Learning from the experience of maternity healthcare workers in Malawi: A qualitative study leading to ten low-cost recommendations to improve working lives and quality of care

Background: In Malawi there are too few maternity healthcare workers to enable delivery of high quality care to women. These staff are often overworked and have low job satisfaction. Skilled maternity healthcare workers are essential to improve outcomes for mothers and babies. This study focuses on understanding the working life experience of maternity staff at district hospitals in Malawi with the aim of developing relevant low-cost solutions to improve working life. Methods: A qualitative study using semi-structured interviews was undertaken in three district hospitals around Malawi’s Capital city. Thirty-one staff formed a convenience sample, purposively selected to cover each cadre. Interviews were recorded, transcribed and then analysed using Interpretative Phenomenological Analysis complemented by Template Analysis to elicit the experience of maternity staff. Results: Staff describe a system where respect, praise and support is lacking. Many want to develop their skills, however, there are barriers to advancement. Despite this, staff are motivated; they are passionate, committed professionals who endeavor to treat patients well, despite having few resources. Their ‘superdiverse’ background and experience helps them build resilience and strive to provide ‘total care’. Conclusions: Improving working lives can improve the care women receive. However, this requires appropriate health policy and investment of resources. There are some inter-relational aspects that can be improved with little cost, which form the ten recommendations of this paper. These improvements in working life center around individual staff (respecting each other, appreciating each other, being available when needed, performing systematic clinical assessments and communicating clearly), leadership (supportive supervision and leading by example) and the system (transparent training selection, training being need driven, clinical skills being considered in rotation of staff). To improve working lives in this way will require commitment to change throughout the health system. Thus, it could help address preventable maternal and newborn deaths. This qualitative study used one-to-one interviews with HCWs and combined two approaches to data analysis to allow a powerful picture of experience to be formed. Interpretative Phenomenological Analysis (IPA) harnesses the lived-experience in an in-depth, bottom-up approach using solely the data from participants. This approach was employed to identify core themes in an intensively-analysed sub-sample of the interviews [12]. Template Analysis(TA) [13] is a top-down approach, which was used to extend and develop these themes across the remainder of the dataset. This combination has been employed in previous studies [14] and benefits from the combined strengths of each approach. TA complements IPA as a flexible means of developing and transferring the coding structure within a larger sample [13, 15]. The study took place near Malawi’s Capital. Participants were recruited from three government hospitals; a district referral center, a district hospital and a community hospital. The district referral center had approximately 15,000 deliveries annually. Care was delivered by consultant and trainee obstetricians, general practitioners, clinical officers, degree level or registered nurses, diploma level nurse midwife technicians, trained nursing auxiliaries and untrained patient and hospital attendants. The district hospital, with approximately 3,700 deliveries annually, had no doctors who deliver obstetric care but had the other cadres. The community hospital with approximately 4,700 deliveries per year had no doctors or nursing auxiliaries. IPA requires participants to have a shared experience, to enable exploration of the common or conflicting ideas within and between cases [16]. Here, the common perspective was working in a government hospital in Malawi. A sampling technique of convenience was used to access those staff who were available when the researcher was present. This was complemented by a purposive approach to ensure different cadres of staff were represented. A sample of six to nine participants was desired for the IPA element of the study as this is the volume of cases for which we felt in-depth experiential analysis was feasible. Beyond that, we wanted to gain a broader perspective of the range of staff and also allow staff who wanted to participate to share their stories. We determined that approximately 10 interviews per site would allow both of these goals to be met. Following ethical approval from the Universities of Malawi and Birmingham, HCWs of all cadres were invited to participate in the study. After obtaining written informed consent, interviews were arranged with staff at a time convenient to them and a unique identifier (pseudonym) allocated. Semi-structured interviews lasting 30–90 min were carried out using a topic guide (Additional file 1), recorded and then transcribed. Participants were invited to receive their transcript, and several requested this, although only one made minor alterations. IPA required a detailed analysis of a small number of cases [12]. Nine cases, with the richest experiential data, spread across sites and cadres of staff were selected. These transcripts were read, re-read, then coded by hand by AMe and in part by ML. Coding focused upon capturing the meanings of important work-related experiences, from the respondents’ perspective. The research team then reviewed the emerging themes and feedback was sought from participants. These themes formed a ‘template’ for the second phase of analysis. This was carried out independently of the IPA, allowing the analysis to be grounded in the lived experience, but also to cope with the volume of data collected. This template was then applied to nine interviews using the qualitative software NVIVO version 10. The sub-themes were modified to incorporate new ideas, before being applied to the remaining dataset. During the application to the remaining dataset, no further themes were added. Each coded theme was explored further. The data was analysed by understanding the distribution of codes across the data. The relationships between themes were then explored. This was carried out by drawing out key ideas from each case and creating individual ‘maps’ of the key themes. Ideas that corroborated or were polarised were identified and considered across cases in addition to within cases this provided the opportunity to develop the contents within each theme more fully [17, 18]. AMe undertook the interviews and analysis. As a medical doctor with a background in obstetrics and gynaecology this PhD student based in the UK, brought a clinical perspective to the analysis. ML provided supervision and triangulation on the developing analysis from the perspective of phenomenological psychology.