Epidemiological Study Design & Statistical Analysis Planning
Design epidemiological studies and statistical analysis plans
What You Can Do
You can systematically design epidemiological studies using evidence-based frameworks for study design selection, develop comprehensive statistical analysis plans with rigorous pre-specification, and assess potential biases and confounding in your study architecture. This skill helps you specify appropriate regression models, calculate sample sizes, plan sensitivity analyses, and document methodological decisions that align with epidemiological best practices.
Features
Compare cohort, case-control, cross-sectional, and experimental designs. Get guidance on selecting the optimal design for your research question, considering feasibility, exposure timing, and outcome measurement.
Create detailed pre-specified SAPs that outline primary analyses, secondary objectives, subgroup analyses, and post-hoc exploration strategies. Reduces analytic flexibility and p-hacking.
Systematically identify selection bias, information bias, confounding, and reverse causation in your study design. Receive specific mitigation strategies for each threat to validity.
Specify appropriate regression models (linear, logistic, Cox proportional hazards, Poisson) with clear rationale for functional forms, interaction terms, and covariate adjustment sets.
Calculate required sample sizes for your primary objective, considering effect sizes, study design, significance levels, and anticipated loss to follow-up or missing data.
Use directed acyclic graphs to identify confounder, mediator, and collider structures. Determine which variables to include or exclude from regression models based on causal structure.
Pre-specify sensitivity analyses to assess robustness of findings to unmeasured confounding, selection bias, measurement error, and alternative analytical choices.
Example Output
Example 1: Study Design Summary
Research Question: Does occupational pesticide exposure increase lung cancer risk?
Recommended Design: Case-control (nested within cohort if possible)
- Cases: Incident lung cancer diagnoses (1990-2025)
- Controls: Cancer-free individuals, stratified by age and gender
- Exposure: Occupational history via employment records
- Confounders: Smoking, family history, other occupational exposures
Example 2: Statistical Analysis Plan
Primary Analysis:
Logistic regression: ln(odds of lung cancer) ~ pesticide exposure + age +
gender + smoking
Model Equation:
log(odds) = β₀ + β₁(exposure) + β₂(age) + β₃(female) + β₄(smoking_pack-years)
Sensitivity Analyses:
1. Restrict to never-smokers only
2. Unmeasured confounding assessment (E-value calculation)
3. Alternative exposure classification (quartiles vs continuous)
4. Multiple imputation for missing smoking data (20 imputations)
Example 3: Bias Assessment Report
Identified Threat: Recall bias in occupational exposure assessment
Severity: Moderate (differential recall possible by case status)
Mitigation Strategy:
- Primary: Use company employment records as gold standard
- Validation: Compare questionnaire response with records
- Sensitivity: Stratified analysis by exposure recall quality
What's Included
- Study Design Decision Tree: Interactive framework to select the optimal design based on your research question, study population, resource constraints, and exposure/outcome timing.
- Statistical Analysis Plan Template: Detailed SAP template with sections for primary objective, statistical methods, subgroup analyses, sensitivity analyses, and model specification rationale.
- Bias Assessment Framework: Structured checklist for identifying selection bias, information bias, confounding bias, and reverse causation with tailored mitigation strategies for each threat.
- Regression Model Specification Guide: Step-by-step guidance for choosing and specifying appropriate models with clear documentation of assumptions, functional forms, interaction terms, and covariate adjustment sets.
- Power and Sample Size Worksheet: Interactive worksheet for calculating sample size requirements with inputs for effect size, study design, significance level, power, and anticipated loss to follow-up.
- Sensitivity Analysis Checklist: Pre-built checklist of sensitivity analyses including unmeasured confounding (E-values), selection bias, measurement error, and alternative analytical approaches.
Who It's For
- Epidemiologists designing new studies
- Biostatisticians developing statistical analysis plans
- Public health researchers planning research protocols
- Clinical trial designers working on safety and efficacy studies
- Academic researchers preparing grant proposals
Best For
- Designing new epidemiological studies from scratch
- Writing pre-specified statistical analysis plans for protocols
- Conducting peer review of study designs and protocols
- Assessing confounding and bias in existing study designs
- Specifying regression models for causal inference studies
You might also like
Prepare comprehensive expert witness reports and testimony that meet Daubert standards for admissibility and withstand cross-examination scrutiny. You'll get structured case analyses integrating psychological assessments, relevant case law citations, and anticipated counterarguments. The skill helps you document your forensic opinions with the legal rigor required in criminal, civil, and family law proceedings.
by skillslib
You design robust motion control systems by generating control architectures, tuning PID and state-space controllers, and analyzing stability margins with detailed performance predictions. You'll get transfer functions, tuned gains, frequency response plots, real-time constraint validation, and code templates ready for implementation in MATLAB or Python.
by skillslib
You can systematically evaluate architectural options by building decision matrices, identifying trade-offs, and documenting design rationale. This skill generates comparison frameworks, risk assessments, and architecture decision records (ADRs) that help you choose between competing approaches based on concrete criteria like performance, scalability, cost, and team capabilities.
by skillslib
You provide a single core insight or perspective, and this skill transforms it into a complete, coordinated thought leadership content suite across every major channel. It generates optimized articles, social threads, email sequences, presentation outlines, and more—all reinforcing the same core message while adapting to each platform's unique audience and format. Every piece maintains strategic consistency while maximizing reach and engagement.
by skillslib
Analyze commercial insurance applications by extracting risk factors, evaluating loss history and industry exposure, and generating structured underwriting recommendations with standardized documentation. You'll reduce review time per application, ensure consistent decision criteria across your team, and maintain compliant audit trails for every underwriting action.
by skillslib
You can systematically diagnose flow problems in your Kanban board by analyzing stage-by-stage work accumulation, cycle time patterns, and throughput trends. The skill generates flow metrics dashboards, identifies capacity constraints, forecasts delivery dates based on historical data, and produces prioritized improvement recommendations with estimated impact—giving you objective evidence to guide WIP adjustments, process changes, and retrospective discussions.
by skillslib
You can map your game's features to certification requirements across ESRB, PEGI, USK, CLASSIND, and platform-specific standards simultaneously. The skill generates compliant test cases that prove adherence to each body's content and technical standards, tracks certification status to identify gaps before submission, and documents evidence trails for audits and appeals. This prevents costly certification rejections (2-8 week delays, $2,500-$10,000+ resubmission fees) by ensuring comprehensive compliance documentation from the QA phase onward.
by skillslib
You can design culturally appropriate survey instruments tailored to international development contexts, analyze mixed-methods data by integrating quantitative findings with qualitative insights, and implement robust data quality management systems for your field operations. Claude helps you create survey templates with adaptive logic, develop validation protocols, and produce data quality reports that highlight issues before analysis begins.
by skillslib