Summary
A before-and-after study, sometimes called a pre-post study, is a type of study in which an outcome is measured before and after an intervention is implemented, and is intended to assess whether the intervention has an effect on the measured outcome. An intervention could be a supplement, medication, or another protocol. For example, people complaining of knee pain receive an intervention of a steroid injection to assess whether this intervention will reduce knee pain, and their pain is measured before the injection and some time after the injection.
A single-group (or single-arm) before-and-after study refers to a uncontrolled study. Certain before-and-after studies have multiple arms of different interventions and/or a control arm receiving no intervention. These are usually categorized as “controlled” before-after studies. Some researchers argue that the uncontrolled before-after study is akin to an observational study because there is no perfect way to determine whether the observed change in the outcome occurred because of the study intervention or because of other factors and/or natural variation in the measured outcome.[1] [2]
An interrupted time series is a type of before-and-after study in which data points are measured several times before and after an intervention. A before-and-after study with only a single measurement of the outcome before the intervention and a single measurement after the intervention has the highest risk of bias. Before-and-after study design is weaker than a randomized interventional study due to the presence of certain biases.
One key bias includes confounder due to extraneous events. For example, if a before-and-after study finds that a diet leads to weight loss, it’s impossible to tell from the study alone whether the diet caused the weight loss, or whether other extraneous variables — like being under the watchful eye of a medical team — may have led to the weight loss. Having a control group can help disentangle the intervention’s effects from these extraneous events.
However, before-and-after studies often don’t have control groups, though they can be introduced. The biggest difference between a controlled before-and-after study and a randomized controlled trial is a lack of randomization, which could introduce bias by not mitigating extraneous factors. As an extreme (and unrealistic) example, imagine that people at a weight loss clinic were nonrandomly assigned to the diet group, and fit athletes were assigned to the control group. The diet would likely have very different effects in these two populations! Randomly assigning these two populations to intervention or control would help mitigate confounding, though this hypothetical unrealistic study would still be far from ideal.[1]
References
- ^Sterne JAC, et alChapter 25: Assessing risk of bias in a non-randomized study.Cochrane Handbook for Systematic Reviews of Interventions version 6.3.(February 2022)
- ^Aggarwal R, Ranganathan PStudy designs: Part 4 - Interventional studies.Perspect Clin Res.(2019)