Value of a small control group for estimating intervention effectiveness: results from simulations of immunization effectiveness studies.

Abstract

To improve evidence for public health practice, the conduct of effectiveness studies by practitioners is needed and may be stimulated if knowledge that smaller than usual samples may provide the same reliability of intervention effect size as larger samples. We examined reliability of intervention effect using computerized simulations of 2000 hypothetical immunization effectiveness studies from an actual study population and by small (30 and 60) and larger (100 and 200) control groups compared with an intervention group of 200 participants. Across simulated studies, the mean intervention effect (14%) and effect sizes were equivalent regardless of control group size and equal to the actual study effect. These results are relevant for similarly designed and executed studies and indicate that studies with smaller control groups can generate valid and accurate evidence for effective public health practice in communities.