readmission-risk-explanation

name: readmission-risk-explanation description: Explain readmission risk drivers using LACE index scoring, clinical and social determinant analysis, and targeted intervention recommendations. Use when assessing 30-day readmission risk, explaining readmission prediction model outputs, designing readmission reduction interventions, or reporting on Hospital Readmissions Reduction Program metrics.

metadata: display_name: "Readmission Risk Explanation" short_description: "Explain 30-day readmission risk factors and drivers" default_prompt: "Explain my readmission risk in simple words and next steps" version: "1.0.0" tags: - healthcare

icon_path: "assets/icon.png"

Readmission Risk Explanation

Overview

Analyze and explain the key drivers of hospital readmission risk for individual patients or patient populations. This skill applies validated scoring models (LACE, HOSPITAL score, CMS risk-standardized models), decomposes risk factors into actionable categories, and generates targeted intervention recommendations to reduce preventable readmissions — aligned with the CMS Hospital Readmissions Reduction Program (HRRP).

When to Use

• Calculating and explaining LACE or HOSPITAL readmission risk scores
• Identifying high-risk patients at discharge for care management enrollment
• Explaining readmission prediction model outputs to clinical teams
• Designing patient-specific readmission prevention plans
• Reporting on HRRP performance metrics and penalty avoidance strategies
• Performing root cause analysis on readmission events

Required Inputs

Methodology

Step 1: LACE Index Calculation

Calculate the LACE score (validated 30-day readmission predictor):

L - Length of Stay (days):

• 1 day = 1 point, 2 days = 2, 3 days = 3, 4-6 days = 4, 7-13 days = 5, 14+ days = 7

A - Acuity of Admission:

• Elective = 0 points, Emergent/Urgent = 3 points

C - Comorbidity (Charlson Comorbidity Index):

• CCI 0 = 0 points, CCI 1 = 1, CCI 2 = 2, CCI 3 = 3, CCI 4+ = 5

E - Emergency Department visits (prior 6 months):

• 0 visits = 0 points, 1 = 1, 2 = 2, 3 = 3, 4+ = 4

Total Score Range: 0-19

Risk Stratification:

• 0-4: Low risk (approximately 2-5% readmission probability)
• 5-9: Moderate risk (approximately 10-15%)
• 10-14: High risk (approximately 20-30%)
• 15-19: Very high risk (approximately 35%+)

Step 2: Multidimensional Risk Factor Analysis

Evaluate risk across five domains:

1. Clinical Factors

• Disease severity and instability at discharge
• Number of active diagnoses (comorbidity burden)
• Medication complexity (polypharmacy: 5 or more medications, high-alert medications)
• Procedure-related complications
• Incomplete treatment course

2. Transitional Care Factors

• Discharge destination (home alone vs. home with support vs. facility)
• Follow-up appointment scheduled within 7 days (yes/no)
• Medication reconciliation completed
• Discharge instructions comprehension verified
• Durable medical equipment arranged

3. Social Determinants of Health (SDOH)

• Social isolation or inadequate caregiver support
• Food insecurity or housing instability
• Transportation barriers to follow-up care
• Health literacy level
• Language barriers

4. Behavioral Factors

• Medication non-adherence history
• Substance use disorders
• Mental health conditions (depression linked to 2x readmission risk)
• Missed appointment history

5. Health System Factors

• Primary care physician engagement
• Care coordination across settings
• Post-discharge outreach protocols
• Prior readmission history (strongest predictor)

Step 3: Risk Driver Prioritization

Rank risk factors by modifiability and impact:

Step 4: Intervention Mapping

Map prioritized risk factors to evidence-based interventions:

Pre-Discharge Interventions:

• Teach-back method for discharge instructions
• Medication reconciliation with pharmacist review
• Schedule follow-up within 7 days (within 48h for high-risk)
• Assess and address SDOH barriers

Transition Interventions:

• Warm handoff to PCP/specialist
• Transitional care nursing (Coleman CTI or Naylor TCM model)
• Medication delivery or assistance program enrollment
• Home health referral if indicated

