Why So Few Value Assessments On Health Services And Procedures, And What Should Be Done?
Editor’s Note
This article is part of a Health Affairs Forefront short series, “Enhancing Value By Evaluating Health Care Services.” The series discusses ways to extend the use of tools for clinical and economic evaluation beyond medical technologies to the services and procedures that account for the bulk of health care spending; the goal is to create a more robust evidence base for the effectiveness and value of health care services. The articles in the series were completed with support for the authors from the Research Consortium for Health Care Value Assessment, a partnership between Altarum and VBID Health, through a grant from the Pharmaceutical Research and Manufacturers of America (PhRMA). PhRMA extended complete independence to Altarum to select researchers and specific topics. Health Affairs retained review and editing rights.
The field of cost-effectiveness analysis (CEA) has grown rapidly but in somewhat lopsided fashion. The number of published cost-per-quality-adjusted life year (QALY) analyses increased from an average of 34 per year from 1990 to 1999, to 484 per year from 2010 to 2014, to 856 per year from 2016 to 2020. However, the analyses have disproportionately focused on prescription drugs, with relatively less attention to health services and procedures. Despite the fact that pharmaceuticals comprise roughly 15 percent of health spending (estimates depend on how the fraction is calculated and range from 10 to 20 percent), 43 percent of CEAs published since 1976 have focused on pharmaceuticals, according to analyses of the Tufts Medical Center’s CEA Registry. While the bulk of health spending comprises services and procedures delivered in hospitals, clinics, and physician offices, only 12 percent of studies have focused on care delivery interventions and 26 percent on medical and surgical procedures. Moreover, the proportion of CEAs devoted to pharmaceuticals has remained steady over time.
The Institute for Clinical and Economic Review (ICER)—the nonprofit organization that conducts clinical and economic analyses of health interventions and serves as a kind of de facto national health technology assessment (HTA) agency—has focused mostly on drug therapies. Since 2015, 47 of 54 of ICER’s evaluations have pertained to prescription drugs (two of the 47 evaluations included both drug and non-drug therapies). Non-drug evaluations included assessments of programs to integrate behavioral health into the primary care setting and intensive diet and lifestyle counseling to prevent diabetes. In 2019, ICER announced that it would bolster its efforts to review non-pharmaceutical interventions. However, since then, only two of 18 of ICER’s completed or ongoing assessments have done so: evaluations of digital health technologies and supervised injection facilities in response to the opioid epidemic.
In this article, we explore these trends, building on prior studies in the area. We examine reasons for the discrepancies and consider potential action to shore up the evidence base on the cost-effectiveness of health services and procedures.
Why The ‘Under-Study’ Of Health Services And Procedures?
Two factors in particular may explain the relative dearth of CEAs on health services and procedures. One is the absence of suitable data. Unlike pharmaceuticals, which are required by the Food and Drug Administration to undergo substantial pre-market testing—generally involving randomized controlled trials (RCTs)—most services and procedures lack such investigations (though there are exceptions). Compared to non-randomized studies, RCTs tend to provide more credible information on the effectiveness of new interventions measured against the current standard of care, an essential ingredient for CEAs.
One can, of course, construct estimates of an intervention’s effectiveness without RCTs, but typically this relies on observational data and requires statistical techniques to control for possible confounders. While the field has advanced methodological approaches (e.g., instrumental variables, propensity score matching) for such analyses, they require more assumptions and uncertainty, and they are more prone to the introduction of improper inferences. Another complication in assessing services and procedures is the role of clinician-related factors (e.g., “learning effects”). It is well documented, for example, that outcomes are often better for providers who perform higher volumes of services and procedures. Such effects may be present with pharmaceuticals but are generally less pronounced.
Information on an intervention’s cost is also generally more readily available for drug than non-drug therapies. Though there is debate about how to estimate drug costs in CEAs, standard sources, such as the Federal Supply Schedule, exist and are recommended by the Second Panel on Cost-Effectiveness in Health and Medicine. In contrast, cost data on services and procedures are often obtained from disparate administrative claims databases or cost accounting systems, and estimates vary widely across regions and settings.
A second factor pertains to the nature of the intervention. Prescription drugs provide a distinct target given therapies’ intellectual property protections. Drug companies that develop safe and effective products are rewarded with temporary monopolies, which shield them from competition. As we have written elsewhere, this outcome is a design feature rather than a “bug” in the system, providing strong incentives for companies to undertake risky investments. However, it also provides incentives to price products above the value delivered, posing a challenge for the health system. CEA offers a means to help ensure that prices align with value. For their part, drug companies have conducted their own CEAs to showcase for payers the value of their products, a phenomenon that has also driven the increase in published CEAs on pharmaceuticals.
