“If I had known, I would have applied”: Poor communication, job dissatisfaction, and attrition of rural health workers in Sierra Leone

Background: Sierra Leone’s health outcomes rank among the worst in the world. A major challenge is the shortage of primary healthcare workers (HCWs) in rural areas due to especially high rates of attrition. This study was undertaken to determine the drivers of job dissatisfaction and poor retention among Sierra Leone’s rural HCWs. Methods: Interviews were conducted with 58 rural and 32 urban primary HCWs in Sierra Leone’s public health sector, complemented by key informant discussions and review of national policy documents. HCW interviews included (1) semi-structured discussion, (2) questionnaire, (3) card sort about HCW priorities, and (4) free-listing of most pressing challenges and needs. Sampling for HCW interviews was stratified purposive, emphasizing rural HCWs. Results: Among 90 HCWs interviewed, 67% were dissatisfied with their jobs (71% rural vs 52% urban) and 61% intended to leave their post (75% rural vs 38% urban). While working and living conditions and remuneration were significant factors, a major reason for rural HCW disenchantment was their inability to access worker rights, benefits, and advancement opportunities. This was caused by HCWs’ lack of knowledge about human resource (HR) policies and procedures, as well as ambiguity in many policies and inequitable implementation. HCWs reported feeling neglected and marginalized and perceived a lack of transparency. These issues can be attributed to the absence of systems for regular two-way communication between the Ministry of Health and HCWs; lack of official national documents with up-to-date, clear HR policies and procedures for HCWs; pay statements that do not provide a breakdown of financial allowances and withholdings; and lack of HCW induction. Conclusions: HCWs in Sierra Leone lacked accurate information about entitlements, policies, and procedures, and this was a driver of rural HCW job dissatisfaction and attrition. System-oriented, low-cost initiatives can address these underlying structural causes in Sierra Leone. These issues likely apply to other countries facing HCW retention challenges and should be considered in development of global HCW retention strategies. Data for this mixed-method cross-sectional study were collected in Sierra Leone in 2014–2015. Data were collected via in-depth interviews with primary HCWs at public-sector health facilities throughout the country (N = 90), key informant discussions with governmental and non-government stakeholders (N = 37), and a review of policy documents and electronic data tools (Fig. 1). Data collection took place prior to Sierra Leone’s Ebola outbreak. Design of health workforce retention study. Abbrev: MOHS = Ministry of Health and Sanitation, HRH = Human Resources for Health Key informant discussions were held with district and national government health officials, managers from non-governmental organizations (NGOs), staff from mission and private health facilities, and other relevant stakeholders. Discussions sought to shed light on systemic gaps undermining health workforce retention. HCW interviews elicited qualitative and quantitative data about their individual professional and personal experiences, challenges, factors influencing their intentions to stay or leave a health posting, and types of changes that would encourage them to stay at their current posts. Interviews included four components: a 30-question survey adapted from previous studies conducted in Uganda and Kenya [26], a semi-structured discussion in which the health workers described their personal experiences, a listing exercise where health workers were asked to free-list in order of priority a maximum of seven of their most important workplace challenges and seven things they would want the government to do to increase their desire to stay at their current posting, and a card sort activity in which health care workers were asked to rank-order 13 cards depicting factors affecting job satisfaction frequently referenced in the general literature. An additional file lists the 13 factors in the card sort and describes the method (see Additional file 2). Interviews were conducted at the health facility in a private place in the language that was most comfortable to the HCW (English, Krio, or both). No personal identifiers were documented, and written consent was obtained before the interview. Interviews lasted approximately 1.5 h and were audio-recorded. The card sort, free lists, and survey responses were documented on a data collection tool. The document review provided additional context to the findings from the HCW interviews. Reviewed documents included the Civil Service Code of Rules and Regulations, Human Resources for Health Policy, Human Resources for Health (HRH) Strategic Plan, Essential Healthcare Package, Primary Healthcare Handbook, and National Health Training Plan. The review also included the electronic and paper tools used to collect and manage HRH data at the district and national levels. We used maximum variation sampling and purposively sampled individuals across five strata, including type of setting, health facility level, health worker cadre, geographical region and district, and supervisory status (Table 1). Facilities were selected with the help of district health management teams (DHMT), and in each district, study investigators prioritized the most “hard-to-reach” health facilities, while attaining a balance of cadres and facility levels. Rural health facilities were oversampled because the study’s primary focus was the rural workforce. The study achieved regional balance, including all 13 districts and 6–8 chiefdoms within each district. Health worker interviews included all cadres involved in primary healthcare at public-sector health facilities (state-registered nurse (SRN), state-enrolled community health nurse (SECHN), community health officer (CHO), community health assistant (CHA), maternal child health aide (MCHA)). Sampling strata for healthcare worker interviews Only one HCW was interviewed at any single rural facility, and a maximum of two HCWs were interviewed at any hospital. Upon reaching each facility, interviewers explained the study to the site supervisor and the chosen HCW, who was then asked if he/she wanted to participate. No HCW refused to participate in the study. Convenience sampling was used for selection of key informants. For the review of key documents and tools, sampling was inclusive of all accessible national documents that were relevant to the health workforce and delivery of primary healthcare services. Given our sampling method and sample size, our study findings are not generalizable to the entire health workforce. Rather, our intention was to provide an in-depth examination and description of dissatisfaction and demotivation among a purposive sample of health care workers so as to better inform policies for improving health workforce retention in Sierra Leone. Audio recordings for the health worker interviews were translated into English where necessary and transcribed. Interview transcripts, national documents, and photos taken at health facilities were imported and managed in Atlas.Ti. We used thematic analysis to generate a nuanced description of health worker perspectives and experiences within the context of facility infrastructure and national policies [27, 28]. Inductive analysis was used to develop primary and axial codes which formed an initial code book. Codes were then modified and refined through “test-coding” of several transcripts. The resulting final codebook was then used with all transcripts. Coded text and images from all data sources were analyzed and synthesized to identify key themes in answer to the research questions. Preliminary results were shared with co-researchers in country, and their feedback was incorporated into the analysis. Quantitative data from card sorts, free-list rankings of challenges and desired interventions, and questionnaires were analyzed using Microsoft Excel and SPSS. The data analyses from the different methods were synthesized and triangulated [29]. Ethical approval for the study was obtained from the Sierra Leone Research and Ethics Committee, as well as the University of Washington Institutional Review Board.