Background: In Tanzania, progress toward achieving the 2015 Millennium Development Goals for maternal and newborn health was slow. An intervention brought together community health workers, health facility staff, and accredited drug dispensing outlet (ADDO) dispensers to improve maternal and newborn health through a mechanism of collaboration and referral. This study explored barriers, successes, and promising approaches to increasing timely access to care by linking the three levels of health care provision. Methods: The study was conducted in the Kibaha district, where we applied qualitative approaches with in-depth interviews and focus group discussions. In-depth interview participants included retail drug shop dispensers (36), community health workers (45), and health facility staff members (15). We conducted one focus group discussion with district officials and four with mothers of newborns and children under 5 years old. Results: Relationships among the three levels of care improved after the linkage intervention, especially for ADDO dispensers and health facility staff who previously had no formal communication pathway. The study participants perceptions of success included improved knowledge of case management and relationships among the three levels of care, more timely access to care, increased numbers of patients/customers, more meetings between community health workers and health facility staff, and a decrease in child and maternal mortality. Reported challenges included stock-outs of medicines at the health facility, participating ADDO dispensers who left to work in other regions, documentation of referrals, and lack of treatment available at health facilities on the weekend. The primary issue that threatens the sustainability of the intervention is that local council health management team members, who are responsible for facilitating the linkage, had not made any supervision visits and were therefore unaware of how the program was running. Conclusion: The study highlights the benefits of approaches that link different levels of care providers to improve access to maternal and child health care. To strengthen this collaboration further, health campaign platforms should include retail drug dispensers as a type of community health care provider. To increase linkage sustainability, the council health management team needs to develop feasible supervision plans.
The study was conducted in Kibaha district in the Coast region of Tanzania in November 2015. According to the 2012 national census, the district has an estimated population of 198,697 inhabitants. The main economic activities include agriculture, livestock keeping, and small businesses. We used a qualitative approach rooted in the principles of grounded theory [13], in which we continued sampling until we reached response saturation. We carried out in-depth interviews with ADDO dispensers, CHWs, and health facility staff. The study participants were purposively sampled to ensure that we obtained adequate information to achieve the study objectives and capture differences in responses among the three groups. In addition, we conducted separate focus group discussions with district officials and mothers of newborns and children under 5 years to gather the opinions of both the supervisors and beneficiaries of health care services. Our interview methods took an inductive approach that allowed participants to report issues related to the experience of working with each other, while we probed for necessary information related to our study objectives [14]. Our sample size was determined using saturation sampling and related to the number of participants who had received training [15]. The original training included 40 ADDO dispensers; however, for this study we interviewed 36, while using the other four to pilot test the data collection tool. The number of CHWs who received training was 85, and based on saturation principles, we aimed to include half of those trained in each ward [15]; in total, we interviewed 45 of them. At health facility level, 15 staff members from all four health care levels were included, with the exception of two who were involved in the pilot exercise. We conducted one focus group discussion with district health officials and four with mothers of newborns and children under five. We based our selection of district officials, who were CHMT members, on their role in supervising the linkage. The group included the district medical officer, district health secretary, reproductive and child health coordinator, district coordinator for chws, acting district pharmacist, acting district dental officer, district family planning coordinator, and district coordinator for neglected tropical diseases. We recruited and trained four experienced research assistants, two girls and two boys. We then piloted the data collection tools in Kibaha with four ADDO dispensers, four CHWs, and two health facility staff members. The in-depth interview guide was revised based on the results of the pilot for use in the actual data collection exercise. The research assistants interviewed two groups of respondents at their workplaces (ADDOs and health facilities). The interview process was flexible enough to allow customers to access services between questions. Community health workers were interviewed at the health facilities where they receive supervision. Focus group discussions with mothers of newborns and children under five also took place at the health facility. The nursing officer in charge selected mothers of healthy newborns who had been in-patients, but who were about to be discharged. Mothers of small children were selected from those who had brought their children in for clinic services. The number of focus group participants ranged from 8 to 12 per group. Each indepth interview lasted for not more than an hour, while focus group discussion ranged between one to 1 hour and a half. All interviews and focus group discussions were conducted in Kiswahili language. The senior social scientist checked the quality of the in-depth interviews by revisiting and interviewing some of the ADDO dispensers, CHWs, and health facility staff. All focus group discussions were facilitated by the senior social scientist, who is the first author, with assistance from a research assistant. We used recording devices; however, the research assistant also took notes, which were then expanded later [16]. All recorded interviews were transcribed. Data was analyzed using NVIVO 9 software [17]. Two persons conducted an inductive thematic analysis of transcripts to come up with codes, and we checked for coding consistency. Additional codes identified through the line-by-line coding were added. We reviewed the list of codes and grouped them into categories and themes for analysis. We analyzed them by comparing themes that related to our study objectives. We did not translate the data in advance; all data were analyzed in Kiswahili. We obtained ethical clearance from the National Institute of Medical Research of Tanzania. An information sheet about the study was drawn up in Kiswahili, explaining why the study was being carried out, by whom, and what it would involve. We assured respondents’ confidentiality. Respondents were asked if they had any questions and whether they agreed to take part in the study. We got the written consent of all respondents. We discussed the study in advance with the responsible district authorities to ensure their support.
N/A