Personalized medicine and treatments for rare diseases — many of which are extraordinarily expensive — are a growing part of the U.S. health care landscape. This trend has important ramifications for the out-of-pocket prices (e.g., copays and coinsurance) that insurers require patients to pay.
A growing body of research says that insurers raise out-of-pocket costs in part to avoid sick enrollees. This problem is likely to get worse as personalized medicine becomes more common, more patients get to choose between competing private health plans, and employers shrink the coverage they offer employees in order to reduce their labor costs. Consequently, if we want new medical innovations to be financially viable for the patients who need it most, health insurance markets need to be regulated to eliminate the perverse financial incentives that limit patients’ coverage.
First, a little background. If insurance markets function as they should — spreading risk across patient populations — then the costs of expensive specialized therapies should be spread across all enrollees, rather than rest on those patients who require treatment. In theory, this would make out-of-pocket payments for affected patients low, with the financial burden of treatments reflected in slightly higher premiums for all. But the increase in premiums due to rare diseases would be small, even for an extraordinarily expensive drug like tisagenlecleucel, which the U.S. Food and Drug Administration approved in August 2017 to treat acute lymphoblastic leukemia in young adults. (It genetically re-engineers the patient’s own T-cells to attack the leukemic cells.) The first cell-based gene therapy approved in the United States, it has been priced at $475,000 for a course of treatment. But acute lymphoblastic leukemia has an annual incidence of approximately 41 cases per 1 million people aged 0 to 14, an age group that comprises less than 20% of the U.S. population. Covering tisagenlecleucel with zero patient cost sharing for these patients would impact insurance premiums by less than 40 cents per enrollee per month.
In reality, patients frequently face significant out-of-pocket costs for expensive specialized medicines, even under otherwise generous insurance policies. This is likely due to what economists call “adverse selection,” or the tendency of sicker, more expensive consumers to choose health insurance plans with more generous coverage. Because sicker patients are more likely to be unprofitable, insurers try to push them toward competitors’ plans by designing plan benefits (e.g., provider networks, drug copays, and prior authorization requirements) to be unattractive to those who need more generous coverage. This flies in the face of how insurance should work. Insurance exists primarily for large losses. Think about having a homeowner’s insurance policy that covers a small burglary but not a major fire.
Recent research has clearly documented how this theory plays out in practice. ACA Marketplace plans in several states have placed commonly prescribed HIV medications, including inexpensive generics, on high cost-sharing tiers, likely in an attempt to avoid attracting enrollees with HIV who may be expensive due to other healthcare utilization. Other research on Medicare Part D and in the ACA Marketplaces has shown that health plans systematically design formularies to limit coverage for drugs used by patients who are predictably unprofitable, such as patients with multiple sclerosis or those with substance-abuse disorders who require specific therapies.
Adverse selection in insurance markets is likely to worsen as personalized medicine grows (as the fixed costs of drug development and production are spread across smaller populations of patients, resulting in higher manufacturer prices for drugs like tisagenlecleucel), and as government public programs shift more individuals to private health plans, rather than offering coverage through one. Today, about 10 million consumers choose from private health plans in ACA Marketplaces. Low-income and disabled Medicaid recipients often choose from competing, private, managed-care plans. And elderly and disabled Americans choose from competing private prescription drug plans for their drug benefits in the Medicare Part D program. In these markets, no insurer has an incentive to provide good coverage for expensive drugs because any insurer who provides even slightly better coverage for these drugs relative to their competitors immediately becomes the “insurer of choice” for anyone taking those drugs, driving up insurer costs, and driving down any profits the firm may be making in this market.
In the employer market, increasing pressures to cut labor costs, combined with the proliferation of new high-cost therapies, provide employers with little incentive to provide generous coverage for these personalized therapies. If employers are faced with managing health care costs and have to decide between raising co-payments for primary care visits (which would affect many of their employees) or dramatically increasing cost sharing for new high-cost therapies (which would affect fewer people), they’ll likely opt for the latter.
These shifting forces present a worrisome scenario for the ability of insurance markets to actually insure Americans against the risk of needing increasingly prevalent high-cost specialized drugs.
To address these issues, some have suggested putting limits on the way insurers design their drug formularies, including caps on out-of-pocket drug costs and limits on insurers’ ability to use prior authorization and step therapy requirements to create barriers to drug utilization. For example, in Medicare Part D there are certain “protected classes” of drugs where insurers are required to cover at least two drugs in the class. New York prohibits the use of “specialty” tiers in formularies altogether.
