DOCUMENT CHANGE RECORD
Version Number	Date	Description
1	8/17/2007	Initial Draft of Requirements Analysis
2	8/27/2007	Draft 2 of Requirements Analysis
3	9/28/2007	Draft 3 of Requirements Analysis

1. INTRODUCTION

Today, effective coordination and delivery of care across care settings is limited by medical records (both paper and electronic) that are setting-specific and by an outdated reliance on manual processes to communicate relevant health care information across care settings. Clinicians have limited access to critical clinical information on individual patients beyond the records of the institution in which he or she provides care. In short, clinicians are effectively “blind” to a patient’s course of treatment once that individual leaves a specific site of care. These limitations can significantly impact a provider’s ability to make optimal care decisions. They also negatively impact the ability to transition patients effectively across care settings, an important issue when considering patients with multiple chronic conditions. As a result of these issues, the transition of patients between settings is characterized by a lack of continuity that directly impacts the quality of care a patient receives.

All stakeholders in the health care industry – from consumers to accrediting organizations – share the need to make informed decisions about quality of care. There is a sense of passion across all stakeholder groups for improving the quality of care for all patients. To that end, each group of stakeholders has initiated actions, largely specific to a narrow business objective and within an individual health delivery segment, to meet their need to measure quality of care. Current segment-specific efforts are important; however, this siloed approach to quality improvement has a serious unintended consequence in that it continues to promote fragmented care delivery. At its best, a fragmented approach cannot address the pervasive challenges related to coordination of care.

The American Health Information Community (AHIC)¹chartered the Quality Workgroup. The Quality Workgroup, chartered in August of 2006, is charged with determining how health information technology (HIT) can be used for the development of quality measures helpful to patients and others in the health care industry. HIT systems are important tools for data collection that can support performance improvement on those quality measures. The members of the Quality Workgroup represent hospitals, hospital associations, health plans, vendor organizations, employers, consumers, and federal agencies.

Broad Charge for the Quality Workgroup:

Make recommendations to the American Health Information Community so that health IT can provide the data needed for the development of quality measures that are useful to patients and others in the health care industry, automate the measurement and reporting of a comprehensive current and future set of quality measures, and accelerate the use of clinical decision support that can improve performance on those quality measures. Also, make recommendations for how performance measures should align with the capabilities and limitations of health IT.

Specific Charge for the Quality Workgroup:

Make recommendations to the American Health Information Community that specify how certified health IT should capture, aggregate and report data for a core set of ambulatory and inpatient quality measures.

The Quality Workgroup envisioned a future based on the collection, aggregation and analysis of longitudinal data to evaluate and improve the quality of patient care across care settings.

Information technology and the sharing of health information across a network of regional health information entities using data from electronic health records (EHR), personal health records (PHR), and strong clinical decision support (CDS) systems will assist providers in ensuring that the right care is delivered to the right patient - every time. Consumers and policymakers will use these same systems to understand how well the nation as a whole and individual providers are doing in improving care and health status in accordance with the national, regional, and local priorities. The Quality Workgroup envisioned a future where transparent reporting of quality performance and quality improvements is used to inform decisions made about patient care. That vision is one of a health care system that is information-driven and patient-focused. Thus, the vision incorporates an expectation that quality measurement and improvement activities will evolve from a site-centric focus to a patient-centric focus. This evolution will require new efforts to collect, aggregate and analyze longitudinal data to evaluate and improve the quality of patient care over defined time periods and across care settings. Moreover, the future state is strongly informed by the perspective that the use of HIT to improve quality is about both reducing reporting burden and driving improvements in care when it is delivered. The Quality Workgroup’s vision articulates a future where HIT:

Contributes to an improved care delivery system – one that is more connected, coordinated, safer and efficient;
Helps patients and providers know exactly how they are performing relative to standard measures of best-practice quality care (both at the point of care and longitudinally); and
Continuously enables quality improvement, including both as it relates to clinical care delivery and to cost efficiency of the systems supporting care delivery.

As discussed in the Institute of Medicine (IOM) report “Performance Measurement, Accelerating Improvement,” quality measurement and reporting across care settings and of long-term outcomes wll require a shift from the current state of assessing and reporting on the delivery of care at one point in time in a specific setting to a future state of assessing quality of care across the spectrum of care settings and over time.² This shift is vital to ensuring that providers have all of the information needed to make informed treatment decisions for their patients. This shift also helps to inform improvements to care delivery across the spectrum of care settings in which a patient may be seen for a given illness.

This view point was recognized in the Quality Workgroup’s vision, along with the assertion that HIT plays a critical role in enabling longitudinal quality assessment and improvement through the integration of data from multiple sources across multiple institutions. A transition from sole-source data collection to multi-source data exchange will occur as the industry increases its use of EHRs and gains experience with data source integration. This transition will necessarily entail far more extensive collaboration between the quality enterprise and the HIT community. Simply digitizing current paper-based records will not be sufficient. However, current progress is slow and there is an opportunity to facilitate this transition through the promotion of effective policies, studies, and other critical activities.

The Quality Workgroup identified the capabilities and policies needed to realize the envisioned patient-centric quality improvement environment and identified key areas for the workgroup to consider for future AHIC recommendations.

This document presents the results of an analysis performed by the Quality Workgroup to define capabilities and policies needed to enable a health care system where patient-centric care and quality improvement are the cornerstones, and where reporting and quality improvement is based on secure and appropriate access to a patient’s longitudinal record of care. The results of the analysis are not intended to prescribe specific solutions. Rather, they are meant to initiate a dialogue within the Quality Workgroup that will lead to their next set of recommendations to the AHIC. Methods used to inform this requirements analysis are described in Appendix A: Methods.

The document is organized into the following sections:

Section 2, Longitudinal Quality Measurement and Improvement: A Case Study. This section sets the context for understanding the requirements of the future quality reporting environment. It illustrates the need for HIT that supports access to longitudinal care records by presenting a chronic illness-based case study that describes how care delivery and quality improvement activities could be improved.
Section 3, Current and Emerging Longitudinal Quality Measurement and Improvement Strategies. In this section, we describe the current strategies aimed at conducting longitudinal analysis. This section also describes emerging strategies to improve the HIT infrastructure used to conduct quality improvement.³
Section 4, Requirements for Realizing the Envisioned Longitudinal Quality Measurement. This section articulates the high-level requirements for realizing the envisioned future state of quality measurement. Requirements are organized into the five key components and include requirements for the Business Case, Measure Development, Data Standards, Data Exchange and Aggregation, and Policy to facilitate future Quality Workgroup activities.
Section 5, Unresolved Critical Path Requirements. Section 5 presents four critical path scenarios that describe various options for the future, where future state requirements are unclear or there where there is disagreement in the industry on what path the future should take.
Section 6, Next Steps. Finally, Section 6 describes the next steps to developing recommendations for the AHIC and Secretary of Health and Human Services (HHS) based on the requirements analysis.

2. LONGITUDINAL QUALITY MEASUREMENT AND IMPROVEMENT: A CASE STUDY

HIT is intended to facilitate the exchange and use of health information to assist providers and ensure the right care is delivered to the right patient – every time. Ideally, all stakeholders would be able to use information generated by transparent and interoperable HIT systems to understand how well the healthcare industry is delivering high quality, efficient, and effective clinical care. HIT can facilitate this assessment at the national, regional, local, or provider level. More importantly, it can facilitate this assessment from a patient-centered perspective, i.e., one that looked across care settings and over time.

Although HIT has been heralded as a critical enabler in driving quality care, efficiency and patient safety, the industry is in the early stages of HIT adoption and HIT penetration across care settings is uneven. Additionally, the drivers, measures, standards, interoperability requirements, and policies that are needed to enable effective HIT-facilitated quality measurement and improvement activities have not yet been established on a consistent basis across the country. Current HIT applications have been developed largely in isolation of quality measurement and improvement organizations, and the latter entities have not yet incorporated the strengths of health IT into their operational processes. As a result, our ability to support quality at the point of care, measure and report on quality, and, most important, to enable longitudinal analysis of quality data across care settings and over time is hobbled.

To showcase the need for enabling longitudinal measurement, this section provides a case study using Diabetes to describing how the care delivery and quality improvement activities are currently enabled through HIT, and then describes how both of those functions could be improved if the envisioned future state is realized.

Diabetes provides an example of a chronic illness for which quality measurement from a longitudinal perspective would significantly increase the ability to manage the patient’s treatment and ensure high quality care is being delivered.

Diabetes is a chronic illness characterized by complications such as diabetic foot disorder and diabetic retinopathy, in addition to a number of co-morbidities such as heart failure and hypertension. It is often managed by multiple providers through multiple visits, and in a variety of care settings.

2.1 THE CURRENT STATE OF DIABETES MANAGEMENT

In this case study, where a multi-disciplinary management team is caring for a diabetic patient, it is critical for each provider to have a comprehensive clinical picture of the patient. In today’s system, a provider has access only to the information that resides in that care setting/institution, as stored in a patient’s medical record and other information relayed directly by the patient. Primary care of diabetic patients requires access to and management of complex clinical information. In today’s environment, the provider typically may not be able to access the course of treatment prescribed by other providers that have seen the patient, the outcome of that treatment, the additional medications prescribed by those providers, and the details surrounding other complications for which that patient may be seeing specialists or other providers of care. Additionally, the primary care provider has no way of knowing that a patient has been admitted for acute episodes unless this information is relayed directly by the patient, and the clinical details associated with an emergency department visit or hospitalization may not be available either. A comprehensive clinical picture that includes critical pieces of information arms the healthcare provider with the relevant points to inform clinical decision-making. Absent a full clinical picture, tests and procedures may be duplicated or diagnoses delayed.

Over the course of time, an individual patient may be routinely seen at his primary care practice, and the medical record at that physician practice would reflect key indicators such as the patient’s glucose management, current medications, and the development of any complications. In the event of an acute event such as diabetic ketoacidosis, the patient may present at an emergency department and subsequently be admitted. For this event, the patient could have a separate medical record that reflects the episode of care that ends with discharge from the hospital.

Today, the level of information shared from this hospital to the patient’s primary care physician is inconsistent and reliant on a number of manual processes. Likewise, for care and treatment of the patient’s kidney failure or heart disease the patient may be followed by a nephrologist and a cardiologist respectively. The level of information shared among these physicians, as well as the availability of information such as the hospitalization, is also inconsistent and often not available in a comprehensive manner to optimize clinical decision making.

The quality measurement and reporting of the care and treatment for this diabetic patient also occurs in a silo today. Individual physicians may report quality indicators that are related to the management of this diabetic patient, while the health plan computes separate and distinct quality indicators. Figure 2 below illustrates the information flows representing the current state diabetes care.

