Retention on Antiretroviral Therapy in National Programs in Low-Income and Middle-Income Countries

Abstract

To the Editors: The study from Cornell et al on a sample of approximately 10% of patients started on antiretroviral therapy (ART) in the public sector in South Africa showed an increasing proportion of patients lost to follow-up (LFU) among patients starting in most recent years.1 These results are of major public health importance considering the risk of HIV drug resistance.2 We agree with the authors that monitoring the retention on ART is complex and interpreting data needs to consider potential bias in classifying patients LFU. Each year, the World Heath Organization and partners are collecting national programs data including retention on ART.3 In 2009, the number of countries reporting and the number of patients assessed almost doubled compared with 2008.3 Among sub-Saharan Africa countries, average retention at 12, 24, and 36 months was 74.5%, 71.9%, and 70.4%, respectively, which is within similar range of the 2008 results.4 Because 2009 data included part of data reported in 2008, a trend analysis is not possible. We observed a large variation between countries; retention at 12 months ranged between 47% and 96%. Such differences cannot be attributed only to program performance, but also reflect the capacity of patient monitoring systems. First, retention is overestimated when calculated exclusively on survival without taking into account patients' LFU and those interrupting treatment. Second, improper recordkeeping of patients transferred from one health facility to another to continue treatment may result in misclassification as LFU, hence underestimating retention. This bias, discussed by Cornell et al, may increase with time as a result of more facilities offering ART resulting in more opportunities for patients to be transferred. In addition, country data may not be representative of the full program when produced from nonrandomly selected sites. In 2009, a review of national monitoring systems conducted in 13 countries in southern and eastern Africa revealed significant variations among countries and health facilities to effectively monitor the outcomes of people on ART. All countries had multiple paper-based and electronic systems that were not interoperable. Data from facilities within the countries have different levels of quality, hence affecting their accuracy and representativeness at the national level. Only seven countries had data quality assessments conducted in the past 5 years (Dick Chamla, personal communication). To document program performance, retention on ART should be produced for each cohort of patients starting in a specific year and also for longer follow-up to assess the long-term retention on ART; in 2009, at a global level, 81% of national ART programs were established for more than 5 years.3 Although accurate data are as critical for patient management than for monitoring programs, open lifelong cohort monitoring for HIV care and ART presents with unprecedented challenges as a result of the growing number of patients and increasing follow-up. We share our colleagues' concern of achieving an accurate, simple, and sustainable monitoring system of patients in HIV care and ART, ascertaining outcomes in patients classified as LFU.5 Jean-Michel Tassie, MD, MPH* Dick Chamla, MD, MPH† Yves Souteyrand, PhD* *Department of HIV/AIDS, World Health Organization, Geneva, Switzerland †Intercountry Support Team, World Health Organization, Harare, Zimbabwe