Odds Ratio and Relative Risk

Odds Ratio and Relative Risk

1. Why measures of association matter in pharmacology research

• An MD Pharmacology thesis, pharmacovigilance signal, or drug-safety/effectiveness study almost always ends in a single number that quantifies how strongly an exposure (a drug, risk factor, or treatment) is associated with an outcome (a disease, adverse event, or response) — the measure of association (Browner Ch.8, pp. "Measures of Association").
• The study design dictates which measure is estimable: a cohort study or clinical trial yields incidence, so it can produce a risk ratio (relative risk), risk difference, rate ratio, or hazard ratio; a case-control study cannot estimate incidence and so is limited to the odds ratio (Browner Ch.8 Table 8.3; Ch.9 pp.5–6).
• A cross-sectional study measures only prevalence, so its association measure is the prevalence ratio — "the cross-sectional equivalent of the risk ratio" (Browner Ch.8, p.3).
• All three frequency measures — risk, odds, rate — share the same numerator (number who develop the dichotomous outcome) and differ only in the denominator; understanding the denominator is the key to choosing and interpreting the association measure (Browner Ch.8, pp.19–20).