Membership of the Faculty of Public Health (MFPH) Research Methods, Epidemiology and Statistics Syllabus

Research Methods, Epidemiology and Statistics syllabus — full chapter and topic list

Expand any chapter to see its topics and sub-topics. This is the whole examinable outline for Research Methods, Epidemiology and Statistics in Membership of the Faculty of Public Health (MFPH), not a summary of it.

1. 1

   Study Design and Epidemiological Methods

   5 topics

   • Descriptive epidemiology
     • Person, place and time
     • Incidence and prevalence measures
     • Mortality and morbidity rates
   • Observational study designs
     • Cross-sectional surveys
     • Case-control studies
     • Cohort studies (prospective and retrospective)
     • Ecological studies and the ecological fallacy
   • Experimental study designs
     • Randomised controlled trials
     • Cluster and stepped-wedge designs
     • Pragmatic vs explanatory trials
   • Natural experiments and quasi-experimental methods
     • Interrupted time series
     • Difference-in-differences
   • Choosing an appropriate study design
2. 2

   Measures of Association and Causal Inference

   5 topics

   • Risk ratios, odds ratios and rate ratios
   • Absolute and relative risk measures
     • Attributable risk and risk difference
     • Number needed to treat
     • Population attributable fraction
   • Bias
     • Selection bias
     • Information and measurement bias
     • Recall and reporting bias
   • Confounding and effect modification
     • Identifying and controlling confounding
     • Stratification and adjustment
   • Causality frameworks
     • Bradford Hill considerations
     • Directed acyclic graphs
3. 3

   Statistical Inference and Analysis

   6 topics

   • Probability distributions and sampling
   • Hypothesis testing
     • P-values and statistical significance
     • Type I and Type II errors
     • Confidence intervals
   • Common statistical tests
     • t-tests and ANOVA
     • Chi-squared tests
     • Non-parametric tests
   • Regression methods
     • Linear regression
     • Logistic regression
     • Poisson and Cox proportional hazards regression
   • Survival analysis and life tables
   • Sample size and power calculations
4. 4

   Diagnostic and Screening Test Evaluation

   4 topics

   • Sensitivity, specificity and predictive values
   • Likelihood ratios and ROC curves
   • Effect of prevalence on predictive value
   • Reliability and validity of measurement
     • Kappa and inter-rater agreement
     • Internal and external validity
5. 5

   Evidence Synthesis and Critical Appraisal

   5 topics

   • Systematic review methodology
     • PICO framing and search strategy
     • PRISMA reporting
     • Risk of bias assessment
   • Meta-analysis
     • Fixed and random effects models
     • Heterogeneity (I-squared)
     • Forest and funnel plots
     • Publication bias
   • Critical appraisal frameworks (CASP, GRADE)
   • Hierarchy of evidence and evidence-based public health
   • Qualitative research methods and appraisal

Research Methods, Epidemiology and Statistics flashcards for Membership of the Faculty of Public Health (MFPH)

19 of 69 cards from the Research Methods, Epidemiology and Statistics deck — real questions with worked answers.

1. In descriptive epidemiology, what three core attributes are used to characterise the distribution of disease in a population?

   Person (who: age, sex, ethnicity, occupation), Place (where: geographic/spatial distribution), and Time (when: trends, seasonality, epidemics). These describe patterns and generate hypotheses about cause.

2. Define incidence and prevalence, and give the formula relating them in a steady state.

   Incidence = new cases per population at risk over a period. Prevalence = existing (new + old) cases at a point/period. In a steady state: $$P \approx I \times \bar{D}$$ where $P$ is prevalence proportion (when small), $I$ is incidence rate and $\bar{D}$ is mean disease duration.

3. Distinguish cumulative incidence (risk) from incidence rate (incidence density), including their formulas.

   Cumulative incidence (risk) $= \frac{\text{new cases}}{\text{population at risk at start}}$ over a fixed period (dimensionless, 0–1). Incidence rate $= \frac{\text{new cases}}{\text{total person-time at risk}}$ (units of per person-time), accounting for varying follow-up.

4. List the main types of observational (non-experimental) study designs in order from weakest to strongest for inferring causation.

   Ecological → cross-sectional → case-control → cohort. Cohort (especially prospective) provides the strongest observational evidence because exposure is measured before outcome, allowing incidence and temporality.

5. In a case-control study, why is the odds ratio used rather than the relative risk?

   Cases are selected by outcome status, so incidence/risk cannot be calculated directly. The odds ratio of exposure can be computed and, when the disease is rare, the OR approximates the relative risk (rare disease assumption).

