claims that during that same time only 35 cases were referred by the Pulaski County coroner." Page: Highlight Page: Highlight Page Current Language: "Arkansas state survey agency officials told GAO that they received and investigated fewer than half of the Pulaski County coroner's referrals." Page: Highlight Page: Suggested Language: "Arkansas state survey agency officials told GAO that they received fewer than half of the referrals claimed to have been made by the Pulaski County coroner. The State survey agency investigated all of the cases that the state survey agency claimed were received from the Pulaski County coroner." Page: 18: Current Language: "According to Arkansas state survey agency officials, the agency did not receive or investigate more than half of the coroner referrals of suspected resident neglect." Page: 18: Suggested Language: "According to Arkansas state survey agency officials, the agency did not receive more than half the number of referrals that the Pulaski County coroner claimed to have made to the state survey agency. The state survey agency did investigate all referrals that the state survey agency claimed were referred by the Pulaski County coroner." Page: 19: Current Language: "Although the Pulaski County coroner told us that he had referred 85 cases of suspected resident neglect from July 1999 through December 2003, Arkansas state survey agency officials said that they had received and investigated fewer than half." Page: 19: Suggested Language: "Although the Pulaski County coroner told us that he had referred 85 cases of suspected resident neglect from July 1999 through December 2003, Arkansas state survey agency officials told us that they had received only 35 referrals from the Pulaski County coroner IN that time period, and investigated all of the 35 referred." The above statements, it should be noted, do not reflect all of the similar statements in the report. Our concern is two fold: 1. These statements can be interpreted to mean that the State Survey Agency (SSA) received all of the claimed 85 referrals, but investigated less than half. This is incorrect. The SSA investigated all referrals received from the Pulaski County Coroner. 2. Related to number 1, above, is the issue of the supposed weight given to the number of referrals that the Pulaski County coroner claims to have made. First, the language of the above examples, and throughout the report, is such that it appears that the GAO is validating the statement of the Pulaski County coroner-and thereby claiming that the numbers of referrals received by the SSA and the MFCU are false. Second, there is no evidence presented in the report to validate such a position. Quite simply, what makes the coroner's claims more credible than either the SSA or MFCU? Why is the Pulaski County coroner's number of referrals accepted at face value? Two separate agencies of the State of Arkansas claim that neither received the eighty-five (85) referrals claimed to have been made by the Pulaski County coroner. Absent some independent verification of the number, the report should be amended to reflect that the GAO was unable to obtain independent verification of the number of referrals claimed by the Pulaski County coroner. It is possible that the GAO considers the following quote from page twenty (20) of the report to be independent evidence to support the claimed number of referrals by the coroner: We found inconsistencies in agency and MFCU recordkeeping. For example, the state survey agency told us that it had received four referrals on the coroner's list but could not provide a copy of any complaint intake forms or the results of its investigations for three of the four referrals. While a MFCU official told us that three other referrals were forwarded to it by the state survey agency, not the coroner, the state survey agency had no record of these referrals. If so, and in the interest of accuracy and fairness, the recordkeeping of the coroner should be examined and any and all flaws documented. Absent such contrast, the quoted text appears to have been inserted for no other reason than to support the GAO's reliance - without independent evidence - of the claimed number of referrals of the coroner. This is particularly the case as the GAO investigator stated to the state survey agency director and staff that the coroner's recordkeeping made retrieval difficult - the impression being that it was both deficient and exceedingly worse than that of the state survey agency. In addition, the report should be amended so that any possible misinterpretation that the SSA did not investigate all referrals it received is removed. What can be established is that the Pulaski County coroner claims to have referred eighty-five cases. The SSA claims to have received thirty-five (35). The MFCU claims to have received fifty (50). There is no evidence that the SSA failed to investigate any case it claims to have received - and no independent evidence that it received more than the thirty-five (35) cases it claims. Finally, the GAO draft report fails to accurately reflect the actual circumstances of the referrals. The GAO provided the Office of Long Term Care with a list of eighty-five (85) names of cases that the Pulaski County coroner claimed to have referred to the state survey agency. As our records reflected that only thirty-five (35) of the names on the list were referrals from the Pulaski County coroner, the Office of Long Term Care at the request of the GAO performed a record review to see whether any of the remaining fifty (50) names appeared. Of the remaining fifty (50) names, twenty-two appeared in the records of the Office of Long Term Care and the Office of Long Term Care had documentation that all of the cases were investigated. While the Office of Long Term Care did not receive the eighty-five referrals claimed by the Pulaski County coroner, the Office of Long Term Care investigated fifty-seven (57) of the 85 names that appeared on the Pulaski County coroner's list. Thirty-five (35) of those cases originated from a referral by the Pulaski County coroner. The remaining twenty-two (22) originated from referrals by another source. However, of the twenty-two (22), nine (9) of the referrals were received prior to the residents' deaths and investigations were completed prior to the deaths. For the language of the GAO draft report to imply that the Office of Long Term Care investigated less than half the referrals is both inaccurate and misleading. It is noteworthy that in 2004 the Office of Long Term Care received six (6) referrals from county coroners; five of these referrals were made by the Pulaski County coroner. Of those five referrals, one (1) was referred to the Office of Long Term Care by the Pulaski County coroner thirteen (13) months after the resident's death; the remaining four were referred to the Office of Long Term Care by the Pulaski County coroner from nine to eleven months after the deaths. Several months before the GAO draft report was completed the Office of Long Term Care informed the GAO that the untimely referrals had been received from the Pulaski County coroner. The GAO investigators advised the Office of Long Term Care director and staff that they were aware that the coroner had not referred these five deaths occurring in 2003. The GAO made the determination not to include these cases in their study or to include any explanation regarding the cases untimely referrals. The Office of Long Term Care never received an explanation regarding any reason for the untimely referrals. It should be noted that the Office of Long Term Care receives thousands of complaints and Incident Reports (reports of specific long-term care facility employee maltreatment of residents) each year. The Office investigates each. To believe that the Office of Long Term Care simply disregarded some of the alleged referrals from one source - absent some independent verification - is not only illogical, it cast doubts on the ultimate conclusions stated in the report. Certainly, the fifty (50) disputed referrals did not constitute a significant increase in workload as a percentage of overall complaints; in addition, thirteen (13) of those disputed death referrals were investigated by the Office of Long Term Care through referral from another source. We would, therefore, appreciate the report being amended to correct this. Compliance with CMS Processes - The Office of Long Term Care will not address issues of whether CMS processes or regulations require change or modification, and will let CMS address those issues. This Office would like to stress, however, that for the time period of this study the Office of Long Term Care has been among the national leaders in citations for Immediate Jeopardy, Substandard Quality of Care, and Actual Harm. We certainly do not claim perfection; we don't honestly believe that can be claimed by any SSA. However, we strive at all times to comply with CMS guidelines and requirements. When we find that we are not meeting those guidelines and requirements, we take steps to do so. A review of the SAEP/SPE performed by CMS annually on SSAs will reveal that the Office of Long Term Care has either fully met CMS' expectations concerning the processes discussed in the report, or took the necessary steps to meet them. While the Office of Long Term Care cannot claim perfection, it is our opinion that the draft report - by addressing issues outside the apparent scope of the study, such as survey predictability - unfairly paints a picture of the state survey agency and its efforts. It should be noted that the State of Arkansas, in addition to its adherence to CMS requirements and guidelines, has taken independent affirmative action to both improve the survey process and to improve quality of care of residents. These include: l. Increase in survey staff positions. In 2001, The Arkansas Department of Human Services made a commitment to increase survey staff numbers for the Office of Long Term Care, and carried through with that commitment. This influx of additional surveyor positions meant that the Office of Long Term Care oversight and compliance determinations of facilities would not only be maintained, but would be strengthened. 2. Increases in the reimbursement to nursing homes. The State of Arkansas has approximately doubled the amount of money paid for the care of residents in nursing homes. A significant percentage of that money - in excess of half - has gone to payment of direct care staff. Facilities are now able to pay direct care staff more, and to provide benefits that they were unable to provide in the past. This effort was made to assist facilities in their retention of staff and to fight the problems that come from both a lack of staff and high turnover. 3. Related to the increased reimbursement, the State of Arkansas mandated minimum staffing requirements through legislation passed in the 2001 legislative session. That law - found in Ark. Code Ann. § 20- 10-1401 et seq. - has resulted in increases in direct care staffing no less than three times. Arkansas nursing home residents are now being served by more direct care staff, and facilities can now provide pay and benefits that will help retain that staff. 4. The Office of Long Term Care, through the Arkansas Department of Human Services, is finalizing a contract with the Arkansas Foundation for Medical Care (AFMC) for an innovative program to provide evaluation and training to both facilities and the Office of Long Term Care. The AFMC is the Quality Improvement Organization (QIO) for the State of Arkansas. Under this contract, the AFMC will be provided data - including survey documentation - to evaluate and locate areas in which it can offer assistance to facilities. Likewise, the AFMC will evaluate the Office of Long Term Care, and provide training to the Office in an effort to further improve the Office's ability to perform its duties in surveying facilities. 5. It is unclear how survey predictability is related to a state law that can result in referrals for complaints. Complaint investigations are conducted based upon strict guidelines that in turn are based on the alleged seriousness and immediacy of harm - these timelines are unrelated to surveys. Nevertheless, it should be noted that the state survey agency has made successful efforts to reduce predictability of surveys. However, the strict guidelines for investigation of complaints means that there is very little that can be done to reduce predictability of complaint investigations. If the allegation is serious, the Office of Long Term Care must investigate within a few days - if a facility is aware that the complaint has been made, it follows that it will be able to estimate when the state survey agency will appear with a high degree of accuracy. Again, however, we do not understand how survey predictability concerns arise from a state law that results in complaint investigations. Thank you for this opportunity to provide our comments and suggestions for this study. If you should have any questions about the statements contained in this letter, please contact me. Sincerely, Signed by: Carol Shockley, Director: [End of section] Related GAO Products: Nursing Home Fire Safety: Recent Fires Highlight Weaknesses in Federal Standards and Oversight. GAO-04-660. Washington, D.C.: July 16, 2004. Nursing Home Quality: Prevalence of Serious Problems, While Declining, Reinforces Importance of Enhanced Oversight. GAO-03-561. Washington, D.C.: July 15, 2003. Nursing Homes: Public Reporting of Quality Indicators Has Merit, but National Implementation Is Premature. GAO-03-187. Washington, D.C.: October 31, 2002. Nursing Homes: Quality of Care More Related to Staffing than Spending. GAO-02-431R. Washington, D.C.: June 13, 2002. Nursing Homes: More Can Be Done to Protect Residents from Abuse. GAO- 02-312. Washington, D.C.: March 1, 2002. Nursing Homes: Federal Efforts to Monitor Resident Assessment Data Should Complement State Activities. GAO-02-279. Washington, D.C.: February 15, 2002. Nursing Homes: Success of Quality Initiatives Requires Sustained Federal and State Commitment. GAO/T-HEHS-00-209. Washington, D.C.: September 28, 2000. Nursing Homes: Sustained Efforts Are Essential to Realize Potential of the Quality Initiatives. GAO/HEHS-00-197. Washington, D.C.: September 28, 2000. Nursing Home Care: Enhanced HCFA Oversight of State Programs Would Better Ensure Quality. GAO/HEHS-00-6. Washington, D.C.: November 4, 1999. Nursing Homes: HCFA Should Strengthen Its Oversight of State Agencies to Better Ensure Quality of Care. GAO/T-HEHS-00-27. Washington, D.C.: November 4, 1999. Nursing Home Oversight: Industry Examples Do Not Demonstrate That Regulatory Actions Were Unreasonable. GAO/HEHS-99-154R. Washington, D.C.: August 13, 1999. Nursing Homes: HCFA Initiatives to Improve Care Are Under Way but Will Require Continued Commitment. GAO/T-HEHS-99-155. Washington, D.C.: June 30, 1999. Nursing Homes: Proposal to Enhance Oversight of Poorly Performing Homes Has Merit. GAO/HEHS-99-157. Washington, D.C.: June 30, 1999. Nursing Homes: Complaint Investigation Processes in Maryland. GAO/T- HEHS-99-146. Washington, D.C.: June 15, 1999. Nursing Homes: Complaint Investigation Processes Often Inadequate to Protect Residents. GAO/HEHS-99-80. Washington, D.C.: March 22, 1999. Nursing Homes: Stronger Complaint and Enforcement Practices Needed to Better Ensure Adequate Care. GAO/T-HEHS-99-89. Washington, D.C.: March 22, 1999. Nursing Homes: Additional Steps Needed to Strengthen Enforcement of Federal Quality Standards. GAO/HEHS-99-46. Washington, D.C.: March 18, 1999. California Nursing Homes: Federal and State Oversight Inadequate to Protect Residents in Homes With Serious Care Problems. GAO/T-HEHS-98- 219. Washington, D.C.: July 28, 1998. California Nursing Homes: Care Problems Persist Despite Federal and State Oversight. GAO/HEHS-98-202. Washington, D.C.: July 27, 1998. FOOTNOTES [1] Ark. Code Ann. § 5-28-204 (Michie 2003). [2] In the absence of autopsy information that