INICET PSM (Community Medicine) Previous Year Questions (2018–2026)

The correct answer is To identify the potential variation between the included studies.

Heterogeneity is the variation in results across the studies in a meta-analysis, beyond what is expected from chance. Exploring it identifies and quantifies the between-study variation and its sources (differences in population, intervention, design or outcome). It is measured with Cochran's Q and the I-squared statistic (I2 of 25, 50 and 75 percent for low, moderate and high heterogeneity). Substantial heterogeneity calls for a random-effects model and subgroup or sensitivity analysis.

The other options are different steps. Assessing risk of bias and judging study quality are critical-appraisal tasks, and estimating the pooled effect size is the main result of the meta-analysis - not the purpose of exploring heterogeneity.

Incorrect Options:

• To assess the risk of bias in the included studies: that is risk-of-bias appraisal, a separate step.
• To determine if the included studies are of high quality: that is quality appraisal, not heterogeneity.
• To estimate the overall effect size of the meta-analysis: that is the pooled estimate, the main output, not the aim of heterogeneity assessment.

High-Yield Facts:

• Heterogeneity = between-study variation in results beyond chance.
• Quantified by Cochran's Q and I-squared (25, 50, 75 percent = low, moderate, high).
• High heterogeneity means a random-effects model is used.
• Explore sources via subgroup and sensitivity analyses.

Figure 1: A forest plot - the spread of individual study estimates around the pooled summary reflects heterogeneity.

Figure 2: A real meta-analysis forest plot reporting heterogeneity as Cochran's Q and Higgins' I-squared at the bottom.

The correct answer is Both A and R are true, and R is the correct explanation of A.

A nested case-control study is carried out within an existing cohort: the cases that arise are compared with controls sampled from those still at risk when each case occurs (incidence-density sampling). Because cases and controls come from a defined cohort with known follow-up and incidence data, the odds ratio obtained directly estimates the relative risk (rate ratio) - without needing the rare-disease assumption of an ordinary case-control study. So A is true, and R (use of cohort incidence data) is precisely why this is possible, so R correctly explains A.

An ordinary case-control study lacks a defined denominator, so it can only approximate the relative risk through the odds ratio. The nested design keeps the efficiency of case-control sampling while gaining the cohort's ability to estimate risk.

Incorrect Options:

• A is true, but R is false: R is actually true and is the reason A holds.
• A is false, but R is true: A is true - nested case-control can estimate relative risk.
• Both A and R are true, but R is not the correct explanation of A: R is in fact the correct explanation.

High-Yield Facts:

• Nested case-control is done within a cohort and uses its incidence data.
• Incidence-density sampling lets it estimate the relative risk or rate ratio.
• It is more efficient and cheaper than analysing the whole cohort.
• Ordinary case-control gives only an odds ratio (approximates RR if disease is rare).

Figure 1: Hierarchy of epidemiological study designs - case-control is an analytic observational design.

Figure 2: A nested case-control study - cases and sampled controls are drawn from within a defined cohort, so its incidence data allow estimation of relative risk.

The correct answer is 64.50%.

The positive predictive value (PPV) is the probability that a person who tests positive truly has the disease: PPV = TP / (TP + FP). Build the 2x2 table from the data. Of the 100 diseased people the test detects 90 (so TP = 90, FN = 10), and among the 900 disease-free people it wrongly flags 50 (FP = 50, TN = 850). The total testing positive is TP + FP = 90 + 50 = 140. Therefore PPV = 90 / 140 = 0.643, i.e. about 64.3 percent - the closest option is 64.5%.

PPV is not an intrinsic property of the test - it falls as disease prevalence falls, because more false positives arise relative to true positives, unlike sensitivity and specificity which are fixed test characteristics. Here the sensitivity is 90/100 = 90 percent, yet the PPV is only about 64 percent because the disease is uncommon. For completeness, the negative predictive value is TN/(TN+FN) = 850/860, about 99 percent.

Incorrect Options:

• 90%: that is the sensitivity, TP/(TP+FN) = 90/100, not the PPV.
• 85.20%: does not correspond to any standard test index for these numbers.
• 15%: far too low - it would imply most positives are false, which is not the case here.

High-Yield Facts:

• PPV = TP / (TP + FP) - here 90/140 = about 64.3 percent.
• Sensitivity = TP/(TP+FN) = 90 percent, a fixed test property.
• PPV rises with prevalence, and NPV falls with prevalence.
• Sensitivity and specificity are independent of prevalence.

Figure 1: PPV = true positives / all positives (TP + FP).

Figure 2: PPV and NPV change with prevalence; sensitivity and specificity do not.