Influence diagnostics for two-component Poisson mixture regression models: applications in public health.

Abstract

In many medical and health applications, Poisson mixture regression models are commonly used to analyse heterogeneous count data. Motivated by two data sets drawn from public health studies, influence diagnostics are proposed for assessing the sensitivity of the fitted two-component Poisson mixture regression models. Under various perturbations of the observed data or model assumptions, influence assessments based on the local influence approach are developed for detecting clusters and/or individual observations that impact on the estimation of model parameters. Results from studies on recurrent urinary tract infections and maternity length of stay illustrate the usefulness of the influence diagnostics.

• Authors: Liming Xiang, Kelvin K W Yau, Andy H Lee, Wing K Fung
• Journal: Statistics in medicine (Stat Med) Vol. 24 Issue 19 Pg. 3053-71 (Oct 15 2005) ISSN: 0277-6715 [Print] England
• PMID: 16149127 (Publication Type: Journal Article, Research Support, Non-U.S. Gov't)
• Copyright: Copyright (c) 2005 John Wiley & Sons, Ltd.

Topics

• Cluster Analysis
• Cohort Studies
• Data Interpretation, Statistical
• Delivery, Obstetric
• Female
• Health Services
• Humans
• Length of Stay
• Poisson Distribution
• Pregnancy
• Public Health
• Regression Analysis
• Retrospective Studies
• Urinary Tract Infections (epidemiology)
• Western Australia (epidemiology)