Skip to main content
The contribution of longitudinal comorbidity measurements to survival analysis
Meyers Primary Care Institute Publications and Presentations
  • C. Y. Wang, Fred Hutchinson Cancer Research Center
  • Laura-Mae Baldwin, University of Washington
  • Barry G. Saver, University of Massachusetts Medical School
  • Sharon A. Dobie, University of Washington
  • Pamela K. Green, University of Washington
  • Yong Cai, University of Utah
  • Carrie N. Klabunde, National Cancer Institute
UMMS Affiliation
Meyers Primary Care Institute; Department of Family Medicine and Community Health
Publication Date
Document Type
Aged; Cause of Death; *Comorbidity; Data Collection; Data Interpretation, Statistical; Female; Geriatric Assessment; Health Services Research; Health Status; Health Status Indicators; Humans; Likelihood Functions; *Longitudinal Studies; Male; Medicare; Multivariate Analysis; Predictive Value of Tests; *Proportional Hazards Models; Research Design; Retrospective Studies; SEER Program; *Survival Analysis; Time Factors; United States
BACKGROUND: Many clinical and health services research studies are longitudinal, raising questions about how best to use an individual's comorbidity measurements over time to predict survival. OBJECTIVES: To evaluate the performance ofdifferent approaches to longitudinal comorbidity measurement in predicting survival, and to examine strategies for addressing the inevitable issue of missing data. RESEARCH DESIGN: Retrospective cohort study using Cox regression analysis to examine the association between various Romano-Charlson comorbidity measures and survival. SUBJECTS: Fifty thousand cancer-free individuals aged 66 or older enrolled in Medicare between 1991 and 1999 for at least 1 year. RESULTS: The best fitting model combined both time independent baseline comorbidity and the time dependent prior year comorbidity measure. The worst fitting model included baseline comorbidity only. Overall, the models fit best when using the "rolling" comorbidity measures that assumed chronic conditions persisted rather than measures using only prior year's recorded diagnoses. CONCLUSIONS: Longitudinal comorbidity is an important predictor of survival, and investigators should make use of individuals' longitudinal comorbidity data in their regression modeling.
DOI of Published Version
Med Care. 2009 Jul;47(7):813-21. Link to article on publisher's site
Related Resources
Link to Article in PubMed
PubMed ID
Citation Information
C. Y. Wang, Laura-Mae Baldwin, Barry G. Saver, Sharon A. Dobie, et al.. "The contribution of longitudinal comorbidity measurements to survival analysis" Vol. 47 Iss. 7 (2009) ISSN: 0025-7079 (Linking)
Available at: