Dataset Information

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

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.

SUBMITTER: Goldstein BA

PROVIDER: S-EPMC5201182 | biostudies-literature | 2017 Jan

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

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

Goldstein Benjamin A BA Pencina Michael J MJ Montez-Rath Maria E ME Winkelmayer Wolfgang C WC

Journal of the American Medical Informatics Association : JAMIA 20160629 1

<h4>Objective</h4>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.<h4>Material and methods</h4>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 dat ...[more]

PMID: 27357832

Similar Datasets

Project description:BackgroundIt is important to find a comorbidity measure with better performance for use with administrative data. The new method proposed by Elixhauser et al. has never been validated and compared to the widely used Charlson method in the Asia region. The objective of this study was to compare the performance of three comorbidity measures using information from different data periods in predicting short- and long-term mortality among patients with acute myocardial infarction (AMI) and chronic obstructive pulmonary disease (COPD).MethodsWe conducted a retrospective cohort study using National Health Insurance claims data (2001-2002) in Taiwan. We constructed the Elixhauser, the Charlson/Deyo, and the Charlson/Romano methods based on the International Classification of Disease, 9th Revision, Clinical Modification codes in the claims data. Two data periods, including the index hospitalization as well as the index and prior 1-year hospitalizations, were used in the analysis. The performances were compared using the c-statistics derived from multiple logistic regression models that included age, gender, race, and whether the patient received surgery or not. The outcomes of interest were in-hospital and 1-year mortality.ResultsThe performance was in the same rank order among both populations regardless of the outcome and data period: Elixhauser > Charlson/Romano > Charlson/Deyo. In predicting in-hospital mortality, the Elixhauser models using information from the index hospitalization performed best, even better than the Charlson/Deyo or Charlson/Romano models using information from the index and prior hospitalizations. Nevertheless, in predicting 1-year mortality, the Elixhauser models using information from the index and 1-year prior hospitalizations performed better than using information from the index hospitalization only.ConclusionsThis is so far the first study to validate the Elixhauser method and compare it to other methods in the Asia region, and is the first to report its differences in data periods between short- and long-term outcomes. The comorbidity measurement developed by Elixhauser et al. has relatively good predictive validity, and researchers should consider its use in claims-based studies.

Project description:BACKGROUND:African children hospitalised with severe febrile illness have a high risk of mortality. The Fluid Expansion As Supportive Therapy (FEAST) trial (ISCRTN 69856593) demonstrated increased mortality risk associated with fluid boluses, but the temporal relationship to bolus therapy and underlying mechanism remains unclear. METHODS:In a post hoc retrospective analysis, flexible parametric models were used to compare change in mortality risk post-randomisation in children allocated to bolus therapy with 20-40?ml/kg 5% albumin or 0.9% saline over 1-2?h or no bolus (control, 4?ml/kg/hour maintenance), overall and for different terminal clinical events (cardiogenic, neurological, respiratory, or unknown/other). RESULTS:Two thousand ninety-seven and 1041 children were randomised to bolus vs no bolus, of whom 254 (12%) and 91 (9%) respectively died within 28?days. Median (IQR) bolus fluid in the bolus groups received by 4?h was 20 (20, 40) ml/kg and was the same at 8?h; total fluids received in bolus groups at 4?h and 8?h were 38 (28, 43) ml/kg and 40 (30, 50) ml/kg, respectively. Total fluid volumes received in the control group by 4?h and 8?h were median (IQR) 10 (6, 15) ml/kg and 10 (10, 26) ml/kg, respectively. Mortality risk was greatest 30?min post-randomisation in both groups, declining sharply to 4?h and then more slowly to 28?days. Maximum mortality risk was similar in bolus and no bolus groups; however, the risk declined more slowly in the bolus group, with significantly higher mortality risk compared to the no bolus group from 1.6 to 101?h (4?days) post-randomisation. The delay in decline in mortality risk in the bolus groups was most pronounced for cardiogenic modes of death. CONCLUSIONS:The increased risk from bolus therapy was not due to a mechanism occurring immediately after bolus administration. Excess mortality risk in the bolus group resulted from slower decrease in mortality risk over the ensuing 4?days. Thus, administration of modest bolus volumes appeared to prevent mortality risk declining at the same rate that it would have done without a bolus, rather than harm associated with bolus resulting from a concurrent increased risk of death peri-bolus administration. TRIAL REGISTRATION:ISRCTN69856593. Date of registration 15 December 2008.

Dataset Information

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

Publications

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

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets