Losses to Follow-up from an Antiretroviral Therapy (ART) Program in the Zomba District of Malawi

Losses to Follow-Up from Antiretroviral Therapy (ART) programs remain a challenge in Malawi. The objectives of this doctoral project were to explore factors that influence loss to follow-up, to describe patterns of follow-up, to identify a definition of lost to follow-up (LTFU) that predicts whether a patient will return to care and to determine the predictors of becoming and remaining LTFU, among patients receiving ART through Dignitas International (DI) in Zomba, Malawi. DI has been working with the Malawi Ministry of Health since 2004 to deliver comprehensive HIV/AIDS care. In collaboration with DI, this project used a mixed-methods approach to explore losses to follow-up from ART. New data collection in the form of concept mapping with Malawian stakeholders was first used to determine why patients on ART become LTFU. Descriptive analyses involving Generalized Estimating Equations (GEE) was used to describe patterns of follow-up among patients who initiated ART from 1 January 2007 to 1 July 2010. Survival functions were plotted to identify the cut-off that can be used to define LTFU. GEE logistic regression was used to model the predictors of becoming and remaining LTFU at each FU visit. In concept mapping, 90 participants consisting of ART patients, ART providers, Health Surveillance Assistants and Zomba District Health Office team members participated. A nine-cluster concept map solution was generated. In the descriptive analyses, n =7,815 patients with n =76,417 follow-up visits were included. Patients did not show a consistent pattern of follow-up over time (e.g., not always late). Plotted survival functions indicated that a LTFU cut-off of ≥9 weeks late was relevant in this context. Based on the GEE models, 5 factors that were explored were consistent in predicting both LTFU risk and a return to care within 12 months of an expected visit (transferring in, duration on ART, World Health Organization stage at ART initiation, duration on ART and being in centralized care). Five variables demonstrated statistical differences across models (female gender, being married, cumulative proportion of non-adherent visits, a ≥10% weight gain and reporting side effects, and being in centralized care). This dissertation points to the dynamic complexities which individuals face during the course of their treatment that affect their ability to return to the clinic for scheduled follow-up. There is a need for a variety of program strategies to adequately address the diverse challenges faced by patients.

Department dc:contributor.department

Dalla Lana School of Public Health

Year dc:date.issued

2013

Author dc:creator

• Rachlis, Beth Stephanie

Advisor dc:contributor.advisor

• Cole, Donald

• Epidemiology
• Global Health

Language dc:language.iso

en_ca

Handle dc:identifier.uri

http://hdl.handle.net/1807/68903

OAI identifier oai:identifier

oai:utoronto.scholaris.ca:1807/68903