Investigational treatments. How strict should we be?

THE BENEFIT LANGUAGE described in the previous article 1 described what might be called evidence- and outcomes-based criteria 2 for coverage: Treatments should be backed by sufficient evidence to determine their effects on health outcomes, and a comparison of the outcomes should show that the is effective, beneficial, and cost-effective. This article will focus on the first of these criteria—that treatments should be backed by solid evidence of effectiveness. In one sense, this criterion should be the least controversial, because it corresponds to the investigational exclusion that has appeared for decades in the contracts of almost all traditional fee-for-service insurers, health maintenance organizations, and government programs. (Henceforth I will use the word plan to include all types of organizations that finance and/or deliver health care, and the word treatment to include all types of health interventions.) Recently, however, this criterion has become far more controversial because of reports