Figure 1. Information Flow in the Current State for a Patient with Diabetes

Image Description: The diagram is divided by a horizontal line into two parts: Care Processes and Data Flows. Situated on the line, at the far left, is an oval labeled Patient. To its right are five boxes labeled Periodic Primary Care Visits, Nephrologist Slash Dialysis, Periodic Cardiologist Visits, ER Visit, and Hospitalization. Periodic Primary Care Visits and Nephrologist Slash Dialysis are each joined to boxes directly below labeled Medical Record Slash EHR, and ER Visit and Hospitalization share a third box labeled the same. Periodic Cardiologist Visits is joined to a box directly below labeled Medical Record. At the bottom of the diagram are two boxes, both labeled Quality Organization. The diagram key translates four types of arrows. Four arrows for Patient Visits Across Care Settings run on the Care Processes side from Patient to Periodic Primary Care Visits, Nephrologist Slash Dialysis, Periodic Cardiologist Visits, and ER Visit. Another such arrow runs on the Care Processes side from ER Visit to Hospitalization. Five arrows for Patient Data Inputs into Care Process run from Patient to the three Medical Record Slash EHR boxes, Medical Record, and a box directly below labeled Pharmacy. Three arrows for Clinical and Administrative Data Flow run from the two Medical Record Slash EHR boxes on the left and Medical Record to a box at the bottom of the diagram labeled Quality Organization. Another such arrow runs from the Medical Record Slash EHR on the right to a second Quality Organization box at the bottom. Three arrows for Feedback Data Flow run from the left Quality Organization box to Periodic Primary Care Visits, Nephrologist Slash Dialysis, and Periodic Cardiologist Visits. Two more such arrows run from the right Quality Organization box to ER Visit and Hospitalization.

2.2 The Future State of Diabetes Management

A key aspect of the ideal future state, as articulated in the Quality Workgroup’s vision, is that the national quality enterprise will be one in which patient-centric care is a reality and HIT is used to enable the collection, aggregation, and analysis of longitudinal data to evaluate and improve the quality of patient care over defined time periods and across care settings.

The ability to collect and evaluate longitudinal data is critical to providing effective care to patients. Patients with chronic illnesses and, in many cases, those with acute illnesses, are usually seen by multiple providers, often concurrently. Coordination of care through the transfer of critical clinical information in the patient’s medical history across settings is vital to ensuring that each provider and patient is able to make informed decisions regarding the patient’s care and treatment. The evaluation of medical treatment from a longitudinal perspective is also crucial to ensuring improvements to the system in a manner that puts the patient first.

Figure 2 presents an example of a diabetes patient in the future state who is seeking care for his or her condition over time and shows the associated flows of data that would result. The patient is managed by a multi-disciplinary team for eye and foot exams, and for monitoring blood levels, such as HbA1c, as well as for management of any number of co-morbid conditions. Additionally, the patient experiences a hospitalization due to diabetes-related complications such as heart disease, stroke, kidney failure, blindness, and/or circulation problems. Information flows interoperably between providers and from providers to the patient (including the patient’s personal health record (PHR)), to the pharmacy, to quality organizations and to health information exchanges (HIEs).

Figure 2. Information Flow in the Future State for a Patient with Diabetes

Image Description:The diagram is divided by a horizontal line into two parts: Care Processes and Data Flows. Situated on the line, at the far left, is an oval labeled Patient. To its right are five boxes labeled Periodic Primary Care Visits, Periodic Ophthalmologist Visit, Periodic Podiatrist Visits, ER Visit, and Hospitalization. Patient is joined to a box directly below labeled PHR. Periodic Primary Care Visits, Periodic Ophthalmologist Visit, and Periodic Podiatrist Visits are each joined to boxes directly below labeled EHR, and ER Visit and Hospitalization share a fourth box labeled the same. Below all of these are two boxes labeled Pharmacy sandwiching a box labeled HIE. The diagram key translates four types of arrows. Four arrows for Patient Visits Across Care Settings run on the Care Processes side from Patient to Periodic Primary Care Visits, Periodic Ophthalmologist Visit, Periodic Podiatrist Visits, and ER Visit. Another such arrow runs on the Care Processes side from ER Visit to Hospitalization. Four arrows for Patient Data Inputs into Care Process run from Patient to the four EHR boxes. Other such arrows run from PHR to the two Pharmacy boxes, from the Pharmacy boxes to HIE, from the two left EHR boxes to the left Pharmacy box, from the two right EHR boxes to the right Pharmacy box, from the four EHR boxes to HIE, and from HIE to a box at the bottom labeled Quality Organization. Three arrows for Clinical and Administrative Data Flow run from the two EHR boxes on the left and the one at the far right to PHR. Five arrows for Feedback Data Flow run from Quality Organization to PHR, Periodic Primary Care Visits, Periodic Ophthalmologist Visit, Periodic Podiatrist Visits, and Hospitalization.

In the future state, the primary care provider would have access to important clinical history from across the care settings at which the patient is being seen. The provider would be armed with the knowledge of what medications had been prescribed by other providers, and would be made aware of any emergency room visits or hospitalizations that took place. This knowledge transfer would be facilitated through interoperable HIT systems including EHRs, PHRs, administrative systems, lab systems, and pharmacy systems. The ability to view information across care settings provides the primary care provider and the patient with a more thorough and detailed understanding of how well the patient is managing his or her illness and what treatment has been prescribed in the past. This ability to view information across care settings will ultimately help the provider make better decisions on what course of action to take to help control the diabetes.

The longitudinal approach to care, facilitated by HIT, can inform system-wide improvement to the quality and efficiency of care.

As the case study illustrates, from a quality measurement perspective, several aspects of the future approach can help inform system-wide improvements to the quality and efficiency of care delivery for diabetic patients. These include:

The ability to assess the processes, outcomes, and efficiency with which care is provided to a diabetes patient across care settings at which he or she may be seen;
The aggregation of that information to understand relevant frequencies, incidence, co-morbidities, outcomes of care processes and costs; and
The identification of critical areas of strength and deficiency.

An added benefit of quality assessment from a longitudinal perspective is that improvements upstream in the patient’s treatment trajectory can result in better outcomes that help reduce costs and improve efficiencies downstream as treatment traditionally becomes more expensive and more complicated. This is particularly true of a chronic condition such as diabetes, which often includes a myriad of other co-morbidities.

An adequate business case for organizations to invest in the infrastructure and programming needed to make longitudinal analysis a reality is fundamental to achieving the future vision. The existence of policies and procedures to facilitate the ability to conduct longitudinal analyses are also critical. The ability to view relevant information from the patient’s PHR, as recorded in the EHRs and administrative systems of the various care delivery institutions at which the patient is seen, and from the pharmacy systems that are used by the patient, are essential to ensuring that the patient is treated in the most appropriate manner possible. Interoperable HIT and HIE facilitate this exchange of information and are foundational components to facilitating effective care coordination and achieving the best possible care delivery. Last, the development of longitudinal measures, the integration of these measures into the HIT systems of the future, and the ability to exchange and aggregate the measure outputs in order to analyze them and inform quality improvement efforts are also necessary for achieving the future state.

3. CURRENT AND EMERGING LONGITUDINAL QUALITY ASSESSMENT CAPABILITIES

As described previously, longitudinal analysis of individual health care encounters across a span of time provides a patient-centered and comprehensive perspective on care provided. Longitudinal assessments allow health care professionals to analyze treatments choices, identify high-risk patients, evaluate the quality of the care, compare provider performance, and manage costs. Longitudinal views of patient care can be created by grouping inpatient, outpatient, lab, and pharmaceutical claims into clinically meaningful episodes of care. These episodes describe a patient's complete course of care and the severity of illness for a single occurrence of an illness or condition. This longitudinal perspective can enable more accurate quality of care measurement and cost assessment.

Longitudinal analysis, often conducted through episode grouping, can being used to:

Calculate the return on investment
Analyze utilization
Manage Costs
Compare provider performance
Identify high-risk patients
Evaluate the appropriateness of treatment options
Identify opportunities for improvement

Current strategies in longitudinal measurement, such as applying episodes of care groupers to claims data, are challenged by poor data quality, gaps in data and a primary focus by current episode of care grouping methodologies on measuring cost rather than adherence to evidence-based care or clinical outcomes. Said another way, current episode of care methodologies do not explicitly address appropriateness of care. The emerging strategies include new and innovative approaches to episode-based, patient-centered, or longitudinal care tracking and measurement and are addressing some of the needs to develop standards, improve the accessibility or availability of required data and expand the focus to include clinical quality.

To provide context in which to think about future requirements and the path to get there, this section describes the current state of the health care industry as it relates to the development of longitudinal analysis capabilities. Specifically,

Section 3.1 describes current state-of-the-art HIT solutions that support longitudinal care and quality improvement.
Section 3.2 describes emerging solutions and strategies that support longitudinal assessment that are being investigated or are in development.
Section 3.3 presents emerging HIT solutions and strategies that will better enable quality measurement and improvement and which can provide a strong foundation for longitudinal quality improvement activities in the future.

3.1 Current HIT Solutions to Support Longitudinal Patient Care and Quality Improvement

Both the Geisinger Health System (Geisinger) and the Intermountain Health Care delivery systems operate health IT solutions that offer clinicians a longitudinal view of their patients.

Geisinger is an integrated delivery system, comprised of hospitals, group practices, and a health plan that serves a geography with a population of approximately 2.5 million individuals. Geisinger has developed a prototype Clinical Decision Intelligence System (CDIS) to drive quality improvements within their system. The CDIS is the foundation of their electronic health record system and it integrates information from a multitude of sources, such as claims, evidence-based medicine guidelines, patient preferences, formularies, finance and operations data. A large digital warehouse allows Geisinger to actively measure clinical trends, identify gaps in care, derive new clinical knowledge, and give providers, researchers, and patients access to clinical information.

Geisinger has successfully demonstrated the impact of CDIS on quality improvement. For example, Geisinger improved care for adults with Type 1 and 2 diabetes through the standardization of clinical practices, automated identification of diabetics and care plan status, automated identification of suboptimal care, automatic order set creation, automatic generation of patient-specific report cards at checkout, automatic outreach to patients, and an influenza/pneumococcal campaign – a chronic disease return visit program. Through these improvements in clinical processes and the use of HIT, Geisinger was able to improve the quality of care by shifting the orientation of care delivery from one that is reactive to one that is anticipatory or proactive. After nine moths of implementing these programmatic improvements, composite quality indicators for patients in their diabetes program, which includes over 25,000 patients, saw a 5.5% increase. These gains are expected to increase further as Geisinger works to systematically standardize these programs and work to ensure they are being effectively implemented throughout their health system.

Intermountain Health Care is a health care delivery system of 21 hospitals (2,200 beds), more than 90 outpatient clinics, an employed physician group, and an insurance plan for patients in Utah and southeastern Idaho. Intermountain provides more than 50 percent of all care delivered in the region it serves. Intermountain began developing electronic medical systems in the 1970’s to support their mission of delivering high-quality patient-centered care. Their current system contains extensive clinical information with two main components, an EHR and a clinical data warehouse.

