Morning Rounds · Numerical Series · PSM Biostatistics
PSM Numericals
A Guide to the Biostatistics & Epidemiology Series
A Guide to the Biostatistics & Epidemiology Series
Seven calculation-focused rounds · 35 questions · NEET-PG / INI-CET / UPSC CMS · +4 / −1 scoring
Biostatistics and epidemiology numericals in Indian PG entrance examinations test two separate skills that candidates often conflate: getting the formula right, and getting the denominator right. A candidate who knows that sensitivity equals true positives divided by all diseased will still drop the mark if they put total population in the denominator instead of the disease-present column. The formula is the easy part. The denominator is where the examination happens — and every round in this series is built around that reality.
The seven rounds progress from the 2×2 table outward. Rounds 01 through 03 establish the foundation: building the table correctly, deriving sensitivity and specificity from it, and then understanding how PPV and NPV behave when prevalence changes. A 95% sensitivity and 95% specificity test — which sounds bulletproof — produces fewer than 1 in 10 true positives when applied to a condition with 0.5% prevalence. That number surprises most candidates the first time they calculate it, and it is the kind of result that examiners return to repeatedly because it exposes whether a candidate genuinely understands the metric or has simply memorised the formula.
Rounds 04 and 05 move into treatment effect measures and outbreak arithmetic. The RRR-without-ARR problem — a 37.5% relative risk reduction that corresponds to an absolute reduction of only 3 percentage points — is the pharmaceutical marketing trap dressed as an exam question. The SAR denominator question — whether to include or exclude the index case — is a one-number difference that changes the answer from 60% to 75% and is tested directly. Rounds 06 and 07 close the series on vital statistics and nutrition, where the traps are multipliers and cut-offs: IMR per 1000 live births, MMR per 100,000, India-specific BMI thresholds that differ from standard WHO, and the Atwater factor for fat that is 9, not 4.
The Seven Rounds
Round 01 · PSM Numerical Series
The 2×2 Table & Its Four Cells ↗
The foundation of every biostatistics question in this series. Five questions build the table from raw numbers, identify what each cell means clinically, fill the table from prevalence plus sensitivity and specificity, calculate the false positive rate and distinguish it from PPV, and trace what happens to all four cells simultaneously when the cut-off is lowered. The column-sum verification method is introduced in the first debrief and applied throughout — the fastest way to catch a cell-assignment error before committing to a rate calculation.
Round 02 · PSM Numerical Series
Sensitivity, Specificity & the False Rates ↗
Sensitivity and specificity calculated from the table, with each denominator made explicit — TP+FN for sensitivity, TN+FP for specificity, both within a single column rather than across rows. The prevalence independence of both metrics is tested directly: the same test applied in a 10% prevalence clinic and a 1% prevalence community survey returns identical sensitivity and specificity — only PPV and NPV change. The round closes on SnNout and SpPin applied to a real screening decision, with the choice between two tests made on sensitivity grounds rather than specificity.
Round 03 · PSM Numerical Series
PPV, NPV & the Prevalence Effect ↗
The round where prevalence becomes clinically dangerous. PPV and NPV are calculated from row denominators — TP+FP and TN+FN respectively — and then watched as prevalence falls from 10% to 1%: PPV crashes from 33% to 4% while NPV rises from 98.6% to 99.9%, with full arithmetic in the debriefs. The series closes on a national screening scenario for a rare autoimmune condition: a 95%/95% test, prevalence 0.5%, positive predictive value 8.7%. Fewer than 1 in 10 positive results represent true disease. A woman told she almost certainly has the condition is almost certainly being misled.
Round 04 · PSM Numerical Series
NNT, NNH, ARR, RRR & the Risk Ratios ↗
Treatment effect measures from first principles. ARR and NNT are calculated together — NNT = 1÷ARR as a decimal, rounded up — then RRR is shown to be arithmetically correct but clinically incomplete without the absolute figures it is derived from. A drug reducing stroke risk from 8% to 5% and a drug reducing risk from 0.2% to 0.125% share the same RRR of 37.5%; their NNTs are 34 and 1333. Relative risk is calculated for a cohort study, odds ratio for a case-control study using the cross-product formula, and NNH for a statin trial where the question of whether NNH exceeding NNT constitutes net benefit is answered: severity of outcomes matters equally, and myopathy and stroke are not equivalent.
