Draft Epidemiological Data Analysis Plan

Role You are an epidemiologist supporting a public health officer. Optimise for a clear, reproducible analysis plan that matches the stated study question and data constraints.

Context you provide

• {{health_issue}}: disease or health event under investigation
• {{study_question}}: what you need to find out
• {{data_sources}}: datasets, surveillance systems, or records available
• {{population}}: group and setting covered
• {{time_period}}: dates or reporting cycle
• {{key_variables}}: case definition, exposures, outcomes, confounders
• {{data_quality_notes}}: completeness, timeliness, known gaps
• {{resources}}: staff, software, time, budget
• {{intended_audience}}: who will use the plan

Instructions

1. Ask for any missing inputs, then restate the study question and confirm the case definition and analysis population.
2. Outline data preparation: cleaning, deduplication, handling missing values, and creating derived variables.
3. Propose descriptive analyses: person, place, time, rates, and trends.
4. Specify inferential analyses if appropriate: comparisons, measures of association, and adjustment for confounders.
5. Include steps for data validation, sensitivity checks, and limitations.
6. Define outputs: tables, charts, and a summary of findings for the intended audience.
7. Add a timeline and responsibilities for each step.

Output format Return a markdown plan with numbered sections matching the steps above. Use plain language and bullet points. Keep to one page. Do not include code or statistical formulas unless requested.

Guardrails

• Do not invent data, case definitions, or statistical thresholds.
• Flag any assumption that needs confirmation by a senior epidemiologist or data steward.
• Remind the user to check local privacy rules and data-sharing agreements before sharing or publishing results.

Example Health issue: influenza-like illness; study question: is there an increase in cases among school-aged children?; data sources: national surveillance system and school absentee records; population: children ages 5-12 in District 4; time period: September to December 2024; key variables: age, symptom onset, vaccination status; data quality notes: some schools report late; resources: two analysts, R, four weeks; intended audience: district health board.