A False Dichotomy: RCTs and Their Contributions to Evidence-Based Public Health

Indeed, triangulating among different information sources and reviewing evidence systematically are the hallmarks of nuanced, ethically conducted, effective public health research. However, we take issue with Shelton’s opening statement that ‘‘randomized controlled trials (RCTs) have limited utility for public health’’ (p. 253). Rigorously designed randomized experiments and other impact evaluation methods have a critical role to play in global public health—not just for answering small-scale questions of efficacy but also for gauging the effectiveness of scaled-up approaches. Setting RCTs in opposition to other systematic approaches for generating knowledge creates a false dichotomy, and it distracts from the more important question that Shelton addresses—namely, which research method is best suited for the question at hand? The choice of a research method is not an either/ or proposition but depends on the particular research question to be answered and the context. So, when do RCTs make sense, what value do they add, and how do they relate to other research methodologies? Laying out a ‘‘theory of change’’ is a critical first step in designing any type of intervention or evaluative research, whether experimental or not; we agree strongly with Shelton on this point. Positing the causal pathway through which an intervention operates should frame all high-quality research. However, we view this as a prerequisite, not a substitute, for rigorous testing. Perhaps most fundamentally, it is still incumbent on global public health practitioners to answer definitively the question, ‘‘Does it work?’’ before asking, ‘‘How can it be made to work practicably at scale?’’ Programs that are assumed to work are sometimes shown—by random assignment—not to work; assumptions made based on common sense may turn out to be incorrect. Rigorous experimental studies continue to generate some of the highest quality evidence for demonstrating the efficacy and effectiveness of interventions, and they continue to demonstrate that expert intuition, although important, is insufficient. Here is one example. Economists have theorized that charging small fees for public health goods (such as insecticidetreated nets, deworming medicine, or water purification solution) could promote use and reduce wastage, thus increasing the efficiency of public subsidies. The logic is that the price would signal to consumers that a product had greater value than a free product, thus encouraging uptake, and also that the ‘‘sunkcost’’ effect (‘‘I’d better use it since I already paid for it’’) would promote use. But a series of RCTs has shown that these assumptions are often incorrect. Charging a nominal fee for such health products substantially lowered uptake, especially among the poor, and distributing them for free did not make them more likely to be wasted by recipients. a Abt Associates Inc., International Health Division, Washington, DC, USA. b Population Council, Lusaka, Zambia. Correspondence to Laurel Hatt ([email protected]).