|
Activity Number:
|
210
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Monday, August 7, 2006 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
WNAR
|
| Abstract - #306042 |
|
Title:
|
Adaptive Poisson Modeling of Medication Adherence in HIV-Positive Methadone Patients
|
|
Author(s):
|
Kevin Delucchi*+ and George Knafl and Nancy Haug and James Sorensen
|
|
Companies:
|
University of California, San Francisco and Oregon Health & Science University and University of California, San Francisco and University of California, San Francisco
|
|
Address:
|
401 Parnassus Ave., San Francisco, CA, 94143-0984,
|
|
Keywords:
|
Poisson ; medication adherence ; longitudinal ; small sample ; clustering ; mixture modeling
|
|
Abstract:
|
The use of adaptive Poisson regression models to capture medication adherence levels from electronic pill-bottle monitoring devices is studied. Data are from a randomized trial designed to improve medication adherence in HIV+ methadone patients. Plots display changes over time and treatment effects consistent with the more standard analysis. Predicted values were clustered via an extensive search among methods and results compared to finite mixture modeling. Clustering found the more extreme cases while mixture modeling reflected levels of adherence. The mixture-modeling grouped subjects into low, intermediate, and high adherence classes while the clustering-method generated compatible, finer adherence groups. Results replicate the primary analytic finding. Improvements in the methodology are shown and factors associated with medication adherence presented. Funding: DA009253 AI057043.
|
