“Because even the person living with HIV/AIDS might need to make babies” – Perspectives on the drivers of feasibility and acceptability of an integrated community health worker model in Iringa, Tanzania

Background: Countries with health workforce shortages are increasingly turning to multipurpose community health workers (CHWs) to extend integrated services to the community-level. However, there may be tradeoffs with the number of tasks a CHW can effectively perform before quality and/or productivity decline. This qualitative study was conducted within an existing program in Iringa, Tanzania where HIV-focused CHWs working as volunteers received additional training on maternal, newborn, and child health (MNCH) promotion, thereby establishing a dual role CHW model. Methods: To evaluate the feasibility and acceptability of the combined HIV/MNCH CHW model, qualitative in-depth interviews (IDIs) with 36 CHWs, 21 supervisors, and 10 program managers were conducted following integration of HIV and MNCH responsibilities (n = 67). Thematic analysis explored perspectives on task planning, prioritization and integration, workload, and the feasibility and acceptability of the dual role model. Interview data and field observations were also used to describe implementation differences between HIV and MNCH roles as a basis for further contextualizing the qualitative findings. Results: Perspectives from a diverse set of stakeholders suggested provision of both HIV and MNCH health promotion by CHWs was feasible. Most CHWs attempted to balance HIV/MNCH responsibilities, although some prioritized MNCH tasks. An increased workload from MNCH did not appear to interfere with HIV responsibilities but drew time away from other income-generating activities on which volunteer CHWs rely. Satisfaction with the dual role model hinged on increased community respect, gaining new knowledge/skills, and improving community health, while the remuneration-level caused dissatisfaction, a complaint that could challenge sustainability. Conclusions: Despite extensive literature on integration, little research at the community level exists. This study demonstrated CHWs can feasibly balance HIV and MNCH roles, but not without some challenges related to the heavier workload. Further research is necessary to determine the quality of health promotion in both HIV and MNCH domains, and whether the dual role model can be maintained over time among these volunteers. TUNAJALI (Swahili for “We Care”) was a large-scale initiative funded over two 5-year cycles (2006-2011; 2012-2017) by the United States Agency for International Development (USAID) to prevent HIV/AIDS and increase access to HIV care, treatment and support services in Tanzania, focusing on both community and facility-based services.13 TUNAJALI II was implemented by Deloitte Consulting Limited (prime partner) and Christian Social Services Commission (technical partner) across 5 regions of Tanzania, namely Dodoma, Iringa, Njombe, Morogoro, and Singida. This study focuses on TUNAJALI II’s community-based program activities in Iringa Region (2012 population: 941 238), located in the Southern Highlands zone of mainland Tanzania, in the districts of Kilolo (2012 population: 218 130) and Iringa Rural (2012 population: 254 032).14 The prevalence of HIV in Iringa Region (9.1%) ranks second highest nationally and is nearly double the national average (5.1%).15 However, coverage estimates for MNCH indicators were higher in Iringa Region than nationally in the 2004-2005, 2010, and 2015 Demographic and Health Surveys – for example, the proportion of live births delivered at health facilities has improved over time both in Iringa (71.8%, 80.4%, to 92.9%) as well as nationally across Tanzania (47.1%, 50.1%, to 62.6%) although remains comparatively lower.16-18 Through TUNAJALI II, local civil society organizations (CSOs) managed over 400 volunteer CHWs in 4 Iringa districts. These village-based CHWs received 12 days of training and are tasked with providing home-based HIV services, following 15 to 100 households each, with a major focus on treatment adherence, retention, and palliative support (Box 1).19 In recent years, the HIV workload of CHWs has decreased as antiretroviral drugs became more available and the number of extremely sick HIV/AIDS patients declined. This led program administrators to suggest that HIV-focused CHWs could absorb additional duties involving MNCH to increase antenatal care (ANC) utilization and health facility deliveries. Therefore, during 2015 approximately half the HIV CHWs working with this program in Iringa Region were trained for 3 weeks using the MNCH curriculum approved by Tanzania’s Ministry of Health, with health promotion topics organized around the recommended timing and frequency of home visits (Box 1).20 These “dual role” CHWs were tasked with both HIV and MNCH services (compared to “single role” CHWs focused on HIV services only). HIV/AIDS Maternal, Newborn and Child Health Abbreviations: CHW, community health worker; ANC, antenatal care. A facility-based HIV healthcare worker provided HIV-focused supervision to both single and dual role CHWs. However, dual role CHWs also reported to a second facility-based healthcare worker for MNCH-related supervision. “Home based care (HBC) focal persons” from the CSO also provided monthly HIV-focused supervision to CHWs. Modest monthly stipends were provided through TUNAJALI II as remuneration to CHWs, which increased from the initial US$17 to US$20 for all CHWs following MNCH training, regardless of whether they were assigned to single or dual responsibilities. This study is part of a larger mixed methods implementation research evaluation of the dual role CHW model. Qualitative and quantitative data were collected from February 2016 to January 2017, with concurrent analysis and triangulation of mixed methods data from June 2016 to March 2017 (Supplementary file 1). Here we present qualitative findings on