Commentary: It’s more than just the price

1. INTRODUCTION

In this issue of HSR, Dowd and Laugesen focus on one aspect of health care that fails the usual criteria for economic efficiency in markets and then suggest ways to remedy that problem. ¹ If prices do not reflect the marginal costs to suppliers—physicians in this case—the wrong mix of services will be produced in response to the distorted prices. Competitive markets reduce price distortion. A well‐designed administratively set fee schedule could also minimize price distortion. Dowd and Laugesen argue that Medicare, which sets the fees it pays physicians, lacks the appropriate information about the marginal costs of physician services. Medicare bases relative prices largely on work effort estimates, typically obtained soon after a procedure is introduced. Physicians often become more expert doing the procedure and adopt new technologies reducing their work effort. Medicare updates fees by relying on the American Medical Association's Specialty Society Relative Value Update Committee (RUC), and there is ample reason to believe this process is ineffective. Dowd and Laugesen argue that Medicare should instead use a specific competitive approach, reference pricing, to inform setting its fees.

It is undoubtedly true that health care prices are distorted, but simply fixing prices is certainly not a sufficient solution to our health care problems. ² Focusing much more narrowly, however, on Dowd and Laugesen's commentary, I have several observations. First, Medicare seems to have done better than private insurers at constraining prices and, furthermore, other proposals addressing distortions in the Medicare fee schedule seem more promising. Second, looking only at individual physician fees ignores what it actually takes to deliver care and will make impossible the competitive strategies Dowd and Laugesen propose. Third, adding some competitive features to health care may be helpful, but not necessarily in the realms they suggest.

2. IS THE FOCUS ON THE MEDICARE PHYSICIAN FEE SCHEDULE MISPLACED?

Dowd and Laugesen introduce their argument for changing the Medicare fee schedule by noting the role of prices in high US health care expenditures. This is curious because the article they quote, by Anderson, Hussey, and Petrosyan notes “…the differential between the prices paid by public and private insurers in the US has increased drastically since [our] 2003 article was written…. Increased attention needs to be paid to the prices paid in the private sector.”(3 p. 89) Thus, Medicare seems to have done a far better job in containing prices than the private sector.

Dowd and Laugesen point out that the RUC uses unreliable surveys of physicians regarding the time and effort needed for each procedure. Surveys may have been the best approach when the fee schedule was developed decades ago, but much more precise and objective data on physician time associated with various tasks is available from time‐stamped electronic health records. ⁴ The specialty societies may argue that such data are not representative but Medicare could require the RUC to demonstrate in each instance why available objective data should not be used. Berenson and Ginsburg offer a series of other suggestions for improving Medicare's process. ⁵ These include phased reductions in fees for certain services Medicare feels are overvalued, adjusting fees to limit inter‐specialty compensation differentials, and legislative reforms to obtain more expert advice beyond the RUC.

Prices do matter, and indeed FFS is not inherently “the problem.” For example, fees for actual time spent or a rough approximation thereof as is reflected in separate billing codes for simple vs complex office visits may not create economic distortions. A more fundamental question is “what should be priced?” Here, the problems in focusing on the Medicare physician fee schedule are clear. Medicare law distinguishes hospital care (Part A) from physician services (Part B) with different payment mechanisms to providers: diagnosis‐related groups (DRGs) for hospital care and RVU‐based payments for physician services in the traditional fee‐for‐service (FFS) Medicare program. These each have separate funding streams. Under Medicare Part C, health maintenance organizations (HMOs) have the ability to reallocate funds across services. However, for patients and physicians preferring a FFS model, one needs a way to set prices.

3. THE PRICE, BUT FOR WHAT?

Dowd and Laugesen dismiss Medicare's experiments in bundling services as efforts to “constrain spending through incentives to constrain the volume of services provided to beneficiaries.” ¹ I believe this reflects a fundamental misunderstanding of what should be the “service” in fee‐for‐service. In most instances, one component of care cannot be obtained (“purchased”) separately from other services. For example, screening colonoscopy is a relatively straightforward procedure that could be a reasonable target for competitive pricing of the type Dowd and Laugesen advocate. The gastroenterologist's fee, however, is just one component—the patient also needs anesthesia and a facility in which the procedure is done. Competition just over the gastroenterologist's fee does not make sense because what would happen if the low‐priced gastroenterologist practiced in a high‐cost facility? Moreover, a well‐organized team can ensure all patients arrive adequately “prepped,” reducing unplanned gaps and lowering overall cost per completed colonoscopy. A bundled payment to a legal entity including all parties would incentivize team effort and efficiency without constraining overall use.

Medicare uses the DRG system to bundle payments to hospitals, providing strong incentives to discharge patients sooner and to efficiently use tests, medications, and other covered resources. Bundling together payments for physician and hospital services could create a similar efficiency incentive but existing Stark and anti‐kickback laws prohibit anything that can be construed as an incentive for physicians to refer to specific hospitals. Medicare, however, has created some exemptions and is considering broadening them which could facilitate the bundling of physician and facility charges. ⁶ , ⁷ It is important to note the difference between strategies to get better prices for a specific bundle of services, such as a screening colonoscopy, and strategies to best achieve a specific patient outcome, such as early detection of colon cancer which may involve a range of tests and different medical specialties. The latter strategies may indeed be intended to constrain the use of “low‐value” services.

4. WHY CONSIDER COMPETITIVE PRICING?

Dowd and Laugesen discuss reference pricing and tiered cost sharing as two approaches to get information in order to modify fee schedules and reduce price distortion. They note that increasing consolidation in the medical care sector limits opportunities for applying competitive strategies. One could use prices elicited in just the competitive markets to inform changes to the national fee schedule. It is unclear, however, why that would be a better strategy for Medicare than beginning with objective data from EHR time‐stamps or using expert panels as suggested by Berenson and Ginsburg.

Dowd and Laugesen note how reference pricing and tiered cost sharing provide strong incentives for physicians to lower fees because patients would seek out those charging less. (This is strategy, however, would not currently work in Medicare because it already sets fees uniformly. Variability in prices, however, is a substantial problem for private payers.) If the goal is shifted from (a) obtaining better information to adjust prices to approximate marginal cost to (b) encouraging patients to get care from lower cost providers, Dowd and Laugesen's suggested approaches offer some promise. Ideally, this would focus on fees for bundles of services that would be “consumed” together. The term “provider” is used to indicate more than just physician services may be involved.

Dowd and Laugesen's preferred competitive strategy is reference pricing, currently relatively uncommon in physician services. Usually, we see the third‐party payer (be it a private plan or Medicare) paying providers either what they charge or something based on a universal fee schedule. The patient's responsibility, after a deductible, is then usually a percentage of what the insurer pays. In Medicare, it is a 20 percent co‐insurance payment. With variability in fees (not the case under Medicare) the incentive for patients to seek out the lower price provider is the fee differential times the co‐payment rate (20 percent). To truly be an incentive, this also assumes the patient actually knows the fees in advance. Under reference pricing, the insurer identifies a fee level acceptable to at least some providers and caps payment to all providers at that level. Patients are responsible for the full additional cost of higher price providers, an incentive five times that of Medicare's current co‐insurance amount on differential fees. This encourages (a) patients to choose less expensive providers, and (b) providers to lower their prices to avoid losing business. Dowd and Laugesen focus on the second benefit, but the first is potentially more important.

Reference pricing (perhaps more accurately, reference reimbursement) is not a new idea—many public agencies and universities cap reimbursements for hotels and meals. Most travelers can choose hotels and select restaurants or entrees to stay within the reimbursement cap. Dozens of websites offer prices and customer assessments and the risks associated with a lower price (quality) hotel or restaurant are relatively small. In contrast, medical care is far from standardized, there are few useful quality measures, and the adverse consequences of poor quality can be significant. In the absence of objective information, moreover, some patients may assume higher prices reflect higher quality because that is their experience in other markets.

An alternative to reference pricing is having the insurer selectively contract with centers of excellence (COEs), usually restricting its insurance coverage to the COEs. (Tiered reimbursement affords patients some reimbursement when seeing nonpreferred providers). Direct negotiation allows contracts to include nonprice factors such as quality of care and accessibility metrics. Zhang, Cowling and Facer ⁸ review the experience of the California Public Employees Retirement System (CalPERS), suggesting the two approaches were similarly effective in shifting patients to the preferred providers.

The implicit messages of the two approaches, however, are quite different. To providers the message of reference pricing is that if their (bundled) fee is at or below the reference price, they will attract patients. In the COE model the message is that if providers continue to meet the criteria to be a COE and accept a negotiated payment, they will attract patients. To patients, the implicit message of reference pricing is that all providers offer comparable services—the only relevant difference is price. With the COE model the implicit message was that the plan (CalPERS) has selected high‐quality providers and therefore offers no reimbursement for providers not meeting its criteria.

