537
High Rates of Death among Patients Lost to Follow-up in Botswana's National ART Program: Implications for Monitoring and Evaluation
G Bisson1, N Ndwapi2, C Rollins1, A Avalos2, R Gross1, S Bellamy1, M Mogorosi2, H Friedman1, D Dickinson3, and Tendani Gaolathe*2
1Univ of Pennsylvania, Philadelphia, US; 2Infectious Diseases Care Clin, Gaborone, Botswana; and 3Independence Surgery, Gaborone, Botswana
Background: While successful scale-up of ART in
resource-limited countries has drastically increased the number of patients on HAART,
many patients are lost to follow-up soon after initiating HAART. If the
subjects lost to follow-up are in fact dead, their exclusion from analyses
would result in overestimates of survival in these programs.
Methods: We performed a retrospective cohort study among
HIV-infected adults consecutively presenting to the Infectious Disease Care
Clinic (IDCC) in Gaborone, Botswana. Patients were eligible if their
registration was recorded in the IDCC database at least 7 months prior to the
end of the study period. Outcomes were recorded according to both passive and
active follow-up. Passive follow-up, classified patients as
dead if their death was recorded anywhere in the clinical record, and lost if
their last contact with the clinic was >30 days past their last visit.
Active follow-up was performed on all lost patients and consisted of telephone
calls and home visits. Χ2 tests
were used to compare proportions, and Kaplan-Meier estimates were used to
compare survival estimates for the passive and active follow-up systems. Cox
proportional hazards models were used to determine risk factors for death after
initiating HAART.
Results: The median duration of follow-up among 410 patients initiating
HAART was 44 weeks (IQR 37 to 49). Passive follow-up categorized 29 of 410 (7%)
as dead and 68 of 410 (17%) as lost; in contrast, active follow-up classified
69 of 410 (17%) as dead and 22 of 410 (5%) as lost. Of the 68 patients, 40 (59%)
categorized as lost in passive follow-up were confirmed dead in active
follow-up (p <0.001). For passive
follow-up, the 52-week Kaplan-Meier survival estimate was 0.93 (95%CI 0.88 to 0.94)
and for active follow-up it decreased to 0.79 (95%CI 0.74 to 0.81). A CD4 cell
count <100 cells/mm3 at baseline (adjusted RH 4.8, 95%CI 2.4 to 9.3)
and male sex (adjusted RH 1.8, 95%CI 1.0 to 2.6) were independently associated
with death within the first 12 months after initiating HAART.
Conclusions: ART programs should implement active follow-up
to determine definitive programmatic response rates and allow better
characterization of risk factors for death. Better outcomes are achieved when
individuals initiated HAART at higher CD4 levels. Further exploration is needed
of the reasons why men are dying at higher rates.
|
