Algorithmic Convergence Across Machine Learning Models Identifies Robust Strategic Priorities for Improving Patient Experience.

Authors

Richard H Savel, Payam Benson, Carmen Collins, Jill Fennimore, Dwight McBee, Kwaku Gyekye, Ije Akunyili, Marc Milano, Ruric Andy Anderson

Published in

American journal of medical quality : the official journal of the American College of Medical Quality. Sep 07, 2026. Epub Sep 07, 2026.

Abstract

Improved patient experience is associated with enhanced clinical outcomes. The Hospital Consumer Assessment of Healthcare Providers and Systems Overall Rating of the Hospital (ORH) can be considered a quality metric for health systems. We hypothesized that machine learning models could identify actionable predictors of a less-than-top-box ORH score across a multihospital system and sought to determine whether these predictors were consistent across model architectures and hospital size.
We analyzed 86 351 Hospital Consumer Assessment of Healthcare Providers and Systems surveys from 12 hospitals within a New Jersey health system (January 2023 to May 2026). Four machine learning models (logistic regression, decision tree, random forest, and XGBoost) were developed to predict a less-than-top-box ORH score. Feature importance was compared across the 3 highest-performing models, and SHapley Additive exPlanations (SHAP) analysis was performed on the XGBoost model. All analyses were repeated separately across hospitals stratified by size (large, medium, and small).
XGBoost, logistic regression, and random forest demonstrated excellent, comparable discrimination (area under the curve 0.871-0.883); the decision tree model underperformed and was excluded. Feature importance and SHAP analysis converged on the same 3 predictor domains: nursing communication, physician communication, and hospital environment, which jointly accounted for 67% of total SHAP importance. This 3-domain pattern remained stable across hospital-size strata, with only modest shifts in relative ranking.
Across multiple machine learning architectures, analytic methods, and hospital sizes, nursing communication, physician communication, and hospital environment consistently emerged as the primary, modifiable drivers of overall hospital rating. These convergent, generalizable findings identify clear strategic priorities for health systems seeking to improve patient experience.

PMID:
42696603
Bibliographic data and abstract were imported from PubMed on 05 Sep 2026.

Read full publication at:
Please sign in to see all details.

Stats

1-terrible, 9-excellent. How would you rate this publication? Sign in in to submit your rating.

Post a comment

You need to be signed in to post comments. You can sign in here.

Comments

There are no comments yet.