Synapse
⌘+K
Synapse
PulseExploreClubsResearchersJournals
Instagram
HomeClubsExplore
April 19, 2021Journal of Medical Internet ResearchOpen Access

Machine learning algorithms achieved high predictive performance for 1-year in-hospital mortality (AUCs 0.92-1.00), use of positive inotropic medication (AUCs 0.85-0.96), and 1-year readmission rates (AUCs 0.63-0.96).

View Full Paper
Ask AI
Bookmark
Share

Why the study?

With the increasing worldwide prevalence of cardiovascular diseases, early prediction and accurate assessment of heart failure risk are crucial to meet clinical demand.

Can machine learning models using electronic health records accurately predict 1-year in-hospital mortality, use of positive inotropic agents, and 1-year readmission in patients hospitalized with heart failure?

Population

13,602 hospitalized patients with newly diagnosed HF at a single center in China

Comparison

Machine learning models predicting outcomes using 79 variables vs each other

Design

Single-center retrospective study

Follow-up

1 year

Authors

HLHaichen LvXYXiaolei YangFWFeng Wang

Discussion

Loading...

Member takes

Overview

May support EHR-based risk stratification in HF; hypothesis-generating, requires prospective validation before practice change.

Structured PICO

Can machine learning models using electronic health records accurately predict 1-year in-hospital mortality, use of positive inotropic agents, and 1-year readmission in patients hospitalized with heart failure?

P
Population
13,602 hospitalized patients with newly diagnosed heart failure (aged >18 years) at a single center in China between December 2010 and August 2018.
I
Intervention
Machine learning predictive models (logistic regression, support vector machine, artificial neural network, random forest, and extreme gradient boosting) using 79 electronic health record variables.
O
Outcome
1-year in-hospital mortality, use of positive inotropic agents, and 1-year all-cause readmission ratehard clinical

Machine learning models utilizing routine electronic health record data can accurately predict mortality, readmission, and inotrope use in hospitalized heart failure patients, potentially aiding clinical risk stratification.

Cite This Study

Lv et al. (2021) studied this question.

synapsesocial.com/papers/6a1a327bf71204b07e3c12f9https://doi.org/10.2196/24996
View Full Paper
Ask AI
Bookmark
Share