It seems like an obvious choice: two equally effective medications, where one is 40 times the cost of the other. Yet doctors’ treatment decisions differed significantly, with each defying expectations in a surprising way.
As it turns out, financial incentives can influence doctors to prescribe a more expensive medicine, but not as much as conventional wisdom might suggest, according to new research from Stanford professor Kate Bundorf. The Stanford scholar found that patients and hospital administrators both wield more influence over these decisions that one might think.
A Tale of Two Treatments
Bundorf, along with Stanford professors Suzann Pershing, Steven Asch, and Laurence Baker, health economists Christine Pal Chee and Todd Wagner, and biostatistician Derek Boothroyd, studied treatment choices of Medicare patients with age-related macular degeneration, the leading cause of vision loss in adults over 50, from 2005 to 2011. Two drug treatments proved successful for preventing, or sometimes even reversing, this condition: ranibizumab (marketed as Lucentis) and the less-expensive bevacizumab (marketed as Avastin). Together, the drugs represent a dominant 98% share of the market.
Under the Medicare system, doctors are compensated on a fee-for-service model and receive additional payments equivalent to 6% of medication costs. This bonus is much higher when based on the $2,000 cost of Lucentis compared with the $50 cost of Avastin, so this provides a potentially strong incentive for doctors to encourage use of Lucentis.
Despite this incentive, only one-third of Medicare patients were prescribed Lucentis, suggesting a different and more complicated relationship than expected. “If it were all about physician incentives,” Bundorf says,“patients would all be taking Lucentis. So clearly there is something else going on.”
While this finding cast doubt on a simple cause-and-effect relationship between physician payments and treatment decisions, it was not yet clear whether behavior was motivated by a different monetary influence, such as value-minded patients. To test this, Bundorf’s team sought to observe how doctors and patients act after taking money out of the equation.
The team turned to the Department of Veterans Affairs, which offered an ideal opportunity to study similar patients and treatment options in an environment where decisions of patients and physicians were independent of financial influences. Subsidized costs at the government-run health system mean patients pay the same amount no matter which medication is prescribed. Doctor compensation is also fixed, with no additional incentives based on medication choice.
“Previously, research had usually been done in the context of Medicare,” Bundorf says, “but we knew almost nothing about what was going on with the VA, which is a very different healthcare system. We thought it would be interesting to make that comparison.”
Surprisingly, findings revealed no strong preference for either medication, with each used by roughly half of the patients. Instead, Bundorf saw another form of influence. The VA, she says, “relies more on what we refer to as administrative mechanisms.” The system relies on a top-down decision-making process, although this can be overruled regionally or by individual doctors. Upon closer examination, the researchers saw that usage patterns varied from region to region, indicating patients were not the primary drivers behind these decisions, and providing evidence of local influence from regional administrators.
“Different VA sites are implementing different degrees of control over physicians and patients,” Bundorf says.
Health Policy Applications
Comparing the two systems, we can see a complex interaction where multiple incentives exert competing influences on doctors, patients, and policymakers. Yet, an important follow-up question remains: Are these mechanisms achieving a desirable outcome?
At present, there is significant room for improvement. For example, Medicare could save a projected $18 billion over the next decade through using the lower-cost treatment exclusively.
Another target for change involves the nature of payments to doctors. Medicare relies too much on volume or quantity-based payment systems, Bundorf says. “We want to think about ways for Medicare to use incentives differently, not just rewarding doctors for doing more or choosing to do the most expensive thing.”
However, eliminating extra payments to doctors altogether may not be the best option. The key is to ensure fundamental tensions are in balance. “Profit seeking is not necessarily bad in the healthcare system — the trick is to create incentives that align the profits of providers with the interests of patients and taxpayers,” Bundorf says.
One mechanism that has already been successful in other contexts is reference pricing. Under this model, an insurer would pay up to a certain amount (usually the lowest-cost option). If patients wanted a more expensive option, they would need to pay the difference. In this example, a patient using Avastin would be fully covered, but one preferring Lucentis would need to pay the difference in cost. This solution could offer patients greater freedom of choice and access to a low-cost option, while controlling aggregate costs to minimize the overall system-wide burden.
“In many ways, the potential for innovation lies on the part of the insurers to figure out how to facilitate the right balance for patients and physicians,” Bundorf says.