Charlson comorbidity index using administrative database in incident PD patients

Area of Science:

• Nephrology
• Epidemiology
• Health Informatics

Background:

• Mortality in end-stage renal disease (ESRD) patients is significantly influenced by comorbid conditions.
• The Charlson Comorbidity Index (CCI) is a recognized predictor of survival in ESRD patients.
• Current CCI assessment often relies on time-consuming medical record reviews.

Purpose of the Study:

• To evaluate the validity of using administrative database International Classification of Disease, 10th revision (ICD-10) codes for calculating CCI.
• To establish a more straightforward method for determining CCI compared to manual review.
• To assess the utility of administrative CCI (A-CCI) in predicting outcomes for peritoneal dialysis (PD) patients.

Main Methods:

• Incident PD patients (N=134) from 1997-2007 were analyzed.
• CCI scores derived from administrative data (A-CCI) were compared with those from medical records review (R-CCI).
• Correlation, sensitivity, and specificity of A-CCI were assessed against R-CCI and patient outcomes.

Main Results:

• A-CCI and R-CCI demonstrated high correlation (r=0.88, p<0.01), with similar mean scores (5.3 vs 5.4).
• A-CCI showed high sensitivity for most comorbidities, though lower for chronic pulmonary and peptic ulcer disease.
• Mortality rates across low, moderate, and high comorbidity groups showed consistent trends between A-CCI and R-CCI, indicating predictive validity.

Conclusions:

• CCI scores derived from administrative ICD-10 data closely mirror those from medical record reviews.
• This administrative approach provides a simple and valid method for predicting outcomes in incident PD patients.
• The A-CCI method offers a practical alternative for risk stratification in ESRD populations.