Open Access Research article

Outcomes and associated risk factors of patients traced after being lost to follow-up from antiretroviral treatment in Lilongwe, Malawi

Ralf Weigel1*, Mindy Hochgesang2, Martin WG Brinkhof3, Mina C Hosseinipour4, Matt Boxshall5, Eustice Mhango6, Brains Nkwazi5, Hannock Tweya5, Maggie Kamlaka5, Frederick Chagwera5 and Sam Phiri5

Author Affiliations

1 Lighthouse Trust at Kamuzu Central Hospital and Ministry of Health, Lilongwe, Malawi

2 Centers for Disease Control and Prevention, Global AIDS Program, Lilongwe, Malawi (2004-2007); Maputo, Mozambique

3 International epidemiological Databases to Evaluate AIDS (IeDEA), University of Bern, Institute of Social and Preventive Medicine, Division of International and Environmental Health, Bern, Switzerland and Swiss Paraplegic Research, Nottwil, Switzerland

4 University of North Carolina Project Lilongwe, Private Bag A104, Lilongwe, Malawi

5 Lighthouse Trust at Kamuzu Central Hospital, Lilongwe, Malawi

6 Ministry of Health, Department of HIV/AIDS, Lilongwe, Malawi

For all author emails, please log on.

BMC Infectious Diseases 2011, 11:31  doi:10.1186/1471-2334-11-31

Published: 27 January 2011



Loss to follow-up is a major challenge of antiretroviral treatment (ART) programs in sub-Saharan Africa. Our objective was to a) determine true outcomes of patients lost to follow-up (LTFU) and b) identify risk factors associated with successful tracing and deaths of patients LTFU from ART in a large public sector clinic in Lilongwe, Malawi.


Patients who were more than 2 weeks late according to their last ART supply and who provided a phone number or address in Lilongwe were eligible for tracing. Their outcomes were updated and risk factors for successful tracing and death were examined.


Of 1800 patients LTFU with consent for tracing, 724 (40%) were eligible and tracing was successful in 534 (74%): 285 (53%) were found to be alive and on ART; 32 (6%) had stopped ART; and 217 (41%) had died. Having a phone contact doubled tracing success (adjusted odds ratio, aOR = 2.1, 95% CI 1.4-3.0) and odds of identifying deaths [aOR = 1.8 (1.2-2.7)] in patients successfully traced. Mortality was higher when ART was fee-based at initiation (aOR = 2.3, 95% CI 1.1-4.7) and declined with follow-up time on ART. Limiting the analysis to patients living in Lilongwe did not change the main findings.


Ascertainment of contact information is a prerequisite for tracing, which can reveal outcomes of a large proportion of patients LTFU. Having a phone contact number is critical for successful tracing, but further research should focus on understanding whether phone tracing is associated with any differential reporting of mortality or LTFU.