All Engines
Financial & Trend

Case Mix Adjustment Engine

Normalize for Disease Burden—Compare Diabetic Population to Healthy Population Apples-to-Apples

The Hidden Acuity Variable

Unadjusted Cost Comparisons

  • Group A: $950 PMPM. Group B: $1,200 PMPM. Which has better care management?
  • Cannot tell if cost difference is efficiency or disease burden (diabetes, cancer, etc.)
  • Trend contaminated by new chronic disease diagnoses vs. actual cost inflation
  • Penalizes employers who hire/retain employees with chronic conditions

Case Mix Adjustment

  • HCC-based risk scoring: each chronic condition adds to population burden
  • Apples-to-apples: $950 PMPM → $1,080 adjusted, $1,200 PMPM → $980 adjusted
  • Clean trend: new diabetes diagnosis impact separated from utilization growth
  • Fair comparison: sicker population managed well vs. healthy population managed poorly

HCC Risk Adjustment Model

python
# Hierarchical Condition Category (HCC) Risk Adjustment hcc_risk_weights = { 'HCC001': 0.302, # HIV/AIDS 'HCC008': 0.331, # Metastatic Cancer 'HCC018': 0.318, # Diabetes with Chronic Complications 'HCC019': 0.104, # Diabetes without Complication 'HCC085': 0.323, # Congestive Heart Failure 'HCC096': 0.368, # Specified Heart Arrhythmias 'HCC108': 0.191, # Vascular Disease 'HCC111': 0.302, # Chronic Obstructive Pulmonary Disease 'HCC134': 0.497, # Dialysis Status (ESRD) # ... 86 total HCCs in CMS-HCC V28 model } def calculate_case_mix_index(population, claims_history): """ Calculate population-level case mix index using HCC model """ total_risk_score = 0 for member in population: # Extract diagnoses from claims history diagnoses = extract_diagnoses(claims_history, member.id) # Map diagnoses to HCCs member_hccs = map_diagnoses_to_hccs(diagnoses) # Calculate member risk score (additive across HCCs) member_risk = 1.0 # Baseline healthy member for hcc in member_hccs: member_risk += hcc_risk_weights[hcc] total_risk_score += member_risk # Population Case Mix Index case_mix_index = total_risk_score / len(population) return case_mix_index def adjust_costs_for_case_mix(actual_costs_pmpm, case_mix_index): """ Normalize costs to a standard case mix of 1.0 (healthy population) """ adjusted_costs = actual_costs_pmpm / case_mix_index return { 'actual_pmpm': actual_costs_pmpm, 'case_mix_index': case_mix_index, 'adjusted_pmpm': adjusted_costs, 'disease_burden_impact': (case_mix_index - 1.0) * 100 } # Example: Two Employers # Employer A (Tech Startup): # - Actual PMPM: $950 # - Case Mix Index: 0.88 (healthier than average) # - Adjusted PMPM: $1,080 ($950 / 0.88) # # Employer B (Manufacturing): # - Actual PMPM: $1,200 # - Case Mix Index: 1.22 (22% sicker than average) # - Adjusted PMPM: $984 ($1,200 / 1.22) # # Conclusion: Manufacturing is actually MORE cost-efficient # when adjusted for their higher disease burden

Acuity Intelligence

Condition Categories
86 HCCs
CMS-HCC V28 model covering all major chronic conditions
Risk Score Range
0.6x to 8.5x
Healthy young adult vs. multi-comorbid dialysis patient
Adjustment Precision
±3% PMPM
Accurate disease burden normalization

Fair Performance Benchmarking

Manufacturing vs. Tech

  • Manufacturer: $1,180 PMPM, Case Mix 1.35 (chronic disease heavy)
  • Tech firm: $920 PMPM, Case Mix 0.82 (young/healthy)
  • Adjusted: Manufacturer $874, Tech $1,122
  • Manufacturer is 22% MORE efficient when disease-adjusted
  • Tech firm has care management opportunity despite lower raw costs

Chronic Disease Trend Isolation

  • 2023: $985 PMPM, Case Mix 1.08
  • 2024: $1,095 PMPM, Case Mix 1.18 (+11% raw trend)
  • 2023 adjusted: $912, 2024 adjusted: $928
  • True trend: +1.8% (not +11%)
  • 9.2% of apparent trend was new chronic diagnoses (aging workforce)

Diabetes Management ROI

  • Diabetic cohort (n=85): $18,500 PMPY vs. $24,000 national benchmark
  • Case Mix Index: 2.15 (HCC018 + comorbidities)
  • Expected costs given acuity: $23,800
  • Actual: $18,500 = 22% better than predicted
  • Disease management program delivering $450K savings

Compare Performance, Not Disease Burden

Adjust for chronic disease acuity using HCC risk scores. Benchmark fairly across different health profiles. Separate new diagnoses from cost trend.

Adjust for Case Mix