Design
A retrospective, qualitative embedded case study of expressions of power during the implementation of a health service intervention in three catchment areas (‘implementing units’) of a rural South African district.
Setting and intervention
One of five districts in a northern province of the country, the study district contains farming areas, small towns and a significant ‘mineral-energy’ complex of mines and coal-fired power stations. At the time of data collection (2017), the district population was around 750,000, the overwhelming majority of whom relied on public health services. Health services are provided in five sub-districts through a mix of hospital, primary health and community based services (Table 1).
Table 1
District profile at time of evaluation (2017)
Population | ~ 750,000 |
Population density | 15.5 people/km2 |
% dependent on the public sector for health care | 92.3% |
Sub-districts | 5 |
Public health sector facilities | 1 Regional Hospital 7 District Hospitals 64 PHC facilities 14 Ward Based Outreach Teams |
Per capita annual PHC expenditure in public health system (2016/17) | R837 (US$58) |
The district was targeted, with others, by the national Department of Health because of high under-5 and neonatal mortality levels, considered to be retarding progress towards achievement of the Millennium Development Goals. In late 2013, a skilled facilitator, who had previously steered programme implementation as a senior manager in another province, was appointed to support the district. From 2014 onwards, he visited the district once a month, scaling down to every two months after three years.
Key elements of the facilitator-led intervention were new coordination structures, established in each of seven catchment areas (district hospital and surrounding facilities), referred to as Monitoring and Response Units (MRU); a system of real-time (48 hour) death reporting, review and response; outreach support from district clinicians and managers; and distribution of evidence-based guidelines. Participants in the MRU, which met monthly, were line managers (referred to as “drivers”), clinician managers (“experts”) and programme managers and information officers (“navigators”), spanning the district hospital, primary health care and community based services. In this regard, the MRU specifically sought to leverage coordinated action on MNCH within the catchment area, crossing official reporting lines which ran in parallel up to the district level. A key principle of the intervention strategy was that no additional funding or external support was to be sourced and that it would rely entirely on better use of existing resources.
By 2017, fairly steep declines in cause-specific under-five mortality, most notably for severe acute child malnutrition, had been recorded in the routine information system of the district, widely attributed by district actors to the effects of the MRU and associated support from district clinicians. The role of the MRU as an intervention in district governance and accountability and the plausible pathways through which it enabled these improved health outcomes are described elsewhere (11).
Sampling and data collection
Although the MRU was a deliberate system strengthening intervention, it was never set up with research or evaluation in mind. Anecdotal evidence prompted interest from an independent research team (the co-authors), who conducted a post-hoc evaluation three years after the start of implementation. In late 2016, the researchers began observing MRU meetings, reviewed available documents and interviewed the intervention facilitator. From the initial data gathered, key intervention stakeholders were identified and an intervention ‘programme theory’ developed, which formed the basis of further data gathering. In April 2017, the co-authors spent a week in the study district conducting a total of 44 interviews with district and sub-district stakeholders. A sub-set of interviews from three MRU catchment areas (hereafter referred to as ‘implementing units’) forms the basis of the analysis presented in this paper. The three implementing units were purposefully selected by district programme managers as representing the spectrum (rather than average) of MRU functioning (high, moderate and low) at the time of the evaluation. The subjective approach to selection was adopted as more objective criteria, such as performance data, failed to reveal any clear patterns.
The three district hospitals ranged in size from 80–143 beds, and were in referral relationships with 8–16 primary health care clinics. A total of 34 actors in the three selected implementing units was interviewed (Table 2). Interviews were set up through the hospital Chief Executive Officer (CEO) with the request to approach the key constituents of the MRU, namely senior and mid-level hospital managers (CEO, nursing service manager, medical manager, maternity and paediatric ward managers, dietitians), primary health care managers, information officers and community outreach team coordinators. The research team worked in pairs, and spent at least one full day in each hospital conducting interviews. Interviews were guided by the programme model, and elements probed included, amongst others, understanding, buy-in to and perceived functioning of the MRU meetings and processes.
Table 2
Actors interviewed in three implementing units
Level | N |
Hospital managers (senior and middle) | 20 |
Primary health care managers | 8 |
Community-based teams | 4 |
Other: emergency services, social worker | 2 |
Total | 34 |
Analysis of data
The original analysis of the full dataset followed the case study approach (18), namely, each unit was first analysed separately and then combined with the others in the district, which was then compared with other districts. A detailed description of the original analysis is described elsewhere (11). Subsequent, secondary analyses have explored specific mechanisms of change, drawing on theories of enabling environments (17) and governance (11). This paper is the last in this series, specifically focusing on actor power.
For the power analysis, interviews from the three implementing units were re-analysed, first by listening to the audio recordings (noting the emotional tone of the interview), followed by immersive re-reading of transcripts, then further coding of data into forms of power. ‘Power over’ was taken as the exercise of formal hierarchical authority in the implementation process; ‘power to’ as perceived knowledge and skills in completing work tasks; ‘power within’ as individualised expressions of autonomy or agency, namely “the ability to make things happen through their own actions” (1); and power ‘with’ as evidence of collective action (joint meetings across spheres, subjective reports coordinated action – formal and informal; linked or not to MRU). Manifestations of power and support for the MRU amongst individual senior, middle and frontline managers were also mapped in a modified stakeholder analysis (19) of each implementing unit. The three units are referred to in the analysis as ‘full’, ‘moderate’ and ‘low’ implementing units, respectively. As the subject matter could be considered politically sensitive, the names of district and catchment areas are deliberately withheld and identifying data kept to a minimum. In the four years since the evaluation was done, there has been turnover of staff in the three catchment sites and the likelihood of quotes being linked to individuals are minimal.
The Standards for Reporting Qualitative Research (SRQR) Checklist is provided as additional file 1 (20).