Choice of design and outcomes in trials among children with moderate acute malnutrition.

There is a need for trials on the effects of food aid products for children with moderate acute malnutrition, to identify how best to restore body tissues and function. The choice of control intervention is a major challenge, with both ethical and scientific implications. While randomized trials are needed, special designs, such as cluster-randomized, stepped-wedged or factorial designs may offer advantages. Anthropometry is widely used as the primary outcome in such trials, but anthropometric traits do not refer directly to specific organs, tissues, or functions. Thus, it is difficult to understand what components of health might be impacted by public health programs, or the underlying mechanisms whereby improved nutritional status might benefit short- and long-term health. Measurement of body composition, specific growth markers and functional outcomes may provide greater insight into the nature and implications of growth failure and recovery. There are now several methodologies suitable for application in infants and young children, e.g., measuring body composition with deuterium dilution, physical activity with accelerometers and linear growth with knemometers. To evaluate the generalizability of the findings from nutrition trials, it is important to collect data on baseline nutritional status.