Scale up and strengthening of comprehensive emergency obstetric and newborn care in Tanzania

Introduction In Tanzania, inadequate access to comprehensive emergency obstetric and newborn care (CEmONC) services is the major bottleneck for perinatal care and results in high maternal and perinatal mortality. From 2015 to 2019, the Accessing Safe Deliveries in Tanzania project was implemented to study how to improve access to CEmONC services in underserved rural areas. Methods A five-year longitudinal cohort study was implemented in seven health centres (HCs) and 21 satellite dispensaries in Morogoro region. Five of the health centres received CEmONC interventions and two served as controls. Forty-two associate clinicians from the intervention HCs were trained in teams for three months in CEmONC and anaesthesia. Managers of 20 intervention facilities, members of the district and regional health management teams were trained in leadership and management. Regular supportive supervision was conducted. Results Interventions resulted in improved responsibility and accountability among managers. In intervention HCs, the mean monthly deliveries increased from 183 (95% CI 174-191) at baseline (July 2014 -June 2016) to 358 (95% CI 328-390) during the intervention period (July 2016 -June 2019). The referral rate to district hospitals in intervention HCs decreased from 6.0% (262/4,392) with 95% CI 5.3-6.7 at baseline to 4.0% (516/12,918) with 95% CI 3.7-4.3 during the intervention period while it increased in the control group from 0.8% (48/ 5,709) to 1.5% (168/11,233). The obstetric case fatality rate decreased slightly from 1.5% (95% CI 0.6-3.1) at baseline to 1.1% (95% CI 0.7-1.6) during the intervention period (not statistically significant). Active engagement strategies and training in leadership and management resulted in uptake and improvement of CEmONC and anaesthesia curricula, and contributed to scale up of CEmONC at health centre level in the country. Conclusions Integration of leadership and managerial capacity building, with CEmONC-specific interventions was associated with health systems strengthening and improved quality of services. At the start of the study, Morogoro region had 15 HCs that were either already offering CEmONC or would be able to do so once staff are trained. This study was a longitudinal cohort study in seven health centres in Morogoro region, Tanzania. Five of these received an intervention and two served as controls to detect secular trends. The HCs in the intervention group were Kibati, Ngerengere, Gairo, Melela and St. Joseph HCs. The hierarchy of health facilities in Tanzania, from bottom to top, includes dispensaries, health centres (HCs), district hospitals, regional hospitals, zonal hospitals and the national specialized hospitals. By design, study health centres had to be far enough from the nearest referral hospital for referral to be a significant challenge for health centre staff and families. Health centres were also chosen to reflect the diversity of funding and governance models for HCs in Tanzania. As described elsewhere, Kibati and Ngerengere HCs had the proper infrastructure for CEmONC services including maternity and neonatal wards, operating theatre and ability to provide blood transfusion but their staff had not received CEmONC training [15]. This group typified the HCs that the government would have to upgrade as it implemented its national goal of 50% of HCs in Tanzania offering CEmONC services by 2020 [3]. Gairo and Melela HCs (publicly funded), and St. Joseph HC (representing a group of faith-based organizations) were already providing CEmONC and were included in the intervention group to study how CEmONC services could be strengthened. There were therefore 5 health centres that underwent the intervention. Mlimba and Mkamba HCs were randomly allocated to the control group from the remaining 5 publicly funded HCs that were already providing CEmONC services. Because of the intentional differences of intervention and control HCs, this study was designed as a before-after intervention where different funding and governance models could be compared. Comparisons of intervention and control HCs are mainly to detect secular trends that could potentially explain the before-after differences observed in the intervention HCs. The study was conceived and designed by the ASDIT team, a multidisciplinary group of researchers at the Tanzanian Training Centre for International Health (Tanzania) and Dalhousie University (Canada), partnering with Morogoro regional administration representing the varied interests of patients, health care providers, healthcare systems and policy makers. The Regional Medical Officer for Morogoro region (GM) was engaged as a public co-investigator and worked as a liaison between the district, regional and national authorities. To identify the facilities, most relevant research topics and meaningful outcomes, we worked with the public co-investigator and administered a leadership and management (L&M) survey customized to care providers. Through workshops and meetings, the project team regularly shared findings with the key stakeholders at district, regional and national levels to provide them with a broader understanding of the project, and the progress and outcomes. In order to develop a set of sound and scientifically derived interventions the project applied principles of operations research to identify and address operational factors that determine maternal and newborn health care in Tanzania [16, 17]. Using evidence-based science on the interventions that work, available material, financial and human resources the project blended medical-based, and leadership and managerial interventions (Fig 1) [18, 19]. Forty-two associate clinicians from the five intervention HCs were trained in teams for three months in CEmONC and anesthesia. Considering the national regulations, assistant medical officers (advanced associate clinicians) from these HCs were trained in CEmONC while clinical officers and nurse-midwives (associate clinicians) were trained in anaesthesia, postoperative care and care of the sick and premature newborn [15]. In Tanzania, assistant medical officers are licensed to perform surgery. The two years of training includes three months in general surgery and three months in obstetrics and gynaecology. The lack of internship program and inadequate supervision after graduation denies them the opportunity to acquire adequate surgical skills in obstetrics. This CEmONC training