Are RCTs the most cost effective choice?

The death of statistician Paul Meier, a proponent of Randomized Controlled Trials (RCT), is as good a point as any to pick up the debate as to whether RCTs should be so heavily promoted as a gold standard in evidence–based medicine.

RCTs are most often used to evaluate treatment safety as well as the efficacy and/or efficiency of health care services and artificial interventions such as pharmaceuticals, medical devices, or surgery as well as behavioral treatments.

The key advantage to such trials is the minimization of allocation bias in the assignment of treatments. They also facilitate blinding of the identity of treatments from investigators, participants, and assessors.

But, in an era where funding for research is becoming more and more scarce are they cost effective for government sponsored research?

NIH director Dr. Francis Collins testified before the Senate in May that due to the cut in funding to NIH in FY 2011, only 17 to 18 percent of NIH grant applications would be funded - the lowest level on record. The Federal Coordinating Council on Comparative Effectiveness Research states that 79% of clinical trials at the NIH are a RCT rather than say an observational study. Of 28,244 completed RCTs in the ClinicalTrials.gov database there are currently 7,708 studies actively recruiting. 

A study in 2006 estimated a mean cost of $12 million per RCT in one branch of the NIH, but concluded the benefits were high. In 2007 a different study questioned the return on investment. Observational studies, which cost considerably less, may produce results similar to RCTs according to among others a 2001 study in the NEMJ. 

As always in medicine, opinions are as frequent as belly buttons, but in an era of shrinking budgets hard questions need to be asked regarding not only how tax dollars are allocated, but the return on investment.