The hospitals currently use a desktop clinical application, the Health Evaluation through Logical Processing (HELP) system, to conduct evaluations of and provide recommendations on the care of each patient. HELP is interfaced to a longitudinal electronic patient record and a Clinical Data Repository (CDR). In both ambulatory and inpatient settings, providers enter visit notes, problems, and medications into the EHR and the information is stored in the clinical data repository. HELP allows users in either inpatient or ambulatory settings to view laboratory results, text reports, and radiology images, regardless of where the care was provided. HELP also contains advanced decision support capabilities for alerts, and privacy safeguards for protecting patient data. HELP, which combines the data from acute hospital encounters within the longitudinal clinical data repository, allows an authorized provider to see a comprehensive picture of patient health information.

A distinct feature of Intermountain’s EHR system was its method of development, which utilized an analytic and process-driven approach. The EHR system was designed to support Intermountain’s key clinical processes in delivering care. Intermountain began by first analyzing their key processes to identify the few processes that affect the largest portion of their patients. Once clinical processes were identified, they followed five steps to create a care process model:

Develop evidence-based best practices with scientific literature.
Blend evidence-based best practices into work flow in the front line.
Generate an outcomes tracking system to track the major intermediate and final outcomes of the disease, the financial outcomes and the service outcomes.
Build these outcomes measures into the EHR.
Build educational materials for clinicians and patients.

Intermountain uses the data from the EHR to manage the system, to advise patients, to plan and organize care, to provide accountability, and to set and achieve goals.

In addition to the efforts of Geisinger and Intermountain Health, other current efforts to offer clinicians a longitudinal view of their patients include software technology that integrates administrative data into patient-centered episodes of care. These episode groupers categorize episodes of care into 500 or more relatively homogeneous types of episodes. This categorization can be used to create a clinical profile of the patient to make more informed decisions for current care and plan for future health care needs. There are two prominent vendor products providing episode grouping software, Episode Treatment Groups (ETGs) by Symmetry/Ingenix, a subsidiary of UnitedHealth Group ⁴and the Medstat Episode Grouper by Thompson Healthcare.⁵ These systems vary in the areas of the number of classification groups, clinical homogeneity of these groups, statistical stability, severity or case-mix adjustment capabilities, and the structure or shifting of episodes.⁶

These software programs are primarily used by payers for physician practice performance measurement. The most common output used to measure physician performance is total episode cost. Total episode cost reflects the total amount paid by the patient and/or insurer for all care provided during an episode. In addition to payers using this data to compare specific patient, physician average, and peer average episode costs, payers are beginning to aggregate their data and to provide accrediting bodies with plan-level reporting of costs. The use of these software programs provides information about patient, provider, and plan patterns of resource use which are primarily measures of efficiency.

Availability of data and lack of standards has made the application of longitudinal quality measurement difficult in many settings. Episode grouping requires access to inpatient, outpatient, lab, and pharmacy claims data which limits application of groupers to settings where providers or payers have access to the entire spectrum of patient claims data. In addition, since current episode groupers are based solely on administrative data, they do not include patient characteristics (social economic status, other health status, or literacy), provider characteristics (such as location of practice), or other clinical information (non-administrative), largely due to the unavailability of these types of data.

3.2 Emerging HIT Strategies to Enable Longitudinal Patient Care and Quality Improvement

There are emerging strategies and ongoing efforts to develop new and innovative approaches to episode-based, patient-centered, and longitudinal care tracking and measurement. The three examples of such efforts described in this section are

Medicare Continuity Assessment Record and Evaluation (CARE)
Continuity of Care Record (CCR)
National Quality Forum (NQF) National Priorities and Goal Setting Project.

These efforts are addressing the need to develop standards for information sharing to facilitate improved delivery of care across care settings.

Medicare Continuity Assessment Record and Evaluation (CARE)

The Deficit Reduction Act of 2005 directed the Centers for Medicare and Medicaid Services (CMS) to develop a Post-Acute Care (PAC) Payment Reform Demonstration. The PAC settings currently employ three different assessment instruments that drive quality measurement and payment: Minimum Data Set (MDS) for Skilled Nursing Facilities (SNFs), the Outcome and Assessment Information Set (OASIS) for Home Health Agencies (HHAs), and the Inpatient Rehabilitation Facility–Patient Assessment Instrument (IRF-PAI) for Inpatient Rehabilitation Facilities (IRFs). Currently there is not a standard assessment in the Long Term Care (LTC) setting. The absence of a standard assessment from hospital discharge through all settings of post-acute care is believed to be one of the major contributors to variations in care, patient outcomes, and cost. There are also disparities in the payment amounts for the same “type” of patient among the three settings.

As part of the PAC Payment Reform Demonstration, CMS is developing a web-based assessment tool to be used at hospital discharge through all post-acute care settings to assess clinical and functional status. This initiative attempts to address challenges related to longitudinal measurement, including the lack of standards for information sharing and the availability of data. The new assessment tool, the CARE, may eventually replace the current tools, i.e., MDS, OASIS, and IRF-PAI. The assessment data gathered through CARE would be collected for all Medicare beneficiaries across the entire span of post-acute care. The new assessment instrument will document various levels of care needs including: administrative information, admission information, current medical items, cognitive status, physical factors, factors affecting outcomes, and discharge items.

This new tool provides the opportunity to encourage new patient-centered quality measures that span the continuum of care during the post-acute phase of an episode, to assess the resources used to treat patients, patient outcomes independent of clinical setting, and to drive new pay-for-performance initiatives.

The Continuity of Care Record (CCR) and Continuity of Care Document (CCD)

The CCR is a joint effort of ASTM International, the Massachusetts Medical Society, the Health Information and Systems Society, and the American Academy of Family Physicians. The aim of this effort is to improve the continuity of care, reduce medical errors, and ensure a minimum level of interoperable health information. The CCR will be a structured transportable set of basic patient information that can be used in the transfer or referral of patients from one provider or setting to another. The CCR is addressing barriers in longitudinal measurement and assessment such as limited standards and data availability.

The basic information captured in the CCR includes: patient and provider information, insurance information, health status, recent care, future care recommendations, and reason for referral or transfer. The information in the CCR will be completed by relevant clinicians and the format of the CCR (xml) is readable by both machine and humans. The developers of the CCR envision that existing EHRs can import and export the information to and from the CCR to allow a seamless transfer of essential information across care settings.

The CCR does not include any laboratory or radiology information that was not included by a clinician. It does not list symptoms but does list diagnoses and reasons for referral or transfer. It does not include a chronology of events as would exist in an EHR. The CCR is expected to prevent medical errors and improve the quality of care by providing essential information to a new health care provider.

The CCD is the harmonization project of HL7 and ASTM endorsed by the Health Information Technology Standards Panel (HITSP). CCD represents the major features of ASTM’s CCR and the HL7’s Clinical Document Architecture, and it is used for the exchange of crucial patient data such as demographics, medications and laboratory results. CCD will enable the rapid transition of CCR content into the more mainstream Clinical Document Architecture representation. Assuming progress is made as planned; it will be included in Version 2.0 of HITSP’s Interoperability Specifications, which are expected to be recognized in December 2007. If this schedule is adhered to, the effective implementation date for the standards will be January 2008.⁷

NQF National Priorities and Goal Setting Project

The NQF project Evaluating Efficiency across Episodes of Care aims to develop a comprehensive measurement framework to assess efficiency, quality, and cost for health care episodes that span care settings. The objectives of this project include creating clear definitions of terminology, establishing a discrete set of domains, and providing guiding principles for implementation. The NQF will:

Identify a subset of priority conditions intended to serve as a starting point for a national effort to measure, report, and improve efficiency across episodes of care.
Establish national performance improvement goals—measurable targets for two priority conditions—over the next three to five years.
Lay out a future vision and research agenda for the evolution of a measurement and monitoring system focused on episode efficiency.

To date, the committee has drafted a measurement framework for “episode efficiency” that includes definitions, domains, and guiding principles; conducted a workshop to further develop the framework among a diverse group of content experts; and selected two priority conditions to serve as operational examples: acute myocardial infarction and low back pain. A third condition, diabetes, is being added.

In the course of these efforts, the NQF identified current methodological issues related to longitudinal measurement. These issues, in relation to data, are: accountability, security, quality of data, data integrity, data collection, aggregation methods, and need for EHR standards. In longitudinal measurement, challenges remain such as the lack of a conceptual regulatory/reimbursement framework for shared accountability within a fragmented delivery system, the inability to attribute of care to specific providers, and privacy concerns for sharing data across multiple care settings.

3.3 EMERGING HIT STRATEGIES TO ENABLE CROSS-ORGANIZATIONAL QUALITY MEASUREMENT AND REPORTING

In addition to the efforts described above, several efforts are under way to better leverage the potential of HIT to drive quality, efficiency and performance, although without a specific focus on enabling patient-centered care. These efforts are nonetheless important in driving quality improvement and are foundational to developing an infrastructure upon which longitudinal analysis can occur. For example, in recent years there have been focused efforts to standardize data elements and sources; to design EHRs to drive quality care; to align quality measurement and reporting with clinician workflow through the evolution of clinical decision support tools,; and to remove current barriers to EHR adoption. The health care industry is beginning to address many of the foundational elements necessary to build a health care system that fully leverages HIT to enable quality, as described below.

