Premium
Predicting Mortality in Older Adults with Kidney Disease: A Pragmatic Prediction Model
Author(s) -
Weiss Jessica W.,
Platt Robert W.,
Thorp Micah L.,
Yang Xiuhai,
Smith David H.,
Petrik Amanda,
Eckstrom Elizabeth,
Morris Cynthia,
O'Hare Ann M.,
Johnson Eric S.
Publication year - 2015
Publication title -
journal of the american geriatrics society
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.992
H-Index - 232
eISSN - 1532-5415
pISSN - 0002-8614
DOI - 10.1111/jgs.13257
Subject(s) - medicine , kidney disease , proportional hazards model , demography , cohort , risk of mortality , retrospective cohort study , population , gerontology , cohort study , renal function , environmental health , sociology
Objectives To develop mortality risk prediction models for older adults with chronic kidney disease ( CKD ) that include comorbidities and measures of health status and use not associated with particular comorbid conditions (nondisease‐specific measures). Design Retrospective cohort study. Setting Kaiser Permanente Northwest ( KPNW ) Health Maintenance Organization. Participants Individuals with severe CKD (estimated glomerular filtration rate <30 mL/min per 1.73 m 2 ; N = 4,054; n = 1,915 aged 65–79, n = 2,139 aged ≥80) who received care at KPNW between 2000 and 2008. Measurements Cox proportional hazards analysis was used to examine the association between selected participant characteristics and all‐cause mortality and to generate age group–specific risk prediction models. Predicted and observed risks were evaluated according to quintile. Predictors from the Cox models were translated into a points‐based system. Internal validation was used to provide best estimates of how these models might perform in an external population. Results The risk prediction models used 16 characteristics to identify participants with the highest risk of mortality at 2 years for adults aged 65 to 79 and 80 and older. Predicted and observed risks agreed within 5% for each quintile; a 4 to 5 times difference in 2‐year predicted mortality risk was observed between the highest and lowest quintiles. The c ‐statistics for each model (0.68–0.69) indicated effective discrimination without evidence of significant overfit (slope shrinkage 0.06–0.09). Models for each age group performed similarly for mortality prediction at 6 months and 2 years in terms of discrimination and calibration. Conclusion When validated, these risk prediction models may be helpful in supporting discussions about prognosis and treatment decisions sensitive to prognosis in older adults with CKD in real‐world clinical settings.