Round 05 · PSM Numerical Series
Attack Rate, SAR & Herd Immunity ↗
Outbreak arithmetic starting with a banquet gastroenteritis, where attack rate denominator is total exposed rather than non-cases, and food-specific attack rates in eaters versus non-eaters identify the vehicle. The secondary attack rate question turns on a single number: the index case is excluded from the denominator, changing the answer from 60% to 75% — and the difference matters because SAR measures transmissibility within a contact group, not across a whole household including the source. Herd immunity threshold is derived from R0 using HIT = 1−1/R0 for measles at R0=15 giving 93.3%, with the critical point that HIT must be maintained continuously rather than achieved once. The round closes on back-calculating R0 from a known threshold, with a reference table of R0 values placing influenza correctly at the low end, not the high end.
Round 06 · PSM Numerical Series
Vital Statistics — IMR, MMR, Fertility & Beyond ↗
Denominators and multipliers for the four most examined vital statistics. IMR uses live births as denominator and 1000 as multiplier. MMR uses live births and 100,000 — the multiplier difference between these two is the most common error in this topic — and is called a ratio rather than a rate because live births are a proxy for the at-risk population, not the at-risk population itself. TFR requires multiplying the sum of age-specific fertility rates by the age-group interval of 5; India's TFR has now fallen to 2.0, below the replacement level of 2.1, for the first time. The dependency ratio question embeds the India demographic dividend context, and the round closes on the CDR age-structure confounding problem — why a district with better health outcomes can have a higher crude death rate than one with worse outcomes, and why standardised rates are mandatory for cross-population mortality comparisons.
Round 07 · PSM Numerical Series · Series Finale
Nutrition Numericals — Calories, BMI & RDA ↗
The series closes on three Atwater factors — carbohydrate 4, protein 4, fat 9 — and every nutrition calculation that flows from them. Total caloric value of a mixed meal, energy deficit arithmetic for weight loss (7700 kcal per kg of adipose tissue, not 9000 kcal which would be pure fat), BMI calculation with the height-in-metres conversion trap and the India-specific cut-offs that begin overweight at 23 rather than 25 because South Asians develop metabolic complications at lower BMI values than the Caucasian populations on which standard WHO thresholds were based. ICMR protein RDA for pregnancy adds 23 g to the weight-based base requirement — distinct from the WHO base figure and the lactation addition. The round closes on macronutrient distribution arithmetic: percentage of total calories converted to kcal, then divided by the correct Atwater factor for that macronutrient — where dividing fat kcal by 4 instead of 9 is the single most common error.
Topics not covered in this series
This series covers the PSM numericals most consistently examined at NEET-PG, INI-CET, and UPSC CMS level but is not exhaustive. Areas outside these seven rounds include: likelihood ratios and pre/post-test probability calculations, Bayesian updating in clinical reasoning, receiver operating characteristic curves and area under the curve, sample size calculations and statistical power, confidence intervals and p-value interpretation, chi-square and t-test mechanics, standardised mortality ratios from indirect standardisation, and the full range of ICMR nutritional survey benchmarks beyond protein and caloric values. Each of these warrants separate treatment and will be addressed in future rounds as examination frequency analysis confirms their priority.
A note for examinees
PSM biostatistics questions at NEET-PG reward the candidate who reads the denominator before touching the formula. Sensitivity and PPV can use the same numerator (TP) and still mean entirely different things because their denominators are different. A 37.5% relative risk reduction and a 3% absolute risk reduction can describe the same trial result; only one of them tells you how many patients you need to treat. A crude death rate of 12 per 1000 can coexist with better health outcomes than a district reporting 7 per 1000 if the age structures differ. In every case, the number is only as useful as the denominator it is sitting on. If any question in this series is factually incorrect, set at the wrong level, or missing a nuance that matters in examination practice, the contact page is open. Corrections sharpen every subsequent round.
Summative Revision
A companion summative revision file covers all seven topics in condensed form — formula reference table, denominator anchors, sensitivity-specificity-PPV-NPV direction table, vital statistics multiplier guide, India-specific cut-offs and benchmarks, and the traps examiners set most often — designed for rapid pre-exam consolidation rather than first-time learning. Open Summative Revision →
Morning Rounds · atsixty.com · Numerical Series · PSM Biostatistics & Epidemiology · Seven rounds · 35 questions · +4 / −1 scoring · NEET-PG / INI-CET / UPSC CMS