Data Standardization: There is limited standardization of data elements, particularly within EHRs, and EHRs are often customized during implementation, resulting in a lack of detailed, standardized implementation specifications for collecting data pertaining to quality metrics in an EHR. Various efforts are underway to begin to standardize data elements within EHRs, as well as to consider data standards and the ability to automate data collection in the development of new quality measures. The National Health Information Network (NHIN) coordinated through the Office of the National Coordinator for Health Information Technology (ONC) is predicated on a platform of interoperability that requires data standardization. To achieve the level of interoperability expected of the NHIN, ONC has been coordinating a number of integrated activities. ONC developed a Quality Use Case that captures the integration of data to support quality measurement, feedback and reporting into EHRs, begins to use quality measures to support clinical decision making, and allows for the aggregation of quality information across multiple providers and entities to support public reporting. The Quality Workgroup included recommendations to accelerate the use of common data elements to enable automated data collection to support a prioritized set of AQA Alliance (AQA)⁸ and Hospital Quality Alliance (HQA) measures in its March 2007 recommendations to the AHIC. HITSP is focused on the harmonization of data standards to meet the requirements outlined in the Quality Use Case and to fill the identified gaps for data elements required to automate the collection of data to support core sets of AQA and HQA quality measures. The standards identified by HITSP will then be used by the Certification Commission for Health Information Technology (CCHIT) during its 2008 certification cycle.
Data Stewardship: Consensus on managing and storing aggregated patient-indexed data is important to the realization of comprehensive HIE. Movement toward comprehensive HIE will rely heavily on resolution of a variety of data stewardship issues. Together, the AQA and the HHS Agency for Healthcare Research and Quality (AHRQ) are working to develop a national data strategy that delineates the role and governance model for data stewardship. Most recently, AHRQ issued a Request for Information (RFI) to gather ideas and thoughts across the industry on requirements for a data steward entity. Following the responses to this RFI, due July 2007, AHRQ will issue a report in October 2007 on the range of viewpoints on data stewardship requirements. The Joint Commission on Accreditation of Healthcare Organizations (The Joint Commission) has also engaged a broad group of stakeholders to identify the critical components of a performance measurement data strategy (this effort is still in progress).
EHR Adoption: Widespread adoption of EHRs is a critical component of supporting quality through HIT. Through the CCHIT, AHIC is working to promote specific EHR products that meet a core set of interoperability requirements. Additionally, policy levers such as modifications to anti-kickback laws and programs such as CMS’s Doctor's Office Quality - Information Technology (DOQ-IT) program which, among other objectives, supports the adoption and effective use of information technology by physicians' office to improve quality and safety, are attempting to address barriers to EHR adoption.
HIEs/Regional Health Information Organizations (RHIOs): There are over 100+ local activities that are currently testing local applications of HIE. State and community-level efforts in HIE will require a national strategy to ensure interoperability and consistency to enable the NHIN’s “networks of networks.” As EHRs evolve, adoption of EHRs increases, and a nationwide interoperable health information network is built, patient data captured as a result of routine care and treatment will be accessible across care settings so that clinicians have the right information at the right time for clinical decision-making.
Better Quality Information to Improve Care for Medicare Beneficiaries (BQI): The AQA, in conjunction with CMS, has implemented six pilots focused on the sharing and use of provider data to publicly report on quality of care. The BQI pilots will be able to aggregate private payer data with CMS claims data to build a more comprehensive picture of provider quality.
Chartered Value Exchanges: To further advance community involvement and participation, HHS is formally supporting the establishment of Chartered Value Exchanges that represent a multi-stakeholder organization that has taken clear action in its community to convene community purchasers, health plans, providers, and consumers to advance the four cornerstones of Value-Driven Health Care. The four cornerstones include: Interoperable HIT, Measure and Publish Quality Information, Measure and Publish Price Information, and Promote Quality and Efficiency of Care. Chartered Value Exchanges will be invited to participate in a nationwide learning network that will be sponsored and maintained by AHRQ.

4. EMERGING FUTURE STATE REQUIREMENTS

This section articulates future requirements as identified through interviews, through testimony, and through secondary sources recommended by subgroup members, interviewees, and referenced through testimony. These future requirements perform several functions. In some cases, they directly address barriers in today’s environment; in other cases, they take the form of enablers that currently exist or are being explored through stand alone or organization specific initiatives; and in yet other cases, they articulate other important characteristics or features that are necessary for HIT to effectively enable patient-centric quality improvement in the future.

The requirements are grouped by five components identified through our review of relevant testimony, interviews, and literature as critical to ensuring the national quality enterprise can conduct patient-centric quality improvement in the future. Table 1 provides a brief overview of each of these key components in the future state, as articulated in the Quality Workgroup vision.

For each component, a discussion of the enablers and barriers seen in the current environment precedes each listing of future requirements. The requirements are further subdivided into those around which there is common agreement throughout the literature reviewed, interviews conducted, and testimony received, and those around which there is less agreement.

Table 1. Key Components of the Future State Vision
Component Future	State Description
Business Case / Incentive Alignment	Measures and guidelines that promote efficient, coordinated care will help galvanize payer’s attempts to implement payment programs that reward effectiveness and coordination of care. These measures and guidelines will incentivize consumers to remain engaged and empowered, to utilize quality information to strengthen their experience, and to take a vested interest in care across settings. Providers will also have access to coordinated guidelines and protocols will be widely used along with the inclusion of multi-practice care processes in quality metrics.
Measure Development	Measure developers will collaborate to facilitate measure harmonization. Vendors will collaborate with the NQF and quality measurement organizations to encourage development and implementation of common conventions and guidelines for measure development. Longitudinal measurement systems will capture the performance of multiple providers caring for a patient, will examine how well care is provided across transitions to different settings (e.g., hospital to nursing home), and most importantly, will evaluate patient outcomes over time. To support improved coordination of care, the national quality enterprise will encourage the alignment of measures across settings as well as the use of interoperable EHRs and PHRs that allow for measurement of care across time and sites of care. The research community will develop guidelines and measures that promote efficient, high quality, coordinated care. The research community will continue to identify gaps and to refine and update existing measures as more information is gathered on coordination of care.
Data Standards	Data collection and aggregation practices will be enabled by measures that use standard definitions of terms to the greatest extent possible. These activities will also be enhanced by the structuring or standardization of documented data either through direct entry of structured information or through focused and standardized free-text searching and parsing techniques.
Data Exchange and Aggregation	Data exchange and aggregation will be facilitated through standardized data sets, and the widespread use of interoperable HIT, including EHRs. Polices and procedures related to collection, storage, aggregation, linkage and transmission of health data will be established with appropriate protection for legitimate use. The establishment of these policies and procedures will allow for the collection of patient-centered data, aggregated across providers and payers to support longitudinal quality measurement at the patient, provider, provider group, plan, and hospital levels.
Policy	A national framework for the use of health data will include a robust infrastructure of policies, standards, and best practices to facilitate the broad and multiple-purpose collection, storage, aggregation, linkage, and transmission of health data with appropriate protections for legitimate use. Appropriate confidentiality protections will be in place for the use of patient data that are in strict compliance with Health Insurance Portability and Accountability Act (HIPAA) regulations. Potential problems of patients opting out of inclusion of their data in a data repository will be addressed and impacts on accurately assessing the quality of care on both the national and community levels will be understood.

The requirements articulated in this section of the report are intended to serve as information points for the Quality Workgroup to consider as they develop the second set of AHIC recommendations. The requirements are predicated on a number of assumptions, including:

EHR technology is the norm for storing and managing clinical information for patients.
Standard-setting and certification bodies have been established to continually develop and update HIT interoperability and methodological standards to enable longitudinal analysis.
The national quality agenda is unified and quality measures are harmonized across care settings.
NQF-endorsed longitudinal measures that facilitate care coordination and patient-centric quality improvement are available for use across care settings.

4.1 BUSINESS CASE/INCENTIVE ALIGNMENT

The annual savings attributable to widespread EHR adoption, including savings resulting from improved safety and disease management, are likely to lie between 7.5 and 30 percent of annual health care spending.⁹ While the aggregate benefits are enormous, the business case for investing in HIT to support direct patient care, quality improvement, and measurement activities is made at the organization level, whether the organization is a private practice, hospital, health system, or payer. The decision is based primarily on a review of the financial benefit of investment in enabling technologies versus the perceived gains. Providers and hospital systems must balance their primary objective of delivering high quality care to patients against the costs related to implementing enabling technologies such as EHRs and CDS to improve how they deliver care, as well as against the costs associated with investing in associated quality measurement and improvement activities. Because longitudinal measurement requires investment in inter-organizational data capture and management strategies, payers and integrated delivery models have greater incentives to make this investment, and have made the most progress as described in Section 3. Table 2 summarizes the key enablers and barriers that impact the business case for HIT investment in today’s environment.

Table 2. Business Case Enablers and Barriers
Enablers	Barriers
Demographic shifts due to increase in elderly and chronically ill populations and the need to better manage the care provided to those populations The increased recognition of the importance of the transitions between care settings and the coordination of care needed across care settings Organizations that have a clear vision regarding their quality and data strategy and commitment to implementing that vision at the organizational level Access to financial resources to support data strategy The development of viable business models for regional HIE networks that support provider EHRs, data aggregators and public reporting entities Community support of a cross-organizaitonal data strategy Vendor support of a cross-organizational data strategy The increasing demand for useful performance information by consumers, providers, and other stakeholders The desire to advance the medical science/technology Informed consumers help drive the quality improvement agenda	Economic pressures – higher costs of doing business, declining reimbursement, limited capital and resources make investment in information technology solutions difficult, especially at the practice level The explosion of performance measurement efforts and associated increased system costs and the burden associated with data collection and reporting Balancing the demand of multiple stakeholders when considering data strategies and implementation of those strategies Management of the business agreements and complex relationships to allow regional data sharing Siloed or organization-specific viewpoints when considering investment in enabling technologies Provider concerns regarding attribution and accountability when considering longitudinal measurement Episode-of-care methodologies that do not take into account prevention and consumer choices No framework for shared accountability beyond integrated systems

As described above, the financial benefits associated with investment in interoperability are substantial, yet there remains a great need to improve the business case for organizations to encourage them to invest in longitudinal analysis and an HIT infrastructure that could enable it. Creating a multi-stakeholder business case is a critical first step to stimulate a cultural shift that will advance individual organizations investment in interoperable HIT the accompanying HIE that is required for longitudinal analysis.

Table 3 below describes the business, technical, and policy-related requirements that are necessary for establishing such a business case in the future state and is followed by a discussion of the areas where future state requirements remain undefined.

Table 3. Business Case-Related Future Requirements
ID	Requirement Description	Common Agreement
1.1	Systems or processes to share, collect, aggregate and report quality, cost of care, and patient experience data will be designed to minimize cost and burden to consumers, physicians’ practices, health plans, and data aggregators. (Based on AQA Data Aggregation Principles document, AQA Data Sharing and Aggregation Principles for Performance Measurement and Reporting document)	Yes
1.2	Sustainable business models will be established for HIE at the regional, state, and national levels that protect privacy and enable appropriate uses of data. (Based on Quality Workgroup vision document)	Yes
1.3	Resources required to establish, maintain and contribute data to HIE databases will be fairly distributed among all who benefit. (Based on AQA Registry Principles document)	Yes
1.4	System design, implementation, and use will minimize costs. (AQA Principles for HIT and Measurement Aggregation document)	Yes
1.5	All participating organizations in a HIE will belong to and comply with agreements of a federation, with clear agreements established to engender the trust that is essential to the exchange of health information. (Based on Markle Technology Principle: Federation, located in the Common Framework Overview and Principles document)	Yes
1.6	Providers will have access to coordinated guidelines and protocols will be widely used along with the inclusion of multi-practice care process in quality metrics. (Addressed in March 2007 Quality Workgroup recommendations to the AHIC)	Yes
1.7	Payment strategies for care provision will reflect shared accountability across organizations. (Based on Quality Workgroup Vision document)	No
1.8	Government, payer, and employer-sponsored initiatives shall provide financial incentives to drive adoption of EHRs and other quality measurement and care coordination enabling technologies. (Based on Quality Workgroup Vision document)	No

4.2 MEASURE DEVELOPMENT

Today, longitudinal measures of quality of care are almost non-existent. Measurement efforts have been driven by what data is currently available, rather than focusing on what is important to measure. The challenge in today’s environment for developing longitudinal measures has been the structure of the health care delivery system: fragmentation across the continuum of care that creates clinical setting-specific silos in data systems, measure development, and standards definitions. Currently, measure development activities focus on provider encounters, occurring in silos. Within these silos, measure developers use differing standards for evidence-grading, differing approaches to measure specifications, and have varying capabilities for measure development and maintenance, resulting in inconsistencies in the way measures are developed, implemented, and maintained.