6. Define a cross-sectional study and state one key limitation regarding causal inference.

   A study measuring exposure and outcome simultaneously in a defined population at one point in time (a 'snapshot'), giving prevalence. Limitation: temporality is usually unclear (cannot tell whether exposure preceded outcome), so causation is hard to establish; prone to reverse causality.

7. What is an ecological study and what is the ecological fallacy?

   An ecological study uses group-level (aggregate) data rather than individual-level data to examine associations. The ecological fallacy is the error of inferring individual-level associations from group-level data, because group correlations may not hold for individuals.

8. What is the defining feature of an experimental study design, and what is the gold-standard experimental design in epidemiology?

   The investigator allocates/assigns the exposure (intervention). The gold standard is the randomised controlled trial (RCT), in which random allocation balances known and unknown confounders between groups.

9. Compare a parallel-group RCT, a crossover trial, and a cluster RCT.

   Parallel: each participant receives only one intervention. Crossover: each participant receives both interventions sequentially (acts as own control; needs washout, assumes no carryover). Cluster: groups (e.g. clinics, villages) rather than individuals are randomised; requires adjustment for intra-cluster correlation.

10. What is intention-to-treat (ITT) analysis and why is it preferred in RCTs?

    ITT analyses participants in the groups to which they were randomised, regardless of adherence or crossover. It preserves randomisation, maintains comparability of groups, avoids attrition/non-compliance bias, and gives a pragmatic (real-world) estimate of effect.

11. Define a natural experiment in epidemiology.

    An observational study exploiting a naturally occurring or externally imposed variation in exposure (not controlled by the investigator) that approximates random allocation — e.g. a policy change, legislation, or natural event affecting some groups but not others.

12. Describe two common quasi-experimental designs and how each strengthens causal inference.

    Interrupted time series: repeated measures before and after an intervention, comparing trend change. Difference-in-differences: compares the change over time in an exposed group with the change in an unexposed (control) group, controlling for time-invariant confounding. Others: regression discontinuity, instrumental variables.

13. What factors determine the choice of an appropriate study design?

    The research question (descriptive vs analytic vs evaluative), outcome frequency (rare disease → case-control; rare exposure → cohort), exposure type, time/resources, ethical constraints (cannot randomise harmful exposures), need for temporality, and required strength of evidence.

14. Define risk ratio (relative risk) and give its formula using a 2×2 table.

    Risk ratio compares cumulative incidence in exposed vs unexposed. With exposed $a/(a+b)$ and unexposed $c/(c+d)$: $$RR = \frac{a/(a+b)}{c/(c+d)}$$

15. Define the odds ratio and give its formula from a 2×2 table.

    The odds ratio is the ratio of the odds of outcome (or exposure) in exposed vs unexposed. From a 2×2 table: $$OR = \frac{a/b}{c/d} = \frac{ad}{bc}$$ (the cross-product ratio).

16. Define the rate ratio and contrast it with the risk ratio.

    Rate ratio (incidence rate ratio) compares incidence rates (cases per person-time): $$\text{Rate ratio} = \frac{I_{\text{exposed}}}{I_{\text{unexposed}}}$$ It uses person-time denominators (handling varying follow-up), whereas the risk ratio uses fixed populations at risk over a defined period.

17. Interpret a risk/odds/rate ratio of 1, greater than 1, and less than 1.

    Ratio $=1$: no association (null value). Ratio $>1$: exposure associated with increased risk (positive association/harmful). Ratio $<1$: exposure associated with decreased risk (protective).

18. Define absolute risk reduction (ARR) and number needed to treat (NNT), with formulas.

    $$ARR = R_{\text{control}} - R_{\text{treated}}$$ (absolute difference in risk). $$NNT = \frac{1}{ARR}$$ the number of patients needing treatment to prevent one additional adverse outcome (round up).

19. Define relative risk reduction (RRR) and explain why it can be misleading without absolute risk.

    $$RRR = \frac{R_{\text{control}} - R_{\text{treated}}}{R_{\text{control}}} = 1 - RR$$ It gives the proportional reduction but conceals baseline risk; a large RRR on a tiny baseline risk yields a small absolute benefit (large NNT).

See more Research Methods, Epidemiology and Statistics flashcards →

Planning Research Methods, Epidemiology and Statistics for Membership of the Faculty of Public Health (MFPH)

Research Methods, Epidemiology and Statistics (Membership of the Faculty of Public Health (MFPH)) FAQ