Comparative effectiveness research as choice architecture: the behavioral law and economics solution to the health care cost crisis.

• Author: Korobkin, Russell
• Position: IV. Relative Value Health Insurance A. The Legal Status of Relative Value Health Insurance 2. Ex Ante Exclusions through Conclusion, with footnotes, p. 549-574

2. Ex Ante Exclusions

   The legal limitations on point-of-treatment utilization review by insurers contrast starkly with the fact that, in most cases, insurers may legally refuse to pay for interventions that are explicitly excluded by the insurance contract. (119) A patchwork of state "mandated benefits" laws requires insurers to cover specified categories of treatments. (120) Pre-ACA federal law includes a handful of private insurance treatment mandates, (121) and the ACA requires that a set of minimum benefits be included in all insurance policies sold in the individual and small-group markets. (122) Beyond these mandates, however, insurers may legally exclude specified interventions from coverage, and courts routinely uphold their right to do so as a matter of freedom of contract. (123)

   Many insurance plans come with pharmaceutical "formularies," for example, whereby drugs in more favored coverage "tiers" require lower patient copayments, and drugs in less favored tiers require higher copayments or are even excluded from coverage altogether. (124) Using an emerging practice known as value-based insurance design ("VBID"), (125) insurance companies and self-insured employers have experimented with offering reduced or even zero copayments for prescription drugs that, when taken as directed, are particularly likely to reduce future health care costs. (126) And, of course, health insurers can and do limit coverage to care provided by hospitals and physicians within their provider network or require greater cost sharing if a patient chooses to go "out of network" for treatment. (127) By limiting drug formularies and practitioner networks to those drugs and providers offering insurers the greatest discounts, insurers can try to use bargaining leverage to negotiate lower prices for covered services.

   Against this background, there is no impediment, in theory, to insurers excluding from coverage treatments that fail to satisfy a cost--benefit test, as long as the exclusions can be adequately specified at the time of contracting. Further, there is no impediment to insurers offering multiple products, priced differently, that exclude from coverage specifically enumerated categories of care.

   2. The Information Problem

      If insurance companies may legally sell health insurance that covers only cost-effective treatments, why does no such product exist in the marketplace? The primary impediment to the sale of health insurance that covers only cost-effective interventions appears to be the difficulty of adequately specifying the relevant coverage exclusions ex ante. (128) There are three related problems:

      First, there is very little solid information about even the basic effectiveness of most medical interventions--according to some estimates, there is scientific evidence for the efficacy of less than half the treatments doctors recommend. (129) Even clinical practice guidelines are notoriously based on consensus opinion rather than scientific fact. (130) There is even less information about the comparative effectiveness of alternative plausible interventions. (131) Even when the law requires a treatment, such as a new pharmaceutical, to obtain regulatory approval before being marketed, its producers usually must demonstrate only that it is safe and effective relative to a placebo rather than comparatively effective vis-a-vis other treatment options for the same condition. This dearth of information makes it extremely difficult for any insurer interested in marketing a policy that covers treatments that satisfy a cost-effectiveness standard to identify ex ante which treatments are, in fact, cost-effective.

      Scholars have long advocated for insurers to contract to provide care that satisfies a well-specified cost--benefit algorithm, which the insurer would then apply at the point of treatment. (132) This creative idea has fallen on deaf ears in the marketplace, probably because the lack of good data would likely subject any insurer's attempt to apply the algorithm to second-guessing, charges of moral hazard, and lawsuits.

      Second, the measures of marginal effectiveness of competing interventions are dynamic; the measures can change quickly when new effectiveness data is produced, when new interventions are developed, or when the market changes (such as when a drug goes off-patent). Even if an insurer could fully specify cost-effective interventions at the time of contracting, the lag time between contracting and use of services would mean that, at the point of treatment, a policy would cover some no-longer-cost-effective interventions and would not cover some now-cost-effective interventions.

      Third, a detailed list of covered and excluded interventions would provide far too much information for boundedly rational consumers to take into account at the time of contracting. Consumers have the working memory to take into account only a handful of attributes when making purchasing decisions, and they almost invariably selectively consider only the most salient product attributes when bombarded with information. (133) Except for patients with significant preexisting conditions, there would be an extremely low probability that any potential condition-intervention pair would become relevant during the policy period. This suggests that consumers are likely to ignore most detailed coverage information. If consumers did not incorporate information provided at the time of contracting into their purchase decisions, the same reverse moral hazard problem associated with post-contractual utilization review would exist: insurers would have a profit incentive to claim to provide cost-effective care but actually not provide even cost-effective care. (134)

   3. CER and Relative Value Ratings

      These informational impediments that prevent insurers from marketing insurance policies that cover only cost-effective treatments can only be overcome with a significant investment in "comparative effectiveness research" ("CER"). The goal of CER is to provide a firmer scientific understanding of the relative clinical benefits of competing medical treatments, services, and interventions. (135) The American Recovery and Reinvestment Act of 2009 (commonly known as the "stimulus bill") provided $1.1 billion to three agencies to conduct CER. (136) The ACA doubled down on this investment, providing $500 million annually beginning in 2013 to 2014. (137)

      The stimulus bill created a federal commission called the "Council for Comparative Effectiveness Research" to coordinate CER among federal agencies and tasked the Institute of Medicine ("IOM") with recommending research priorities. (138) The IOM quickly provided a list of 100 "top priority" topics for CER, including many studies that would explicitly compare alternative treatments for common medical conditions. (139) The ACA then changed the administrative structure, replacing the Council with a private nonprofit organization called the Patient-Centered Outcomes Research Institute ("PCORI"). (140) PCORI is now charged with setting CER priorities. (141) Its governing board includes government officials and representatives of various stakeholder groups, such as patients, physicians, insurers, and manufacturers of drugs and medical devices, but it is required to ensure peer review of the research it funds and may appoint expert advisory panels. (142)

      How the information generated by CER is used is critical to its potential to help rationalize health care spending. One way CER can reduce health care costs is by supplying providers with better information about which treatments either do not work at all or provide no marginal benefits relative to cheaper interventions. As President Obama put the point in 2009, "[I]f there's a blue pill and a red pill, and the blue pill is half the price of the red pill and works just as well, why not pay half price for the thing that's going to make you well?" (143)

      As impeccable as the logic of this point may be, however, eliminating treatments that have absolutely no marginal benefit is unlikely to significantly "bend the curve" of health care costs (that is, reduce the rate of health care inflation). But CER also has the potential to help reduce the provision of care that has a positive expected benefit but is not justified by its cost. By facilitating understanding not just of the absolute effectiveness of treatments but also of their cost-effectiveness, CER can provide the informational basis necessary for private insurers to sell RVHI.

      For CER to facilitate RVHI, its findings should be used to assign scores to potential medical interventions for different conditions based on marginal costs and marginal benefits. I call such scores "relative value ratings," and I propose that they range from a high score of "1" (extremely cost-effective) to a low of "10" (not at all cost-effective), although other scales would be plausible as well. As an illustration of how the ratings scale would work, consider the following three examples:

      * Standard treatment regimens for cardiovascular disease are understood as one of the great success stories of improved medical technology in the second half of the twentieth century. In 2004, health economist David Cutler estimated that the expected lifespan of an average forty-five-year-old would increase by 4.5 years as a result of this technology, at a total cost of about $30,000. (144) This intervention--or set of interventions--would likely earn the highest possible relative value rating of "1" for patients with relevant symptoms.

      * At the other end of the relative value spectrum, consider an intervention that harkens to President Obama's example of the two different colored pills with identical effectiveness and radically different prices. According to an executive of a health insurance company, the brand-name ache medication, Minocin PAC, retails for $668 per month, which is $618 more than the generic equivalent. The brand-name product is...