A cascade model of mentorship for frontline health workers in rural health facilities in Eastern Uganda: Processes, achievements and lessons

Background: There is increasing demand for trainers to shift from traditional didactic training to innovative approaches that are more results-oriented. Mentorship is one such approach that could bridge the clinical knowledge gap among health workers. Objectives: This paper describes the experiences of an attempt to improve health-worker performance in maternal and newborn health in three rural districts through a mentoring process using the cascade model. The paper further highlights achievements and lessons learnt during implementation of the cascade model. Methods: The cascade model started with initial training of health workers from three districts of Pallisa, Kibuku and Kamuli from where potential local mentors were selected for further training and mentorship by central mentors. These local mentors then went on to conduct mentorship visits supported by the external mentors. The mentorship process concentrated on partograph use, newborn resuscitation, prevention and management of Post-Partum Haemorrhage (PPH), including active management of third stage of labour, preeclampsia management and management of the sick newborn. Data for this paper was obtained from key informant interviews with district-level managers and local mentors. Results: Mentorship improved several aspects of health-care delivery, ranging from improved competencies and responsiveness to emergencies and health-worker professionalism. In addition, due to better district leadership for Maternal and Newborn Health (MNH), there were improved supplies/medicine availability, team work and innovative local problem-solving approaches. Health workers were ultimately empowered to perform better. Conclusions: The study demonstrated that it is possible to improve the competencies of frontline health workers through performance enhancement for MNH services using locally built capacity in clinical mentorship for Emergency Obstetric and Newborn Care (EmONC). The cascade mentoring process needed strong external mentorship support at the start to ensure improved capacity among local mentors to provide mentorship among local district staff.

