clinical-decision-support

Clinical Decision Support - Clinical Decision Support Document Generation

Skill Overview

Generate professional clinical decision support (CDS) documents for pharmaceutical companies, clinical research organizations, and medical decision-makers, including biomarker-stratified patient cohort analyses and treatment recommendation reports graded by the GRADE evidence system.

Applicable Scenarios

1. Drug Development and Clinical Trials

Generate subgroup analysis documents for Phase 2/3 trials, companion diagnostic development reports, and regulatory submission materials for FDA/EMA during new drug development. Support biomarker-stratified analyses, survival curve plotting, and hazard ratio calculations to meet documentation needs across the drug development lifecycle.

2. Medical Affairs and Scientific Engagement

Produce KOL (key opinion leader) educational materials, medical strategy documents, and advisory board presentations for medical affairs teams. Create evidence-based treatment algorithms and clinical pathway diagrams to support conference presentations and manuscript preparation.

3. Clinical Guideline Development and Real-World Evidence Research

Assist professional societies in developing clinical practice guidelines and generate recommendations graded using the GRADE system. Support cohort analyses based on real-world data (RWE) to produce comparative effectiveness and outcomes study reports.

Core Features

Biomarker-Stratified Analyses

Support patient cohort analyses based on molecular subtyping (e.g., GBM mesenchymal–immune-activated vs. proneural), genomic alterations (mutations, copy number variations, gene fusions), IHC markers, and PD-L1 scoring. Automatically generate Kaplan-Meier survival curves, waterfall plots, and forest plots; calculate hazard ratios, P-values, and 95% confidence intervals.

GRADE Evidence-Graded Treatment Recommendations

Produce evidence-based clinical guidelines that meet international standards, using the GRADE system to grade recommendation strength (1A/1B/2A/2B/2C) and evidence quality. Built-in decision-algorithm flowchart generation enables creation of treatment pathways based on biomarker status and clinical criteria.

Professional-Format Output and Compliance Support

All documents output in LaTeX/PDF format with compact 0.5-inch margins, suitable for regulatory submissions and academic publication. Built-in HIPAA de-identification, ICH-GCP compliance checks, and confidentiality statements ensure documents meet pharmaceutical industry regulatory requirements.

Frequently Asked Questions

What is the difference between the Clinical Decision Support skill and the Treatment Planning skill?

The Clinical Decision Support skill focuses on analyses and evidence synthesis at the population level, suitable for drug development, clinical trials, and guideline development. The Treatment Planning skill is used for bedside care planning and treatment protocols for individual patients. If you need cohort analyses, guideline generation, or preparation of regulatory submission materials, use this skill; if you are creating a care plan for a single patient, use the Treatment Planning skill.

What statistical analyses and charts are supported in the generated documents?

The skill supports comprehensive statistical analyses, including Kaplan-Meier survival curves with log-rank tests, Cox regression, Fisher’s exact test, and more. It can automatically generate waterfall plots (best response), forest plots (subgroup analyses), swimlane plots (patient timelines), and TikZ decision-algorithm flowcharts. All charts are publication-grade and meet medical journal submission standards.

How do you ensure the generated documents meet regulatory requirements?

The skill includes multiple compliance features: HIPAA Safe Harbor removal of 18 identifiers, ICH-GCP alignment, and confidentiality statement templates. Documents use professional 0.5-inch margin formatting and include full statistical methodology and references. All terminology uses standard medical vocabularies such as SNOMED-CT and LOINC to ensure professionalism and accuracy.