Comparative Effectiveness: One Size Doesn’t Fit All

No sooner had the Obama administration committed a billion dollars to comparative effectiveness research than the critics began laying out their concerns: such research is a prelude to rationing, they said; it threatens to thwart doctors’ and patients’ abilities to make their own decisions. It will transfer too much power to government bureaucrats and treat medical practice like a cookbook.
Now that the Institute of Medicine has issued its priorities for comparative effectiveness research (CER), I will look at a common criticism: that it acts as if medical care is a “one size fits all” enterprise, and thereby forces policy makers to make blunt decisions that will unjustifiably harm people who don’t respond to medical interventions the way an “average” person would respond. This concern is legitimate, but an intelligent use of CER should allow us to avoid this fate.
If your life, like mine, has been touched by breast cancer, then you probably share my hope that researchers will find new treatments to reduce the harms of this awful illness. But if you also share my concern for the fiscal solvency of our nation, you might also be disturbed at the high price of some new cancer treatments.
Consider a drug like Avastin: a treatment that increases life expectancy of patients with some metastatic cancers by interrupting blood flow to the tumors. Avastin can cost more than $100,000 per patient, and in some cancers leads to an increase of only two months in median survival. Two months for $100,000—a steep price to pay.
With medical costs consuming an increasing portion of government budgets, and with U.S. businesses struggling to offer employees healthcare coverage, many experts contend that we cannot afford treatments that bring such modest benefits at such a startling price.
How might comparative effectiveness research inform such issues? CER strives to provide information to guide decision making. A comparative effectiveness study might evaluate the cost effectiveness of competing breast cancer treatments. Or it might not analyze cost at all, and focus instead on estimating the relative impact that alternative treatments have on people’s quality and quantity of life.
In neither of these cases would CER, on its own, show us whether to use these treatments. Like its name suggests, CER promises to provide decision makers with information on the relative effectiveness of common medical interventions, so that government payers, insurance companies, doctors and, yes, patients can spend their health care dollars more wisely.
To understand the “one size doesn’t fit all” criticism, let’s suppose that a new drug increases median survival in patients with metastatic breast cancer by two months. That doesn’t mean that it increases everyone’s survival by two months. It might have no effect on the majority of patients, harm a small minority, and bring huge benefits to another minority.
CER, by lumping all patients into one group, would ignore these important differences. And if policymakers, unimpressed by this two-month figure, decided not to pay for this drug, some patients will lose a chance at these huge benefits.
This criticism of CER, however, overlooks more nuanced ways decision makers can potentially use CER information. With the right data, CER can improve medical decision-making by splitting patients into relevant groups, rather than lumping them into a single group.
For example, if we know in advance that patients who meet certain criteria stand to gain much more than other patients, then CER is a tool to help identify this subgroup. A treatment that costs $600,000/life year across all patients may be much more cost effective in a specific subgroup of patients.
A treatment that brings no benefit to the majority of patients but a substantial benefit to a minority of patients could very well deserve to play an important role in the treatment of that subgroup of patients. CER can potentially identify such subgroups. Indeed, if our country starts emphasizing comparative effectiveness in making treatment coverage decisions, it will give researchers in academia and in industry an incentive to find out which patients stand to benefit the most from various healthcare interventions.
On the other hand, if we do not know in advance who will benefit from a specific treatment and who will be harmed – if we can’t, for instance, figure out who will gain years rather than months of survival from the drug – then the only rational way to decide whether to use such a treatment is to assume that each patient is roughly the same and has the same chance of benefit and harm as all other patients.
If only 5 percent of patients benefit from a certain treatment, and we don’t know who those patients are upfront, then we have to assume that any given patient receiving that treatment stands a 5 percent chance of benefiting. And then we have to decide, as a society, whether that 5 percent chance of benefit is worth the costs – both medical and financial – of that treatment.
It would be unwise to use CER to lump together the unlumpable: the long-term survivors from those destined to die soon regardless of treatment. But rather than dismiss CER for treating everyone as if they are average, we should fund the kind of research that will identify who stands to benefit the most from the health care available to them.
View the original post at The Hastings Center.

Leave a Reply

PeterUbel