New! Sign up for our free email newsletter.
Science News
from research organizations

Data from full hospital stay not much better at predicting risk for readmission than data from first day, researchers find

Date:
March 31, 2016
Source:
UT Southwestern Medical Center
Summary:
Culling more detailed clinical data from electronic health records throughout a hospital stay did not substantially improve predictions about who was more likely to be readmitted, an analysis showed, suggesting further studies will be needed to help build effective analytical tools that can help predict outcomes and readmissions.
Share:
FULL STORY

Culling more detailed clinical data from electronic health records throughout a hospital stay did not substantially improve predictions about who was more likely to be readmitted, an analysis by UT Southwestern researchers showed, suggesting further studies will be needed to help build effective analytical tools that can help predict outcomes and readmissions.

The study used detailed data in the electronic medical record on hospital complications, and vital sign and lab test trajectories. Surprisingly, researchers found that data from a patient's entire hospital stay did not meaningfully improve the accuracy of predicting readmissions within 30 days compared with using data collected from just the first 24 hours of admission.

"Our group's previous research found that using clinical data from the first day of admission was more effective in predicting hospital readmissions than using administrative billing data," said lead author Dr. Oanh Nguyen, Assistant Professor of Internal Medicine and Clinical Sciences at UT Southwestern. "So we expected that adding even more detailed clinical data from the entire hospitalization would allow us to better identify which patients are at highest risk for readmission. However, we were surprised to find that this was not the case."

The study suggests that other factors influence who does poorly after discharge well beyond the severity of someone's illness and treatment during hospitalization. "More 'big data' alone did not make much of a difference. Better models for predicting readmissions will require 'better data' on things like psychosocial and behavioral factors that are not currently captured in electronic health records," said Dr. Ethan A. Halm, Chief of the William T. and Gay F. Solomon Division of General Internal Medicine, and Chief of the Division of Outcomes and Health Services Research in the Department of Clinical Sciences at UT Southwestern.

The study is believed to be the first to rigorously measure the additive influence of in-hospital complications, clinical trajectory, and stability at discharge on the risk of 30-day hospital readmission, the researchers reported.

Researchers examined discharges between November 2009 and October 2010 from six diverse hospitals in North Texas. In roughly 13 percent of nearly 33,000 admission cases reviewed, patients were readmitted during the following 30 days to one of 75 hospitals in North Texas, researchers found. The researchers then compared models to predict who would be readmitted, comparing data from the first day of admission to the hospital (first-day model) to data from throughout the entire hospital stay (full-stay model). They found that the full-stay model performed only marginally better.

However, the researchers did find that both the full-stay and first-day risk prediction models they developed based on electronic health record data were better than several existing models at predicting who would be readmitted. Researchers suggested that predictive models tailored to predict readmissions for specific diseases versus all-comers might be even more effective, and these head-to-head comparisons are areas of current investigation.

Researchers are attempting to develop better analytical tools that will flag patients more likely to be readmitted in order to develop interventions to reduce that likelihood, Dr. Nguyen said. The study also underscored the need to collect other types of social and behavioral data that may influence readmissions, such as substance abuse, support structures, or whether patients took the medicine prescribed or adhered to follow-up visits as recommended.

The study, which appears in the Journal of Hospital Medicine, was supported by the UT Southwestern Center for Patient-Centered Outcomes Research led by Dr. Halm, Professor of Internal Medicine and Clinical Sciences, who holds the Walter Family Distinguished Chair in Internal Medicine in Honor of Albert D. Roberts, M.D. The Center is supported by a $5 million grant from the federal Agency for Healthcare Research and Quality and seeks to assess the benefits and harms of different preventive, diagnostic, therapeutic, and health delivery system interventions to inform decision-making, highlighting comparisons and outcomes that matter to people.


Story Source:

Materials provided by UT Southwestern Medical Center. Note: Content may be edited for style and length.


Journal Reference:

  1. Oanh Kieu Nguyen, Anil N. Makam, Christopher Clark, Song Zhang, Bin Xie, Ferdinand Velasco, Ruben Amarasingham, Ethan A. Halm. Predicting all-cause readmissions using electronic health record data from the entire hospitalization: Model development and comparison. Journal of Hospital Medicine, 2016; DOI: 10.1002/jhm.2568

Cite This Page:

UT Southwestern Medical Center. "Data from full hospital stay not much better at predicting risk for readmission than data from first day, researchers find." ScienceDaily. ScienceDaily, 31 March 2016. <www.sciencedaily.com/releases/2016/03/160331082849.htm>.
UT Southwestern Medical Center. (2016, March 31). Data from full hospital stay not much better at predicting risk for readmission than data from first day, researchers find. ScienceDaily. Retrieved March 28, 2024 from www.sciencedaily.com/releases/2016/03/160331082849.htm
UT Southwestern Medical Center. "Data from full hospital stay not much better at predicting risk for readmission than data from first day, researchers find." ScienceDaily. www.sciencedaily.com/releases/2016/03/160331082849.htm (accessed March 28, 2024).

Explore More

from ScienceDaily

RELATED STORIES