Overview of Study Design

Evidence-based medicine rests on a hierarchy of study designs, each suited to different research questions and each carrying its own set of strengths and limitations. Understanding this hierarchy is fundamental to critically appraising the medical literature.

The GATE frame — Graphic Appraisal Tool for Epidemiological studies — provides a visual scaffold for describing the main elements of any epidemiological study. It incorporates the PECOT framework: Population, Exposure/Intervention, Comparison, Outcome, and Time. Together, these tools help clinicians rapidly identify what a study is asking, who it studied, and what it found.

Study designs are broadly divided into observational and experimental categories. Observational studies — including cross-sectional, cohort, and case-control designs — observe participants without assigning exposures. Experimental studies, most notably randomised controlled trials (RCTs), assign participants to exposures or interventions, enabling causal inference.

Cross-sectional studies measure exposure and outcome at a single point in time, providing prevalence data. They are quick and inexpensive but cannot establish temporal sequence — we cannot know whether the exposure preceded the outcome.

Cohort studies recruit participants free of the outcome of interest, classify them by exposure status, and follow them forward in time. They can establish incidence, calculate relative risk and risk difference, and examine multiple outcomes from a single exposure. They are ideal for common outcomes and for studying rare exposures. Limitations include loss to follow-up, the healthy worker effect when occupational groups are selected, and the time and expense required for outcomes that develop slowly.

Case-control studies start from the outcome: cases have the disease; controls do not. The investigator then looks backward to determine whether exposure was more common in cases than controls. The measure of association is the odds ratio. Case-control studies are efficient for rare diseases and transient exposures, but are susceptible to recall bias and cannot directly calculate incidence.

RCTs assign participants at random to intervention or control groups, controlling for both known and unknown confounders. Genuine clinical equipoise — authentic uncertainty about which arm is superior — is an ethical prerequisite. The strength of an RCT lies in its ability to establish causality; limitations include cost, loss to follow-up, non-adherence, and concerns about generalisability.

Internal validity — the degree to which results accurately reflect the true relationship in the study population — depends on controlling chance (random error), bias, and confounding. External validity, or generalisability, asks whether results can be applied beyond the study population.

Bradford Hill's causal guidelines provide criteria for inferring causation from an observed association: strength, consistency, specificity, temporality, biological gradient (dose-response), plausibility, coherence, experimental evidence, and analogy.

The National Ethics Advisory Committee (NEAC) ethical principles for research can be remembered with the mnemonic A DJ RIB: Autonomy, Distributive justice, Justice, Respect, Integrity, and Beneficence.