University of California, San Francisco Logo

University of California, San Francisco | About UCSF | Search UCSF | UCSF Medical Center

Home > Global Health Literature Digest > Overestimates of Survival
Overestimates of survival after HAART: implications for global scale-up efforts
Global Health Sciences Literature Digest
Published June 2, 2008
Journal Article

Bisson GP, Gaolathe T, Gross R, Rollins C, Bellamy S, Mogorosi M, et al. Overestimates of survival after HAART: implications for global scale-up efforts. PLoS ONE 2008 Mar 5;3(3):e1725.

Objective

To measure the effect of loss to follow-up on estimates of survival and risk factors for death following initiation of HAART

Study Design

Retrospective and prospective cohort

Setting

The first public health clinic for infectious disease in Gaborone, Botswana to offer HAART to persons with AIDS

Participants

HIV-infected adult patients who registered for care and were initiated on HAART between February 2003 and August 2003

Methods

Eligible patients were identified retrospectively and their in-patient and out-patient medical and pharmacy records were reviewed to collect demographic, laboratory, pharmaceutical, and clinical data. Patients were classified as alive, dead, or lost to follow-up. Patients were classified as alive and in-care if they were in between clinic visits at the time of review and were known to have filled their HAART prescription within the past 30 days. They were classified as dead if their deaths were recorded in any of the records reviewed. They were classified as lost to follow-up if their last contact with the clinic or pharmacy was more than 30 days after their last scheduled visit.

Lost patients were prospectively traced by telephone calls and, if necessary, home visits. If three calls to the patient, or to contacts listed in the patient's records, were unsuccessful, a Setswana-speaking study nurse made a home visit to locate the patient or a family member. When a transfer of care to another clinic was noted in the records, ascertainment of vital status was attempted. If vital status (alive or dead) could not be determined, the patient remained classified as lost to follow-up.

The date of initiating HAART and the date of outcome (alive, dead, or lost) were used as the starting and ending points of observation. Patients who were alive and in care or lost to follow-up were censored at the date of their last clinic visit. The Kaplan-Meier product limit methods were used to calculate one-year survival estimates using outcomes before and after tracing. That is, the estimates after tracing reclassified persons as alive and in care or dead if tracing provided updated information and the survival estimates were calculated using the original outcomes and again using updated outcomes. The differences in these survival estimates were compared using the log-rank test. Unadjusted and adjusted risk factors for death were measured using Cox proportional hazards models. Variables examined included gender, weight, history of antiretroviral therapy, median baseline CD4 counts and viral loads, hemoglobin levels, presence of tuberculosis, and the initial HAART regimen. Variables that were statistically significant (P<0.05) in the unadjusted analysis or that were thought to be associated with death were included in a multivariable model. Differences in the relative hazards between the models using the initial outcomes and the updated outcomes after tracing 20% or greater were considered to be evidence of bias associated with inaccurate ascertainment of death.

Results

During each wave, participants received HIV testing, along with pre- and post-test counseling, to determine HIV prevalence and to estimate HIV incidence in the community. A structured questionnaire was administered by trained interviewers to determine key demographics and injecting behaviors. Aside from the variations in the recruitment strategies, data collection methods were the same for all sites.

Conclusions

Patients who are lost to follow-up can add a substantial bias to the survival estimates and risk factors for death following initiation of HAART. Specifically, they are likely to have died but are misclassified as lost; therefore, they are censored in survival analysis rather than analyzed as an event. This result produces substantial bias on the outcome of HAART and suggests that death rates following initiation of HAART reported by studies in developing countries may be underestimated. Efforts to increase ascertainment of death should be implemented to give accurate measures of mortality and risk factors for death associated with HAART.

Quality Rating

This study used a somewhat representative sample of patients attending HAART clinics in sub-Saharan Africa. The comparison groups are the same, with reclassification of status after tracing. The ascertainment of "exposure" was collected from all relevant sources of clinic information and reclassification was done using comprehensive tracing methods. The assessment of the outcome was done by study staff, but considering the outcome that was measured and the source of this information, this is not likely to have biased the results. The patients were adequately followed but the duration of follow-up was relatively short, particularly for the outcome of interest.

In Context

The most important measure of the success of HAART is its effect on mortality, and thus reliable data on outcomes following treatment are necessary. Previous studies of survival after initiating HAART have had high proportions of the cohort lost to follow-up.(1) The effect that this has on survival estimates depends on both the degree of the loss and the absolute number of deaths in this group. In Malawi, tracing found that 50% of patients who were lost to follow-up had died.(2) However, the proportion of patients on HAART who were lost was small, and so the end result was that tracing and updating of deaths did not alter the survival estimates substantially. Several public HAART care clinics in Africa have documented high rates of loss to follow-up, and the death rates reported from these sites may have been underestimated.(3,4,5,6,7,8,9,10)

Programmatic Implications

Tracing appears to be a feasible and effective method to improve the ascertainment of death. HAART monitoring programs should implement tracing to more accurately assess both survival and other clinical outcomes.

References

  1. Rosen S, Fox MP, Gil CJ. Patient retention in antiretroviral therapy programs in sub-Saharan Africa: a systematic review. PLoS Med. 2007;4:e298.
  2. Yu Jk, Chen SC, Wang KY, Chang CS, Makombe SD, et al. True outcomes for patients on antiretroviral therapy who are "lost to follow up" in Malawi. Bull World Health Organ 2007;85:550-4.
  3. Stringer JS, Zulu I, Levy J, Stringer EM, Mwango A, et al. Rapid scale up of antiretroviral therapy at primary care sites in Zambia: feasibility and early outcomes. JAMA 2006;296:782-93.
  4. Djomand G, Roels T, Ellerbrock T, Hanson D, Diomande F, et al, Virologic and immunologic outcomes and programmatic challenges of an antiretroviral treatment pilot project in Abidjan, Côte d'Ivoire. AIDS 2003;Suppl 3:S5-15
  5. Ferrandini L, Jeannin A, Pinoges L, Izopet J, Odhiambo D, et al. Scaling up of highly active antiretroviral therapy in a rural district of Malawi: an effectiveness assessment. Lancet 2006;367:1335-42.
  6. Hosseinipour MC, Neuhann FH, Kanyama CC, Namarika DC, Weigel R, et al, Lessons learned from a paying antiretroviral therapy service in the public health sector at Kamusu central hospital, Malawi: 1-year experience. J Int Assoc Physicians AIDS Care (Chi Ill) 2006;5:103-8.
  7. Weidle PJ, MalambaS, Mwebaze R, Sozi C, Rudundo G, et al. Assessment of a pilot antiretroviral drug therapy programme in Uganda: patients' response, survival, and drug resistance. Lancet 2002;360:34-40.
  8. Macharia DK, Chang LW, Lule G, Owili DM, Tesfaledet G, et al. Antiretroviral therapy in private sector of Nairobi, Kenya: a review of the experience of five physicians. AIDS 2003;17:938-40.
  9. Wools-Kaloustian K, Kimaiyo S, Diero L, Siika A, Sidle J, et al. Viability and effectiveness of large-scale HIV treatment initiatives in sub-Saharan African: experience from western Kenya. AIDS 2006;20:41-8.
  10. Braistein P, Brinkhof MW, Dabis F, Schechter M, Boulle A, et al, Mortality of HIV-1 infected patients in the first year of antiretroviral therapy: comparison between low-income and high-income countries. Lancet 2006;367:817-24.