Pathways of care for HIV infected children in Beira, Mozambique: Pre-post intervention study to assess impact of task shifting

Background: In 2013, Mozambique implemented task-shifting (TS) from clinical officers to maternal and child nurses to improve care for HIV positive children < 5 years old. A retrospective, pre-post intervention study was designed to evaluate effectiveness of a new pathway of care in a sample of Beira District Local Health Facilities (LHFs), the primary, local, community healthcare services. Methods: The study was conducted by accessing registries of At Risk Children Clinics (ARCCs) and HIV Health Services. Two time periods, pre- and post-intervention, were compared using a set of endpoints. Variables distribution was explored using descriptive statistics. T-student, Mann Whitney and Chi-square tests were used for comparisons. Results: Overall, 588 HIV infected children (F = 51.4%) were recruited, 330 belonging to the post intervention period. The mean time from referral to ARCC until initiation of ART decreased from 2.3 (± 4.4) to 1.1 (± 5.0) months after the intervention implementation (p-value: 0.000). A significant increase of Isoniazid prophylaxis (O.R.: 2.69; 95%CI: 1.7-4.15) and a decrease of both regular nutritional assessment (O.R. = 0.45; 95%CI: 0.31-0.64) and CD4 count at the beginning of ART (O.R. = 0.46; 95%CI: 0.32-0.65) were documented after the intervention. Conclusions: Despite several limitations and controversial results on nutrition assessment and CD4 count at the initiation of ART reported after the intervention, it could be assumed that TS alone may play a role in the improvement of the global effectiveness of care for HIV infected children only if integrated into a wider range of public health measures. A retrospective pre-post intervention study was conducted [7]. The LHFs’ inclusion criteria defined by the study protocol included: 1) to have implemented HHS since June 2012; 2) shifting to Option B+ and OSM in June 2013, including ART administered to children within the ARCC. Of the 15 LHFs of the Beira District, 5 (33.3%) met the previous criteria and were enrolled in the study: Ponta Gêa, Munhava, Macurungo, Nhaconjo, Mascarenha. All the HIV infected children < 5 years old accessing the ARCC of one of the 5 LHFs between June 2012 to May 2014 were identified through the Antiretroviral (ARV) registries and were recruited in the study. In particular, HIV Exposed Infants (HEI) were identified as HIV infected according to WHO guidelines: through a polymerase chain reaction using a dried blood spots test (PCR-DBS) if less than 18 months old, or through a HIV rapid test if older than 18 months. Since health data of all children accessing the ARCC of the LHFs who met the study inclusion criteria were available, no sample size calculation was performed. The health district-level intervention of interest, introduced in Beira by the Local Health Authorities in June 2013, consisted in the TS of the care and treatment of HIV infected children from clinical officers to MCH nurses to ensure a continuum of care within the ARCC. As the intervention was introduced in June 2013, two time periods lasting 365 days each were compared: one pre-intervention, from June 2012 to May 2013 (T1), and one post-intervention, from June 2013 to May 2014 (T2). Figure 1 shows the pathways of care provided in pre- and post- intervention time periods. Pathways of care adopted in pre- and post- intervention time periods within the Beira’ HIV healthcare service In the pre-intervention pathway of care (pre-TS), all HEIs were followed at the ARCC from the first month of life to the definition of their HIV status. Any child identified as HIV positive was then referred to a HHS integrated within the OPD of the same LHF receiving HIV care and treatment from a clinical officer, occasionally present in the structure, while mothers were under the responsibility of a nurse. In the post-intervention pathway of care (post-TS) all HEIs were monitored at the ARCC until the definitive diagnosis of their HIV status was determined, while any child identified as being infected was hosted in the ARCC until they turned 5 years old. Therefore, HIV infected children and their mothers were monitored and treated by the MCH nurses according to TS. Only after 5 years old could the children, together with the mothers, be referred to the OPD-HHS. We compared the post intervention group (post-TS) with the pre intervention group (pre-TS) by using the following set of endpoints as a proxy to assess the impact of the new pathway of care, including TS implementation, on the effectiveness of taking charge of patients: Further, we considered CD4 count at the beginning of ART as an indicator of the process of taking charge of patients’ care. Individual-level, routinely collected data were extracted from ARCC registries and HHS patients’ files. Patients’ information were further enriched with pharmacy files. Record linkage was possible through a unique code assigned to the patient. Data extraction was performed between January and March 2016 by trained personnel of “Doctors with Africa” and the research Centre of Infectious Disease (CIDI) of the Catholic University of Mozambique (UCM). Data entry and management was in charge of CIDI research and was performed by using Microsoft Access 2013. The variables distribution was explored by frequency tables and descriptive statistics such as mean and median. Comparison of the means of the two groups was done by using the t-student test, when data presented a normal distribution, while comparison of medians was performed using the nonparametric Mann Whitney test. The Chi-square test was used to compare the two groups by categorical variables, while Fisher’s exact test was applied when the expected values were below or equal to 5. Crude Odds Ratios (ORs) were calculated to measure the association between endpoints and intervention (independent variable), being ORs values > 1 predictive of a more effective care. Data were analyzed by IBM SPSS Software 23 version (IBM Corp., Armonk, NY, USA). All p-values were two-sided and p-value < 0.05 was considered statistically significant. Ethical approval of the protocol was achieved and (as this study used secondary data) informed patient consent was not required.