A researcher conducts a cross-sectional study and calculates a Prevalence Odds Ratio (POR) of 3.2 for obesity and hypertension. A colleague argues this overestimates the true association compared to a cohort study. Under which condition is the POR a reasonable approximation of the Incidence Rate Ratio (IRR)?
- A When the disease duration is similar between exposed and unexposed groups ✓
- B When the disease is rare (prevalence < 5%)
- C When the study is very large (n > 10,000)
- D When all confounders have been adjusted using logistic regression
Correct answer: A. When the disease duration is similar between exposed and unexposed groups
Explanation
Prevalence = Incidence × Duration; therefore, POR approximates IRR only when disease duration is similar in exposed and unexposed groups (so duration does not confound the prevalence-incidence relationship). If exposed individuals have longer disease duration (e.g., they survive longer with the disease), prevalence will be disproportionately higher in the exposed, inflating POR above the true IRR. The 'rare disease assumption' applies specifically to the OR from case-control studies approximating RR, not to cross-sectional POR.
Reference: Park's Textbook of Preventive and Social Medicine, 27th ed.
High-yield for: NEET PGINI-CETNExTFMGEUSMLEPLABMRCP
Written and medically reviewed by the StethoPrep medical team.