Clinical Research
This skill supports the design of prospective clinical studies before submission: it helps select and classify endpoints (primary/key-secondary/exploratory, with surrogate-endpoint flagging), estimate sample size and power for two-arm designs (means, proportions, survival), and score a study plan for feasibility, returning a GO/GO-WITH-CONDITIONS/REDESIGN/NO-GO phase-gate decision. Every output is an estimate with a named human owner (clinician, biostatistician, regulatory owner), never clinical fact or a finished protocol.
✨ What it does
- Endpoint selector scores candidate endpoints across 5 dimensions and classifies them as PRIMARY / KEY-SECONDARY / EXPLOR
- Sample size estimator computes power/sample size for two-arm means, proportions, and survival designs, with dropout infl
- Phase-gate scorer returns a 0-100 feasibility score and a GO / GO-WITH-CONDITIONS / REDESIGN / NO-GO verdict with named
- Onboarding questionnaire captures defaults (profile, alpha, power, dropout, named owners) that customize tool behavior.
- Opt-in autoresearch bridge can iteratively improve a study plan against the feasibility score.
- Outputs always include an 'ESTIMATE' banner and named human owner.
🎯 When to use it
- When choosing a primary endpoint and needing to defend it against surrogate-endpoint scrutiny.
- When you need a defensible first sample-size estimate for a protocol synopsis.
- When a study plan needs a feasibility read before a phase-gate review.
- When pressure-testing whether planned enrollment is achievable given the eligible population and sites.
🚀 How to use
Trigger: /ar:setup
Trigger by asking to design a clinical study, select endpoints, estimate sample size, or score feasibility. Run the onboarding questionnaire once before starting to set defaults. Then run the scripts with appropriate inputs. Example prompts:
Help me choose a primary endpoint for a phase 2 device trial and estimate the sample size.
Run a feasibility phase-gate review on this study plan.
Inputs: a protocol synopsis (template provided), endpoints JSON, study plan JSON, and CLI flags for design parameters.
📄 Output: A gate packet containing the protocol synopsis, endpoint classification, sample size estimate, and feasibility verdict with named owners.
📦 Add this skill to Claude Code
# 1. Get the skills repo
git clone --depth 1 https://github.com/alirezarezvani/claude-skills /tmp/claude-skills
# 2. Copy this skill into your project (or ~/.claude/skills for all projects)
mkdir -p .claude/skills
cp -r /tmp/claude-skills/research-ops/skills/clinical-research .claude/skills/clinical-researchSkill source: research-ops/skills/clinical-research/SKILL.md
⚠️ Good to know
Outputs are estimates requiring confirmation by a biostatistician; the skill does not replace regulatory submission (use ra-qm-team) or final protocol development.
❓ FAQ
Can an unvalidated surrogate endpoint be a primary endpoint?
No, the endpoint selector enforces a penalty and prevents an unvalidated surrogate from anchoring a primary endpoint.
What does the phase-gate scorer return?
It returns a feasibility score from 0 to 100 and a verdict of GO, GO-WITH-CONDITIONS, REDESIGN, or NO-GO, along with named owners who must sign.
How do I customize the skill's defaults?
Run the onboarding questionnaire (python3 scripts/onboard.py) to set your development area, alpha, power, dropout, and named owners; answers are saved and used automatically.
🤖 Overview, features, install steps and FAQ were generated from the project's SKILL.md on Sep 4, 2026. Always check the original source before running commands.