|
Determinants of waiting times for ART in the Free State Province, South Africa: prospective cohort study with retrospective database linkage
M. May1, S. Ingle1, V. Timmerman2, E. Kotze3, K. Uebel2, M. Bachmann4, J. Sterne1, M. Egger5, L. Fairall2, IeDEA Southern Africa
1University of Bristol, Department of Social Medicine, Bristol, United Kingdom, 2University of Cape Town, Lung Institute, Cape Town, South Africa, 3Free State Department of Health, Bloemfontein, South Africa, 4UEA, Public Health & Health Services Research Unit, Norwich, United Kingdom, 5University of Bern, Institute of Social and Preventative Medicine, Bern, Switzerland
Background: ART programmes should follow patients from enrolment, since waiting times for treatment impact mortality. In the Free State Province ARV Programme (FSAP), incomplete recording of pre-treatment CD4 counts is a barrier to monitoring. We investigated factors associated with delays in starting ART using data from FSAP augmented by CD4 counts obtained from the National Health Laboratory Service (NHLS).
Methods: Patients aged 15+years, enrolled between 5/04 and 12/06 were followed until 12/07. The FSAP and NHLS databases were deterministically matched using South African ID numbers, gender and date-of-birth. Laboratory-supplied CD4 counts were used to define patient eligibility for ART (CD4< 200cells/mm3). Weibull models were used to estimate associations expressed as time ratios (TR) of median waiting times. Explanatory variables included were: health district, clinic type, clinic location (rural/peri-urban), CD4 at eligibility, gender, age and year of enrolment.
Results: CD4 counts were available in 77.2%(22,182/28,732) of enrolled patients compared with 49.4%(14,207/28,732) before database linkage. Of the former, 61%(13,554/22,182) were eligible for ART. Of these, 13,249(98%) with complete data were included in analyses, 3,475(26%) of whom died whilst waiting for treatment with median time to death 3.0 (IQR; 1.1-6.8) months. The median waiting time for treatment was 4.0 (IQR; 2.1-7.2) months. Waiting times varied by a factor of 2 across districts and were halved in combined assessment and treatment clinics compared with assessment only clinics. Waiting times were shorter for enrolments in 2006 v. 2004 (TR 0.71; 95%CI 0.67-0.75), for women (TR 0.87;0.83-0.91) and decreased with decreasing CD4 count (p trend 0.04).
Conclusions: Delayed initiation of ART is a major cause of death in HIV-infected patients. Database linkage improves programme monitoring. Probabilistic matching with laboratory data and death registries might further decrease loss-to-follow-up and improve ascertainment of death. Waiting times have decreased over time, but further improvements might be possible through integration of assessment and treatment.
Download the e-Poster
|