Epidemiology and Biostatistics Training Course

Participants can calculate and interpret core measures of disease frequency and association.

• Person-time denominators constructed for dynamic populations and follow-up periods
• Cumulative incidence calculated from new cases and population at risk
• Point and period prevalence estimated using defined numerator populations
• Mortality rates, case fatality and proportional mortality compared appropriately
• Risk ratio, rate ratio and odds ratio interpreted correctly
• Attributable risk and population attributable fraction linked to prevention priorities

Participants can select a study design that fits a defined public health question.

• Cross-sectional studies estimate prevalence and exposure-outcome patterns efficiently
• Cohort designs establish temporality through measured follow-up populations
• Case-control studies sample outcome groups for rare disease investigation
• Randomised controlled trials assess intervention effects under allocated exposure
• Ecological studies use group-level data with appropriate interpretation limits
• Design selection balances validity, feasibility, ethics, cost and timeliness

Participants can prepare a sampling and measurement approach for a health study.

• Simple random, systematic and stratified sampling compared for representativeness
• Cluster sampling accounts for geographical and service-delivery groupings
• Sample size inputs include effect size, power, alpha and attrition
• Exposure variables operationalised using valid definitions and measurement scales
• Outcome ascertainment methods assessed for sensitivity and specificity
• Variable dictionaries document coding, units, permissible values and missingness

Participants can prepare, summarise and visualise a health dataset for initial analysis.

• Data cleaning identifies duplicates, impossible values, inconsistencies and coding errors
• Missing data patterns assessed before selecting a handling approach
• Frequency tables summarise categorical variables using counts and percentages
• Mean, median, standard deviation and interquartile range selected appropriately
• Epidemic curves display case onset patterns and possible exposure windows
• Spot maps and choropleth maps reveal geographical clustering patterns

Participants can use probability concepts and distributions to interpret health data uncertainty.

• Probability rules explain chance variation in observed health outcomes
• Normal distributions support interpretation of continuous clinical measurements
• Binomial distributions model binary outcomes within fixed populations
• Poisson distributions model counts of events over person-time
• Sampling distributions explain standard errors and repeated-sample variation
• Central limit theorem supports inference from sufficiently large samples

Participants can estimate population parameters and interpret common statistical tests.

• Point estimates distinguish sample statistics from unknown population parameters
• Confidence intervals express estimate precision and plausible parameter values
• Null and alternative hypotheses stated for public health comparisons
• Type one and type two errors considered during study planning
• T-tests compare means under stated distributional assumptions
• Chi-square tests assess associations between categorical health variables

Participants can identify threats to validity and plan methods to address them.

• Selection bias arises when study inclusion relates to exposure or outcome
• Information bias results from misclassification, recall errors or observer variation
• Confounding assessed through causal relationships, not statistical significance alone
• Directed acyclic graphs clarify plausible causal pathways and adjustment sets
• Stratified analysis compares crude and adjusted association estimates
• Effect modification reported when associations differ between defined subgroups

Participants can build and interpret a linear regression model for a continuous outcome.

• Linear regression models mean differences in continuous outcome measurements
• Regression coefficients interpreted with units, reference groups and uncertainty
• Categorical predictors entered through indicator variables and reference categories
• Residual plots assess linearity, normality and constant variance assumptions
• Influential observations examined using leverage and Cook's distance
• Adjusted predictions communicate model findings to non-technical audiences

Participants can model and interpret associations with a binary health outcome.

• Logistic regression estimates odds of binary health outcomes
• Adjusted odds ratios interpreted against clearly stated reference categories
• Model specification includes clinically relevant covariates and confounders
• Interaction terms test whether associations differ across population groups
• Goodness-of-fit assessed using calibration and discrimination measures
• Predicted probabilities translate logistic model results into decision information

Participants can analyse follow-up data where outcome timing and censoring matter.

• Time-to-event outcomes define origin, event, censoring and follow-up periods
• Kaplan-Meier curves estimate survival probabilities over observed follow-up
• Log-rank tests compare survival experiences between exposure groups
• Cox proportional hazards models estimate adjusted hazard ratios
• Proportional hazards assumptions checked using plots and residual-based tests
• Competing risks considered when alternative events prevent outcome occurrence

Participants can interpret surveillance data and recommend proportionate public health action.

• Surveillance objectives define data sources, reporting frequency and intended action
• Passive, active and sentinel surveillance compared for disease monitoring
• Standard case definitions support consistent reporting across reporting sites
• Threshold methods distinguish expected variation from unusual disease activity
• Completeness, timeliness and representativeness assess surveillance system performance
• Weekly incidence reports combine trends, denominators, uncertainty and action recommendations

Participants can structure an outbreak investigation using descriptive and analytical methods.

• Outbreak verification confirms excess cases against expected baseline occurrence
• Line lists organise person, place, time, exposure and outcome data
• Epidemic curve shapes suggest point-source, continuous or propagated transmission
• Attack rates compare illness occurrence among defined exposure groups
• Retrospective cohort studies test suspected exposures in closed populations
• Case-control designs investigate outbreaks where source populations remain uncertain

Participants can assess the value and limitations of screening and diagnostic tests.

• Two-by-two tables organise test results against reference standards
• Sensitivity and specificity quantify test performance among diseased groups
• Positive and negative predictive values depend on disease prevalence
• Likelihood ratios update diagnostic probability after positive or negative tests
• Receiver operating characteristic curves compare alternative test thresholds
• Wilson and Jungner principles assess screening programme suitability

Participants can appraise quantitative health evidence and judge its relevance to decisions.

• Research questions assessed against population, exposure, comparator and outcome definitions
• STROBE reporting items identify omissions in observational study reports
• CONSORT reporting items assess randomised trial transparency and completeness
• Internal validity judged through selection, measurement and confounding risks
• Forest plots interpret effect estimates, confidence intervals and study weights
• GRADE criteria rate certainty for intervention and policy recommendations

Participants can prepare a study protocol and statistical analysis plan for review.

• Protocol objectives specify primary, secondary and exploratory research questions
• Analysis populations define inclusion, exclusion and follow-up rules
• Outcome measures include definitions, timing, scales and data sources
• Statistical analysis plans pre-specify models, covariates and subgroup analyses
• Missing data procedures state assumptions, reporting and sensitivity analyses
• Results tables and figures planned before database lock

Participants can present epidemiological findings accurately for technical and decision-making audiences.

• Results narratives distinguish observed findings from causal interpretations
• Absolute and relative measures presented together for decision relevance
• Tables follow clear denominators, labels, units and footnote conventions
• Charts selected according to variable type, comparison and audience purpose
• Uncertainty communicated through intervals, assumptions, limitations and sensitivity findings
• Recommendations link evidence strength to feasible public health actions

Participants can build, review and present a complete epidemiological analysis and decision brief.

• Define a priority public health question using available case data
• Construct case definitions, variables, denominators and analytical population rules
• Select descriptive, comparative and regression methods for stated objectives
• Calculate measures, interpret outputs and document analytical assumptions
• Review bias, confounding, missing data and limitations before conclusions
• Present findings, action recommendations and an evidence-based decision brief