Electronic health records, with their promise of tracking hundreds millions of people's medical data, could be a boon for those looking to create near-individual-level predictions for an intervention's benefit. "Bringing to bear the large databases for these problems will be part of the solution," Kravitz says.

But, cautions Ioannidis, these storehouses of data points will not necessarily be an adequate substitute for rigorously run trials. "We need to carefully validate the accuracy of the information they provide," he says.

The move for more individualized cost-effectiveness assessment has already started to influence the way trials are designed, Kravitz notes. Researchers are putting more emphasis on stratifying various groups in their study populations. But those larger and more complex trials hinge on funders that also must be convinced that more personalized effectiveness is going to be, if not required in the future, at least broadly desired.

Market solutions?
The larger financial burden consumers are likely to carry for their own care might end up eventually helping to bring about more demand for individual effectiveness estimates. With only knowledge of the average success rate for a surgical procedure (and full insurance coverage), a patient is likely to pick the one that delivers the best average outcome or has the lowest average risk regardless of cost—if the patient has the luxury of choosing at all (often payers and providers will make the choice based on average cost per quality-adjusted life year for the general population).

As patients are paying more out-of-pocket for care, interventions might be up for tougher scrutiny. Drug companies and doctors that have been able to market their goods and services directly to consumers could face competition to demonstrate that their offerings really will bring the best—and most economical—outcomes for a particular profile of patient.

But Kravitz suggests that there is only so much the average patient is actually going to assess when faced with different costs and so-called quality-adjusted life years. "There's going to need to be a lot of work on how to communicate the implications of these numbers to ordinary people," Kravitz says. "They're kind of difficult for economists and doctors to grasp," he adds. And even if perfect information were possible, people are not always coolly logical when it comes to their own or a loved one's healthcare decisions.

In an ideal market, more detailed predictions about how well an advertised drug or surgery might be expected to work for an individual consumer could theoretically drive at least slightly better, cheaper treatments. "At the moment consumers are at the mercy of corporate structures" without a sense of their place on the patient outcome curve, Ioannidis says. So making people more aware of "individualized cost-effectiveness will hopefully make their choices more rational—or at least better informed." And it will be up to providers, payers and policy makers to use the new information to make larger care decisions more rational, too.

Comments