Propensity score matching in the absence of randomized controlled trials : a case study on the effects of breastfeeding on childhood obesity

Randomized controlled trials (RCTs) are not always possible when investigating the effects of a treatment. This can be due to ethical reasons, for example, it would be unethical to randomly allow or disallow mothers to breastfeed their children due to the variety of health benefits to both mother and child, which that are known to come about from breastfeeding. This case study, illustrates one way in which observational data can be used in the absence of an RCT to investigate the effects of breastfeeding on BMI during childhood. It will outline the important things to consider when using propensity score matching (PSM) to mimic an RCT.

The case study will first provide an overview of the research question, the reasons that an RCT could not be used and reasons for the use of PSM. It will then provide further information on the decisions made whilst exploring the data and carrying out the analysis, problems encountered during the data analysis as well as lessons learned during the research.