In contrast, services and procedures generally lack property rights. The health system has evolved other payment mechanisms (e.g., diagnosis-related groups for hospital care and fee schedules for physician visits) to reimburse for services and procedures, and to align prices with costs if not always with value. To be sure, individuals and groups who champion particular services or procedures have their own incentives to conduct CEA to persuade payers about the value of certain interventions, but the inducements tend to be more limited than those regarding pharmaceuticals .
Whatever the reason, the absence of more CEAs on services and procedures is concerning, because the disproportionate application of CEAs shines a brighter light on one corner of the health care landscape than others and thus could lead to a misallocation of resources. For example, because cost-effectiveness ratios are less likely available for services and procedures, it could mean more low-value care is provided in those areas.
What To Do?
Several steps would help the field. ICER should conduct more CEAs on health services and procedures, and, as noted, the organization has announced that it will do so. Still, there is a limit to what a privately funded organization can accomplish. Economic evaluations are public goods and thus best funded by the public sector, which would bring accountability and resources to the effort. Existing government agencies (e.g., the National Institutes of Health, Agency for Health Care Research and Quality, and the Veterans Health Administration) could conduct or fund CEAs of non-drug interventions, though there are practical limits given their missions and priorities. Other agencies, such as the Patient Centered Outcomes Research Institute and the Centers for Medicare and Medicaid Services, are logical candidates, but they are prohibited by statute and precedent from conducting CEAs. Legislative changes would likely be required and would face a difficult path.
A new, publicly funded HTA organization would be a valuable addition. As we and others argued in a recent report, the new entity could prioritize important understudied areas, including evaluations of health services and procedures. Importantly, the organization would be advisory, disseminating information for the marketplace rather than mandating decisions. Many HTA bodies worldwide have established programs to assess clinical and economic evidence for devices, services, and procedures, including CADTH in Canada and IQWiG in Germany. England’s National Institute for Health and Care Excellence (NICE) maintains a separate work stream devoted to public health and social care interventions.
With or without a new HTA entity, private payers in the US have an important role to play. Payers have initiated value-based programs targeting services and procedures. For example, value-based insurance design (VBID) efforts—which provide incentives to encourage use of “high-value” services and discourage “low-value” care—have generally focused on tests, services and procedures rather than drugs. These endeavors have sometimes been informed by efforts from groups such as Choosing Wisely and other organizations that have identified low-value services to be avoided. However, these initiatives have generally not relied on CEAs to determine value; instead they have based identification of low-value services on measures of clinical benefit or ill-defined assessments of whether a service’s costs are excessive. According to one recent study, explicit consideration of economic value was reported for only 15 percent of over 1,000 low-value care interventions identified by five prominent organizations (US Preventive Services Task Force, Choosing Wisely American College of Physicians, American College of Cardiology/American Heart Association, and American Society of Clinical Oncology.)
Without formal assessment of value, payers cannot properly assess opportunity costs and thus reveal the trade-offs involved in choosing one intervention over another; therefore, they cannot maximize population health given available resources. A service or procedure deemed “low value” based on its high cost may actually be cost-effective given its clinical benefit; conversely, one deemed “high value” based on clinical benefit might still be overpriced. Recent health insurer decisions to cease coverage of proton beam therapy for certain cancers, or to lower the price of robotic surgery, show the value of using more explicit cost-effectiveness considerations.
Improvements in data and methods would also help. While the field has seen rapid growth in electronic health records, administrative claims databases, registries, and other “real-world” data sources, it has lacked standards for measuring costs and capturing health outcomes across platforms and geographic settings. Greater standardization and common definitions would enable more precise and replicable estimates to inform CEAs of services and procedures. Regulators and stakeholder groups have begun issuing guidance for real-world evidence in terms of appropriate data sources, methods, and research questions to inform drug development and approval decisions. Similar guidance from HTA bodies on services and procedures would bolster use of real-world data in these areas.
A number of other actions would also help. More RCTs and well-conducted observational studies on the effectiveness of health services and procedures would provide more reliable inputs for CEAs. More federal funding for such work would be required. A multi-stakeholder panel could convene to consider criteria for prioritizing non-drug topics for CEAs, based on factors such as population size, spending levels, uncertainty in existing care, and “disruptive” potential. The panel might identify areas for methodologic standards for evaluating non-drug interventions. For example, compared to evaluations of pharmaceuticals, it is often more difficult to isolate the clinical benefits attributable to a service or procedure provided as part of a multi-component program (e.g., hypertension management). Finally, journal editors could play a role by giving higher priority to CEAs of non-drug therapies.
The importance of these steps should not be underestimated. Understanding the potential cost-effectiveness of all types of health care interventions is critical for improving the efficiency of the health system and optimizing population health.
Authors’ Note
The authors are members of the Center for the Evaluation of Value and Risk in Health at the Institute for Clinical Research and Health Policy Studies at Tufts Medical Center. The center receives funding from government, private foundation, and pharmaceutical industry sources. Drs. Neumann and Ollendorf have consulted with pharmaceutical companies on issues related to health economics and outcomes research.