(294) Development and Validation of Elixhauser Summary Comorbidity Score for Use with ICD-10-CM Coded Data Among Older Adults

Saturday, August 27, 2022

8:00 AM – 6:00 PM CEST

Location: Hall C

Publication Number: 968

Presenting Author(s)

Hemalkumar B. Mehta, PhD
– Assistant Professor, Bloomberg School of Public Health, Johns Hopkins University, Baltimore, Maryland, United States

Co-Author(s)

Shuang Li, PhD
– The University of Texas Medical Branch, United States
Huijun An, MS
– Research Analyst, Johns Hopkins Bloomberg School of Public Health, MD, United States
James Goodwin
– The University of Texas Medical Branch, United States
G. Caleb Alexander, MD
– Professor of Epidemiology and Medicine, JHSPH Center for Drug Safety & Effectiveness, Baltimore, Maryland, United States
Jodi Segal, FISPE
– Professor of Medicine, Johns Hopkins University, Baltimore, Maryland, United States

Background: A summary comorbidity score derived for older adults may better measure comorbidities in observational studies.

Objectives: We developed a summary Elixhauser comorbidity score (S-Elixhauser) to predict 30-day, in-hospital and 1-year mortality in older adults using the 38 comorbidities developed by the Agency for Healthcare Research and Quality (AHRQ) and compared its performance with existing comorbidity measures in older patients and in three disease-specific populations (heart failure, chronic obstructive pulmonary disease and diabetes mellitus).

Methods: This retrospective cohort study used Medicare claims for fee-for-service beneficiaries from 2017 to 2019 and included Medicare beneficiaries hospitalized in 2018 (n=899,844) and three disease-specific hospitalized patient cohorts. We used logistic regression and separately derived weights for 38 comorbidities to predict 30-day, in-hospital and 1-year mortality. We internally validated and calibrated the S-Elixhauser, and compared its performance against the individual Elixhauser comorbidity indicators (38 comorbidities), the AHRQ-derived weighted summary score, the Charlson comorbidity indicators (17 comorbidities) and the Charlson summary score in predicting 30-day, in-hospital and 1-year mortality, using c-statistics.

Results: The S-Elixhauser score was well-calibrated and showed modestly greater discrimination in predicting 30-day mortality (c-statistic 0.705, CI 0.703-0.707) than the AHRQ Elixhauser score (c-statistic 0.688, CI 0.686-0.690), the Charlson comorbidity indicators (c-statistic 0.696, CI 0.694- 0.698) and the Charlson summary score (c-statistic 0.690, CI 0.688-0.693), but less discrimination than the individual Elixhauser comorbidity indicators (0.711, CI 0.709-0.713). Similar findings were observed for in-hospital and 1-year mortality. In three disease-specific populations, the S-Elixhauser score performed equally well or better than the AHRQ Elixhauser score and both Charlson comorbidity measures.

Conclusions: The newly derived S-Elixhauser comorbidity score predicts short-term and long-term mortality in older adults with a single measure. This may be a valuable tool to measure comorbidity burden in observational studies of older adults.