|Home | About | Journals | Submit | Contact Us | Français|
Tremendous progress has been made in the development of practical quality of care measures that are now applied in a variety of health care settings. In this issue of HSR, Swaminathan et al. use Healthcare Effectiveness Data and Information Set (HEDIS) measures to examine the stability of health maintenance organization quality ratings over time, and Scanlon et al. use these measures to assess if market competition improves quality.
These two important research questions could not have been addressed without established and tested quality measures, and a strong foundation of quality of care research. While this foundation is laudable, we would like to draw attention to a conspicuous problem for most nationally available measures of quality, i.e. they focus primarily on the underuse and misuse of health services, not their overuse. Underuse refers to the lack of provision of necessary or needed care (e.g., withholding aspirin in a patient with coronary artery disease). Misuse of care refers to the provision of wrong care (e.g., prescribing a subtherapeutic dose of a medication). Overuse, however, refers to providing unnecessary care (e.g., echocardiogram in a young healthy patient) or care where the risk exceeds the benefits (e.g., carotid endarterectomy in an otherwise healthy patient with asymptomatic carotid disease). Unnecessary use of health care services is not just a cost issue but also a health care safety problem as patients may suffer complications from procedures that they did not need in the first place. Nevertheless, among the 27 HEDIS Effectiveness of Care measures proposed for Medicare in 2009, almost all address underuse or misuse of health services.
Recent years have seen the funding and advancement of a research agenda designed to develop the tools necessary to assess the quality of care. While there is still much more work that needs to be done, it is notable that most recent advances in characterizing and measuring quality focus on underuse of health care services (McGlynn et al. 2003). While overuse had garnered significant research attention in 1980s and 1990s (Brook et al. 1986; Brook, Park, and Chassin 1990), it has been supplanted with population-based concerns about high versus low utilization in different geographic areas or high versus low utilization physicians (Hayward et al. 1994; Wennberg and Fisher 2002; Fisher et al. 2003; Baker, Fisher, and Wennberg 2008). However, there is little evidence that conclusively finds that higher utilization rates in an area indicate overuse. To the contrary, past research that identified inappropriate use of procedures found no evidence that it occurred more often in high versus low utilization/cost areas (Chassin et al. 1987; Leape et al. 1990).
Despite the important cost, quality, and safety implications of overuse, this branch of quality research has moved from the center stage over a decade ago to the periphery today. We believe that this is a result of research, cultural, and political challenges.
There are many complexities and obstacles in assessing both the underuse and overuse of health services, but arguably, overuse is a much more complex problem. Underuse of health services is generally more recognizable. A hospital can readily ascertain all the acute myocardial infarctions in the past month and determine the proportion of cases that should have received an aspirin or beta blocker at discharge. Researchers can use a plethora of publicly available databases to examine the proportion of diabetics who reported an eye exam or a foot exam in the past year. However, estimating overuse is more difficult. For example, a health plan cannot easily determine whether a child receiving a tympanostomy tube for treating otitis media with effusion was “overuse.” To assess appropriateness, at least one year's worth of medical records documenting the number of episodes and duration of ear infections is necessary (Keyhani et al. 2008).
Finally, the absence of national guidelines covering the appropriate use of services in various clinical scenarios is an impediment to routine assessment of overuse. National guidelines generally apply to few conditions and even then to a narrow set of circumstances. For example, there is little randomized clinical trial evidence that covers all the clinical scenarios in which a patient with symptomatic coronary disease presents to the doctor.
Investigators who pioneered the RAND Appropriateness method (Brook et al. 1986; Chassin et al. 1987; Brook, Park, and Chassin 1990) to measure the appropriateness of health care services attempted to address the lack of specificity of national guidelines. For example, the current national guideline adopted by three leading medical societies on the treatment of otitis media state that a child with bilateral hearing loss and 120 days of ear effusions is a candidate for receipt of ear tubes (Rosenfeld et al. 2004). However, this guideline does not address a child with a past history of ear disease with 90 days of effusion, unilateral hearing loss, and/or speech delay. The RAND Appropriateness method attempts to overcome this limitation in national guidelines by using a two-round modified Delphi process to integrate literature with expert opinion into explicit criteria to generate an exhaustive and mutually exclusive list of potential clinical scenarios to represent the range of circumstances that might present to a clinician. These clinical scenarios are rated as appropriate, uncertain, and inappropriate through an iterative scoring process by a panel of experts (Brook et al. 1986; Chassin et al. 1987). While this method provided the research foundation necessary to address overuse, the further development and implementation of this research has largely been abandoned in the United States.