establishes the cause of death, we were unable to determine the extent to which unacceptable care may have contributed directly to individual deaths. See GAO, California Nursing Homes: Care Problems Persist Despite Federal and State Oversight, GAO/HEHS-98-202 (Washington, D.C.: July 27, 1998). [3] See GAO, Nursing Homes: Sustained Efforts Are Essential to Realize Potential of the Quality Initiatives, GAO/HEHS-00-197 (Washington, D.C.: Sept. 28, 2000) and Nursing Home Quality: Prevalence of Serious Problems, While Declining, Reinforces Importance of Enhanced Oversight, GAO-03-561 (Washington, D.C.: July 15, 2003). [4] Starting in August 2003, Missouri nursing homes were required to report resident deaths to county officials, such as coroners. The Missouri law, however, does not require coroner investigations of the deaths. See Mo. Ann. Stat. § 198-071 (West 2004). [5] A list of related GAO products is at the end of this report. [6] Arkansas has two state survey agencies--the Office of Long Term Care in the Department of Human Services and the Division of Health Facility Services in the Department of Health. The former is responsible for surveying nursing homes and the latter surveys other providers who participate in Medicare and Medicaid, such as hospitals and home health agencies. In this report, we use the term state survey agency to refer to the Office of Long Term Care. [7] CMS generally interprets these requirements to permit a statewide average interval of 12.9 months and a maximum interval of 15.9 months for each home. [8] Death investigations often vary considerably by jurisdiction (whether state, county, district, or city). Some states use a medical examiner (21 states and the District of Columbia), some use a coroner (11 states), and some use a mixed system of medical examiners and coroners (18 states). Medical examiners and coroners are responsible for investigating sudden or violent deaths and for providing accurate, legally defensible determinations of the causes of these deaths. Generally, medical examiners are licensed physicians and are appointed, while coroners need not be physicians and are elected. [9] When enacted, the Arkansas law required a referral if there was reasonable cause to suspect that the resident died of abuse, sexual abuse, or neglect. In 2003, the law was amended to substitute maltreatment for these terms. Coroner referrals did not actually characterize the specific nature of each finding in relation to one of the statutory categories for referral. In the absence of such characterization, we characterize each referral under the law as based on a finding of neglect. [10] Most states have laws that require suspicious or unusual deaths (or those for which the cause is unknown or unnatural) to be reported to a state or local authority, and some specifically require the reporting of deaths resulting from abuse or neglect. Prior to 1999, Arkansas law required the reporting of cases in which there was reasonable cause to suspect that any adult had died of abuse, sexual abuse, or negligence. [11] MFCUs were authorized by the Medicare-Medicaid Anti-Fraud and Abuse Amendments, Pub. L. No. 95-142 §17, 91 Stat. 1175, 1201-1202 (1977). Currently, 47 states and the District of Columbia participate in the Medicaid fraud control grant program. [12] According to the state survey agency, only four referrals were received from coroners outside of Pulaski County, and we excluded these from our analysis. We did not contact Arkansas's 74 other coroners to determine whether any additional referrals were sent. Although assessing the effectiveness of the state's law was beyond the scope of our review, MFCU officials told us that few other coroners investigate nursing home resident deaths and that nursing homes may not be reporting all deaths to their local coroners as the state law requires. For example, MFCU officials told us that there were eight deaths in one home in the course of 1 month that were not reported to the coroner or investigated and at least one decedent was sent to a funeral home owned by the coroner. The Arkansas statute does not provide sanctions for failure to report nursing home deaths to coroners or for coroners' failure to investigate reported deaths. They also told us that all but two of the state's 75 county coroners are elected; therefore, most state coroners are not accountable to other county or state officials. The Pulaski County coroner is appointed by the county's chief executive officer. [13] Two of the coroner's three staff members are licensed paramedics. [14] Although the referrals sometimes identified multiple care problems, we attempted to identify the primary cause for each of the coroner's 86 referrals. Overall, 88 percent of decedents with pressure sores had stage III/IV pressure sores or necrotic or gangrenous tissue (see table 3). Fifty-seven percent of decedents with pressure sores had three or more pressure sores. [15] The risk factors for pressure sores include confinement to a bed or chair, inability to move, loss of bowel or bladder control, poor nutrition, and lowered mental awareness. Actions to prevent pressure sores include repositioning the patient every 1 to 2 hours; using a special pressure-relieving mattress or chair pad; placing pillows or wedges between the knees and ankles and under legs to keep the patient's heels