The CalPERS experiment was relatively limited and the criteria used to select COEs were not very sophisticated. There is no reason, however, that CalPERS could not over time demand increasingly challenging measures of quality such as the adenoma detection rate—the proportion of cases in which something is found, an indication of the care with the procedure is done. ⁹ Simple reference pricing is an invitation for some providers to lower marginal costs by “cutting corners.” The COE model pressures providers to lower costs subject to the constraint of maintaining high quality. Recent recalls of generic drugs manufactured abroad with minimal FDA oversight points to the dangers of focusing only on price. ¹⁰

5. IMPLEMENTING A COMPETITIVE APPROACH TO PRICING IN MEDICARE

Dowd and Laugesen seek to correct the well‐known problems in the current Medicare fee schedule. If Medicare were to use reference pricing for the small number of procedures in which it might be applicable, the affected proceduralists (eg, gastroenterologists) would certainly claim it is unfair to target them while ignoring proceduralists for whom reference pricing is clinically implausible (eg, trauma surgeons). The reaction by Medicare FFS patients is also not difficult to predict. Nearly, everyone using Traditional Medicare has implicitly chosen not to be in a Medicare Advantage plan offering better benefits at lower overall cost. Those patients want the freedom of choice and predictability of Medicare FFS. Reference pricing might be good for Medicare, but for patients in Traditional Medicare it is clearly a “take‐away.”

Even if reference pricing survived politically, implementation is likely to be a challenge. CalPERS was ideally situated; it targeted procedures that amounted to a few thousand cases a year across the state, potentially concentrating them in 40‐60 hospitals. Being in the select group meeting the reference price (or Centers of Excellence criteria) would mean more patients, but not nearly enough to overwhelm any hospital's capacity. If expanded to all Medicare hospitals, however, reference pricing could bring so many more patients that a hospital (let alone a physician) would face demand it couldn't meet.

6. REFERENCE PRICING—WHERE MIGHT IT BE APPLIED?

Even though Medicare has contained prices better than private insurers, the relative prices in the Medicare fee schedule should be re‐examined. It is probably better, however, to muster the political will for a complete fee schedule overhaul than to focus just on the small number of services to which reference pricing might be applied. The ideas put forward by Berenson and Ginsburg offer an excellent starting point. ⁵ Beyond adjustments to the fee schedule, however, Medicare's experimentation with value‐based payments should be enhanced. They currently just offer increased payments to providers meeting certain quality criteria. To steer patients to those higher value providers Medicare could forgive part or all the deductible and co‐insurance for those services while requiring Medicare supplemental plans to pass those savings through to the patient. This lowers the cost to Traditional Medicare enrollees making good choices, rather than increasing the costs to patients not making informed choices, as with reference pricing.

Outside of the Medicare program, many insurers (and large self‐insured employers such as CalPERS) already negotiate fees with hospitals and sometimes also with physician groups. They then steer patients to those selected providers, albeit rarely using reference pricing. The Centers of Excellence model offers important advantages for select bundles of elective procedures and services such as chemotherapy in which the decision of “whether” can be made separately from “where” to have it. Beyond pricing, the quality focus of the COE model may reveal those who develop improved processes that can be adopted by others. ²

Reference pricing, however, may offer advantages when there is wide variation in prices and that information is readily available. For example, pharmaceutical benefit managers often use tiered pricing to encourage the use of generic drugs, with a second tier for lower‐priced branded drugs in a therapeutic class and a higher co‐payment for other brand‐named drugs in the class. Manufacturers of drugs in the third tier have no incentive to lower their prices to the PBM. Reference pricing in the third tier, however, can force competition among those drugs. Robinson ¹¹ offers a useful summary of the potential advantages, disadvantages, and complications in applying reference pricing to pharmaceuticals. His review is particularly valuable since it includes examples of how affected parties adapted over time to tiered pricing, reducing its effectiveness. Similar adaptive responses would undoubtedly happen with reference pricing were it to be applied to physicians.

7. CONCLUSION

Prices do matter. Attention, however, needs to be paid to what is being priced (bundles are better than individual services) and ensuring that quality is maintained or enhanced because patients are generally unable assess quality by themselves. We should also think carefully about simpler ways to get information for administratively set prices and when each type of competitive strategy is appropriate to “move patients” to lower cost, high‐quality settings.