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

Comparing alternative models: log vs Cox proportional hazard?

Anirban Basu1, Willard G Manning, John Mullahy

  • 1Harris School of Public Policy, The University of Chicago, Chicago, IL 60637, USA.

Health Economics
|August 24, 2004
PubMed
Summary
This summary is machine-generated.

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

Biomonitoring of nutritional and toxic elements in raw and cooked rice in the Health Effects of Arsenic Longitudinal Study, Bangladesh.

Ecotoxicology and environmental safety·2026
Same author

Influenza Antiviral Prescriptions in the US Inpatient and Outpatient Settings During the 2015-2022 Influenza Seasons.

Influenza and other respiratory viruses·2026
Same author

PERsonalised Medicine for Intensification of Treatment (PERMIT) in type 2 diabetes mellitus: a target trial emulation from routine data.

Health technology assessment (Winchester, England)·2026
Same author

Interplay Between the Gut-Microbiota-Brain Axis and the Immune System in Viral Neuroinvasion and CNS Pathogenesis.

Comprehensive Physiology·2026
Same author

From sensing to shaping: microglial responses in the pathogenesis of viral encephalitis.

Current opinion in immunology·2026
Same author

Cardiac Resynchronization Therapy: A Review.

JAMA·2026

Health economists should consider gamma regression with a log link for skewed health expenditure data. This method shows more robustness than OLS or Cox proportional hazards regression in econometric analyses.

Area of Science:

  • Econometrics
  • Health Economics
  • Statistical Modeling

Background:

  • Health economists frequently use log-linear models (OLS, GLM) for skewed health outcome data like expenditures and length of stay.
  • Cox proportional hazards regression is emerging as a less parametric alternative, even without censoring.
  • The econometric performance of these models with skewed data requires further examination.

Purpose of the Study:

  • To assess the econometric bias of alternative estimators for right-skewed health outcome data.
  • To compare the performance of Cox proportional hazards regression against traditional models.
  • To identify the most robust estimation method for skewed health expenditure and length of stay data.

Main Methods:

  • Examined the econometric performance of various estimators under different data-generating mechanisms.

Related Experiment Videos

  • Compared Ordinary Least Squares (OLS) on ln(y), generalized linear models (GLM), and Cox proportional hazards regression.
  • Evaluated the bias of these models when applied to right-skewed data.

    • Main Results:

    • No single alternative estimator performed best across all examined conditions.
    • Gamma regression with a log link demonstrated greater robustness to varying data-generating mechanisms compared to OLS on ln(y) and Cox regression.
    • Consistent estimation of E(y|x) using Cox regression necessitates adherence to the proportional hazards assumption.

    Conclusions:

    • Gamma regression with a log link is a robust choice for analyzing skewed health expenditure and length of stay data.
    • The proportional hazards assumption is critical for the validity of Cox regression in estimating conditional means.
    • Careful model selection is essential for accurate econometric analysis of skewed health outcome data.