If reference‐based benefit designs work, why are they not widely adopted? Insurers and administrators not doing enough to address price variation

The article by Siniako and Mehrota in the current issue of Health Services Research ¹ —“Association of a National Insurer's Reference‐Based Pricing Program and Choice of Imaging Facility, Spending, and Utilization”—adds to a small but consistent literature demonstrating that referenced‐based pricing can reduce health spending for specific services by prompting patients to use lower cost providers or products.1, 2 Reference‐Based Benefit Design (hereafter referred to as RBBD) originated in European countries for the purpose of controlling health care spending due to price variation for pharmaceuticals within the same therapeutic class.2 By setting a reimbursable price within each therapeutic class that was at the low end of the price distribution, national health insurance programs provided an incentive for manufacturers and distributors to lower their price or risk losing significant market share.

Reference‐based pharmacy benefits have been adopted by some payers in the United States, but are not commonplace in public, self‐insured, and commercial risk‐bearing insurance programs, including the Affordable Care Act's health insurance exchange program. Robinson et al ⁴ examined the outcome of a RBBD program initiated by the RETA Trust, a national association that purchases health care for a membership‐based religious organization. The RETA Trust set up a RBBD program where the allowed amount the insurance plan would cover for drugs within a therapeutic category was based on the lowest priced drug in that category. Patients choosing more expensive drugs were required to pay the entire difference out of pocket. In a well‐designed evaluation, RBBD patients were more likely to select lower price drugs than patients in a control group, resulting in lower spending per prescription and overall savings for RETA, but also higher average out‐of‐pocket spending because some patients chose drugs with prices above the reference price. ⁴

Despite its origins in pharmaceuticals, RBBD has been more frequently applied to services and treatments rather than drugs in the United States. For example, in a recent review article, Robinson, Brown, and Whaley report that RBBD is being used for knee and shoulder surgery, cataract removal, colonoscopy, outpatient laboratory tests, and expensive diagnostic imaging such as computed tomography (CT) and magnetic resonance imaging (MRI). ² All analyses took advantage of benefit design changes initiated by self‐insured employers (eg, Safeway Corporation), public employee and pension programs (eg, CalPERS), or employer or union trusts such as RETA. The impact of RBBD has been examined relative to control groups of similar beneficiaries—often enrollees of a large insurance plan or a comparable employer or union group. Methodologically, studies use a difference in difference design to compare changes in outcomes over time (ie, pre‐post changes) for beneficiaries in the RBBD group relative to the control group. Outcomes of studies examining the impact of RBBD typically look at pre‐post changes in the probability of choosing a lower priced service provider, changes in overall spending for the relevant service(s) in the insured population, changes in beneficiary out‐of‐pocket costs, and changes in complication rates or quality measures. Uniformly, as reported by Robinson, Brown, and Whaley, RBBD has been shown to have its intended effect, meaning significant savings for the insured population of interest as market share shifted to lower price providers (eg, estimates range from a 9‐percentage point increase in use of low‐price providers for cataract removal and CT scans to a 19‐percentage point increase in use of low‐price providers for outpatient laboratory tests). Shifting market share to lower priced providers resulted in an 11 percent‐21 percent reduction in prices paid, which would translate into billions of dollars of savings if RBBD was applied widely across the United States. ²

Sinaiko and Mehrotra ¹ examined the impact of a RBBD implemented by three self‐insured employers in 2015, relative to a sample of adults with similar PPO coverage and provider network, but without the RBBD option. The focus for the RBBD was advanced imaging, including CT and MRI, with the RBBD applied to the facility fee but not the professional fees. Using 12 months of prior claims data, as well as 23 months of claims data after the benefit design change, the authors estimated the impact on facility choice, net procedure price, spending, and advanced imaging utilization. Sinaiko and Mehrota found a 9.3‐percentage point increase in the use of lower price imaging facilities over the 23‐month RBBD period, and a $101.05 decrease in net prices paid for imaging services in the second year (but not in the first year). Out‐of‐pocket spending was significantly higher in the first year of the RBBD but not in the second year, and there was no difference in utilization of CT scans or MRIs between the treatment and control groups after the RBBD. The use of lower priced facilities resulted in an 8.1 percent decrease from baseline expenditures for these imaging procedures for these self‐insured employers. ¹

Given the consistency of results on RBBD from published studies, a puzzling question is why RBBD is not used more frequently to curtail health care spending.

1. INSURERS AND THIRD‐PARTY ADMINISTRATORS DO NOT FACILITATE AWARENESS OF PRICE VARIATION OR MAKE REFERENCE PRICING EASY TO IMPLEMENT

The significant price variation that exists in health care markets could be reduced by more aggressive price negotiations by insurers, without adopting RBBD or similar benefit designs. If insurers identified price variation within markets for comparable services, particularly relatively homogenous commodities such as laboratory tests or diagnostic imaging, they could negotiate away the variation to the point where pricing would converge to the equivalent of a reference price. In the economic theory of the firm, price would equal the marginal cost of providing the commodity or service, and no single supplier would have enough market leverage to raise price because there would be plenty of supply from other providers to meet the market demand. Some might argue that these services are not entirely homogenous because they can vary on specific dimensions such as convenience of access, ability to share results with providers’ electronic medical record systems, or coordination within a given provider group. If true, the market may be considered “monopolistically competitive” rather than “perfectly competitive,” with product and service differentiation allowing for some degree of price differentiation, but not nearly the degree of variation currently observed.