Current measurement of quality performance relies largely on claims data because more robust electronic clinical information is generally lacking and performance information is insufficient for payers to determine true under- and over-utilization of services. The market perceives the cost of quality reporting to be higher than the benefits, but generally supports movement toward value-based payment.

Despite these challenges in the current system, some integrated delivery networks (IDNs) and payers are currently aligning and utilizing data from multiple sources. In addition, there are efforts to standardize a web-based patient assessment tool to be used across the post-acute spectrum of care (CMS CARE project; see page 11). Standardization efforts across the continuum of care are critical for sharing information between information technology systems and across providers. EHRs remain the key tool for data collection and provision of evidence-based care. Table 4 below presents the enablers and disablers that exist in today’s environment as they relate to longitudinal measure development.

Table 4. Measure Development Enablers and Barriers
Enablers	Barriers
Use of episodes-of-care groupers that include risk adjustment algorithms Use of standard quality definitions throughout industry Use of standard algorithms throughout the industry Providers and clinical leader engagement in the measure development and endorsement process	Lack of measures that span the entire continuum of care Measurement often dictated by what is measurable rather than what should be measured Limited set of national consensus measures Robust measures not yet developed for all provider specialties Limited clinical detail in claims data Limited standardized clinical information across EHR products Few integrated or episode-based metrics Redundancy in reporting Most measures are aggregated at patient, provider, or encounter level

Harmonization of existing measures; development of methodologies and associated measures for assessing quality across the continuum of care; identification of necessary data elements and integration of data elements into EHRs are a few of the areas that must be addressed in order for measure development to effectively support longitudinal and patient-centric quality improvement. Table 5 below describes business, technical, and/or policy-related requirements related to measure development in the future state.

Table 5. Measure Development-Related Future Requirements
ID	Requirement Description	Common Agreement
2.1	Measure developers will produce harmonized measure sets in line with national goals that are defined at the data element level to allow uniform data collection and aggregation across disparate systems. (Based on Quality Workgroup Vision document)	Yes
2.2	The NQF will establish common conventions and guidelines for measure development and maintenance, including HIT standards and specifications. (Based on Quality Workgroup Vision document))	Yes
2.3	The NQF, in partnership with measure developers, will build consensus around national goals for quality measurement that will drive the development of longitudinal measures. (Based on Quality Workgroup Vision document)	No
2.4	Tested methodologies will be employed to define care processes associated with specific illnesses and to define related longitudinal measures that will help improve the quality of the care delivered through those processes. (Based on Testimony, Interviews)	Yes
2.5	To the greatest extent possible, longitudinal quality measurement will be structured to accurately reflect all services that are accountable in whole or in part for the performance measured. Attribution to providers or organizations should be explicit and transparent. (Based on AQA Consumer Principles document) [Also included as Requirement 1.6]	No

4.3 DATA STANDARDS

Data standards are an essential component to successful longitudinal measurement. Standards are needed to ensure interoperability of EHR, PHR, and other clinical and administrative system products. HITSP, a public-private partnership focused on developing consensus-based standards, is the primary organization in today’s environment responsible for the development of these standards at the national level. Working in partnership with HITSP, CCHIT serves as the nation’s primary recognized certification body for EHR products. Together, both these organizations help drive the standardization of HIT across the health care industry. However, in addition to standards that help promote interoperability for HIT, standards are needed for specific uses for these systems as they relate to longitudinal quality measurement and improvement, such as for matching patient records, aggregating data, and data exchange over time and across care settings.

While progress is being made on the development of standards and common nomenclature for EHRs, and administrative and clinical HIT systems, there is a lack of broad application of standards-based data definitions. Achieving standardization will help facilitate the sharing and use of clinical information to improve the quality of care delivery. However, even where standard nomenclatures exist, the determination of which codes should be used to describe care can vary significantly from provider to provider. This, in turn, impacts the ability to determine if the care provided relates to a given episode or clinical condition, which in turn impacts the ability to accurately capture data related from a longitudinal perspective. Additionally, clinical documentation continues to be unstructured, non-standardized and potentially inaccurate. While we are in the early stages of EHR development and adoption, there are limitations on the ability of EHRs to facilitate quality measurement and capture appropriate clinical documentation, resulting in an opportunity lost to use this valuable information to facilitate quality improvement. Table 6 below presents information on how current use of standards, lack of standards, or activities to develop standards enables or disables longitudinal quality measurement and improvement.

While progress is being made on the development of standards and common nomenclature for EHRs, and administrative and clinical IT systems, there is a lack of broad application of standards-bas ed data definitions. Even where standard nomenclatures exist, the determination of which codes should be used to describe care can vary significantly from provider to provider. This, in turn, impacts the ability to determine if the care provided relates to a given episode or clinical condition, which in turn impacts the ability to accurately capture data related from a longitudinal perspective. Additionally, clinical documentation continues to be unstructured, non-standardized and potentially inaccurate. While we are in the early stages of EHR development and adoption, there are limitations on the ability of EHRs to facilitate quality measurement and capture appropriate clinical documentation. The table below presents information on how current use of standards, lack of standards, or activities to develop standards enables or disables longitudinal quality measurement and improvement.

Table 6. Data Standards Enablers and Barriers
Enablers	Barriers
CCHIT and HITSP potential to promote standards development and certification for health IT as it relates to quality measurement and reporting Increased standardization of performance measures and collection processes through measure developer harmonization efforts Use of standard data aggregation algorithms for quality measures Increased levels of collaboration between providers, purchasers, consumers and accreditors to produce uniform standards for sharing and aggregating health data and public reporting Increasing availability of language processing programs to extract data from text or narrative reports The potential for EHRs to serve as a key tool for standardized data collection and provision of evidence based care Software systems help minimize human and system error in data entry and clinical documentation Increasing use of technologies and methodologies for encrypted identification to enable the data re-identified, while ensuring confidentiality is preserved when doing aggregate or population level analyses	Lack of standards for data definition and aggregation Lack of defined interfaces to allow the aggregation (Vision) Application of non-standard data definitions to data sources Inconsistent formats for data collection both within and across organizations Application of non-standard data elements, templates, formats across different data sources One-dimensional or fragmented efforts No consensus on appropriate use of taxonomy for many data elements, Unstructured clinical documentation and non-standardized nomenclature Use of inaccurate clinical data from manual documentation Exclusion of working diagnoses in the medical record Gaps in quality management capabilities of EHRs Lack of ability to capture free text data from an EHR Inaccurate capture of provider-entered data into an electronic health record Limited standardized clinical information across EHR products Lack of patient identified data available in large warehouses to conduct research Inability to integrate multiple IT systems within an organization No standards for defining patient-provider relationships No standards for defining visits and relevant care processes for an individual episode or illness Need for a national patient identifier or methodology for matching patient records Patient record matching is expensive Lack of widely agreed upon standards for data stewardship

One of the most significant areas where progress needs to be made to enable the future state is in the area of data standards, however significant work remains. Standards must ultimately be developed and established across a number of different areas in order to enable effective and efficient longitudinal analysis capability. Standards must cover EHR, PHR, and CDS technologies; the methodologies and algorithms used to match patient records; ensure interoperability to enable data exchange and aggregation; and facilitate care coordination and care delivery. The table below describes business, technical, and/or policy-related requirements related to development and use of data standards in the future state.

Table 7. Data Standards-Related Future Requirements
ID	Description	Common Agreement
Electronic Health Record Standards
3.1	EHR certification will include basic standards to ensure interoperability with other systems. (Based on Quality Workgroup Vision document)	Yes
3.2	EHR certification will include basic standards for that define standardized data elements needed for longitudinal quality measurement, care management, and quality improvement. (Based on Quality Workgroup Vision document)	Yes
3.3.	EHR certification will include basic standards for structured capture of narrative documentation and create parameters around all free text documentation. (Based on Testimony, Interviews, Various documentation)	Yes
Personal Health Record Standards
3.4	PHR certification will include standards to ensure interoperability with EHRs and CDS Systems. (Based on Quality Workgroup Vision document)	Yes
3.5	PHR certification will include basic standards that define standardized data elements that could be used for quality reporting and improvement activities. (Based on Testimony, Interviews)	Yes
Clinical Decision Support Standards
3.6	CDS system certification will include standards to ensure interoperability with other systems including EHRs and PHRs. (Based on Quality Workgroup Vision document)	Yes
3.7	Standard development organizations will establish standards for CDS that allow for the application of applied collaborative research into best practice. (Based on Quality Workgroup Vision document)	Yes
Patient Record Matching Standards
3.8	Standards will be developed for recording and storing patient information relative to name, date of birth, address and other demographic information that may be used to match medical records. (Based on Draft Unpublished Document)	Yes
3.9	Standards will be developed for identifying institutions through which care was delivered so that information can be matched to medical records. (Based on Testimony, Interviews)	Yes
3.10	Standardized criteria, methodologies, and algorithms will be used to link and match patient records throughout various linked databases for performance measurement while protecting patient privacy. (Based on Draft Unpublished Document, Testimony, Interviews)	Yes
3.11	Methodologies/algorithms for linking and matching patient records will use multiple attributes for matching patients and records. (Based on Subgroup Input)	No
3.12	Methodologies/algorithms for linking and matching patient records will have little tolerance for error and shall avoid false positive matches at all costs. (Based on Subgroup Input, Markle Technology Principle: Accuracy located in the Markle Technology Principles document)	Yes
3.13	Methodologies/algorithms for linking and matching patient records will have the capacity for the correction of common errors in a timely manner. (Based on Subgroup Input)	Yes
3.14	Methodologies/algorithms for linking and matching patient records will have rapid/near instantaneous processing times. (Based on Subgroup Input)	Yes
HIT Interoperability Standards
3.15	Standards for networked technologies will enable near real time information exchange across systems, care settings, and among providers, payers, consumers and other stakeholders as deemed appropriate. (Based on Testimony, Interviews, Quality Workgroup Vision document)	Yes
3.16	Uniform operating rules and standards for the exchange and aggregation of quality and efficiency data used in both the public and private sectors will be established for the purposes of performance measurement and reporting. (Based on AQA Data Sharing and Aggregation Principles document and National Health Data Stewardship Entity document)	Yes
3.17	Uniform operating rules and standards for the exchange and aggregation of quality and efficiency data used in both the public and private sectors will be established for the purposes of establishing effective care coordination across settings. (Based on Interviews, Testimony, AQA Data Sharing and Aggregation Principles document, and National Health Data Stewardship Entity document)	Yes
3.18	Open networks, standards, and protocols will ensure that compatibility, connectivity, and interoperability characterize the systems used for quality measurement and improvement. (Based on AQA Principles for HIT and Measurement Aggregation document)	Yes
3.19	Stacked data structures and master variable dictionaries will be used to manage all systems housing linkable patient data. (Based on Testimony)
3.20	Certification for HIT systems will include an assessment of adherence to established interoperability standards. (Based on Quality Workgroup Vision document)	Yes
3.21	Certification for HIT systems will include all technical standards that are adoption ready, including what is referred to as the “operable” set: HL7 v2.x data interchange standard, the HL7 Reference Information Model, the DICOM standard for imaging, the NCPDP SCRIPT prescription drug 2 information standard, the LOINC vocabulary for laboratory tests, the IEEE/CEN/ISO 1073 medical device communication standard, the ASC X12 administrative transaction standard, HL7 Data Types, Clinical Document Architecture, and the HL7 Clinical Context Management Specification. (Based on Connecting for Health Key Findings document)	Yes
3.22	Government, payer, and employer sponsored initiatives will promote the use of open, non-proprietary standards for interoperability and data exchange. (Based on Interviews, Testimony)	No
Care Coordination Standards
3.23	A core set of data elements common to all care settings will be automatically captured in relevant HIT systems and transferred across care settings upon provider request. (Based on Subgroup Input, Continuity of Care Concept Paper)	No
3.24	A unique set of data elements to individual care settings will be automatically captured in relevant HIT systems and transferred across care settings upon provider request. (Based on Subgroup Input, Continuity of Care Concept Paper)	No
3.25	Software applications for care management (e.g. EHRs, practice management systems, registries) will make standardized quality, performance, and efficiency measurement a routine by-product of their use. (Based on AQA Principles for HIT and Measurement Aggregation document)	Yes
3.26	Software applications for care management (e.g., EHRs, practice management systems, registries) will be designed to enable the merger of their data with others for the purpose of facilitating quality improvement efforts or the production of standardized quality, performance and efficiency measurement. (Based on AQA Principles for HIT and Measurement Aggregation document)	Yes
Clinical Registry Standards
3.27	Publicly available protocols that encompass common nomenclature, data definitions, data collection, sample size, sampling and data transfer protocols (when appropriate) and reporting format will be standardized among institutions reporting to the registry. (AQA Registry Principles)	No
3.28	When data on the same procedure(s) is transmitted and collected by multiple registries (e.g., different specialty societies collecting data on the same clinical procedure), data collection and submission procedures will follow the same data field definitions, protocols, and methodology. Ideally, national standards should be used or developed whenever possible. (AQA Registry Principles)	No
3.29	Clinical data registries shall be populated with clinical and administrative data as required to appropriately measure quality. (AQA Registry Principles document)	Yes
3.30	Data submitted to clinical registries shall enable comparative reporting to inform choices, quality improvement or quality assessment. (AQA Registry Principles document)	Yes
3.31	Data sources for clinical registries shall be multi institutional with agreed upon policies for data submission. (AQA Registry Principles document)	Yes
3.32	Registries shall have an infrastructure to support and maintain expansion in data capacity and analysis of trends. (AQA Registry Principles document)	Yes
3.33	Registries should be able to accept the electronic transfer of data. (AQA Registry Principles document)	Yes
3.34	Health data registries shall be built to minimize human factor errors and entry. (AQA Registry Principles document)	Yes

4.4 DATA EXCHANGE AND AGGREGATION

While the increased recognition of the value of longitudinal performance measurement data has prompted the development of quality databases across care settings, several barriers exist that are significant barriers to successful implementation of effective data exchange and aggregation of quality data. Some of these barriers include the limited availability of the necessary data, the lack of data standards to enable sharing and aggregation across sources, and the disparate definitions and coding choices for the care processes associated with specific episodes.

Information capture is, by the nature of care delivery, setting-specific, existing in silos. However, even within a unique care setting, matching and aggregating data from across data sources can be problematic. Independent health care systems such as Intermountain Health or IDNs such as Kaiser Permanente have made progress in integrating and aggregating data across the continuum of care and over time, and using that data to achieve critical quality improvements. Collaborations among providers at the local and regional level, RHIOs and BQIs are exploring locally grown methodologies for effective data exchange and aggregation across organizations. Quality alliances among providers, purchasers, consumers and accreditors are beginning to produce uniform standards for sharing and aggregating health data for quality improvement and for public reporting. However efforts to enable these activities from a longitudinal perspective are few and far between. Disease registries an important source of longitudinal data as well, however uniform data standards and rules for sharing and aggregating the information contained within them are not in place, thus enabling their integration with data from other sources to inform quality improvement is difficult.

Table 8 below summarizes the key enablers and barriers to data exchange and aggregation.

Table 8. Data Exchange and Aggregation Enablers and Barriers
Enablers	Barriers
Increased recognition of the value of performance measurement data across care settings Availability and accessibility of data from multiple sources Integrated delivery systems, payer, and large health systems’ ability and business case to integrate data across settings Increasing availability of scalable open source software development to reduce costs of multiple approaches to data aggregation Standardization across HIT systems to enable data exchange between systems and across providers EHRs and PHRs role as key tools for standardized data collection and the provision of evidence based care Collaboration among local and regional health care providers, RHIOs, and BQI sites to explore and test models for effective data exchange and aggregation Presidential executive order requiring Federal agencies and their contractors to promote the use of interoperable HIT so that data can easily be shared.	No single data source to provide a complete view of a patient's care across settings Lack of data standards to support data exchange across systems Limited access to laboratory data, PBM data, EHR data both within organizations and across organizations Continued paper-based data storage Low EHR adoption Inability to aggregate data from across multiple HIT systems due to interoperability issues Inability to integrate multiple IT systems within an organization due to interoperability issues Costs associated with developing interfaces between systems Inability to aggregate data from across multiple care settings Use of proprietary technologies Limited technology infrastructure capabilities of those receiving or viewing the information Access to broad band across regional locations, particularly in rural areas HIE operational in less than 10 regions across the United States Disease registries are not integrated into HIE efforts, though they house longitudinal data Current quality metrics do not require, and thus do not incentivize, reporting across organizations Liability issues, need for more explicit privacy policy development, and lack of business case limit exploration of effective data exchange and aggregation strategies Limited ability to analyze cost/utilization across episodes

Data exchange and aggregation both within and across care institutions are limited in large part due to the limitations in existing technology and infrastructure, interoperability standards, the resulting use of organization or locality-specific methodologies for exchanging and aggregating data, and the difficulty in developing business agreements to exchange data across institutions. The table previously presented in Section 3.3 outlined the requirements related to interoperability standards necessary to promote data exchange and aggregation. This section describes the business, technical, and/or policy-related requirements related to other aspects data exchange and aggregation in the future state, including requirements related to health information exchange, data stewardship, and the role of record locator services (RLS).

Table 9. Data Exchange and Aggregation-Related Requirements
ID	Requirement Description	Common Agreement
HIE
4.1	Data exchange and aggregation will be compliant with privacy, confidentiality and other applicable rules, while ensuring that providers, plans, other data contributors, and consumers have necessary and appropriate access to useful information. (Based on the AQA Data Sharing and Aggregation Principles for Performance Measurement and Reporting document)	Yes
4.2	Those submitting data to an HIE will be accountable for the accuracy and completeness of their data. (Based on the AQA Data Sharing and Aggregation Principles for Performance Measurement and Reporting document)	Yes
4.3	HIE will operate under the following principle: Only the minimum number of rules and protocols essential to widespread exchange of health information shall be specified as part of a Common Framework. It is desirable to leave to the local systems those things best handled locally, while specifying at a national level those things required as universal in order to allow for exchange among subordinate networks. (Based on Markle, Technology Principle: Make it “Thin,” located in the Markle Technology Principles document)	Yes
4.4	Any hardware or software will be used for HIE as long as it conforms to a common framework of essential requirements. The HIE network should support variation and innovation in response to local needs. The HIE network must be able to scale and evolve over time. (Based on Markle Technology Principle: Flexibility, located in the Common Framework Overview and Principles document)	Yes
4.5	There will be feedback mechanisms to help organizations to fix or “clean” their data in the event that errors are discovered [in regards to HIE]. (Markle, Technology Principle: Accuracy, located in the Common Framework Overview and Principles document)	Yes
4.6	Aggregation will be compliant with privacy, confidentiality and other applicable rules, while ensuring that providers, plans, other data contributors, and consumers have necessary and appropriate access to useful information. (AQA Data Sharing and Aggregation Principles for Performance Measurement and Reporting, located in the Common Framework Overview and Principles document)	Yes
4.7	Those submitting data will be accountable for the accuracy and completeness of their data. (AQA Data Sharing and Aggregation Principles for Performance Measurement and Reporting, located in the Common Framework Overview and Principles document)	Yes
4.8	HIE models will take into account the current structure of the health care system, taking advantage of what is deployed today where possible and appropriate. (Based on the Markle Technology Principle: Avoid Rip and Replace, located in the Common Framework Overview and Principles document)	Yes
4.9	HIE will be predicated on a federated model that protects the privacy and security of information and allows accurate and timely access to information. This includes a common set of policies, procedures, and standards to facilitate reliable, efficient sharing of health information among authorized users, is required. These standards or practices specify when patient information can be shared, which information can be shared, and how the information can be used. (Based on NHIN RFI Collaborative Response)	Yes
4.10	HIE will be predicated on a decentralized model that protects the privacy and security of information and allows accurate and timely access to information. It facilitates the transfer of selected information from one end-point system to another (not necessarily the source system), as is required for providing care and supporting informed patient participation in care. (Based on NHIN RFI Collaborative Response)	Yes
4.11	HIEs will develop polices to encourage consumer education around: a) security; b) how it works; c) patient permission; d) who has access; e) the benefits of HIE to the patient and physician (Based on National Committee on Vital Health Statistics (NCVHS) Testimony)	No
Care Coordination (through HIE)
4.12	Clinical and administrative patient data will be transferred across care settings when patients transition from one care setting to the next and shall be made available to the provider treating the patient. (Based on Interviews, Testimony, Subgroup Input)	Yes
4.13	All providers responsible for providing care to an individual patient will be able to access that patient’s clinical and administrative records across sites to ensure a full (longitudinal) understanding of the patient’s medical history and facilitate deliver of appropriate care. (Based on Interviews, Subgroup Input)	Yes
4.14	Longitudinal data on patients will be filtered for exclusion criteria for each case identified as eligible for a quality measure. (Based on Quality Use Case)	Yes
4.15	Clinicians and hospitals will collaborate to share pertinent information in a timely manner that promotes patient safety and quality improvement. (AQA Reporting Principles document)	Yes
Data Stewardship
4.16	The data stewardship entity will be responsible for ensuring all nationally established standards related to data exchange and aggregation are adhered to all institutions and data under its purview and shall established more detailed operational standards or rules as appropriate. (Based on Interviews)	No
4.17	The data steward entity will be objective in its decision making; weigh carefully the views of its constituents in developing concepts and operating rules and standards; bring about needed changes in ways that minimizes disruption to current aggregation efforts; review the effects of past decisions and interpret, amend or replace operating rules, standards and processes in a timely fashion when such action is indicated; follow an open, orderly process for setting policies, operating rules and standards that precludes placing any particular interest above the interests of the many stakeholders who rely on health care information. (Based on AQA Data Sharing and Aggregation Principles document and National Health Data Stewardship Entity document)	No
4.18	The data stewardship entity will address various data aggregation issues including required characteristics of aggregators (e.g., they should be trusted and respected entities), transparency of aggregation processes, control and ownership rights of the data, potential liability within data aggregation processes, and issues that arise when competing aggregation efforts are in a single market area; should ensure that the experience of existing aggregation efforts are leveraged. These responsibilities will be conducted in accordance with any relevant national standards. (Based on AQA Data Sharing and Aggregation Principles document and National Health Data Stewardship Entity document))	No
4.19	The data stewardship entity will set policies, rules and standards for collecting public and private sector data from relevant stakeholders, including providers, employers, health insurance plans and others based on an agreed-upon measurement set; should assess the pros and cons of using data derived from administrative data (e.g., claims, pharmacy and lab data), medical record review and surveys, and develop policies that prioritize data sources based on various dimensions. These responsibilities will be conducted in accordance with any relevant national standards. (Based on AQA Data Sharing and Aggregation Principles document and National Health Data Stewardship Entity document)	No
4.20	The data stewardship entity will address at what specific level(s) data should be aggregated (e.g., individual physician level or group practice level). When making this determination, should consider sample size issues and physician/practice identifier issues. These responsibilities will be conducted in accordance with any relevant national standards. (Based on AQA Data Sharing and Aggregation Principles document and National Health Data Stewardship Entity document)	No
4.21	The data stewardship entity will set methodological rules and standards for aggregating data, including those addressing risk adjustment, measure weights and sample size. These responsibilities will be conducted in accordance with any relevant national standards. (Based on AQA Data Sharing and Aggregation Principles document and National Health Data Stewardship Entity document)	No
4.22	The data stewardship entity will set data analysis rules and standards. These responsibilities will be conducted in accordance with any relevant national standards. (Based on AQA Data Sharing and Aggregation Principles document and National Health Data Stewardship Entity document)	No
4.23	The data stewardship entity will set policies, rules and standards to ensure that the validity of the data submitted is independently audited. These responsibilities will be conducted in accordance with any relevant national standards. (Based on AQA Data Sharing and Aggregation Principles document and National Health Data Stewardship Entity document)	No
4.24	The data stewardship entity will recommend allowable and non-allowable uses of data, based on current law. These responsibilities will be conducted in accordance with any relevant national standards. (Based on AQA Data Sharing and Aggregation Principles document, National Health Data Stewardship Entity document, and the AHRQ Data Steward Request for Information)	No
4.25	The data stewardship entity will specify who should have access to data and applicable limitations, such as confidentiality and privacy rules; should consider policies which allow contributors, including both public and private sector entities, to have access to their own data as well as information which allows them to compare their data against benchmarks. These responsibilities will be conducted in accordance with any relevant national standards. (Based on AQA Data Sharing and Aggregation Principles document and National Health Data Stewardship Entity document)	No
4.26	The data stewardship entity will develop guiding principles for public reporting and reporting back information to clinicians. Screening processes to ensure valid reporting also should be addressed. These responsibilities will be conducted in accordance with any relevant national standards. (AQA Data Sharing and Aggregation Principles document and National Health Data Stewardship Entity document)	No
Record Locator Service
4.27	Record Locator Services will be used to help providers locate relevant records across care institutions for any patients he or she is treating. (Based on Subgroup Input)	No
4.28	Standards will be developed for identifying institutions through which care was delivered so that information can be matched to medical records. (Based on Interviews, Testimony) [Same as Requirement 3.9]	Yes
4.29	Methodologies/algorithms for linking and matching patient records [through a Record Locator Service] will have rapid/near instantaneous processing times. (Based on Subgroup Input) [Similar to Requirement 3.14]	No

4.5 POLICY

NNational, state, and organizational-level policies play a critical role in governing today’s quality enterprise. Practices related to informed consent, patient privacy, data identification and de-identification practices, and information security are governed through HIPAA, common law, state laws, institutional review boards, data safety and monitoring boards, and organizational governing committees. The establishment of policies related to these topics is critical to ensuring the ability of organizations to exchange data, aggregate it, and share it for different end uses including longitudinal analysis and patient-centric quality improvement, while at the same time encouraging transparency, managing liability and protecting privacy. Current payment practices and business models do not promote investment in longitudinal and patient-centric quality improvement across organizational entities. Many argue these issues can only be resolved with the development of national or state policies that incentivize organizations to collaborate and exchange information with a goal of improving care from a patient-centric perspective.

Pilot and grant programs to invest in the exploration of data exchange strategies are often initiated through government funding. These include programs such as RHIOs/HIE’s, CMS demonstration projects, BQI, NHIN, and the Value Exchange program. The results of these programs may serve to provide the evidence needed to advance or change policies at the organizational, state, or national level to promote longitudinal and patient-centric quality improvement. Table 10 outlines policy-based enablers for quality measurement, specifically longitudinal measurement, and barriers that could be overcome by the development of such policies or policy-based initiatives.

Table 10. Policy-Related Enablers and Barriers
Enablers	Barriers
The Presidential executive order requiring Federal agencies and their contractors to promote the use of interoperable HIT so that data can easily be shared. Mandatory reporting programs at the Federal and state levels that include a robust list of measures spanning issues of patient safety and quality Initiation of pilot projects that provide leadership for a national framework (ex: NHIN, Value Exchanges, BQI) and act as learning laboratories to link public and private data sets and assess clinical quality, cost of care and patient experience Organizational policies governing the use of encrypted identifiers so the receivers of the information (providers, hospitals, plans) can re-identify the data while ensuring confidentiality is preserved when doing aggregate or population level analyses	Lack of alignment of reimbursement with quality performance Lack of alignment of reimbursement with care coordination Privacy concerns around the sharing patient identified data across organizations Privacy concerns often prevent aggregation of identified patient data The need for clear policies to define the use of patient-identified data in large data warehouses HIE is operational in less than 10 regions to enable data sharing for Quality Reporting Gaps in regulations and practices relating to privacy/security and secondary use of data Lack of standardized mechanisms for external reporting including data stewardship Limited access to broad band across regional locations, particularly in rural areas Limited EHR infrastructure within underserved communities without assistance and associations with larger organizations

Policies are needed at the organizational, community, state, and national level to facilitate the ability to conduct longitudinal analyses through data exchange and aggregation across institutions. They are also needed to facilitate the use of data (both point of care data and longitudinal data), while at the same time maintaining appropriate privacy and security safeguards for the sharing and use patient data. Many interviewees indicated that lack of a detailed legal framework will be a critical barrier to enabling longitudinal analysis in the future. Table 11 describes future requirements related to critical policy issues that need to be addressed in the future state that have been articulated by thought-leaders in the health care industry.

Table 11. Policy-Related Future Requirements
ID	Description	Common Agreement
5.1	All HIEs, including in support of the delivery of care and the conduct of research and public health reporting, will be conducted in an environment of trust, based upon conformance with appropriate requirements for patient privacy, security, confidentiality, integrity, audit, and informed consent. (Markle, Technology Principle: Privacy and Security, located in the Common Framework Overview and Principles document)	Yes
5.2	Systems or processes to share, collect, aggregate and report quality, cost of care and patient experience data will be compliant with privacy, confidentiality and other applicable rules, while ensuring that providers, plans, other data contributors, and consumers have necessary and appropriate access to useful information. (AQA Data Aggregation Principles document)	Yes
5.3	There will be a general policy of openness about developments, practices, and policies with respect to personal data. Individuals should be able to know what information exists about them, the purpose of its use, who can access and use it, and where it resides. (Markle, Policy Principle: Openness and Transparency, located in the Common Framework Overview and Principles document)	Yes
5.4	Individuals will control access to their personal information. (Markle, Policy Principle: Individual Participation and Control, located in the Common Framework Overview and Principles document)	Yes
5.5	Personal data will not be disclosed, made available, or otherwise used for purposes other than those specified. (Markle, Policy Principle: Use Limitation, located in the Common Framework Overview and Principles document)	Yes
5.6	Personal data will not be disclosed, made available, or otherwise used for purposes other than those specified. (Markle, Policy Principle: Use Limitation, located in the Common Framework Overview and Principles document)	Yes
5.7	Individuals will have the right to: • Have personal data relating to them communicated within a reasonable time (at an affordable charge, if any), and in a form that is readily understandable; • Be given reasons if a request (as described above) is denied, and to be able to challenge such denial; and • Challenge data relating to them and have it rectified, completed, or mended. (Markle, Policy Principle: Individual Participation and Control, located in the Common Framework Overview and Principles document)	Yes
5.8	Personal data will be protected by reasonable security safeguards against such risks as loss or unauthorized access, destruction, use, modification, or disclosure. (Markle, Policy Principle: Security Safeguards and Controls, located in the Common Framework Overview and Principles document)	Yes
5.9	Entities in control of personal health data will be held accountable for implementing information use and safeguard practices. (Markle, Policy Principle: Accountability and Oversight, located in the Common Framework Overview and Principles document)	Yes
5.10	Legal and financial remedies will exist to address any security breaches or privacy violations. (Markle, Policy Principle: Remedies)	Yes
5.11	A organization (provider or HIE) will incorporate a governing body to develop, maintain, track, and enforce access and control rules (Testimony)	Yes
5.12	Payment strategies for care provision will reflect shared accountability across organizations. (Based on Quality Workgroup Vision document) [Also included as Requirement 1.7]	No
5.13	Government, payer, and employer-sponsored initiatives will provide financial incentives to drive adoption of EHRs and other quality measurement and care coordination enabling technologies. (Based on Quality Workgroup Vision document) [Also included as 1.8]	No

5. UNRESOLVED CRITICAL PATH REQUIREMENTS

The requirements articulated in Section 4 describe critical features needed to enable the envisioned future state, one where longitudinal analysis is commonly used to improve care delivery, and the ability to collect, share, and report information is facilitated by HIT. However, as noted throughout the Section 4, there are several areas where there is a lack of common agreement regarding those requirements or where requirements have not been articulated. We reviewed these areas and identified a core set of topic areas that are part of the critical path to achieving the future state.

