5141.0: Wednesday, October 24, 2001: 12:30 PM-2:00 PM

Oral Session

Statistical Modeling Applications

Statistical modeling of health data is ubiquitous, but the literature is dominated by a relatively few model forms applied in relatively standard ways. The papers in this session involve statistical modeling techniques that are not yet commonly seen in health science journals, applied to a very diverse collection of public health topics. Methods include Poisson regression analysis of rates with quasi-likelihood dispersion adjustment, zero-inflated Poisson regression, system dynamics modeling, and parallel process latent growth modeling. These are applied, respectively, to landmine incidents, automotive accidents with elderly drivers, court referrals to batterer intervention programs for domestic violence, and depression level among crack-cocaine users.
See individual abstracts for presenting author's disclosure statement.
Learning Objectives: At the conclusion of the session, the participant should be able to: 1. Describe the components of a zero-inflated Poisson model, and data analytic features that call for its use. 2. Discuss the impact of overdispersion on conventional Poisson regression analysis of automotive accidents involving elderly drivers, and how generalized Poisson regression corrects for this problem. 3. Identify complex relationships involving feedback, delays, and non-linearity, and describe one approach to modeling systems with such relationships. 4. Describe the features of a parallel process latent growth model for multiple outcome variables, and its application to studying the trend of depression with changes in cocaine use over time.
Presider(s):Craig D. Turnbull, PhD
12:30 PMAnalyzing landmine incidents via zero-inflated Poisson models
Lawrence H. Moulton, Aldo A. Benini, Charles E. Conley, Shawn Messick
12:50 PMGeneralized Poisson modeling of automobile accidents involving elderly drivers
John T. Wulu, Felix Famoye, Karan P. Singh, Gerald McGwin
1:10 PMEffectiveness of court mandated referrals: A system dynamics model and simulation
Peter S. Hovmand, MSW
1:30 PMLongitudinal study of crack-cocaine use and depression: a parallel process latent growth model
Jichuan Wang, Harvey Siegal, Russel Falck, Robert Carlson
Sponsor:Statistics
Cosponsors:Injury Control and Emergency Health Services
CE Credits:CME, Health Education (CHES), Nursing, Pharmacy, Social Work

The 129th Annual Meeting of APHA