Thus, it is somewhat of a mystery as to why insurers and third‐party administrators (TPAs) have not been able to achieve price reductions through contracting. One hypothesis is that insurers wish to contract with most providers to offer broad networks in their plan offerings, and thus hesitate to take a hard line in negotiations so as to avoid the possibility of excluding a provider from its network. This explanation is weak, however, since all competing insurers are operating with the same information about price variation and should be negotiating aggressively on behalf of their beneficiaries, self‐insured employers, and public plan sponsors. A second potential explanation is that insurers do not know the full extent of price variation, especially for particular types of services, and thus do not appreciate the variation in net prices that results from negotiation based on charge discounts. This explanation seems more plausible, especially since the federal government and other regulators are pushing the need for government intervention to facilitate more transparency on health care pricing. ¹² Further evidence supporting this explanation comes from the fact that almost all examples of RBBD reported in the literature were undertaken by self‐insured employers or large benefit trusts that administer insurance programs for employees and retirees. These employers and program administrators identified significant price variation in their claims data sets, and designed and implemented RBBD programs to reduce this variation. Some exceptions exist, like the RBBD program implemented by a large national insurer reported in this issue, ¹ but the important question is why insurers and TPAs are not serving as better agents for their customers and forcing prices to converge downward for this set of “shoppable services.”

2. COMMUNICATING HOW THE RBBD WORKS TO BENEFICIARIES IS CONFUSING AND MUST BE CUSTOMIZED AT THE LOCAL MARKET LEVEL

RBBD is predicated on reducing price variation by creating a threat to the market share of high‐priced providers. For these designs to work, there must be capacity in the market for services to move to lower priced providers and transparent information about price and quality. ¹³ For many commodities, such as laboratory or imaging services, the production function should allow for most providers to accommodate increased volume. Also required for implementation of a successful RBBD is clear communication to plan beneficiaries about which services are covered, how the beneficiary can compare provider prices, and the consequences of choosing providers that charge above the established reference price. This communication must be customized to each specific market for purposes of identifying where local providers stand relative to the reference price, which often varies from market to market. Most of the published RBBD studies have involved customized education and communications led by the plan sponsor. However, the management time costs of these activities are significant, and efficiencies could be realized if insurers developed communication tools that could be utilized by all plan sponsors.

3. CONTRACTING AND PAYMENT PROBLEMS DUE TO FEE‐FOR‐SERVICE REIMBURSEMENT

The preceding paragraphs have directed blame at insurance companies and TPAs for not addressing price variation in health care markets and for not pushing more widespread use of RBBD. While not a reason to exonerate insurers and TPAs, the historic fee‐for‐service (FFS) system of payment for health care services in the United States has not helped, because it requires literally thousands of billing codes and code combinations. As such, contracting typically tends to be more global—often based on a discount of provider charges—rather than service‐specific. This broad approach to negotiating on percent discount on charges masks the true magnitude of price variation and thus cannot force prices toward a lower bound. Until our society contracts and pays for health care differently, we will most likely continue to live with unwarranted price variation that will require approaches such as RBBD to “clean up” the inefficiencies of FFS reimbursement.

4. CONCERNS ABOUT EXPOSING EMPLOYEES AND PLAN BENEFICIARIES TO HIGH OUT‐OF‐POCKET COSTS

In another published study, Sinaiko et al ¹⁴ described why more employers are not implementing RBBD programs. Most importantly, administration of such programs is time consuming and difficult for employers to initiate and explain to their employees, and employers are concerned about RBBD programs exposing employees to potentially high out‐of‐pocket costs. In fact, the complexity and costs of administering RBBD programs should motivate insurers and TPAs to offer such programs as turnkey solutions so employers can realize the benefits of RBBD programs without having to research, plan, implement, and monitor these programs themselves. With respect to concerns about exposure to high out‐of‐pocket spending, this argument seems disingenuous because employers have gravitated to high deductible health plans (HDHPs) in an effort to generate more cost consciousness among beneficiaries, despite the fact that such plans push more first‐dollar out‐of‐pocket exposure to employees. As the literature on HDHPs has shown, these plans may discourage any use of services, whereas RBBD programs offer a safer choice: whether to use a low‐price provider for zero or low out‐of‐pocket costs versus whether to use a relatively high‐priced provider with additional out‐of‐pocket consequences. ¹⁴ , ¹⁵

5. SUMMARY

RBBD programs exist because there is significant unwarranted price variation in health care markets for tests, procedures, and services that are relatively homogenous. In theory, insurers and TPAs should be able to negotiate away this price variation on behalf of their clients, which include patients, sponsors of public insurance programs, and self‐insured employers. To date however, insurers and TPAs have not shown the ability to reduce unwarranted price variation. This phenomenon is somewhat puzzling, but may be related to the way health care discounts are contracted in a FFS system, as well as the desire to offer broad provider networks. As such, self‐insured employers and insurance plan sponsors have turned to demand‐side changes involving benefit design, with RBBD being one such option. ⁸ , ¹¹ , ¹⁶ , ¹⁷ While there are some critics of RBBD, ¹⁸ , ¹⁹ most published studies have shown RBBD to be effective at shifting utilization to lower priced providers, resulting in significant reductions in overall spending for RBBD‐covered services. However, the use of RBBD is not widespread and unfortunately unlikely to grow unless insurers and TPAs, including large national insurers, make RBBD easier to implement in local markets by developing communications strategies and digital tools that do not require significant investment and management from plan sponsors. Of course, a more direct path to reducing unwarranted price variation would be for insurers and TPAs to negotiate more aggressively with providers, avoiding the need to administer complex benefit designs such as RBBD.