Despite the demonstrated value of using the RAND method to create detailed guidelines, there are many cultural obstacles to using this method to assess practice. The publication of research attempting to identify overuse can be met with a degree of skepticism and outright hostility from physicians. Such research seems to imply the implicit suggestion that overuse is a manifestation of physicians' self-interested behavior to generate income. While everyone recognizes the system failures evident in underuse and misuse of health services, overuse is misconstrued as a problem arising from both physician's integrity and autonomy rather than arising from system failures. While there is a clear profit motive to promote greater use of procedures such as coronary interventions in hospitals and diagnostic procedures in outpatient settings, there are undoubtedly system failures that contribute to overuse such as poor communication between primary care providers and specialists, the absence of a unified medical record, lack of guidelines, or patient preference among many other potential factors. A greater balance of investigating root causes is needed, including consumer demand, as well as greater attention to factors outside of a physician's control that lead to overuse.
There are many political barriers to advancing the research agenda on overuse. One well-known illustration occurred when the Agency for Health Care Policy and Research (AHCPR), now known as the Agency for Healthcare Research and Quality (AHRQ), published guidelines that suggested that nonsurgical approaches were recommended in the initial management of acute back problems. The guidelines and underlying research supplying evidence led to lobbying efforts from the North American Spine Society, which felt that its scope of practice was threatened (Deyo et al. 1997). The end result was that the House of Representatives passed a resolution in 1996 for zero funding for AHCPR. The budget for the Agency was restored in the Senate after significant efforts by the research advocates. However, this experience led to the creation of a newly named Agency, with a mission that largely abandoned its role in guideline development.
Without political support and funding, an agenda of quality research that includes overuse cannot move forward. Despite the frequent policy discussions on the importance of reducing unnecessary care and the interest in comparative effectiveness, the U.S. government avoids creating national comprehensive practice guidelines. Specialty societies are attempting to create guidelines recognizing that many procedures are used inappropriately. While a few medical societies (Klocke et al. 2003; Epstein et al. 2008) have incorporated the RAND Appropriateness method, many guidelines that are currently issued apply to a narrow set of circumstance or are organized based on levels of evidence without concrete recommendations. It is difficult to see how these guidelines can be translated into measures that can be used to assess practice. Without good measures, overuse in the U.S. health system cannot be dealt with in the same way that underuse is currently addressed through a plethora of private, state, and federal initiatives.
Physicians and their associations are not the only political barrier to moving the field of overuse forward. There are overt political pressures from special interests, such as the medical device industry, to use expensive tests of unclear value. A notable example was pressure on the Center for Medicare and Medicaid Services (CMS) to pay for PET scans to diagnose Alzheimer's dementia (Weiss). CMS and the Alzheimer's Association had concluded that PET scans were not useful in diagnosis. However, political connection of the PET scan manufacturer to Senator Ted Stevens (Weiss 2004) ultimately led CMS to modify its earlier decision and reimbursing PET scans in a narrow clinical circumstance. CMS is constantly subject to this type of political pressure, which also feeds the unnecessary use of health services.
Finally, overuse research is costly. AHRQ, which was an important funding agency of this type of research, no longer has the funding necessary to move this research agenda forward. Development of guidelines for a single procedure using RAND's method for determining appropriateness can be expensive, e.g., convening a national panel of experts over 2 days can vary in cost from $200,000 to $500,000. Nonetheless, the costs of such panels are minuscule in the face of the documented unnecessary use of health services that leads to billions of dollars of unneeded expenditures every year. Until there is a clear recognition by policy makers that significant resources are required to develop any type of quality measures, including measures of overuse or appropriateness, this research will not occur.
While the challenges enumerated are great, there are many reasons to invest in this research and move this facet of the quality agenda forward. The alternatives to directly addressing overuse are the use of blunt measures such as increasing patient's cost sharing. Increasing cost sharing would sacrifice needed care to limit unnecessary care. The RAND Health Insurance Experiment demonstrated that patients, when exposed to high cost sharing, do not discriminate between necessary and unnecessary care (Chernew and Newhouse 2008).
Addressing overuse can ensure access to effective and appropriate care while eliminating unnecessary care. Without political will accompanied by significant funding of this research, examining overuse will continue to be the forgotten stepchild of quality of care research. Ironically, while this research was pioneered in the United States, investigators in other countries seem to have taken on its mantle and continue this type of work ( Junghans et al. 2007; Hemingway et al. 2008). The importance of overuse research is clear—necessary leadership to tackle this complex and thorny problem is lacking. We can do better and we must for quality as well as cost reasons.
Joint Acknowledgment/Disclosure Statement: This project was not directly supported by any external grants or funds. Drs. Keyhani and Siu are supported by grants from the VA HSR&D Service and the National Institute on Aging.
Disclaimers: We have no conflicts of interest to report.
Additional supporting information may be found in the online version of this article:
Please note: Blackwell Publishing is not responsible for the content or functionality of any supplementary materials supplied by the authors. Any queries (other than missing material) should be directed to the corresponding author for the article.