Jove
Visualize
Contact Us
JoVE
x logofacebook logolinkedin logoyoutube logo
ABOUT JoVE
OverviewLeadershipBlogJoVE Help Center
AUTHORS
Publishing ProcessEditorial BoardScope & PoliciesPeer ReviewFAQSubmit
LIBRARIANS
TestimonialsSubscriptionsAccessResourcesLibrary Advisory BoardFAQ
RESEARCH
JoVE JournalMethods CollectionsJoVE Encyclopedia of ExperimentsArchive
EDUCATION
JoVE CoreJoVE BusinessJoVE Science EducationJoVE Lab ManualFaculty Resource CenterFaculty Site
Terms & Conditions of Use
Privacy Policy
Policies

Related Experiment Videos

Improving models to predict care utilization using machine learning: a retrospective observational study.

Christopher Kitchen1, Talan Zhang1, Klaus Lemke1

  • 1Johns Hopkins University, 2024 E Monument St, Baltimore, US.

JMIR Formative Research
|June 2, 2026
PubMed
Summary
This summary is machine-generated.

Related Concept Videos

Methods of Documentation VI: Case Management Model01:15

Methods of Documentation VI: Case Management Model

The case management model is a multidisciplinary approach that involves healthcare professionals from diverse disciplines, such as physicians, nurses, therapists, social workers, and pharmacists, working collaboratively to address the various needs of patients. Each healthcare professional brings unique expertise and perspectives, contributing to a more comprehensive understanding of the patient's condition and tailoring treatment plans accordingly.
For example, a patient with a chronic illness...

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by
Same author

An examination of the availability and characteristics of social needs data in the electronic health records: a path to social data harmonization and standardization at Johns Hopkins medicine.

JAMIA open·2026
Same author

Improving Models to Predict Care Utilization Using Machine Learning: Retrospective Observational Study.

JMIR formative research·2026
Same author

Charting the future of ACMI: a report from the 2025 ACMI symposium.

Journal of the American Medical Informatics Association : JAMIA·2026
Same author

Identifying Clinical Diagnostic Trajectories Associated With Suicide Death Using Temporal Sequence Mining of Linked Claims and Mortality Data.

medRxiv : the preprint server for health sciences·2026
Same author

Conceptualization and development of the Maryland suicide data warehouse to improve mortality prediction.

JAMIA open·2026
Same author

Evaluating the role of cost and fairness in machine learning models of suicide risk.

Scientific reports·2026
Same journal

Effects of Virtual Reality on Postoperative Pain Management Following Minimally Invasive Gynecologic Surgery: Randomized Controlled Trial.

JMIR formative research·2026
Same journal

Prediction of Clinically Significant Depressive Symptoms at 2-Year Follow-Up in Older Adults: Machine Learning Study Using the English Longitudinal Study of Ageing.

JMIR formative research·2026
Same journal

Awareness, Educational Needs, and Curriculum Preferences Regarding AI and Medical Big Data Education Among Clinical Medicine Undergraduates: Cross-Sectional Survey Study.

JMIR formative research·2026
Same journal

Stakeholder Experiences With the Pneumococcal Conjugate Vaccine Chatbot as a Complementary Capacity-Building Tool for Frontline Health Workers in India: Qualitative Study.

JMIR formative research·2026
Same journal

Acceptability and Perceived Usefulness of a Digital Gambling Harm Minimisation Tool: A Cross-Sectional Study.

JMIR formative research·2026
Same journal

Knowledge Graphs Based on Meta-Analysis Papers Improve the Quality of Case Formulation: Mixed Methods Design.

JMIR formative research·2026
See all related articles

Machine learning models show slight improvements over traditional logistic regression for predicting healthcare costs and hospitalizations. While XGBoost offers enhanced performance, logistic regression remains a strong choice for interpretability and efficiency in clinical risk estimation.

Area of Science:

  • Healthcare analytics and predictive modeling.
  • Application of artificial intelligence and machine learning in clinical settings.

Background:

  • Artificial intelligence (AI) and machine learning (ML) are increasingly used in healthcare for risk estimation.
  • The Johns Hopkins Adjusted Clinical Groups (ACG) system is a widely adopted method for categorizing clinical risk factors.
  • This study evaluates novel statistical approaches against established systems like ACG.

Purpose of the Study:

  • To assess the performance of the ACG system using machine learning algorithms.
  • To compare ML model performance against traditional logistic regression for predicting hospitalization and cost.
  • To determine if ML can enhance ACG-related workflows through improved performance metrics.

Main Methods:

  • Retrospective observational study using a cross-validation framework.

Related Experiment Videos

  • Modeled all-cause hospitalization and elevated total cost (top 5th percentile).
  • Tuned hyperparameters for XGBoost, random forest, and elastic net using AUROC and F1 optimization via grid search.

    • Main Results:

    • XGBoost demonstrated statistically significant, though small, improvements in cross-validated AUROC and F1 scores over logistic regression.
    • Logistic regression achieved an average ROC of 0.886 (cost) and 0.841 (hospitalization).
    • XGBoost achieved higher F1 scores for cost (0.411) compared to logistic regression (0.367), while logistic regression performed better for hospitalization (0.341 vs. 0.328).

    Conclusions:

    • Logistic regression is highly suitable for clinical risk prediction tasks prioritizing efficiency and interpretability.
    • ML models offer diverse performance trade-offs, with XGBoost showing potential for specific use cases like cost prediction.
    • Health systems must align model selection with specific needs and performance expectations, considering factors like class imbalance.