The MANIFEST Project was implemented in three districts: Kamuli, Kibuku and Pallisa in Eastern Uganda from 2013 to 2015 [16]. The total population of the three districts was 1,106,100 [17]. The three districts had a total of 30 health facilities (27 Health Center IIs, two Health Center IVs and one general hospital). The project employed a comprehensive intervention that was comprised of two main components (community mobilisation and empowerment and health-systems strengthening) that have been described in detail elsewhere [16]. The health-systems component consisted of a three-pronged approach for improving the quality of maternal and newborn health services. The three main elements included strengthening leadership for maternal and newborn health (MNH) at district and facility level, motivation of health workers and mentorship. Mentorship was designed as a follow on from refresher trainings in Emergency Obstetrics and Newborn Care (EmONC), so as to maintain gains from the formal classroom teaching as well as strengthen competencies for EmONC for frontline health workers. This would be pivotal in delivering maternal and newborn health care (MNHC) equitably by enhancing access to high-quality care at the lowest referral health facilities, which was the overall goal of MANIFEST. Between 2013 and 2015, the MANIFEST project conducted a cascade mentorship model in 12 health facilities: four Comprehensive Emergency Obstetric and Newborn Care (CEmONC) and eight Basic Emergency Obstetric and Neonatal Care (BEmONC) health centres across three districts in rural Eastern Uganda. The cascade mentoring process was implemented in two phases: the initial mentoring of local mentors by central mentors and the actual mentoring process by local mentors. A total of 36 mentorship visits was carried out. This cascade mentoring process was driven by a theory of change that was modelled around the health-systems’ building blocks described by the World Health Organization [2]. The main aspects of the mentorship process are summarised below. A refresher training of health workers from August to October 2013 preceded the first mentoring phase. A total of 204 frontline health workers who offer maternal and newborn health services and members of the District Health Team members (69 in Kamuli, 52 in Kibuku and 83 in Pallisa) participated in the trainings. The training manual was adapted from the Advances in Labour and Risk Management tool (ALARM, 2008) and HBB + programmes recognised by Ministry Of Health for maternal and newborn care. Each training lasted five days and the content was tailored towards addressing the main causes of maternal and newborn death. Didactic lectures, group discussions and demonstrations were the main methods of training. Pre-and post-written assessments were done as a means of assessing the knowledge gained from the training. Selection of local mentors was done by the district health teams based on criteria that included work experience, performance in the pre- and post-test, qualification and participation in maternal and newborn health activities within the district. Initially 12 local mentors were selected per district to make a total of 36 local mentors. However, only 16 mentors remained active throughout the project. In October 2013, all the 36 selected potential mentors were invited for a further three-day orientation in mentorship. The training focused on the characteristics of a mentor, how to conduct a mentorship session, linkage of mentorship with supportive supervision, practical skills in emergency obstetric care, newborn resuscitation and management of intrapartum haemorrhage. There was a six-month lapse after the training and initiation of the first practical mentoring phase. Three to four mentoring teams were established per district. Each team was comprised of three to four members. Each team was assigned to a health centre and schedules and durations of mentoring sessions were drawn. A mentorship handbook was developed to guide the mentors through the mentorship process, in addition to a logbook that was to be used by the mentees to keep a record of activities done and any challenges encountered. Expert obstetricians and paediatricians (external mentors) conducted the mentoring of local mentors. The first mentoring phase lasted six months. The main purpose of initial mentoring sessions was to demonstrate how to conduct mentorship. The central mentors therefore led the first four sessions, while the local mentors led the last two. After each mentorship session, feedback was given, during which the mentorship process was assessed. After the first mentorship phase, a qualitative evaluation was conducted to assess progress, identify problems and suggest solutions. The evaluation showed that health workers were more confident in the maternal-care component, but less confident in newborn care. Use of the mentee logbook was also noted to be particularly poor. These findings provided a basis for redesigning the second phase of mentorship. The mentee logbook was redesigned, with more focus on newborn care, by reconstituting the central team of mentors to include a pediatrician. Additionally, a decision was made to establish a resuscitation corner where there was none, in addition to treatment units for sick and pre-term babies in each of the referral facilities. The resuscitation corner consisted of a mattress laid on a hard surface, newborn resuscitation equipment, an oxygen source, source of warmth (blanket, overhead bulb), infection-control facilities and treatment algorithms. The spaces for pre-term babies and sick newborn care consisted of equipment, drugs and supplies for advanced care (intubation, incubator, continuous positive pressure ventilation, alternative feeding options and specialised drugs). It was also noted that some of the health workers who had been selected as mentors were not suitable to act as mentors. Therefore, a decision was made to re-evaluate them and to maintain only those considered capable for the second phase of mentorship. During the second phase of mentorship, the health centres designated for mentorship were expanded to include facilities that provided basic emergency obstetric and newborn care. Two additional health centres of level III were selected from each of the three mentorship districts. Selection of health centres was done in consultation with the district health office. Health centres which handled large volumes of maternal and newborn cases were also included. The local mentors who had been involved in the mentorship activities were evaluated and categorised as ‘consistent and confident to mentor others’, ‘consistent but not confident to mentor others’ and ‘confident but constrained by hierarchical problems’. The mentors were evaluated in the broad areas of personal demeanour, attitude towards mentorship and confidence, suitability as a mentor in terms of knowledge/skills for EmONC, the actual conduct of mentorship, ability to teach and demonstrate skills for EmONC, ability to identify areas for mentorship at facility entry and during the visit, ability to give appropriate feedback, consistency and their own self-assessment as a mentor. The 16 mentors out of the 36 who were considered consistent and confident to mentor others continued with the mentorship process. After six months, the final evaluation for the mentorship was conducted. This was conducted as part of the end-line evaluation for the programme. Table 1 provides a summary of the sessions that were covered during the mentorship visits. Key sessions covered during the mentorship visits. Key ⱡ indicates that a knowledge and skills drill was done Newborn resuscitation corners were established first in the higher-level centres such as hospital, HC IV and III. In Uganda, the district health system is organised in four tiers: health centres of level II (HCII); III (HCIII); IV (HCIV); and hospital. HC II provides ambulatory care including antenatal care and delivery; HC III provides, in addition, inpatient services and laboratory diagnostics. HC IV offers caesarean operations and blood transfusion in addition to level III care services, while the hospital is a district referral facility and provides oversight and leadership to the lower levels of care that are HC IV, III and II. The resuscitation corners were established in each of the referral centres in anticipation of referred pre-term and sick newborn babies. Data for this paper were obtained through qualitative semi-structured interviews collected from July to October 2015. The interviews were designed to answer the specific research questions described in the introduction section. The data was collected until a point of saturation was reached at interview number 18; thereafter a decision was made to stop collecting any more data. The key informants for this study were selected purposively based on their role within the district and the health facility as local mentors under the MANIFEST project. They included district health-team members, health facility in-charges and maternity in-charges of different facilities. Seven were from Kamuli, six from Pallisa and five from Kibuku districts, respectively. In terms of professional training, they included a mix of nurses (six), midwives (six), medical officers (two), clinical officers (two) and environmental officers (two). Their years of service ranged from three to 35, with an average of 13 years. Lastly, two-thirds [12] of these were female. The interviews were recorded using a digital recorder and transcribed verbatim to maintain their initial meaning. Data was analysed manually using the thematic analysis technique [18]. Four of the authors (JA, RMA, MT and RMK) independently read the interview transcripts and coded them in relation to the study objective. All the other authors then reviewed the codes through an iterative process, which yielded agreement on the codes to be further developed. From the codes, themes and subthemes were developed consistent with the objectives of the study through a similarly iterative process that involved all the authors. An example of how the themes were developed is summarised in Table 2. Lastly, through a process of reflection, the external mentors, who include some of the authors (JA and JBK), made self-reflections, which focused on lessons learnt for future implementation. The process of data analysis is summarised in the analysis train in Table 2. Example of the analysis process from text to themes. Ethical approval to carry out the study was obtained from Makerere University School of Public Health Higher Degrees, Research and Ethics Committee. Before interviews, written informed consent was obtained from each key informant.

N/A

• Dima Health