feasibility and acceptability, with reference to quantitative data in associated publications. In Iringa Rural and Kilolo, 6 types of respondents were selected for interviews using several sampling approaches (Table 1). At the prime partner’s regional office, the director, HIV technical officer, and reproductive and child health technical officer were interviewed. At the CSOs, all staff members with program responsibilities were interviewed, including the director, project coordinators, and HBC focal persons. CHWs and their supervisors were interviewed from 20 facilities (10 per district) among the 71 facilities involved across the two districts. Purposive sampling methods were used to achieve maximum variation in facility and CHW characteristics, including: facility type (health center and dispensary) and ownership (public, private faith-based); location; availability of HIV care and treatment services; monthly service utilization for ANC and deliveries; number of CHWs per facility; type of CHWs (single vs. dual role); and Gender of CHWs. The locations of sampled dispensaries and health centers are presented in Figure 1 (notably clustered toward the center owing to the region’s geography with Ruaha National Park to the West and Udzungwa Mountains National Park to the East). Map of Iringa Region in the Southern Highlands of Tanzania and Location of the 20 Sampled Facilities in Iringa Rural and Kilolo Districts. Notes: Dispensaries: D1 to D15; Health Centers: HC1 to HC5. Maps were generated using QGIS Version 2.18.21 Country, regional and district boundaries were sourced from Tanzania’s National Bureau of Statistics.22 Coordinates for health facilities were sourced from online health facility registry maintained by Tanzania’s MoHCDGEC.23 Abbreviations: IDI, in-depth interview; CHWs, community health workers; CSO, civil society organization; MNCH, maternal, newborn, and child health. Semi-structured in-depth interviews (IDIs), 30-60 minutes long, were conducted in Swahili by 6 trained Tanzanian research assistants. IDI topic guides were organized by several a priori themes: Training and guidance for integration of HIV-MNCH tasks; CHW role expansion from HIV to MNCH; CHW performance; workload balance; feasibility and acceptability; and recommendations for improvement. Interviews were conducted over 3 weeks during February 2016, early in the implementation phase: 3 months after MNCH training in Iringa Rural and 8 months after MNCH training in Kilolo. The CSOs convened regularly scheduled monthly meetings with CHWs at central locations to share information, discuss challenges, collect monthly HIV summary reports, and assess data quality. Where possible, the research team scheduled CHW interviews to coincide with these monthly meetings. Visits to selected facilities were scheduled to interview supervisors, as well as CHWs not reached during the monthly meetings. Following each data collection event, the research team held debriefing sessions with the team leaders on main points given by respondents and emergent ideas and themes. Research assistants also submitted a form with summary notes on each IDI, including interview setting and quality, key summary points, and any new information. In addition, information from program documentation (eg, quarterly reports) and field observations (eg, monthly CSO meetings and visits to facilities, district health offices, and CSO officers) was compiled to further inform understanding of HIV and MNCH roles and program supervision and oversight. Audio-recorded interviews were transcribed by 6 research assistants, 3 of whom had conducted the IDIs. Personal identifiers were removed from transcripts to ensure confidentiality. To assess transcription quality, a Tanzanian research scientist listened to audio-recorded interviews while reading along with the transcript for 2 files per transcriptionist. Interviews were translated from Swahili into English. Translation quality was assessed using the first and seventh documents submitted by each translator, reading each Swahili paragraph followed by the English translation for the full document. Qualitative data management, coding, and thematic analysis were performed using the web-based software, Dedoose.24 Familiarization with field notes and interview transcripts supported the initial preliminary coding structure, based on presumptive topics in the interview guides and emerging themes identified during the familiarization phase. The structure of the preliminary codebook was independently tested by two primary coders experienced with CHW programs. Each coder deductively applied codes from the preliminary codebook to two selected transcripts. They then met to compare code agreement, examining line-by-line for discrepancies. Through this process, the coders reached consensus by discussing modifications to code definitions and agreeing on the definition of newly emergent codes. Four additional transcripts were independently coded and compared to assess agreement. Codes were further refined and added or deleted as new themes became apparent.25 The final codebook contained 68 codes within 6 thematic areas. Remaining transcripts were divided between the two primary coders. Using a thematic analysis approach, code report excerpts were organized in data display matrices to chart key findings and illustrative quotes by IDI respondent type, including descriptive, text-based summaries for each key finding. Data matrices helped identify recurrent patterns and themes and facilitated comparison of diverse perspectives by respondent type. Key findings are presented in the context of understanding program feasibility and acceptability,26 using a subset of codes (Table 2). Interview data, field observations, and program documents were used to triangulate and compare implementation features of the HIV and MNCH roles. Abbreviations: MNCH, maternal, newborn, and child health; CHWs, community health workers; CSO, civil society organization.