 |
The 130th Annual Meeting of APHA |
|
The 130th Annual Meeting of APHA |
Anindya De, PhD, Emory University, 1520 Clifton Rd., NE, Atlanta, GA 30322-4207, 404-639-4909, and2@cdc.gov, Laura P. Kimble, PhD, RN, School of Nursing, Emory University, 1520 Clifton Rd., NE, Atlanta, GA 30044, and Dyanne D. Affonso, PhD, Faculty of Nursing, University of Toronto, 50 St. George Street, Toronto, ON M5S 3H4, Canada.
Multi-site international studies are necessary to address health related issues on a global scale, however statistical analysis of study findings can be challenging. The presence of nonconstancy of error variance, presence of intra-class correlation within study sites, and non-constancy of slopes of covariates violate the assumptions of commonly used regression analysis and analysis of covariance statistical models. The purpose of this paper is to demonstrate the usefulness of hierarchical linear modeling (HLM) in analyzing data from a longitudinal multi-site international study of post-partum depression. The sample included 800 women who had recently had their first child. The 10 study sites were located in the countries of Italy, Australia, South Korea, Taiwan, Sweden, U.S.A., India, Guyana, and Finland. Depression data were collected at 5 and 11 weeks post-partum using the Beck Depression Inventory (BDI). A three level HLM analysis was carried out. At the within-subject level, observations on an individual were modeled using a linear relationship with the time variable. At the between-subject level, age was used as a covariate in the linear models for coefficients of the model in the previous level. At the site-level, a dichotomous country characteristic was used to specify whether the site is in a country with western culture or not. Results showed that there is a significant negative impact of time and age on the depression score. Western sites also have a negative effect on depression level. There are some significant variations in initial status and rate of change among the subjects and the sites.
Learning Objectives:
Keywords: International Health, Statistics
Presenting author's disclosure statement:
I do not have any significant financial interest/arrangement or affiliation with any organization/institution whose products or services are being discussed in this session.