Skip to main page content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2017 Jan;24(1):176-181.
doi: 10.1093/jamia/ocw057. Epub 2016 Jun 29.

Predicting mortality over different time horizons: which data elements are needed?

Benjamin A Goldstein  1   2 , Michael J Pencina  3   2 , Maria E Montez-Rath  4 , Wolfgang C Winkelmayer  5
Affiliations

Predicting mortality over different time horizons: which data elements are needed?

Benjamin A Goldstein et al. J Am Med Inform Assoc. 2017 Jan.

Abstract

Objective: Electronic health records (EHRs) are a resource for "big data" analytics, containing a variety of data elements. We investigate how different categories of information contribute to prediction of mortality over different time horizons among patients undergoing hemodialysis treatment.

Material and methods: We derived prediction models for mortality over 7 time horizons using EHR data on older patients from a national chain of dialysis clinics linked with administrative data using LASSO (least absolute shrinkage and selection operator) regression. We assessed how different categories of information relate to risk assessment and compared discrete models to time-to-event models.

Results: The best predictors used all the available data (c-statistic ranged from 0.72-0.76), with stronger models in the near term. While different variable groups showed different utility, exclusion of any particular group did not lead to a meaningfully different risk assessment. Discrete time models performed better than time-to-event models.

Conclusions: Different variable groups were predictive over different time horizons, with vital signs most predictive for near-term mortality and demographic and comorbidities more important in long-term mortality.

Keywords: ESRD; Electronic Health Records; hemodialysis; predictive modeling.

PubMed Disclaimer

Figures

Figure 1
Figure 1
Cohort selection. All DaVita patients who were alive at day 30 of dialysis were eligible. Patients had to have 1 year of Medicare coverage and 6 months of Part D coverage prior to index date.
Figure 2
Figure 2
(a) C-statistics for different variable sets over different time horizons. The model with all variables performs the best. The most dynamic variables (eg, vitals) are most predictive in the near term, and the more stable variables (eg, comorbidities) are most predictive in the short term. (b) C-statistics after excluding different variable sets. Removal of any 1 variable set does not lead to meaningful differences in model performance.
Figure 3
Figure 3
Comparison of models based on discrete time and time-to-event analyses. The discrete time models perform best.
See this image and copyright information in PMC

References

    1. Mani S, Ozdas A, Aliferis C, et al. Medical decision support using machine learning for early detection of late-onset neonatal sepsis. J Am Med Inform Assoc. 2014;21(2):326–336. - PMC - PubMed
    1. Hivert M-F, Dusseault-Bélanger F, Cohen A, Courteau J, Vanasse A. Modified metabolic syndrome criteria for identification of patients at risk of developing diabetes and coronary heart diseases: longitudinal assessment via electronic health records. Can J Cardiol. 2012;28(6):744–749. - PubMed
    1. Eapen ZJ, Liang L, Fonarow GC, et al. Validated, electronic health record deployable prediction models for assessing patient risk of 30-day rehospitalization and mortality in older heart failure patients. JACC Heart Fail. 2013;1(3):245–251. - PubMed
    1. Matheny ME, Miller RA, Ikizler TA, et al. Development of inpatient risk stratification models of acute kidney injury for use in electronic health records. Med Decis Mak Int J Soc Med Decis Mak. 2010;30(6):639–650. - PMC - PubMed
    1. U.S. Renal Data System, USRDS 2014 Annual Data Report: Atlas of End-Stage Renal Disease in the United States. National Institute of Health, Bethesda, MD: National Diabetes and Digestive and Kidney Diseases; 2014.

Publication types