Multi-level zero-inflated poisson regression modelling of correlated count data with excess zeros

Lee, Andy H.; Kui Wang; Scott, Jane A.; Yau, Kelvin K. W.; McLachlan, Geoffrey J.; Wang, Kui
February 2006
Statistical Methods in Medical Research;Feb2006, Vol. 15 Issue 1, p47
Academic Journal
journal article
Count data with excess zeros relative to a Poisson distribution are common in many biomedical applications. A popular approach to the analysis of such data is to use a zero-inflated Poisson (ZIP) regression model. Often, because of the hierarchical study design or the data collection procedure, zero-inflation and lack of independence may occur simultaneously, which render the standard ZIP model inadequate. To account for the preponderance of zero counts and the inherent correlation of observations, a class of multi-level ZIP regression model with random effects is presented. Model fitting is facilitated using an expectation-maximization algorithm, whereas variance components are estimated via residual maximum likelihood estimating equations. A score test for zero-inflation is also presented. The multi-level ZIP model is then generalized to cope with a more complex correlation structure. Application to the analysis of correlated count data from a longitudinal infant feeding study illustrates the usefulness of the approach.


Related Articles

  • Arterial blood gas and pulse oximetry in initial management of patients with community-acquired pneumonia. Levin, Kenneth P.; Hanusa, Barbara H.; Rotondi, Armando; Singer, Daniel E.; Coley, Christopher M.; Marrie, Thomas J.; Kapoor, Wishwa N.; Fine, Michael J.; Levin, K P; Hanusa, B H; Rotondi, A; Singer, D E; Coley, C M; Marrie, T J; Kapoor, W N; Fine, M J // JGIM: Journal of General Internal Medicine;Sep2001, Vol. 16 Issue 9, p590 

    Objective: To identify the factors associated with the use of arterial blood gas (ABG) and pulse oximetry (PO) in the initial management of patients with community-acquired pneumonia (CAP) and arterial hypoxemia at presentation.Participants: A total of 944 outpatients...

  • Matching conditional and marginal shapes in binary random intercept models using a bridge distribution function. Zengri Wang; Thomas A. Louis // Biometrika;Dec2003, Vol. 90 Issue 4, p765 

    Random effects logistic regression models are often used to model clustered binary response data. Regression parameters in these models have a conditional, subject-specific interpretation in that they quantify regression effects for each cluster. Very often, the logistic functional shape...

  • Comparison of subject-specific and population averaged models for count data from cluster-unit intervention trials. Young, Mary L.; Preisser, John S.; Qaqish, Bahjat F.; Wolfson, Mark // Statistical Methods in Medical Research;Apr2007, Vol. 16 Issue 2, p167 

    Maximum likelihood estimation techniques for subject-specific (SS) generalized linear mixed models and generalized estimating equations for marginal or population-averaged (PA) models are often used for the analysis of cluster-unit intervention trials. Although both classes of procedures account...

  • Hospital readmission in general medicine patients: a prediction model. Hasan, Omar; Meltzer, David O.; Shaykevich, Shimon A.; Bell, Chaim M.; Kaboli, Peter J.; Auerbach, Andrew D.; Wetterneck, Tosha B.; Arora, Vineet M.; Zhang, James; Schnipper, Jeffrey L. // JGIM: Journal of General Internal Medicine;Mar2010, Vol. 25 Issue 3, p211 

    Background: Previous studies of hospital readmission have focused on specific conditions or populations and generated complex prediction models.Objective: To identify predictors of early hospital readmission in a diverse patient population and derive and validate a...

  • Results of the Medications at Transitions and Clinical Handoffs (MATCH) study: an analysis of medication reconciliation errors and risk factors at hospital admission. Gleason, Kristine M.; McDaniel, Molly R.; Feinglass, Joseph; Baker, David W.; Lindquist, Lee; Liss, David; Noskin, Gary A. // JGIM: Journal of General Internal Medicine;May2010, Vol. 25 Issue 5, p441 

    Background: This study was designed to determine risk factors and potential harm associated with medication errors at hospital admission.Methods: Study pharmacist and hospital-physician medication histories were compared with medication orders to identify unexplained...

  • Hospital utilization and costs in a cohort of injection drug users. Palepu, Anita; Tyndall, Mark W.; Leon, Hector; Muller, Jennifer; O'Shaughnessy, Michael V.; Schechter, Martin T.; Anis, Aslam H.; Palepu, A; Tyndall, M W; Leon, H; Muller, J; O'Shaughnessy, M V; Schechter, M T; Anis, A H // CMAJ: Canadian Medical Association Journal;8/21/2001, Vol. 165 Issue 4, p415 

    Background: Many injection drug users (IDUs) seek care at emergency departments and some require hospital admission because of late presentation in the course of their illness. We determined the predictors of frequent emergency department visits and hospital admissions among...

  • Semantic memory and depressive symptoms in patients with subjective cognitive decline, mild cognitive impairment, and Alzheimer's disease. Lehrner, J.; Coutinho, G.; Mattos, P.; Moser, D.; Pflüger, M.; Gleiss, A.; Auff, E.; Dal-Bianco, P.; Pusswald, G.; Stögmann, E. // International Psychogeriatrics;Jul2017, Vol. 29 Issue 7, p1123 

    Background: Semantic memory may be impaired in clinically recognized states of cognitive impairment. We investigated the relationship between semantic memory and depressive symptoms (DS) in patients with cognitive impairment.Methods: 323 cognitively healthy controls and...

  • Alcohol consumption and major depression: findings from a follow-up study. Wang, Jian Li; Patten, Scott B; Wang, J; Patten, S B // Canadian Journal of Psychiatry;Sep2001, Vol. 46 Issue 7, p632 

    Objective: To investigate whether alcohol consumption predicts major depressive disorder episodes (MDEs) in the general population.Method: The respondents without depression (n = 12,290) in the longitudinal cohort of the Canadian National Population Health Survey (NPHS)...

  • Hierarchical Regression for Analyses of Multiple Outcomes. Richardson, David B.; Hamra, Ghassan B.; MacLehose, Richard F.; Cole, Stephen R.; Haitao Chu // American Journal of Epidemiology;9/1/2015, Vol. 182 Issue 5, p459 

    In cohort mortality studies, there often is interest in associations between an exposure of primary interest and mortality due to a range of different causes. A standard approach to such analyses involves fitting a separate regression model for each type of outcome. However, the statistical...


Read the Article


Sorry, but this item is not currently available from your library.

Try another library?
Sign out of this library

Other Topics