Considering how much money we spend on health care in the U.S., we might hope that we allocate a good chunk of it toward evaluating the impact of the health policies we have in place. Unfortunately, this is rarely the case.

Austin Frakt explores U.S. health policy evaluation (or lack thereof) in one of his recent posts on The New York Times’ The Upshot, and Aaron Carroll walks us through these findings on Health Care Triage. They both report that less than a tenth of a percent of U.S. health care spending is directed toward policy evaluation. That means that most of the programs and policies in place in the United States aren’t backed by evidence.

Once in a blue moon, researchers have the opportunity to conduct a randomized controlled trial to investigate the effects of health policy. The 1970s RAND health insurance experiment was one of those opportunities; it found that, overall, the people who paid for a share of their health care costs used fewer health services and did not experience negative impacts on their health. This finding has been driving the payment structure of health insurance ever since.

The Oregon Medicaid Experiment is the other classic randomized trial in health policy, and though it has been extremely influential in suggesting that expanded access to Medicaid decreases financial hardship and rate of depression, some of its other findings aren’t statistically significant. Most other health policy RCTs are scarce and often underpowered.

There’s a lot of talk about restructuring our health care system, but what we need urgently is to direct some of our health care dollars toward rigorously evaluating our policies. We can’t wish our policies into effectiveness; we need evidence to do that.

Read Austin’s post and watch Aaron’s video here: