Synapse
⌘+K
Synapse
PulseExploreClubsResearchersJournals
Instagram
HomeClubsExplore
March 4, 2026Medical Care

ANN models perform comparably to logistic regression in predicting costs, readmissions, and preventable hospitalizations.

View Full Paper
Ask AI
Bookmark
Share

Why the study?

The relative benefits of artificial neural networks compared with traditional methods in health services research remain unclear.

Do Artificial Neural Network models improve the prediction of high Medicare costs, 30-day readmissions, and preventable hospitalizations compared to logistic regression in Medicare beneficiaries aged 65 and older?

Population

254,748 Medicare beneficiaries aged 65 and older

Comparison

Logistic regression vs artificial neural network models

Design

Cross-sectional study

Key result

Logistic regression and ANN models showed comparable predictive performance for high Medicare costs, 30-day readmissions, and preventable hospitalizations among older adults.

Authors

JCJie ChenSJSeyeon JangMWMin Qi Wang

Discussion

Loading...

Member takes

Overview

ANN offers no advantage over logistic regression in older Medicare patients; reinforces traditional models as sufficient for claims-based risk prediction.

Key Points

  • This research examines health care utilization and costs among older adults, comparing logistic regression and ANN models.
  • Conducted a cross-sectional analysis using Medicare claims and CAHPS surveys from 2018 to 2022.
  • Included 254,748 community-dwelling Medicare beneficiaries aged 65 and older.
  • Assessed model performance using metrics like AUC, sensitivity, specificity, PPV, NPV, and Brier scores.
  • Chronic conditions such as heart disease and depression predicted higher Medicare costs.
  • Factors like poor self-rated health and lower educational attainment correlated with increased readmissions.
  • Both logistic regression and ANN models showed comparable performance across key outcomes.

Structured PICO

Do Artificial Neural Network models improve the prediction of high Medicare costs, 30-day readmissions, and preventable hospitalizations compared to logistic regression in Medicare beneficiaries aged 65 and older?

P
Population
254,748 community-dwelling Medicare beneficiaries aged 65 and older
I
Intervention
Artificial Neural Network (ANN) models
C
Comparator
Logistic regression models
O
Outcome
Prediction of high Medicare costs (top 25%), 30-day readmissions, and preventable hospitalizations (PQIs)

Artificial neural networks and traditional logistic regression models demonstrate comparable performance in predicting healthcare utilization and costs among older adults using structured Medicare data.

Cite This Study

Chen et al. (2026) studied this question. Logistic regression and ANN models showed comparable predictive performance for high Medicare costs, 30-day readmissions, and preventable hospitalizations among older adults.

synapsesocial.com/papers/69a7ccc3d48f933b5eed8955https://doi.org/10.1097/mlr.0000000000002305
View Full Paper
Ask AI
Bookmark
Share