What does this research mean for the field?

A random forest machine learning model accurately predicts all-cause mortality risk in patients with concurrent type 2 diabetes mellitus and hypertension with superior discriminative performance compared to other algorithms and logistic regression. Novelty: ClaimNovelty.METHODOLOGICAL. Consensus alignment: ConsensusAlignment.NEUTRAL.

What question did this study set out to answer?

To develop and validate machine learning algorithms for predicting all-cause mortality in adults with concurrent type 2 diabetes mellitus and hypertension.

September 25, 2026JMIR Medical InformaticsOpen Access

Random forest outperforms logistic regression and other models predicting all-cause mortality with ~0.87 AUC.

Q: What is the clinical evidence from this study?

Study design: Cohort. Population: Type 2 diabetes mellitus combined with hypertension (n=2428). Intervention: Random forest model vs. Light gradient boosting machine, decision tree, extreme gradient boosting, and logistic regression. Primary outcome: All-cause mortality prediction (AUC 0.873, 95% CI 0.856-0.891).

View Full Paper

Ask AI

Bookmark

Why the study?

Accurate risk prediction models for all-cause mortality in patients with type 2 diabetes mellitus combined with hypertension remain limited.

Population

2428 adult participants with T2DM and hypertension from NHANES 1999-2018

Comparison

Random forest vs light gradient boosting machine vs decision tree vs extreme gradient boosting vs logistic regression models

Design

National cohort study

Follow-up

Median 6.75 years (IQR 4.20-8.90)

Key result

A random forest model demonstrated superior performance for predicting all-cause mortality compared to other machine learning models and logistic regression (AUC 0.873; 95% CI 0.856-0.891).

Authors

WDWenlong Ding LFLei Fang CFCunming Fang

Discussion

Member takes

Overview

May support risk stratification in diabetes with hypertension; leaves open prospective validation before clinical adoption.

Key Points

To develop and validate machine learning algorithms for predicting all-cause mortality in adults with concurrent type 2 diabetes mellitus and hypertension.
Analyzed data from the National Health and Nutrition Examination Survey (1999–2018) linked to mortality records through December 31, 2019, including 2,428 adults aged ≥20 years with concurrent type 2 diabetes mellitus and hypertension followed for a median of 6.75 years.
Developed and compared five predictive models—random forest, light gradient boosting machine, decision tree, extreme gradient boosting, and logistic regression—evaluating discriminatory performance via area under the receiver operating characteristic curve (AUC), calibration plots, and decision curves.
Over the median follow-up period of 6.75 years, 719 of the 2,428 participants (29.6%) died from all causes.
The random forest model showed the highest discriminative ability with an AUC of 0.873 (95% CI, 0.856–0.891), outperforming extreme gradient boosting (AUC 0.792), light gradient boosting machine (AUC 0.785), logistic regression (AUC 0.783), and decision tree (AUC 0.732).
Age, race, chronic kidney disease, body mass index, and blood urea nitrogen emerged as the most influential predictive features for mortality risk.

Study Design

Type

Cohort (n=2,428)

Multicenter

Yes

Structured PICO

Population

2,428 adults aged ≥20 years with concurrent type 2 diabetes mellitus and hypertension, followed for a median of 6.75 years.

Exposure

Machine learning-based prediction models (random forest, light gradient boosting machine, decision tree, extreme gradient boosting, and logistic regression)

Outcome

All-cause mortalityhard clinical

Main Result

Effect estimate: AUC 0.873 (95% CI 0.856-0.891)

A random forest machine learning model accurately predicts all-cause mortality in patients with concurrent type 2 diabetes and hypertension, outperforming other algorithms.

Cite This Study

Ding et al. (2026) conducted a cohort in Type 2 diabetes mellitus combined with hypertension (n=2,428). Random forest model vs. Light gradient boosting machine, decision tree, extreme gradient boosting, and logistic regression was evaluated on All-cause mortality prediction (AUC 0.873, 95% CI 0.856-0.891). A random forest model demonstrated superior performance for predicting all-cause mortality compared to other machine learning models and logistic regression (AUC 0.873; 95% CI 0.856-0.891).

synapsesocial.com/papers/6ab60f69406bf401c1467f56 https://doi.org/10.2196/85557

View Full Paper

Ask AI

Bookmark

September 25, 2026JMIR Medical InformaticsOpen Access

Random forest outperforms logistic regression and other models predicting all-cause mortality with ~0.87 AUC.

View Full Paper

Ask AI

Bookmark

Why the study?

Accurate risk prediction models for all-cause mortality in patients with type 2 diabetes mellitus combined with hypertension remain limited.

Population

2428 adult participants with T2DM and hypertension from NHANES 1999-2018

Comparison

Random forest vs light gradient boosting machine vs decision tree vs extreme gradient boosting vs logistic regression models

Design

National cohort study

Follow-up

Median 6.75 years (IQR 4.20-8.90)

Key result

A random forest model demonstrated superior performance for predicting all-cause mortality compared to other machine learning models and logistic regression (AUC 0.873; 95% CI 0.856-0.891).

Authors

WDWenlong Ding LFLei Fang CFCunming Fang

Discussion

Member takes

Overview

May support risk stratification in diabetes with hypertension; leaves open prospective validation before clinical adoption.

Key Points

To develop and validate machine learning algorithms for predicting all-cause mortality in adults with concurrent type 2 diabetes mellitus and hypertension.
Analyzed data from the National Health and Nutrition Examination Survey (1999–2018) linked to mortality records through December 31, 2019, including 2,428 adults aged ≥20 years with concurrent type 2 diabetes mellitus and hypertension followed for a median of 6.75 years.
Developed and compared five predictive models—random forest, light gradient boosting machine, decision tree, extreme gradient boosting, and logistic regression—evaluating discriminatory performance via area under the receiver operating characteristic curve (AUC), calibration plots, and decision curves.
Over the median follow-up period of 6.75 years, 719 of the 2,428 participants (29.6%) died from all causes.
The random forest model showed the highest discriminative ability with an AUC of 0.873 (95% CI, 0.856–0.891), outperforming extreme gradient boosting (AUC 0.792), light gradient boosting machine (AUC 0.785), logistic regression (AUC 0.783), and decision tree (AUC 0.732).
Age, race, chronic kidney disease, body mass index, and blood urea nitrogen emerged as the most influential predictive features for mortality risk.

Study Design

Type

Cohort (n=2,428)

Multicenter

Yes

Structured PICO

Population

2,428 adults aged ≥20 years with concurrent type 2 diabetes mellitus and hypertension, followed for a median of 6.75 years.

Exposure

Machine learning-based prediction models (random forest, light gradient boosting machine, decision tree, extreme gradient boosting, and logistic regression)

Outcome

All-cause mortalityhard clinical

Main Result

Effect estimate: AUC 0.873 (95% CI 0.856-0.891)

A random forest machine learning model accurately predicts all-cause mortality in patients with concurrent type 2 diabetes and hypertension, outperforming other algorithms.

Cite This Study

synapsesocial.com/papers/6ab60f69406bf401c1467f56 https://doi.org/10.2196/85557

Also Consider

Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context:

View Full Paper

Ask AI

Bookmark