August 2, 2021PLoS ONEOpen Access

The best machine learning model outperformed the TIMI risk score for predicting in-hospital (AUC 0.88 vs 0.81), 30-day (AUC 0.90 vs 0.80), and 1-year (AUC 0.84 vs 0.76) mortality (p < 0.0001 for all).

View Full Paper

Ask AI

Bookmark

Why the study?

Conventional risk scores for predicting short- and long-term mortality after STEMI are often not population specific.

Does a machine learning algorithm improve short- and long-term mortality prediction compared to the TIMI risk score in Asian patients with STEMI?

Population

Multi-ethnic Asian STEMI patients in Malaysia (6299 in-hospital, 3130 30-day, 2939 1-year)

Comparison

Machine learning algorithms vs TIMI risk score

Design

Registry-based prediction model development study

Follow-up

Up to 1-year

Authors

SMSorayya Malek KIKhairul Shafiq Ibrahim RSRaja Ezman Raja Shariff

Discussion

Member takes

Overview

ML models may enhance mortality risk stratification in Asian STEMI patients; leaves open whether they outperform TIMI prospectively or alter management.

Structured PICO

Does a machine learning algorithm improve short- and long-term mortality prediction compared to the TIMI risk score in Asian patients with STEMI?

Population

6,299 Asian patients with ST-elevation myocardial infarction (STEMI) from the National Cardiovascular Disease Database for Malaysia registry (6,299 for in-hospital, 3,130 for 30-days, and 2,939 for 1-year model development).

Intervention

Machine learning algorithms for mortality prediction using up to 50 clinical variables.

Comparator

Conventional Thrombolysis in Myocardial Infarction (TIMI) risk score.

Outcome

In-hospital, 30-day, and 1-year mortality prediction (assessed by area under the receiver operating characteristic curve [AUC]).hard clinical

Machine learning algorithms significantly improve short- and long-term mortality prediction compared to the conventional TIMI risk score in Asian patients with STEMI.

Cite This Study

Malek et al. (2021) studied this question.

synapsesocial.com/papers/6a7d6974c66da78fab8b36a3 https://doi.org/10.1371/journal.pone.0254894

View Full Paper

Ask AI

Bookmark

August 2, 2021PLoS ONEOpen Access

The best machine learning model outperformed the TIMI risk score for predicting in-hospital (AUC 0.88 vs 0.81), 30-day (AUC 0.90 vs 0.80), and 1-year (AUC 0.84 vs 0.76) mortality (p < 0.0001 for all).

View Full Paper

Ask AI

Bookmark

Why the study?

Conventional risk scores for predicting short- and long-term mortality after STEMI are often not population specific.

Does a machine learning algorithm improve short- and long-term mortality prediction compared to the TIMI risk score in Asian patients with STEMI?

Population

Multi-ethnic Asian STEMI patients in Malaysia (6299 in-hospital, 3130 30-day, 2939 1-year)

Comparison

Machine learning algorithms vs TIMI risk score

Design

Registry-based prediction model development study

Follow-up

Up to 1-year

Authors

SMSorayya Malek KIKhairul Shafiq Ibrahim RSRaja Ezman Raja Shariff

Discussion

Member takes

Overview

ML models may enhance mortality risk stratification in Asian STEMI patients; leaves open whether they outperform TIMI prospectively or alter management.

Structured PICO

Does a machine learning algorithm improve short- and long-term mortality prediction compared to the TIMI risk score in Asian patients with STEMI?

Population

Intervention

Machine learning algorithms for mortality prediction using up to 50 clinical variables.

Comparator

Conventional Thrombolysis in Myocardial Infarction (TIMI) risk score.

Outcome

In-hospital, 30-day, and 1-year mortality prediction (assessed by area under the receiver operating characteristic curve [AUC]).hard clinical

Machine learning algorithms significantly improve short- and long-term mortality prediction compared to the conventional TIMI risk score in Asian patients with STEMI.

Cite This Study

Malek et al. (2021) studied this question.

synapsesocial.com/papers/6a7d6974c66da78fab8b36a3 https://doi.org/10.1371/journal.pone.0254894

View Full Paper

Ask AI

Bookmark