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

Using hierarchical modeling to measure ICU quality.

Laurent G Glance1, Andrew W Dick2, Turner M Osler3

  • 1University of Rochester, School of Medicine and Dentistry, 601 Elmwood Avenue, Rochester, NY, 14642, USA. Laurent_Glance@urmc.rochester.edu.

Intensive Care Medicine
|October 10, 2003
PubMed
Summary

Related Concept Videos

You might also read

Related Articles

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

Sort by
Same author

Beyond the Cuff: Rethinking How We Monitor and Manage Intraoperative Hypotension.

Anesthesiology·2026
Same author

A Novel Method to Apply Race and Ethnicity Observation for Nursing Home Residents Using Multiple Medicare Administrative Datasets.

Medical care·2026
Same author

Hospital-Level Variation in Transcatheter vs Surgical Aortic Valve Replacement Among Patients Younger Than 65 Years.

The Annals of thoracic surgery·2026
Same author

In Response.

Anesthesia and analgesia·2026
Same author

Trends and Disparities in Post-acute Care Utilization After Hospitalization for Sepsis in the United States: A Systematic Review.

Medical care·2026
Same author

Hospital Employees View Patient Safety Culture Differently According to Their Role.

Journal of patient safety·2025

Hierarchical and conventional models for Intensive Care Unit (ICU) quality outlier identification showed strong agreement. This confirms the reliability of both methods for assessing hospital performance in critical care settings.

Area of Science:

  • Critical Care Medicine
  • Health Services Research
  • Biostatistics

Background:

  • Identifying Intensive Care Unit (ICU) quality outliers is crucial for improving patient outcomes.
  • Conventional logistic regression and hierarchical modeling are statistical approaches used for this purpose.
  • Evaluating the agreement between these modeling techniques is essential for validating their use in multi-institutional databases.

Purpose of the Study:

  • To compare the performance of hierarchical modeling versus conventional logistic regression in identifying ICU quality outliers.
  • To assess the agreement between these two statistical approaches within a large multi-institutional dataset.

Main Methods:

  • Retrospective analysis of the Project IMPACT database (1997-1999).
  • Inclusion of 40,435 adult patients from 55 ICUs meeting criteria for the Simplified Acute Physiology Score II (SAPS II).

Related Experiment Videos

  • Customization of SAPS II using both conventional logistic regression and hierarchical (random coefficients) models to identify ICU outliers based on observed vs. expected mortality ratios.

    • Main Results:

    • Both hierarchical and non-hierarchical models demonstrated excellent discrimination (C-statistics of 0.870 and 0.865, respectively) and calibration (Hosmer-Lemeshow statistics).
    • The hierarchical model accounted for between-hospital variability.
    • Analysis using the kappa statistic revealed almost perfect agreement between the two models in identifying high-performance and low-performance ICU outliers.

    Conclusions:

    • Statistical models derived from both non-hierarchical and hierarchical approaches demonstrate excellent agreement in identifying ICU quality outliers.
    • These findings support the use of both methods for assessing and comparing ICU performance across multiple institutions.