Premium
Development of comorbidity score for patients undergoing major surgery
Author(s) -
Mehta Hemalkumar B.,
Yong Shan,
Sura Sneha D.,
Hughes Byron D.,
Kuo YongFang,
Williams Stephen B.,
Tyler Douglas S.,
Riall Taylor S.,
Goodwin James S.
Publication year - 2019
Publication title -
health services research
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.706
H-Index - 121
eISSN - 1475-6773
pISSN - 0017-9124
DOI - 10.1111/1475-6773.13209
Subject(s) - comorbidity , medicine , cohort , statistic , logistic regression , retrospective cohort study , medicaid , emergency medicine , cohort study , statistics , health care , mathematics , economics , economic growth
Objective To develop and validate a claims‐based comorbidity score for patients undergoing major surgery, and compare its performance with established comorbidity scores. Data Source Five percent Medicare data from 2007 to 2014. Study Design Retrospective cohort study of patients aged ≥65 years undergoing six major operations (N = 99 250). Data Collection One‐year mortality was the primary outcome. Secondary outcomes were hospital mortality, 30‐day mortality, 30‐day readmission, and length of stay. The comorbidity score was developed in the derivation cohort (70 percent sample) using logistic regression model. The comorbidity score was calibrated and validated in the validation cohort (30 percent sample), and compared against the Charlson, Elixhauser, and Centers for Medicare and Medicaid Services Hierarchical Condition Categories (CMS‐HCC) comorbidity scores using c ‐statistic, net reclassification improvement, and integrated discrimination improvement. Principal Findings In the validation cohort, the surgery‐specific comorbidity score was well calibrated and performed better than the Charlson, Elixhauser, and CMS‐HCC comorbidity scores for all outcomes; the performance was comparable to the CMS‐HCC for 30‐day readmission. For example, the surgery‐specific comorbidity score ( c ‐statistic = 0.792; 95% CI, 0.785‐0.799) had greater discrimination than the Charlson ( c ‐statistic = 0.747; 95% CI, 0.739‐0.755), Elixhauser ( c ‐statistic = 0.747; 95% CI, 0.735‐0.755), or CMS‐HCC ( c ‐statistic = 0.755; 95% CI, 0.747‐0.763) scores in predicting 1‐year mortality. The net reclassification improvement and integrated discrimination improvement were greater for surgery‐specific comorbidity score compared to the Charlson, Elixhauser, and CMS‐HCC scores. Conclusions Compared to commonly used comorbidity measures, a surgery‐specific comorbidity score better predicted outcomes in the surgical population.