Table 2.

Overview of research methods

Stakeholder mapping	The objective of this tool was to identify the key stakeholders who influence or are knowledgeable about HRH polices and their implementation. It was conducted in two countries as in Zimbabwe and Cambodia the topic was thought to be sensitive and not suited to group mapping. In Sierra Leone it was conducted at national level, whereas in Uganda, the exercise was done separately at national and district level. Male and female participants were drawn purposively from key constituencies (e.g. donors, Ministry of Health, Ministry of Finance, professional associations, NGOs, political stakeholders) and asked to brainstorm key stakeholders in HRH along two axes (influence and interest) and discuss changes through time
Document review	The objective of this tool was to describe the HRH policies, the reasons for their introduction, how they had been implemented, and any effects of the policy changes over the selected period, and the extent to which a focus on gender is included. The documents selected typically included national health strategic plans; national health workforce development plans; mid-term reviews of the health workforce development plans; health policy interventions on HRH, such as the incentive schemes; policies on remuneration (e.g. salaries, allowances, pensions, regulation of additional earnings); policy documents on recruitment (placement, promotion, retirement, and training of health workers); documents of organizations working in HRH; and academic studies or evaluations relating to health worker incentives. These documents were analysed using a thematic framework which was shared with the key informant interviews (KII)
Key informant interviews	Key informant interviews were undertaken to explore KI perceptions of health worker incentive policies, their evolution in the post conflict period, their implementation and effects. KIs from national down to local level were purposively selected, according to their knowledge of the focal topics. The interviews were semi-structured and focused on the following topics: Challenges for health worker attraction, retention, distribution and performance, post-conflict and at present and any differences for women and men health workers How policies had responded to these challenges Implementation experiences, constraints and lessons Their understanding of the effects of past policies Current thinking on reform options and priorities The interviews were tape recorded and noted after gaining permission from the participants. The interviews took place in a private place acceptable to the interviewee, such as their office. Thematic analysis using NVIVO (and ATLAS Ti in Uganda) was carried out on transcribed (and sometimes translated) texts. The analysis started from an agreed coding structure, shared with the document review, but with flexibility to alter according to the themes arising in the KII
Life histories (LH) with health workers	Life histories were deployed to explore health workers’ perceptions and experiences of their working environment, how it has evolved and factors which would encourage or discourage them from staying in post in remote areas and being productive. These were conducted with health workers meeting specific criteria (including gender, length of service in the area, to capture experiences of conflict and post-conflict periods) in selected health care facilities in the study areas using an open-ended topic guide. They were encouraged to produce visual aids, such as timelines. Life histories are arguably particularly conducive to gender analysis as participants are enabled to narrate in their own voices their experiences of work (and war or fragility) and how gender shaped their experiences (Ssali et al. 2016).The topic guide covered the following areas: How they became health workers Their career path since, and what influenced it, including the role of gender What motivates/discourages them to work in rural areas and across different sectors Challenges they face in their job and how they cope with them Conflict related challenges and how they coped Their career aspirations Their knowledge and perceptions of recent and current incentives. The life histories in the RinGs projects in Zimbabwe and Cambodia included a specific gender lens and included probes on issues relating to implementation of equal opportunities policies and legislation and how gender relations, expectations and norms at the household, organizational and socio-cultural levels affect health workers’ access to training, promotion and career advancement opportunities.The interviews and analysis followed the same procedures as for the KII.
Analysis of routine staffing data	The objective of this tool was to analyse trends in health worker availability, distribution, attrition, and performance during the post-conflict period. Existing human resource and selected service utilization data was collated from national, regional/district or facility sources (whichever were judged to be most reliable and complete). This was only completed for Cambodia and Sierra Leone; in Uganda, HRH data analysis was not included in the original study protocol, while in Zimbabwe it was included but not completed because of gaps in the HRH datasets. For the other two countries, data was collated for the defined periods using structured data extraction forms and analysed to describe the trends in health workers supply, distribution and output during the post-conflict period. In Cambodia, more extensive efficiency analysis was undertaken (Ensor, So et al. 2016)The indicators included numbers and trends over time for: staffing numbers for key cadres by gender and proportion of filled posts (where available); staff to population ratios; staff to output ratios; attrition rates (staff lost per year); and other relevant indicators, such as absenteeism
Health workers incentives survey	This was undertaken to understand the current incentive environment facing key health workers, their characteristics and the factors which motivate and demotivate them (to provide a quantitative measure to complement the analysis of the life histories). For this, a structured questionnaire was used to collect data from defined key cadres of health workers in face-to-face interviews. The study population included key cadres of health workers, with especial focus on those who are hard to retain. The sample size was based on the total number of workers in each category in the selected study areas, with a smaller proportion chosen for larger groups. Sampling was clustered by facility and non-random (small numbers available in each category and area meant that convenient sampling has to be used)The questionnaire focused on the following topics: Health worker characteristics Current earnings from different sources Current working patterns—public/private mix, other sources of income, dual practice, etc. Working hours and workload Perceptions of working environment and factors which motivate/demotivate and how these have changed over time Willingness to work or stay in rural areas The data were checked in the field, double entered, cleaned and analysed using SPSS or STATA software. Analysis was done according to cadre, region, type of facility, sector (in Zimbabwe) and gender