program was designed to strengthen surgical skills taking into consideration that they were expected to work independently in remote HCs. The curricula for CEmONC and anaesthesia were built on training programs for associate clinicians previously delivered at the St. Francis Referral Hospital [20]. To reinforce knowledge and skills, post-training activities included eHealth strategies, quarterly supportive supervision visits and continuous mentorship. The eHealth strategies included the offline eLearning modules and tele-consultation. For tele-consultation, care providers at the intervention HCs were also linked with obstetricians, a paediatrician and an anaesthetist based at St. Francis Referral Hospital. Mentorship and supportive supervisory visits were done every three months and included clinical audits and data collection for C-sections, maternal deaths and morbidities, fresh stillbirths, early neonatal deaths and methods of anaesthesia [15]. Mentorship activities focused on identified areas of substandard care. The project team designed capacity-building workshops in L&M and onsite mentorship geared at equipping health managers with essential knowledge and skills on leading change. These were basic principles and strategies in leading change that would improve performance and CEmONC services at their health facilities. The workshops were conducted in 2018 and 2021 and involved participants from 20 primary health facilities, i.e., the 5 intervention health centres and 15 satellite dispensaries, members of the district and regional health management teams specifically the district medical officers and other district health personnel. These dispensaries referred patients with medical complications to their respective health centres. Quality improvement plans developed after the 2018 “Big Results Now” (BRN) Star Rating assessment were used to mentor (onsite) the health facility health management teams and jointly address the gaps identified. The BRN star rating is a government system that measures the performance of various healthcare facilities aimed at improving quality of healthcare [21]. Since prior research has shown that engaging workplace teams in leadership development programmes is critical to success [22], onsite mentorship was a major component of the ASDIT intervention. All data were collected by the research team. The data on CEmONC services were collected concurrently with supportive supervisory visits as described above. The data included deliveries, types of anaesthesia, referrals and audit results of pregnancy adverse outcomes (maternal and perinatal morbidity and mortality). These were obtained from the working log books at each centre. Data on L&M were collected at baseline in 2018 and at the end of the study in 2021 using validated tools i.e., the “Big Results Now” Star Rating assessment and L&M survey tools [21, 23]. The BRN tool assesses the following domains: 1) health facility management (12 indicators); 2) use of facility data for service improvements (6 indicators); 3) staff performance management (5 indicators); 4) organization of services (8 indicators); 5) handling of emergencies/referral (7 indicators); 6) client focus (4 indicators); 7) social accountability (7 indicators); 8) facility infrastructure (14 indicators); 9) infection prevention and control (11 indicators); 10) clinical services (13 indicators); and 11) clinical support services (23 indicators) (S1 Table). The BRN system rates health facilities from 0 to 5 stars depending on the quality of services provided. The BRN star rating is based on the score of the minimum scoring domain and not the total or average marks. A score of 0–19% is graded no stars, 20–39% one star, 40–59% two stars, 60–79% three stars, 80–89% four stars, and 90–100% five stars [21]. The target of the government improvement initiative was to have 80% of primary health facilities rated with three stars or more by 2017–18. Three stars implied that the facility was performing at a minimum required standard and the domain scoring the least scored 60–79%. The “Leadership and Management” (L&M) survey primarily used Likert scales to assess data on care providers’ perceptions on L&M competencies, focusing on the following domains: team climate of facilities; staff role clarity; and job satisfaction (Table 1). CEmONC costs were collected from the health centres, Tanzania Medical Store Department and non-governmental organizations that had upgraded health centres for CEmONC services provision. Several uptake strategies were employed to enhance uptake of key interventions. These included engagement of key decision makers and the regional and council management teams throughout the project implementation period. The team conducted biannual national and regional stakeholders’ meetings and provided updates of the project during the quarterly regional maternal and child mortality audit meetings. Using Stata (version 15), one-way ANOVA and Chi-square tests were used to assess the impact of the intervention model by comparing outcomes during the baseline (July 2014—June 2016) and intervention (July 2016—June 2019). A one-way ANOVA test was used to determine the statistical differences of the mean monthly deliveries and mean score of the BRN key domains. Chi-square tests were used for the obstetric case fatality rates and proportions of justified C-sections within the intervention and control health centres. Confidence intervals were set to 95%. Ethical approval was granted by the National Institute for Medical Research (NIMR) of Tanzania with Ref. No. NIMR/HQ/R.8a/Vol.IX/1986, Dalhousie University Institutional Review Board and the Tanzania Commission for Science and Technology (COSTECH) with Ref. No. CST/ AD.69/227/2015. Permission to conduct research in these facilities was obtained from the regional and district local governments. Informed written consent for the L&M survey was obtained from all participants. Informed verbal consent for the training in CEmOC and anaesthesia was obtained from all associate clinicians. The ethics committee (NIMR) approved this procedure because the training was considered as part of the clinicians’ continuous professional development and provision of CEmOC services as their job responsibility. There was no need for patient’s consent because this study used anonymized patient data that was already being collected as part of the routine operation of the health centres. All methods were performed in accordance with the relevant guidelines and regulations.