Skip to main content
ClaudeWave
Skill843 repo starsupdated 4d ago

clinical-trial

The clinical-trial skill assists with designing and analyzing randomized controlled trials and clinical studies by providing structured guidance on sample size calculation, randomization and blinding strategies, endpoint selection, interim analysis planning, and CONSORT reporting standards. Use this skill when users request help with RCT design, Phase I-IV trial planning, power analysis, statistical analysis plans, or compliance with clinical trial reporting guidelines.

Install in Claude Code
Copy
git clone --depth 1 https://github.com/beita6969/ScienceClaw /tmp/clinical-trial && cp -r /tmp/clinical-trial/skills/clinical-trial ~/.claude/skills/clinical-trial
Then start a new Claude Code session; the skill loads automatically.

SKILL.md

## When to Trigger

Activate this skill when the user mentions:
- Clinical trial design, RCT, randomized controlled trial
- Sample size calculation, power analysis for trials
- CONSORT, STROBE, SPIRIT guidelines
- Phase I, II, III, IV trials
- Primary/secondary endpoints, composite endpoints
- Interim analysis, adaptive trial design, futility
- Blinding, randomization, allocation concealment
- Intention-to-treat (ITT), per-protocol analysis

## Step-by-Step Methodology

1. **Define the research question** - Specify PICO (Population, Intervention, Comparator, Outcome). Determine if superiority, non-inferiority, or equivalence design is appropriate.
2. **Select trial phase and design** - Choose phase (I: safety/dose, II: efficacy signal, III: confirmatory, IV: post-market). Consider parallel, crossover, factorial, or adaptive designs.
3. **Primary endpoint selection** - Define primary outcome (must be clinically meaningful). Specify measurement timing and minimal clinically important difference (MCID).
4. **Sample size calculation** - Specify alpha (typically 0.05, two-sided), power (typically 80-90%), expected effect size, and dropout rate. Use appropriate formula for the endpoint type (continuous, binary, time-to-event).
5. **Randomization and blinding** - Recommend randomization method (simple, block, stratified, minimization). Specify blinding level (open-label, single, double, triple).
6. **Statistical analysis plan** - Pre-specify primary analysis method (t-test, chi-square, log-rank, mixed models). Define interim analysis schedule with alpha-spending function (O'Brien-Fleming, Lan-DeMets).
7. **Reporting** - Follow CONSORT for RCTs, STROBE for observational, SPIRIT for protocols. Include flow diagram, enrollment numbers, and all pre-specified analyses.

## Key Databases and Tools

- **ClinicalTrials.gov** - Trial registration and results
- **Cochrane Library** - Systematic reviews of trials
- **FDA / EMA guidance documents** - Regulatory requirements
- **nQuery / PASS / G*Power** - Sample size software
- **CONSORT / SPIRIT checklists** - Reporting standards

## Output Format

- Sample size as a table: assumptions, formula, per-arm and total N.
- Trial design as a structured summary: phase, design type, arms, blinding, duration.
- CONSORT flow diagram description with all participant numbers.
- Statistical analysis plan with pre-specified primary and secondary analyses.

## Quality Checklist

- [ ] PICO clearly defined
- [ ] Primary endpoint is clinically meaningful and measurable
- [ ] Sample size assumptions explicitly stated with references
- [ ] Dropout rate factored into sample size
- [ ] Randomization and blinding method specified
- [ ] Multiple comparison adjustment for secondary endpoints
- [ ] Interim analysis boundaries pre-specified if applicable
- [ ] ITT and per-protocol populations defined
- [ ] Regulatory guideline alignment noted (FDA/EMA/ICH)