off of the bed; cleaning skin as soon as possible after incontinence; and providing appropriate nutritional support. [16] All but 5 of the 27 homes referred by the coroner were located in Pulaski County. The residents from these 5 homes died in a Pulaski County medical facility and, as a result, were referred by the Pulaski County coroner. Three of the 27 homes with coroner referrals have since closed. [17] The body of a resident who died in this same home prior to enactment of the 1999 Arkansas law was exhumed and the decedent was found to have suffocated while tied to his nursing home bed. [18] One of the 19 homes is a federal facility operated by the Department of Veterans Affairs and is not subject to surveys by the state survey agency. [19] We excluded from our analysis cases for which a coroner's referral was not received but the state survey agency indicated it had conducted an investigation, primarily complaints filed by family members or others. We excluded such cases because the focus of our analysis was the state's disposition of coroner referrals, not a broader review of the state's disposition of all complaints, regardless of source. Nine of the survey agency's non-coroner complaint investigations were conducted prior to the residents' deaths and may not have raised concerns similar to those identified in the coroner's referrals. Elsewhere in the report, we acknowledge seven of the survey agency's non-coroner complaint investigations that involved allegations similar to the coroner's. [20] To help both the state survey agency and the MFCU identify all coroner referrals made since July 1999, we provided a list that we developed using the Pulaski County coroner's files. Both agencies used this list to identify coroner referrals they received but were unable to locate all 86 referrals. [21] Five of the 27 homes, where the coroner identified 10 cases of potential neglect, had no state survey agency or MFCU investigations. [22] In March 2004, the coroner began requesting signed receipts. [23] In addition, the state survey agency substantiated two non-coroner complaints for decedents the coroner said he referred but which agency officials indicated were not received. In one case, a family member filed a complaint 6 days after a resident's death with allegations similar to those in the coroner's referral. The resident broke both hips when she fell out of bed. The state survey agency investigated the family member's complaint twice. According to state survey agency officials, a review of the initial investigation, which cited misuse of restraints at the less than actual harm level, indicated the need for another investigation. The second investigation cited two actual harm deficiencies for shortcomings in resident assessment and failure to prevent accidents. In the other case, state surveyors were at the nursing home when a resident, attempting to burn off his restraints, set himself on fire. Surveyors cited the home with several deficiencies at the immediate jeopardy level. [24] Although the state survey agency recommended termination of this home in October 2000, CMS's Dallas regional office imposed a directed plan of correction that included requirements that the home reduce the number of Medicare and Medicaid residents by 50 percent within about 2 weeks and hire independent third-party consultants in the areas of nursing services, pharmacy services, medical records and documentation, behavioral intervention, and quality assurance, as well as correct all conditions of immediate jeopardy. This approach gave the home significant leeway in returning to compliance. For example, the state survey agency was given the discretion to keep the home open if it showed good faith in removing immediate jeopardy. However, the home did not meet the terms of the directed plan of correction and thus was terminated in early November 2000. The home reopened under new ownership, new management, and a new name in July 2001 but did not begin receiving Medicaid payments until June 2002. [25] The decedents' deaths occurred from March 25, 2000, through April 13, 2000, and the state survey agency received the coroner's referral for all four cases on April 25, 2000. [26] A similar October 1999 complaint by family members was not substantiated. Overall, at least 25 percent of the decedents referred by the coroner were also the subject of complaints by family members or others. [27] The state survey agency recommended a $10,000 civil monetary penalty. CMS reduced the penalty to $2,000, which the facility paid. [28] Although federal guidance sets a high threshold of immediate jeopardy for citing past noncompliance, the Arkansas state survey agency's complaint investigation guidance indicates that past noncompliance may be cited whenever the violation resulted in actual harm or immediate jeopardy to a resident. [29] Nationwide, past noncompliance appears to be rarely used, cited in less than 1 percent of standard surveys and less than 1 percent of complaint investigations. During the last 4 standard surveys for each nursing home nationwide, 204 instances of past noncompliance were cited on about 63,000 surveys. Overall, about half of the state survey agencies cited past noncompliance. The Arkansas state survey agency accounted for about 10 percent of such citations. [30] The state survey agency investigated