PUBH5010 Epidemiology Methods and Uses

How PUBH5010 is assessed

What PUBH5010 covers

Ten teaching blocks → one open-book exam-ready map. Each links to its free chapter guide.

The final exam is open book with a supplied “Helpful Formulae” sheet, so the equations are handed to you and the marks live in judgement — which measure, which bias, which direction, and the one plain-English sentence a layperson would understand. It is also a hurdle: you must score at least 50% on the exam itself to pass the subject.

This guide drills exactly that recurring answer shape: name the design → build the 2×2 → pick & compute the measure → interpret in plain words.

Worked example · free

Build the 2×2 and pick the right measure — the Section-B centrepiece, mark by mark

Key terms

Risk ratio (RR)

The risk of the outcome in the exposed group divided by the risk in the unexposed group, R₁/R₀, computed straight from the 2×2 a/(a+b) over c/(c+d). It is the right measure for a cohort or RCT, where risk can be observed directly; RR = 1 means no association. Interpret it as “twice the risk”, never “two times higher”.

Odds ratio (OR)

The odds of exposure (or outcome) in one group over the other, ad/bc on the 2×2. It is the measure a case-control study licenses, because cases and controls are sampled on the outcome so risk cannot be read off directly. For a rare outcome the OR approximates the RR; for a common one it overstates it.

Confounding

A third factor associated with both the exposure and the outcome, and not on the causal pathway between them, that distorts the observed association. It must meet three criteria, and the examiner wants the direction of the distortion — whether it pulls the estimate away from or toward the null — not just the label.

Non-differential misclassification

Measurement error that is the same in the groups being compared (e.g. exposure mis-measured equally in cases and controls). For a binary exposure it characteristically biases the estimate toward the null — it dilutes a real association — which is why it is the safer direction to predict in Section C.

Sensitivity and specificity

Properties of a screening or diagnostic test read down the columns of the diagnostic 2×2: sensitivity = the proportion of true cases the test catches; specificity = the proportion of true non-cases it correctly clears. They are intrinsic to the test; predictive values (PPV/NPV), read across the rows, also depend on the prevalence of disease.

PUBH5010 FAQ

Questions students ask in each PUBH5010 chapter

How do I tell a cohort from a case-control study?

Ask what defined the groups. If people were grouped by their exposure and then followed forward to see who got the outcome, it is a cohort (group by exposure, measure RR). If people were chosen because they already had the outcome (cases) or did not (controls), and exposure was asked about afterwards, it is a case-control (group by outcome, measure OR). The direction of sampling, not the timing of data collection, is the tell. Chapter 01 · Study Types and the Design Tree

What's the difference between prevalence and incidence?

Prevalence counts all existing cases (old and new) at a point in time, as a proportion of the population — it measures the burden of disease. Incidence counts only new cases over a period — it measures the risk or rate of getting the disease. A long-lasting disease can have high prevalence but low incidence; a quickly-fatal or quickly-cured one can have high incidence but low prevalence. Chapter 02 · Measures of Frequency

How do I know whether to use RR or OR?

Let the design decide. A cohort or RCT lets you observe the risk of the outcome in each exposure group, so you can and should compute the risk ratio (and risk difference). A case-control study samples on the outcome, so risk is not observable and you must use the odds ratio. Reporting an OR from cohort data is a common slip the exam tests for. Chapter 03 · Measures of Association and the 2×2 Table

How is selection bias different from confounding?

Selection bias comes from how people entered or stayed in the study — the analysed sample is not a fair window on the study base. Confounding comes from a third variable mixed into a correctly-sampled comparison. The practical difference is fixability: confounding can often be adjusted for in the analysis if you measured the confounder; selection bias usually cannot, because the needed information was never sampled. Chapter 04 · Selection Bias

What are the three criteria for a confounder?

A confounder must be (1) associated with the exposure, (2) an independent risk factor for the outcome, and (3) not on the causal pathway between exposure and outcome. If a variable lies on the pathway (a mediator), adjusting for it would wrongly remove part of the real effect; that distinction is a favourite exam trap. Apply all three to the specific variable named in the question. Chapter 05 · Confounding, Effect Modification and Standardisation

What's the difference between validity and reliability?

Validity is about accuracy — does the instrument measure the true value (for a yes/no measure, captured by sensitivity and specificity)? Reliability is about consistency — does it give the same answer when repeated? A scale can be reliable but invalid (consistently 2 kg too high) or valid on average but unreliable (right on average, noisy each time). Epidemiological bias mostly flows from invalidity, especially misclassification. Chapter 06 · Measurement Error and Misclassification

What order should I appraise a study in?

Use a fixed pipeline so you never miss a step under time pressure: name the design and hypothesis; assess chance (confidence interval, p-value, precision); then the three validity threats — selection bias, measurement/information bias, and confounding — each with its likely direction; then weigh which matters most and give a calibrated bottom line. The fixed order is itself worth marks because it shows disciplined reasoning. Chapter 07 · Critical Appraisal

Why does randomisation let an RCT claim causation?

Because allocating treatment by chance balances confounders — crucially including ones you never measured or thought of — across the arms, so any outcome difference can be attributed to the treatment rather than to pre-existing group differences. Observational designs can only adjust for confounders they measured; randomisation handles the unknown ones too, which is why a well-run RCT gives the strongest causal evidence. Chapter 08 · Randomised Controlled Trials

What's the difference between sensitivity/specificity and predictive values?

Sensitivity and specificity are read down the columns (true disease status) and are intrinsic properties of the test — they do not change with how common the disease is. Predictive values (PPV, NPV) are read across the rows (test result) and answer the clinically useful question 'given this result, how likely is the disease?' — and they depend heavily on prevalence. The columns-vs-rows rule keeps the four quantities straight. Chapter 09 · Screening and Diagnostic Tests

What makes a review 'systematic'?

An explicit, reproducible method: a pre-specified question, a documented and comprehensive search strategy, clear inclusion and exclusion criteria applied consistently, and a structured critical appraisal of every included study. This transparency — anyone could repeat the process and get the same set of studies — is what separates a systematic review from a narrative review, where selection and emphasis are at the author's discretion and prone to bias. Chapter 10 · Systematic Reviews, Meta-Analysis and Causation

How to study for the exam

Make the 2×2 table your reflex: nearly every Section-A and Section-B prompt reduces to “name the design, build the 2×2, pick the measure it licenses, then say what the number means in plain words.” Because the exam is open book with a supplied formula sheet, do not waste revision copying equations — instead drill the decisions: cohort/RCT → RR & RD; case-control → OR; a third factor linked to both exposure and outcome → confounding (check the three criteria and call the direction); error the same in both groups → non-differential → bias toward the null.

Tab the book so the right page opens in seconds — open-book speed is a skill you practise. And rehearse the lay-sentence stems (“twice the risk”, never “two times higher”), because the interpretation marks are the ones most students drop. Since the exam is a 50% hurdle, banked interpretation and measure-choice marks are the safest marks in the subject.