Persistent shortcomings in our knowledge of the actual effects of programs and interventions are due to a gap in both the quantity and methodological quality of impact evaluations. Impact evaluations are studies that measure the impact on health directly attributable to a specific program or policy, regardless of other potential explanatory factors. The aim of this manuscript is to present the types of epidemiological studies that can be employed in impact evaluations of nutrition programs. Study designs are presented according to the level of inference needed for the evaluation (adequacy, plausibility and probability). Examples of program evaluations using cross-sectional, quasi-experimental, and randomized controlled trials are discussed. The advantage of the observational studies over randomized controlled trials when the aim is to assess programs effectiveness in real world conditions is highlighted. Impact evaluations are urgently needed to subsidize decision makers on what programs should continue to receive investments and what programs need to be reformulated.
Feeding; Program evaluation; Evaluation studies as topic; Nutrition public health; Public policies