but did not substantiate non- coroner complaints for five decedents the coroner said he referred and agency officials indicated they did not receive. The allegations in the non-coroner complaints were similar to those contained in the Pulaski County coroner's referrals. In one case, the survey agency referred the complaint to the MFCU that requested an exhumation of the decedent's body for an autopsy. Before the autopsy results were obtained, the survey agency determined that the complaint was unsubstantiated. In a second case, the survey agency received the complaint alleging a fall 3 weeks before the resident's death; the complaint was investigated 6 months after the resident's death but without the benefit of the coroner's photos of the decedent's bruises. For a third case, nursing home staff filed two complaints before the resident's death alleging poor pressure sore care. When he died, the resident had 12 pressure sores, but again, surveyors lacked the coroner's photos of the decedent. [31] Although neither the MFCU nor the state survey agency substantiated the alleged neglect for 8 of the same 14 referrals, we believe that several factors raise questions about the thoroughness of some MFCU investigations. In 2000, MFCU investigators were authorized to declare cases inactive and some cases were closed on the basis that medical records documented the receipt of necessary care, without a thorough review of the records by a registered nurse. (The MFCU now employs two nurse investigators who typically perform a review of medical records intended, in part, to identify inconsistencies and gaps in documentation of resident care.) In addition, the MFCU did not pursue every case it received, citing the difficulty of proving that neglect by a facility was the direct, natural, or probable cause of a resident's condition and because the agency's resources were limited. [32] To support its "not substantiated" finding, the state survey agency cited several factors, including documentation that the facility was following the plan of care, the fact that the pressure sores were reported to have developed in the hospital, or that the family wanted to be conservative in the care provided. Because of concern about the basis for some "not substantiated" findings, we asked our expert to review seven cases in which the seriousness of the decedents' conditions as documented in the coroner's photos raised a question about the validity of the conclusions reached during the state survey agency's investigations. This assessment was based on a review of the various investigative reports, medical records we obtained, and photos of decedents taken by the coroner. All of the decedents had serious pressure sores, and four referrals involved two nursing homes. In two of the seven cases reviewed, our expert found that there was not enough documentation to draw a definitive conclusion. [33] Pressure sores can be painful. For example, a physician more than quadrupled the amount of pain medication for one decedent over about a two and one-half month period because of pressure sores at the base of her spine. We found that pain management was a problem in other coroner referrals. For example, the medical records associated with one coroner referral noted that the resident had complained to her daughter of foot pain. When the daughter removed her mother's shoe and sock she found bloody toes from pressure sores that the home had failed to document. Two other decedents did not receive pain medication as prescribed. [34] The state's February 2000 survey was conducted to allow this nursing home to again serve Medicaid beneficiaries. The home had been terminated from participation in the Medicaid program in January 2000 for poor performance after an October 1999 survey that found actual harm and immediate jeopardy deficiencies in quality of care. [35] GAO, Nursing Homes: Complaint Investigation Processes Often Inadequate to Protect Residents, GAO/HEHS-99-80 (Washington, D.C.: Mar. 22, 1999). [36] See Appendix III, amended attachment I to CMS comments. [37] CMS guidance instructs state survey agencies to establish complaint prioritization time frames for serious complaints in terms of working days, not calendar days. If a complaint judged to be immediate jeopardy was received on a Saturday, the survey agency would not be expected to initiate its investigation until Tuesday, 4 days after receipt of the complaint. [38] State survey agency officials were unable to identify the investigation priority for 2 of the 36 coroner referrals. However, over 3 months elapsed between the time the state survey agency received and investigated one of these referrals. For the second referral, the survey agency could not identify the date of receipt, but nevertheless completed its investigation within 12 working days of the resident's death. [39] Our analysis includes 35 of the 36 coroner referrals because the survey agency was unable to provide the date of receipt for 1 referral. [40] GAO/HEHS-98-202, GAO/HEHS-00-197, and GAO-03-561. [41] See GAO-03-561. This analysis was based on states' most recent surveys in OSCAR as of April 9, 2002 and represents a reduction from prior surveys when about 45 percent of Arkansas's standard surveys were predictable (38 percent nationwide). [42] In contrast, fewer surveys nationally were predictable for the former (13 percent) than the latter (21 percent) reason. [43] GAO/HEHS-98-202. [44] GAO/HEHS-00-197. [45] GAO/HEHS-98-202. [46] In 1999, we reported that CMS guidance on past noncompliance did not require the imposition of a sanction, even for a deficiency that contributed to the death of a resident. CMS concurred with our recommendation to revise its guidance and on May 28, 2004, instructed state survey agencies to impose a civil monetary penalty when citing past noncompliance. See GAO, Nursing Homes: Additional Steps Needed to Strengthen Enforcement of Federal Quality Standards, GAO/HEHS-99-46 (Washington, D.C.: Mar. 18, 1999). [47] No plan of correction is required because the deficiency is assumed to have been corrected and no longer exists. However, CMS could require the facility to document how it discovered the deficient practice and the corrective action it took. [48] A portion of CMS's comments was based on tables presented in attachment 1 to its comments. Because the tables did not accurately reflect the coroner's cases discussed in our report, CMS submitted an amended attachment 1 which we have substituted for the original. CMS, however, did not make corresponding changes on pages 6 and 7 of its comments. [49] GAO-03-561. [50] CMS officials told us that the pressure sore guidance is expected to be released before the end of 2004. [51] Our elapsed time calculation differs from that of CMS because we relied on copies of signed receipts provided by the coroner. These receipts indicated that the state survey agency received all of these referrals either on April 13, 2004, or on April 14, 2004, rather than on the dates indicated by CMS in amended attachment 1 to their comments. We believe that the approximately 2-week disparity between the dates shown on the signed receipts and the dates that the survey agency said it received four of these referrals raises a question about how promptly the survey agency registers complaints in its tracking system. Because the coroner did not begin requiring signed receipts for referrals of suspected neglect until March 2004, we were unable to determine if there were similar delays in registering the 36 coroner referrals received prior to 2004. [52] The coroner informed us that these five referrals were delayed while awaiting final autopsy reports, which can take 8 to 9 months to complete. [53] These averages and ranges differ from those CMS provided in its comments because CMS included the five 2004 coroner referrals that were outside the scope of our review. [54] Arkansas state survey agency officials told us that they did not investigate one coroner referral they had received. We excluded this referral from those received by the survey agency. [55] Although the state survey agency said it received coroner referrals for 2 of the 5 cases, we excluded the two from our analysis of referrals investigated by the state survey agency because it could provide no documentation of its investigation, including the outcome. [56] See GAO/HEHS-98-202, GAO/HEHS-00-197, and GAO-03-561. GAO's Mission: The Government Accountability Office, the investigative arm of Congress, exists to support Congress in meeting its constitutional responsibilities and to help improve the performance and accountability of the federal government for the American people. GAO examines the use of public funds; evaluates federal programs and policies; and provides analyses, recommendations, and other assistance to help Congress make informed oversight, policy, and funding decisions. GAO's commitment to good government is reflected in its core values of accountability, integrity, and reliability. Obtaining Copies of GAO Reports and Testimony: The fastest and easiest way to obtain copies of GAO documents at no cost is through the Internet. GAO's Web site ( www.gao.gov ) contains abstracts and full-text files of current reports and testimony and an expanding archive of older products. The Web site features a search engine to help you locate documents using key words and phrases. You can print these documents in their entirety, including charts and other graphics. Each day, GAO issues a list of newly released reports, testimony, and correspondence. GAO posts this list, known as "Today's Reports," on its Web site daily. The list contains links to the full-text document files. To have GAO e-mail this list to you every afternoon, go to www.gao.gov and select "Subscribe to e-mail alerts" under the "Order GAO Products" heading. Order by Mail or Phone: The first copy of each printed report is free. Additional copies are $2 each. A check or money order should be made out to the Superintendent of Documents. GAO also accepts VISA and Mastercard. Orders for 100 or more copies mailed to a single address are discounted 25 percent. Orders should be sent to: U.S. Government Accountability Office 441 G Street NW, Room LM Washington, D.C. 20548: To order by Phone: Voice: (202) 512-6000: TDD: (202) 512-2537: Fax: (202) 512-6061: To Report Fraud, Waste, and Abuse in Federal Programs: Contact: Web site: www.gao.gov/fraudnet/fraudnet.htm E-mail: fraudnet@gao.gov Automated answering system: (800) 424-5454 or (202) 512-7470: Public Affairs: Jeff Nelligan, managing director, NelliganJ@gao.gov (202) 512-4800 U.S. Government Accountability Office, 441 G Street NW, Room 7149 Washington, D.C. 20548: