Monday, June 22, 2009

Commentary: Meaningful Use

“The American Recovery and Reinvestment Act authorizes the Centers for Medicare & Medicaid Services (CMS) to provide a reimbursement incentive for physician and hospital providers who are successful in becoming “meaningful users” of an electronic health record (EHR). These incentive payments begin in 2011 and gradually phase down. Starting in 2015, providers are expected to have adopted and be actively utilizing an EHR in compliance with the “meaningful use” definition or they will be subject to financial penalties under Medicare.” Source: Meaningful Use: A Definition, Recommendations from the Meaningful Use Workgroup to the Health IT Policy Committee, June 16, 2009.

The Office of the National Coordinator for Health Information Technology (ONC) is seeking comments on the preliminary definition of “Meaningful Use,” as presented to the HIT Policy Committee on June 16, 2009.

The definition and measurement of "meaningful use" in the context of health information technology (HIT) implementation has emerged as a critical policy discussion - particularly in light of the need to establish benchmarks for incentive payments mandated in ARRA. The most important difficulties in this discussion result from confusion in other definitions including electronic health records (EHRs) and health information exchange (HIE). Although CCHIT has apparently made progress in presenting criteria for certification of both EHRs and HIEs, their clear definition remains elusive and subject to ongoing debate. In their call for public comment, the HIT Policy Committee proposes a Preamble: Meaningful Use – A Definition and a Meaningful Use Matrix structured according to health care policy priorities. The ultimate goal of meaningful use - as stated in the Preamble – is to “enable significant and measurable improvements in population health through a transformed health care delivery system.” While this broad goal encompasses the entire system, health information technology seems narrowly defined as an Electronic Health Record (EHR), excluding any discussion of infrastructures required to support health information exchange and other core functions of EHRs. The challenge in this framework is to demonstrate how the adoption of an EHR can be associated with emergence of a transformed health care delivery system for improved population health. What are the dimensions of this goal and how should they be operationalized and measured within a particular health care context? The Matrix is an attempt to identify health care policy priorities on five axes – regardless of context: (1) Improve quality, safety, efficiency, and reduce health disparities; (2) Engage patients and families; and (3) Improve care coordination; and (4) Improve population and public health; and (5) Ensure adequate privacy and security protections for personal health information. Each of these axes yields a set of care goals with more specific objectives and measures to be applied in 2011, 2013, and 2015.

Unfortunately, the logic of the Matrix is highly inconsistent. The five policy axes are certainly not orthogonal, but must be interrelated in different ways among complex care delivery systems. For example, improving population and public health would be closely associated with reduction in health disparities, and improved coordination would certainly be linked with better health care quality. Some of the goals listed for particular policy priorities seem misplaced while corresponding objectives and measures seem unrelated or indistinguishable. For example, the goal to “Report to patient registries” in Priority #1 would seem more appropriate to the public health priority (#4). The goal to “Apply clinical decision support at the point of care” (under priority #1) appears also as an objective in 2013 and 2015. Logically the objectives and measures in the Matrix should represent more specific instances within each stated goal. Clinical decision support is a very complex problem inadequately presented as a required component in the determination of meaningful use.

The Matrix time frame in general poses a challenge as the measures for achievement of objectives appear to lose definition (or disappear entirely) from 2011 to 2015, when they are mostly “to be determined” (tbd). (What happened to measurement of health disparities in 2015? Are they supposed to have been resolved by then?) In my opinion there are several reasons for the inadequacy of the Matrix framework.
(1) The logic of the Matrix assumes the existence of infrastructures required for HIT implementation – including the adoption of EHRs. As I have already commented on the proposal for regional extension centers, it is unrealistic to assume that existing policies to foster formation of RHIOs and HIEs will effectively promote a Nationwide Health Information Network (NHIN) without sustained public investment at the federal level. Essential institutions excluded from the ONC discussion of meaningful use include educational and research providers. The eventual NHIN infrastructure - including education and research institutions - will have a very substantial impact on the evolving parameters of HIT and EHR meaningful use in health care delivery, especially for health information exchange and interoperability.
(2) Another invalid assumption is that HIT or EHRs – implemented at the physician, group practice or hospital level - will reduce costs or improve the quality of health care without very significant national system redesign. HIT implemented in the existing fragmented system may only result in greater redundancies and inefficiencies – as well as higher costs. (See the report by the CBO: Evidence on the Costs and Benefits of Health Information Technology, 2008. See also the very provocative article by A. Gawande in the New Yorker: The Cost Conundrum: What a Texas town can teach us about health care, 2009)
(3) Information technologies evolve very quickly and new innovations as yet unforeseen may radically affect the field of biomedical informatics and the redesign of the U.S. health care system, especially in this time of crisis. Since the logical structure of the Matrix appears fatally flawed, it is unlikely that this framework would survive even relatively small technology-driven adjustments - to be useful as a robust theory would. This framework would therefore not be effective in motivating the individual or institutional decision-maker to make the sustained financial and human resource investments required for effective EHR implementation and meaningful use.

Like the ONC, professional associations such as AMIA and HIMSS appear to have avoided issues related to operational definition of meaningful use by considering EHR adoption as an essentially binary variable and emphasizing the association between presence of EHR technology and health care outcomes. There should be more careful attention to the complex health care processes linking HIT with such outcomes. According to AMIA, criteria for meaningful use of EHRs should focus on "clinical endpoints achieved," and the "relationship between, and effectiveness of, key EHR functions and performance on quality measures over time." (See AMIA letter to James Scanlon dated April 30, 2009) HIMSS states unequivocally that "quality measures are a by-product of the successful implementation of CCHIT-certified EHR technology, not separate initiatives." (See lines 74-86: Definition of Meaningful Use of Certified EHR Technology for Hospitals - Approved by the HIMSS Board of Directors on April 24, 2009) They formulate an untested assumption of a unique causal link between certified EHR usage and quality - for which there is no logical or empirical justification - except apparently to promote the value and necessity of CCHIT certification.

Unfortunately, there is no evidence base to associate HIT or EHR usage (certified, meaningful or otherwise) with many of the measures of general health care quality outcomes appearing in the Matrix. The low rate of EHR adoption in the U.S. renders quantitative research on the relationship between EHR technology and health care quality impossible at this time. One recent study by Li Zhou et al. to appear in JAMIA attempts to examine EHR functions and performance based on data from the Healthcare Effectiveness Data and Information Set (HEDIS) - only to report that there is no statistical association between usage of EHRs and such performance data. Although this result is probably due to numerous methodological weaknesses - including sample sizes and construct definitions - the study suggests to me that reductionist research designs cannot capture the complexity of EHR meaningful use or the possible causal effects on quality of health care. Furthermore, had this study demonstrated such a statistically significant association, there would have been no validity to any claim of causality. Many unmeasured and uncontrolled contextual variables may co-vary to explain quality of health care.

The lack of a research program and evidence base demonstrating the positive health care outcomes promised with implementation of HIT or EHRs has serious consequences for the motivation to adopt these technologies as well as the determination of financial incentives as mandated through Medicare and Medicaid. Where research evidence for meaningful use of HIT and EHRs does not exist and cannot be generated in the proposed framework, the program of incentives will be wasteful of public funds not only through ineffective incentive payments but also through vulnerability to legal challenge.

Rather than attempting to prescribe and incentivize implementation of HIT and EHRs, the government should directly offer public services funded through sustained federal investment - such as the NHIN - that will shape meaningful use of these technologies. An incentive for example, might be access to technical support services freely available through linkage to the NHIN.

No comments: