Charlson comorbidity index using administrative database in incident PD patients.

Aims: Mortality risks in ESRD (end stage renal disease) patients are related to comorbid diseases. The Charlson Comorbidity Index (CCI) has been reported to be a strong predictor of survival in incipient ESRD patients. The authors studied the validity of CCI using administrative database according to International Classification of Disease, 10th revision (ICD-10) codes, to devise a more straightforward method of determining CCI than that based on medical records review. Subjects and methods: Incident peritoneal dialysis patients (N = 134) were enrolled from 1997 through 2007. We compared CCI scored by the administrative database (A-CCI) with CCI scored by medical records review (R-CCI). These CCI scores and patients' outcomes were analyzed. Results: For all patients, mean A-CCI and R-CCI were 5.3 ± 2.1 (range 2 - 11) and 5.4 ± 2.1 (range 2 - 11), respectively. High correlation was found between A-CCI and R-CCI scores (r = 0.88, p < 0.01). The sensitivity of A-CCI was high (0.57 - 1.00) for nine comorbidities, but sensitivities for chronic pulmonary disease and peptic ulcer disease were poor (< 0.50). However, specificity was excellent for most comorbidities. Three comorbidity groups were established by tertile ranking: low comorbidity (score = 2 - 4), moderate comorbidity (score = 5 - 6), and high comorbidity (score= 7-11 ). The mortality rates were; 7.17, 15.96, and 23.07/100 patient-years by A-CCI, and 6.69, 13.58 and 28.16/100 patient-years by R-CCI, respectively. Conclusion: CCI scores from administrative database using ICD-10 were similar to CCI scores by medical records review. This method is simple and valid to predict the outcomes of incipient PD patients.