Post-Discharge Interventions (within 72 hours):

• Phone call by RN to assess symptoms, medication adherence
• Confirm follow-up appointment and transportation
• Remote patient monitoring for applicable conditions
• Community health worker engagement for SDOH

Ongoing Interventions (30-day window):

• Care management enrollment for high-risk patients
• Chronic disease self-management education
• Behavioral health integration
• Palliative care referral if appropriate

Step 5: Explanation Generation

Produce a clear, stakeholder-appropriate explanation:

• For clinicians: Quantitative risk score with key clinical drivers and targeted interventions
• For case managers: Actionable risk factors with specific care transition recommendations
• For quality leaders: Population-level risk trends, HRRP performance, and intervention effectiveness
• For patients/families: Plain-language explanation of what to watch for and how to prevent return visits

Output Specification

The output includes:

index_admission: admission_date, discharge_date, primary_diagnosis (description, icd10, ms_drg), discharge_disposition

lace_score: total score, risk_tier, and individual component breakdowns

risk_factors: organized by domain (clinical, transitional, sdoh, behavioral, system) with factor, detail, severity, and modifiable flag

priority_drivers: ranked list with factor, domain, contribution_to_risk, modifiability

interventions: target_risk_factor, intervention description, evidence_basis, owner, timing

readmission_probability: estimated percentage

explanation_narrative: clear prose explanation tailored to audience

Analysis Framework

HRRP Applicable Conditions

CMS penalizes excess readmissions for these conditions/procedures:

1. Acute myocardial infarction (AMI)
2. Heart failure (HF)
3. Pneumonia
4. COPD
5. Total hip/knee arthroplasty (THA/TKA)
6. Coronary artery bypass graft (CABG)

Root Cause Analysis for Readmission Events

When analyzing actual readmissions, classify by:

• Potentially preventable: Related to index condition, inadequate transition, missed follow-up
• Planned: Scheduled readmission (staged procedure, chemotherapy)
• Unrelated: New condition unrelated to index admission

Examples

Input: 68-year-old female, discharged after HF exacerbation. LOS: 5 days. Emergent admission. Charlson: 4. ED visits in 6 months: 3. Lives alone. On 8 medications. No follow-up scheduled at discharge.

LACE: L=4 + A=3 + C=5 + E=3 = 15 (Very High Risk)

Priority Drivers: (1) No follow-up scheduled — immediate intervention required. (2) Polypharmacy with complex regimen. (3) Lives alone — no caregiver support. (4) High prior ED utilization suggesting inadequate outpatient management.

Recommended Interventions: Schedule cardiology follow-up within 48 hours, pharmacist medication reconciliation and teach-back, home health referral with telemonitoring, social work SDOH assessment, 48-hour post-discharge RN call.

Guidelines

1. LACE is a screening tool, not a clinical decision — use it to prioritize intervention intensity, not to deny care
2. Modifiable factors take priority — focus interventions on factors that can be changed
3. SDOH are clinical factors — treat social determinants with the same rigor as clinical risk factors
4. Document interventions delivered — track which interventions were implemented and their outcomes
5. Consider health equity — ensure risk models and interventions do not perpetuate disparities

Validation Checklist

• LACE score components are calculated correctly from documented values
• All five risk domains are assessed (clinical, transitional, SDOH, behavioral, system)
• Risk factors are classified by modifiability and impact
• Each priority driver has at least one mapped intervention
• Interventions have clear ownership and timing
• Explanation is appropriate for the target audience
• HRRP applicability is noted for relevant diagnoses

HIPAA Compliance Notes

• Readmission risk data often includes sensitive information (behavioral health, substance use, SDOH) requiring enhanced privacy protections
• Risk scores shared across care settings must comply with minimum necessary standard
• Patient consent may be required for SDOH data sharing depending on state law
• De-identify risk data used for population analytics and program evaluation
• Ensure risk model outputs are not used in discriminatory ways that violate civil rights protections