These types of policies can be effective at limiting patient cost-sharing for these treatments, but they can also can make it hard for insurers to correctly limit access to low-value drugs. Policies like protected drug classes can protect too much if they apply equally to the higher- and lower-value drugs within a class. In principle, any constraints on insurers’ formulary designs can limit the extent to which an insurer can appropriately steer consumers away from high-cost therapies toward lower-cost alternatives with similar efficacy.
Others have suggested a return to risk-rated insurance premiums, where insurers are allowed to charge the sick higher premiums than the healthy. This would help fix insurers’ incentive problem by making all consumers equally attractive to cover, but it would be fraught with distributional concerns as the sick would pay markedly higher premiums and many would go uninsured.
Rather, what we need to make out-of-pocket costs manageable and new personalized treatments accessible are policies that shift insurers’ financial incentives and counteract adverse selection. These policies must (1) maintain “community rating” of premiums, which pool risk across the entire population within an area and charge individuals within that area the same premium, and (2) make high-cost consumers profitable instead of unprofitable. Risk adjustment and reinsurance are the most commonly used policies for shifting insurers’ financial incentives in this way. When used correctly, they can incentivize health plans to attract and care for sick, expensive patients, including those patients prescribed high-price drugs. These policies are widely used in the ACA Marketplaces, Medicaid, Medicare Part D, Medicare Advantage, as well as health insurance markets in the Netherlands, Switzerland, and Germany.
Under risk adjustment, the government transfers money away from health plans with healthy populations and toward plans that attract sick, costly patients, removing the financial incentive of health plans to skim the healthy and avoid the sick. We now have ample evidence across multiple market settings that, for conditions where risk adjustment is working well (i.e., where risk adjustment causes the revenues an insurer receives for a patient with a particular condition to be close to the insurer’s expected cost for that patient), patients face lower drug cost-sharing regardless of upstream manufacturer prices. Thus, a $475,000 tisagenlecleucel treatment need not in principal expose a patient to any financial strain.
Unfortunately, risk adjustment is not currently designed to deal with the advance of personalized medicine. The problem for tisagenlecleucel, and any new medical technology that dramatically increases the cost of treating a condition, is the costs of the drug are not yet incorporated into risk-adjustment formulas, due to two- to five-year lags in the data used to calibrate these formulas. Worse, they may never be incorporated, as conventional wisdom in risk adjustment eschews including rare diseases in the algorithm.
If we want patients to be able to access to new therapies, risk adjustment models have to be updated more frequently to incorporate technological advances. Health-plan payment systems also have to go beyond risk adjustment transfers for broadly-classified common chronic diseases (for which current risk-adjustment policies do quite well) and pay more attention to smaller diseases that are treated by personalized therapies.
Another important regulatory shift that could normalize good insurance coverage for rare-disease treatments is backstopping health plans with mandatory reinsurance programs — like the one that was recently terminated in the ACA Marketplaces. Reinsurance pays for individual outlier patients whose drug utilization might otherwise make them exceptionally unprofitable for a health plan to enroll. It is useful in dampening the financial pressure to avoid patients whose conditions have not yet been entered into the risk-adjustment system or for whom the risk-adjustment calibration is out of date. While insurers often choose to buy private reinsurance, only mandatory, universal community-rated reinsurance will guarantee that the costs of these high cost therapies are pooled across all insurers, thereby eliminating the incentive for any one insurer to avoid enrolling patients who are likely to need these treatments.
No solution is perfect. It is administratively difficult to constantly re-optimize risk adjustment, and generous reinsurance can remove an insurers’ incentive to efficiently manage care and control spending, especially if they have to engage in aggressive negotiations with drug manufacturers over drug list prices.
However, if we do not develop and implement tools for pooling the financial cost of new high-cost treatments, access to them will be cut off, even for patients with generous insurance policies. It is critical that we get ahead of this problem by developing health-plan payment systems that get insurers to compete to provide patients with access to effective therapies rather than incentivize them to avoid patients whose health may hang in the balance.
Michael Geruso is an assistant professor of economics at University of Texas at Austin and a faculty research fellow at the National Bureau of Economic Research.
Anupam B. Jena
Anupam B. Jena is the Ruth L. Newhouse associate professor of health care policy at Harvard Medical School, an internist at Massachusetts General Hospital, and a faculty research fellow at the National Bureau of Economic Research.
Timothy J. Layton
Timothy J. Layton is an assistant professor of health care policy a Harvard Medical School and a faculty research fellow at the National Bureau of Economic Research.
Referenced path does not exist