Half of medical students have test anxiety, and the highest figures are in this region

Design

Systematic review and meta-analysis, 35 studies, random-effects generalised linear mixed model

Population

13,489 medical students across multiple regions

Primary outcome

Pooled prevalence of test anxiety

Effect

49.8% (95% CI 39.9-59.6), I-squared 98.2%; 44.9% excluding three visual analogue scale studies; South-East Asia 59.3%

Thirty-five studies covering 13,489 medical students were pooled using a random-effects generalised linear mixed model with logit transformation, with risk of bias assessed on a version of the Newcastle-Ottawa Scale adapted for cross-sectional studies.

Pooled prevalence of test anxiety was 49.8% (95% CI 39.9-59.6), with I-squared of 98.2% and tau-squared of 1.33. Removing the three studies that used a visual analogue scale brought it to 44.9%. Only the assessment tool reached significance in subgroup analysis, but the numerical pattern is consistent: higher in women (52.6% against 46.3%), higher in first-year students (58.6%) than fifth-year (47.9%), and highest in South-East Asia at 59.3%, ahead of the Middle East at 54.1%. Studies conducted during examinations reported 55.4%, against 41.5% afterwards. Publication year showed no association with prevalence.

For psychiatrists in India this is the closest thing to a local number in today's literature, and it is the highest regional figure in the analysis. The practical implication is about where a service sits rather than what a clinician prescribes: if half of a medical school has test anxiety, the appropriate response is a route into help that does not require a diagnosis, not an expansion of clinic slots. The timing finding also matters - measuring during examinations and after them gives different answers, so a service audit taken in a quiet week will understate the problem.

• Ask when a student's symptoms occur relative to examinations before framing them as generalised anxiety
• First-year students reported the highest rates; direct outreach there rather than assuming seniority is harder
• Distinguish test anxiety from an anxiety disorder - it is situation-specific by definition
• If you audit a student population, record when in the academic year you measured
• South-East Asia had the highest regional estimate, which makes this directly relevant to Indian medical schools

Why it matters

A prevalence near 50% makes this a service design question, not a referral question.

Don't overread it

Heterogeneity is near total; the pooled figure is not a precise population estimate.

The statistics, in plain English

An I-squared of 98.2% alongside a tau-squared of 1.33 means these 35 studies are not estimating one underlying number. The confidence interval of 39.9% to 59.6% describes the uncertainty around an average that may not correspond to any real population. The subgroup analyses make the point: prevalence shifted from 41.5% to 55.4% depending only on when in the examination cycle students were asked, and dropped from 49.8% to 44.9% when three studies using a different measurement scale were removed. Read this as evidence that test anxiety is very common rather than as evidence that it affects exactly half of students.

Read the rest in the app

You have read your two free briefings this month. The app carries all 27 specialties, every morning, free — and this finding is waiting in it.

Get Daily Doctor — free

Scan to keep reading on your phone. No account needed to start.