How to Choose a Study Design

Abstract

In conducting a clinical study, one must determine the timing of the study. A cohort study may be either retrospective or prospective. A case-control study is retrospective and compares subjects with a particular condition to controls without that condition by evaluating events which preceded the development of the condition. A cross-sectional study examines subjects at one particular moment in time. Once the timing of a study is determined, the type of control group is selected. One may present an uncontrolled series or use historical or concurrent controls. The prospective, concurrently controlled series is the strongest of the above designs. The randomized clinical trial is the strongest study design and is employed to avoid the errors that are potentially present in non-randomized studies (Chapter 9). It is a prospective, concurrently controlled study design in which subjects are randomly allocated to two or more treatment modalities. The protocol specifies the population source, admission criteria, enactment of the procedure, and outcome determination for the study.