Abstract:
|
The HIV care cascade is a conceptual model used to describe the steps that patients with HIV must take to achieve viral suppression. Policy makers and clinical researchers are interested in rates of progression through each stage of the cascade, and in individual characteristics that predict linkage or retention. Analytic methods typically focus on the targeted aspects of the cascade using a single cohort, or on simulation of compartment models that rely on heterogeneous inputs from descriptive statistics or published studies. To capture the advantages of using cohort data with the complexity of a compartmental model, we have developed a state-space modeling (SSM) approach. The SSM is parameterized in terms of probability of transition between states, and allows regression-based estimation to assess covariate effects on retention, disengagement and re-entry into care, side-door entry, and mortality. The modeling approach is demonstrated by analyzing longitudinal data on 48,947 patients enrolled in AMPATH (Kenya) between 2008 and 2012.
|
ASA Meetings Department
732 North Washington Street, Alexandria, VA 22314
(703) 684-1221 • meetings@amstat.org
Copyright © American Statistical Association.