Table 1.

Advantages and disadvantages of quasi-experimental studies

Conceptually simple before-and-after designs with or without external controlsSecular trends in treatment may affect outcome independent of the intervention
Better suited for small sample sizesPassage of time may interfere with presumed exposure/intervention and outcome relationship
Repeated assessment of the same individual familiarizes patients with instruments and may change behavior
Regression to the mean may minimize extreme preintervention values unrelated to the intervention itself
Potential selection bias if external control group not the same as the intervention group
Potential selection bias if loss to follow-up, dropouts, or attrition is different between comparator groups
Rigorous timing of outcome ascertainment must be ensured