home / skills / a5c-ai / babysitter / population-health-stratification

This skill helps stratify patient populations by risk using claims, clinical data, and social determinants to prioritize care management interventions.

npx playbooks add skill a5c-ai/babysitter --skill population-health-stratification

Review the files below or copy the command above to add this skill to your agents.

Files (1)
SKILL.md
2.2 KB
---
name: population-health-stratification
description: Stratify patient populations by risk level using claims data, clinical data, and social determinants to prioritize care management interventions
allowed-tools: Read, Grep, Write, Edit, Glob, Bash, WebFetch
---

# Population Health Stratification

Stratify patient populations by risk level using claims data, clinical data, and social determinants to prioritize care management interventions.

## Overview

This skill enables risk stratification of patient populations for care management. It encompasses data analysis, risk modeling, segment identification, and intervention prioritization to target resources effectively.

## Capabilities

### Risk Assessment
- Claims-based risk scores
- Clinical risk factors
- Utilization patterns
- Social determinants
- Predictive modeling

### Data Analysis
- Multi-source integration
- Pattern identification
- Cohort analysis
- Trend tracking
- Outcome correlation

### Stratification Models
- Rising risk identification
- High-risk patient flagging
- Condition-specific cohorts
- Utilization tiers
- Intervention matching

### Resource Targeting
- Care management allocation
- Intervention prioritization
- Program matching
- Outreach planning
- Impact projection

## Usage Guidelines

### Stratification Process
1. Define population scope
2. Aggregate data sources
3. Apply risk algorithms
4. Validate stratification
5. Create patient segments
6. Match interventions
7. Monitor outcomes

### Risk Factors
- Chronic conditions
- Prior utilization
- Medication complexity
- Social needs
- Care gaps

### Intervention Matching
- High-risk: Intensive care management
- Rising-risk: Targeted outreach
- Low-risk: Wellness programs
- Condition-specific: Disease management
- Social needs: Community resources

## Integration Points

### Related Processes
- Population Health Management Program
- Clinical Pathway Development
- Service Line Strategic Planning

### Collaborating Skills
- care-transition-coordination
- clinical-workflow-analysis
- quality-metrics-measurement

## References

- Population health frameworks
- Risk stratification methodologies
- AHRQ population health tools
- ACO quality metrics

Overview

This skill performs population health stratification to rank patients by risk using claims, clinical records, and social determinants. It produces actionable segments and intervention priorities so care teams can focus resources where they will have the most impact. The output supports operational workflows for care management and program planning.

How this skill works

The skill ingests multi-source data (claims, EHR, social determinants) and normalizes it for analysis. It calculates risk scores and patterns using configurable algorithms, identifies cohorts (high-risk, rising-risk, condition-specific), and recommends intervention matches. Results include segment lists, prioritization rationale, and outcome tracking metrics for monitoring and validation.

When to use it

  • Prioritizing patients for care management and case reviews
  • Identifying rising-risk individuals for early intervention
  • Allocating limited care coordination resources across populations
  • Designing targeted outreach or disease-management programs
  • Measuring impact of interventions and adjusting program design

Best practices

  • Define the population scope and business goals before modeling
  • Combine claims, clinical, and social data to improve accuracy
  • Validate models with historical outcomes and clinical review
  • Regularly re-evaluate thresholds and update models for drift
  • Map stratification outputs to concrete interventions and workflows

Example use cases

  • Flagging high-cost, high-utilization patients for intensive case management
  • Detecting patients with rising risk to trigger outreach and prevent hospitalization
  • Creating condition-specific cohorts (diabetes, CHF) for tailored disease programs
  • Prioritizing social needs referrals by combining SDoH flags with utilization risk
  • Projecting program impact by simulating resource allocation across tiers

FAQ

What data sources are required?

At minimum, claims and clinical encounter data are needed; adding social determinants and pharmacy data improves risk detection and intervention matching.

How are risk thresholds chosen?

Thresholds should align with care-management capacity and program goals; choose empirically using historical outcomes and clinician input, and update periodically.