Background: A major obstacle facing many lower-income countries in establishing and maintaining HIV treatment programmes is the scarcity of trained health care providers. To address this shortage, the World Health Organization has recommend task shifting to HIV-infected peers.
Methods: We designed a model of HIV care that utilizes HIV-infected patients, community care coordinators (CCCs), to care for their clinically stable peers with the assistance of preprogrammed personal digital assistants (PDAs). Rather than presenting for the standard of care, monthly clinic visits, in this model, patients were seen every three months in clinics and monthly by their CCCs in the community during the interim two months. This study was conducted in Kosirai Division, western Kenya, where eight of the 24 sub-locations (defined geographic areas) within the division were randomly assigned to the intervention with the remainder used as controls.Prior to entering the field, CCCs underwent intensive didactic training and mentoring related to the assessment and support of HIV patients, as well as the use of PDAs. PDAs were programmed with specific questions and to issue alerts if responses fell outside of pre-established parameters. CCCs were regularly evaluated in six performance areas. An impressionistic analysis on the transcripts from the monthly group meetings that formed the basis of the continuous feedback and quality improvement programme was used to assess this model.
Results: All eight of the assigned CCCs successfully passed their training and mentoring, entered the field and remained active for the two years of the study. On evaluation of the CCCs, 89% of their summary scores were documented as superior during Year 1 and 94% as superior during Year 2. Six themes emerged from the impressionistic analysis in Year 1: confidentiality and "community" disclosure; roles and responsibilities; logistics; clinical care partnership; antiretroviral adherence; and PDA issues. At the end of the trial, of those patients not lost to follow up, 64% (56 of 87) in the intervention and 52% (58 of 103) in the control group were willing to continue in the programme (p = 0.26).
Conclusion: We found that an antiretroviral treatment delivery model that shifted patient monitoring and antiretroviral dispensing tasks into the community by HIV-infected patients was both acceptable and feasible.
Trial registration: ClinicalTrials.gov ID NCT00371540.