Annual Meeting Recordings are now available for purchase
300317
Bayesian Longitudinal Actor-Partner Interdependence Models for HIV-Impacted Family Interventions
Wednesday, November 19, 2014
: 8:30 AM - 8:50 AM
Li-Jung Liang, PhD
,
Department of Medicine Statistics Core, University of California, Los Angeles, Los Angeles, CA
Li Li, PhD
,
Department of Epidemiology- School of Public Health, University of California, Los Angeles
The concept of multilevel intervention assumes that there are complex interconnecting relations between members of the family. One consequence of this interdependence is that the mutual impact that the family members have on each other could affect their outcomes. Complex interdependencies lead to statistical challenges: (1) longitudinal data, including both outcomes and predictors of interest, collected from multiple members of a family are likely to be correlated, and (2) one person’s predictor may influence not only that person’s outcome, but also the outcomes of their family members. It is important to develop analytical methods to properly analyze such complex data. We develop several Bayesian longitudinal actor-partner interdependence models (APIMs), which are multilevel regression models with predictors of actor and partner, to address these challenges and estimate the intervention effects of interest. This approach allows us to properly account for the multiple levels of correlations within families and also estimate the actor and partner effects using functions of shared random effects. The proposed approach is applied to a study on HIV-impacted families in China.
Learning Areas:
Biostatistics, economics
Public health or related research
Social and behavioral sciences
Learning Objectives:
Analyze multiple types of longitudinal outcomes from multilevel family intervention studies using the Bayesian APIM approach
Keyword(s): Biostatistics, HIV Interventions
Presenting author's disclosure statement:Qualified on the content I am responsible for because: I am an Assistant Professor at the UCLA School of Medicine Statistics Core, and I am a Biostatistician. I have experience applying Bayesian statistics, multilevel modeling, longitudinal and repeated measures methodologies, and survival models to a broad range of health-related research problems, including HIV impacted-family intervention, substance abuse, health services research, and public health studies.
Any relevant financial relationships? No
I agree to comply with the American Public Health Association Conflict of Interest and Commercial Support Guidelines,
and to disclose to the participants any off-label or experimental uses of a commercial product or service discussed
in my presentation.