This section presents several scenarios that describe various options for the future state for areas that are part of the critical path to the future state, and discusses the strengths and weaknesses of each option. These scenario include:

High versus Low EHR Adoption
The use of a National Patient Identifier versus matching algorithms
The locus for patient records de-identification
The locus of data aggregation

The intent of this section, as with the requirements summarized in Section 4, is to foster discussion and dialogue regarding recommendations that the Quality Workgroup could put forth to help drive the industry towards the envisioned future state.

5.1 HIGH VS. LOW EHR ADOPTION

The requirements in Section 4 are based on an assumption that in the future state, EHR use is the norm for health care providers across all care settings. The ability to achieve high penetration of interoperable EHRs is impacted by many factors; political, financial, technical, and cultural. Thus, considering the implications of low versus high EHR penetration can provide valuable input on important interim requirements to consider. Table 12 presents strengths and weaknesses to consider for either scenario.

Table 12. Future Scenario: High versus Low EHR Adoption
Low EHR Adoption	High EHR Adoption
Description
Quality improvement will be based primarily on a hybrid of claims and clinical data.	Quality improvement will be based primarily on clinical data.
Strengths
A limited degree of quality assessment is possible using hybrid claims and clinical data.	Assessment of quality based on clinical data is the gold standard.
For the limited number of nationally-endorsed quality measures, which use claims data, consistency in interpretation and transparency in reporting has already been established.	EHRs can be integrated with other data sources, supporting capabilities such as HIE resulting in more comprehensive clinical information in the hands of clinicians (e.g., availability of a longitudinal patient record is possible with high EHR adoption).
Weaknesses
Unevenness in measurement ability due to variations in available data sources across organizations.	Clinical decision support tools are still evolving, to fully take advantage of high EHR adoption
Unable to link disparate data sets together to ensure richer clinical picture at point of care. Inability to truly assess quality from claims data (though efficiency can be assessed).	Standards in data transfer, data content and use of data within EHRs still evolving.
Claims are not always filed for all visits.
Claims data integrity is sometimes questionable.

5.2 USE OF A VOLUNTARY UNIQUE PATIENT IDENTIFIER VERSUS MATCHING ALGORITHMS

The ability to match patient records is integral to enabling HIE. Use of a voluntary unique patient identifier is a subject that has come up repeatedly in today’s HIE debates and is one that is filled with many political overtones. Some states have implemented voluntary patient identifier programs, however for the most part, existing mechanisms for patient record matching rely on the use of organization-level identifiers and matching algorithms. There are opportunities to learn from public health agencies on methods to match patient records across disparate sources. By improving accuracy in matching (less false positive and / or negatives), patient safety is enhanced and the likelihood for redundant tests is decreased. Additionally, with more efficient matching, costs for providers and others to maintain accuracy in their local files decreases.

Table 13 presents an overview of the strengths and weaknesses to both approaches: use of a voluntary unique patient identifier versus matching algorithms.

Table 13. Future Scenario: High versus Low EHR Adoption
Voluntary Unique Patient Identifier	Matching Algorithms
Description
Patient-record matching will be accomplished through the use of voluntary unique patient identifiers at the state or regional level. Limited matching will still be needed.	Patient-record matching will be accomplished through the use of matching algorithms.
Strengths
Methods for matching will be easier to implement because only one major system will be in use.	Public concern over privacy and security will be tempered with use of algorithms and established stewardship of data.
Record matching may be possible at the state or regional level, depending on the level of implementation.	Development of matching algorithms and a core set of matching requirements or principles can be used to facilitate record matching without significant system changes.
Weaknesses
Safeguards for national patient identifier theft and protection are necessary if such a program is implemented.	Accuracy of matching and ensuring the “right” patient match is dependent on application of algorithms.
Public concerns over privacy and security may derail establishment of national patient identifier.	Multiple algorithms will need to be developed and maintained to accommodate the varied record identification methodologies used across diverse systems.
Information technology system changes will be needed across any states or regions where an identifier program is implemented.

5.3 LOCUS OF PATIENT RECORDS DE-IDENTIFICATION

Patient record de-identification is necessary to ensure privacy is protected when data are used for purposes other than direct clinical care. However, when data are used for quality measurement purposes, in order to provide feedback at the patient-level, a re-identification process must also occur. The locus of that de-identification/re-identification activity may occur at several different levels in the future and each option has implications for the ability to aggregate data across sites and data sources, such as: who is privy to the information necessary to de-identify and re-identify records; who can provide patient-level feedback; and how long will the feedback loop take. Table 14 describes how this activity would occur at three different levels and also describes strengths and weaknesses related to each option.

Table 14. Future Scenario: Locus of Patient Record De-Identification
Provider/Care Delivery Organization	HIE	Quality Organization
Patient data is de-identified at the care delivery organization prior to submitting data to an HIE or other entity for aggregation.	Patient data is received from care settings and de-identified in a database or RLS as managed by the HIE. HIE will serve as the owner of the quality improvement feedback loop to providers and care delivery organizations. HIE will also serve as the steward for the data it maintains.	Patient-identified data is received from HIE for translation into quality measures. Quality organization will serve as the owner of the quality improvement feedback loop to providers and care delivery organizations. De-identification of data occurs at Quality organization prior to reporting.
Strengths
Patient privacy protections are greatest when the only place there is identified data is at the care delivery organization.	Patient records can be linked across care settings; a longitudinal patient medical record is possible	Patient records can be linked across care settings.
Data held within a care delivery organization is bound and protected by HIPAA regulations.	Longitudinal analysis is possible; data can be used to support longitudinal quality measurement.	Longitudinal analysis is possible.
	Established protocols are uniformly followed by HIEs today and data are appropriately safeguarded in de-identification and re-identification activities.	Data can be used support to support longitudinal quality measurement.
	Data stewardship and privacy policies are already established and there is transparency in the process, to minimize public concerns over privacy.	Patient-level feedback loop to clinicians is streamlined because Quality Organizations can provide feedback directly to the providers and care delivery organizations without going through the HIE to re-identify data.
	Submitting organizations do not have burden of ensuring anonymity of patient data.
Weaknesses
Ability to link patient records across care settings is constrained.	Patient-level feedback may be delayed since data must be de-identified, sent to the Quality Organization, returned in the form of usable metrics, re-identified, and sent back to the providers and care delivery organizations.	Additional rules, standards, and agreements are needed to protect patient privacy since de-identified data will accessible by both the HIE and Quality Organization.
Ability to use the data to support longitudinal analysis or quality measurement is inhibited.	Mechanisms and standards for de-identification and re-identification must further evolve to meet industry-wide needs.	Secondary use of patient-level clinical data may be limited when managed by Quality Organization as compared to HIE.
Applies additional burden on provider settings to de-identify data prior to transmission/submission to various entities.
Introduces risk to integrity of data de-identification and risk of security breaches when de-identification responsibility is decentralized across individual care delivery organizations.

5.4 LOCUS OF DATA AGGREGATION

Data aggregation is necessary to develop quality metrics and perform any degree of population level analysis. The level at which data aggregation occurs, if at all, has implications for the infrastructure, rules, policies that are needed and types of analyses that are possible. Table 15 presents the strengths and weaknesses when considering three different approaches: no aggregation, local/regional aggregation, and national aggregation.

Table 15. Future Scenario: Locus of Data Aggregation
None	Local/Regional Aggregation	National Aggregation
Description
Data are not aggregated. An RLS is the primary mechanism for exchanging data across sites.	Data aggregation occurs at the local/state level. Data are stored in multiple regional databases.	Data aggregation occurs at the national level. Data are stored in a central database.
Strengths
Data can be used to facilitate care delivery.	Data can be used to facilitate longitudinal quality improvement initiatives at the local/regional level.	Data can be used to facilitate longitudinal quality improvement initiatives at multiple levels (local, regional, national)
Provider organizations can request and manage point-to-point transfer and receipt of patient data for established purposes such as medication reconciliation.	Strategies for matching patient data can be established at the local/regional level, building on current local/regional efforts and leveraging local innovations.	New opportunities exist for nationwide application of rich clinical data such as “electronic epidemiology,” adverse event tracking, and various other secondary uses of clinical data.
Standards for data exchange, security of data, established sharing agreements, and patient privacy protections may be easier to establish at the local/regional level; a business case for doing so may also be easier to establish and minimize public concerns over privacy.	New opportunities exist for nationwide application of rich clinical data such as “electronic epidemiology,” adverse event tracking, and various other secondary uses of clinical data.
Partnerships in quality improvement may be encouraged at local level as payors, providers and patients work together through local/regional collaboratives.
Weaknesses
Data cannot be easily used to facilitate longitudinal quality improvement initiatives.	National strategy/framework for alignment is necessary (and as yet undeveloped) to build from local efforts and enable population analyses at national level.	A national strategy will need to be developed to match patient data.
There is a need for policies and standards for security and data sharing at the source system or care delivery organization level.	There is potential for inconsistencies across localities or regions as aggregation methods may differ.	One set of standards will need to be used to manage data exchange, ensure security of data, establish data sharing agreements, and ensure patient privacy is protected.
A back up strategy will be needed if source systems go down or are unavailable.	Individual localities and regions are subject to local pressures, policies and needs and may threaten overall standardization and application of data.	Database scale may be an issue.
Transfer or exchange of information is the responsibility of the care delivery organization.	The ability to manage and facilitate data exchange may be an issue.
Provider and community pushback for patient data storage outside the community.
Public concerns over ownership of patient data, and security concerns over protection of that data

6. NEXT STEPS

Based on the requirements analysis, several topic were identified as potential areas where the Quality Workgroup may have an opportunity to develop recommendations. These topics focus on areas where improvements could facilitate the ability of HIT to support (longitudinal) quality measurement and improvement. They include areas where there may be an opportunity to inform changes to the overall environment and also areas that are more HIT-specific. The list was prioritized by the requirements analysis subgroup by loosely considering the following three factors:

The Quality Workgroup’s ability to affect significant change and/or drive the industry
Whether or not the topics are part of the critical path to the future state
What other initiatives are underway related to the topics.

Topic 1: Data Aggregation

Data aggregation is necessary to develop quality metrics and perform any degree of population level analysis. The level at which data aggregation occurs, if at all, has implications for the infrastructure, rules, policies that are needed and types of analyses that are possible. There is a lack of such common policies, procedures, and standards for data aggregation. Other issues related to data aggregation include a determination of at which locus data aggregation should occur (not at all, locally/regionally, nationally), and the debate of using a RLS versus databases that can aggregate data.