Majority of patients lost to follow-up still alive one year later, Ugandan study shows

A study of patients who failed to return to their HIV clinic in Uganda found that the majority were still alive after one year, contrary to assumptions, and in many cases had either transferred to other clinics or faced transport and access difficulties in getting to the clinic, according to research by Elvin Geng and colleagues in a paper published in advance online by the Journal of Acquired Immune Deficiency Syndromes.

An estimated 15-40% of patients are lost to follow up within the first year of starting antiretroviral therapy. Without understanding the reasons why this happens and what happens to these patients subsequently it is impossible to have a true picture of the effect of global scale-up.

Effective scale up of antiretroviral treatment requires outcome evaluation of those being treated. This provides an understanding of the impact of such programmes as well as the necessary evidence for improving future strategies.

To date evaluation of LTFU have been limited to epidemiological analysis. Death as well as high risk for death is, in most studies, the inferred outcome. Direct interview, a strategy to find out from the patients or those close to them why they no longer attend clinic, is seldom used. The authors note that the specific causes of death associated with LTFU have not been identified and are key to targeting and finding those at highest risk.

The common assumption, note the authors, is that all losses to follow-up are adverse effects, but this ignores potential positive outcomes. For example, patients may transfer to treatment centres closer to their homes.

The authors undertook a cohort study of all HIV-infected adults attending the Immune Suppression Syndrome (ISS) clinic in Mbarara, Uganda who began ART between January 1, 2004 and September 30, 2007. 92% of the 1.1 million in the district served by the clinic reside in rural areas.

Of the 3,628 patients evaluated over close to four years 829 became lost to follow-up (defined by six months absence from clinic). This corresponded to a cumulative incidence of LTFU of 39% (95% CI: 37-42) three years after starting ART.

The authors tracked down and interviewed a representative sample (15% or 128) of those patients LTFU or close informants in the community to determine reasons for and outcomes of LTFU.

Information was available for 87% (111). 43% (48) were found and of the remaining 57% (63) a close informant was found. The median time between last clinic visit and tracking of the patient was 11.6 months (IQR 9.4-14.3) and 33 kilometres (IQR 4-66) the median distance from home to clinic.

71% (79) of the sample were found to be alive, of whom 48 were directly interviewed and an informant interviewed for the remaining 31.

The main reasons for absence from clinic (LTFU) of those directly interviewed according to the authors are social or structural and included: a lack of transportation for 50% (95% CI: 35-65) and distance for 42% (95% CI: 28-57).

Poverty as well as child care and employment issues were also given as fairly common reasons for failure to return to clinic. The authors note that accessing health in Africa is just one of many competing needs. The authors argue these socio-structural factors contrast sharply with prevailing Western models which focus on individual motivation and behaviours.

However 40 of the 48 (83% 95% CI: 70-93) directly interviewed had attended a different health care facility in the previous three months and 34 of the 48 (71%, 95% CI: 56 to 83) had taken ART within the last 30 days. These factors signify positive outcomes contrary to commonly held beliefs that all LTFUs result in adverse outcomes.

A high proportion of the sample died (32) with a cumulative one year incidence of 36%, of which the majority (30) died shortly after their last clinic visit.

The authors suggest that those who died within the first months following their last visit died from conditions that developed whilst they were still in care rather than after care ended. This means not all deaths are due to LTFU. Improvement in outcomes, the authors note, will come with improved point-of-care diagnostics and effective treatment, both of opportunistic conditions and HIV.

Predictors of death following the last clinic visit included: older age - each 10 year increase in age increased the rate of death 2.04 fold (P=0.03); low blood pressure - the presence of mean arterial blood pressure <75 mmHg gave a 2.97 increased rate of death (P=0.02); a central nervous syndrome raised the rate 2.86 fold (P=0.03) and a low pre-ART CD4 cell count - each increase of 50 cells/mm³ reduced the rate of death by 38% (P=0.02).

is the extent to which the findings are applicable to other settings remains unknown, but the authors note that the data do provide proof of concept.

Specific factors are found to be predictive of subsequent death following last clinic visit. This will enable clinics to determine who to find, and who to prioritise for community follow-up, following their failure to return to clinic.

The wide-ranging reasons for loss-to-follow-up (LTFU) identified in this study support a sampling-based, site-specific approach to understand why these occur note the authors. While this approach is not a retention tool, in and of itself, it can provide valuable information to improve retention and identify those factors associated with poor outcomes (following LTFU).

The authors note several limitations. The diverse nature of treatment programmes across sub-Saharan Africa precludes results being generalised to other clinic settings. They suggest that analysis of LTFUs be part of routine monitoring and evaluation.

Also, while the sample was unselected and consecutive it was not a formal random sample and could in theory have biases. In addition they cite the lack of extensive in-depth interviews as well as a comparison group suggesting that findings are not definitive. A high number of missing CD4 counts in the analysis of predictors of death may not be random and could potentially lead to bias.

The authors conclude that a sampling-based approach is both a feasible and cost-efficient means for individual clinics to understand the reasons for LTFUs “and can be used to target outreach efforts to the right patients at the right time.”