Health worker attrition at a rural district hospital in Rwanda: a need for improved placement and retention strategies

Introduction The shortage and maldistribution of health care workers in sub-Saharan Africa is a major concern for rural health facilities. Rural areas have 63% of sub-Saharan Africa population but only 37% of its doctors. Although attrition of health care workers is implicated in the human resources for health crisis in the rural settings, few studies report attrition rates and risk factors for attrition in rural district hospitals in sub-Saharan Africa. Methods We assessed attrition of health care workers at a Kirehe District Hospital in rural Rwanda. We included all hospital staff employed as of January 1, 2013 in this retrospective cohort study. We report the proportion of staff that left employment during 2013, and used a logistic regression to assess individual characteristics associated with attrition. Results Of the 142 staff employed at Kirehe District Hospital at the start of 2013, 31.7% (n=45) of all staff and 81.8% (n=9) of doctors left employment in 2013. Being a doctor (OR=10.0, 95% CI: 1.9-52.1, p=0.006) and having up to two years of experience at the hospital (OR=5.3, 95% CI: 1.3-21.7, p=0.022) were associated with attrition. Conclusion Kirehe District Hospital experienced high attrition rates in 2013, particularly among doctors. Opportunities for further training through Rwanda’s Human Resources for Health program in 2013 and a two-year compulsory service program for doctors that is not linked to interventions for rural retention may have driven these patterns. Efforts to link these programs with rural placement and retention strategies are recommended.


Introduction
Developing countries, particularly in rural areas, suffer from a shortage and uneven distribution of human resources for health [1][2][3][4]. Globally, 50% of the population resides in rural settings but roughly 25% of doctors and 38% of nurses work in these areas [5].
Sub-Saharan Africa, which bears a quarter of the global disease burden [4] and has 63% of her population living in rural settings [6], only has 3% of the world's trained health workforce [4] with 37% of doctors and 49% of nurses and midwives working in the rural areas [7]. The same pattern exists in Rwanda where with 82% of the population residing rurally, only 12% of physicians and 42% of nurses serve in rural areas [8]. This dearth and disparity in the distribution of health care workers exacerbates health inequities in the rural areas [8][9][10].
Health care worker attrition is a large driver of the rural human resources for health crisis [5]. In 2010, the World Health Organization recommended policies for attraction, recruitment and retention of health care workers in the rural settings [5].
Recommendations included recruitment of health care workers with a rural background, inclusion of rural health needs in training curricula, and use of mandatory rural service [5]. Rwanda implemented some of these initiatives including a compulsory service program for newly graduate doctors that began in 2010, and required two years of service at district hospitals after receiving medical practice license.
Rwanda's third health sector strategic plan outlined initiatives to reduce the overall shortage of health workers, as well as improve their quality [11]. Improvement in management, rational deployment, equitable distribution and retention of health care workers were identified as necessary steps to ensure health workers were adequately available in the rural settings [11]. In 2012, through the Human Resources for Health program, Rwanda began the implementation of a national health worker training program to improve the quantity, diversity and quality of health workers [9]. Data collection and analysis: We retrieved attrition and individual characteristics data from human resources files, payrolls, contracts and employee's curriculum vitae. We described the study population and rates of attrition, defined as leaving employment any time during 2013, stratified by type of staff and staff demographics.
Because of the use of routinely collected data, data on some variables were unavailable for some staff; in such cases, we reported the reduced sample size. We used Fisher's exact test to Consent for participation was not applicable.

Results
As of January 1 2013, there were 142 salaried health care workers at Kirehe District Hospital (Table 1). Doctors represented 7.7% (n=11) of the staff and 60.6% (n=86) were nurses and midwives.
The majority of the staff were male (54.9%, n=78) and in time-  Table   2).

Discussion
Kirehe District Hospital experienced high levels of health care worker attrition in 2013, 31.7% overall and 81.8% for doctors. These rates were considerably higher than the only other study we identified exploring attrition at district hospitals in East Africa, which reported a 3.0% overall and 4.0% for doctors in Kenya [3]. While the Kenyan study might have had lower rates of attrition due to inclusion of district hospitals in both urban and rural settings, the high rural attrition rates in our study reflects the findings of a multicountry study in Africa and Asia noting low intentions of medical students to work in rural settings [12]. In a web-based survey in Rwanda, 60% of district health managers reported human resource challenges such as staff shortages and high rates of attrition as significant barriers in the health sector performance [13].
In our study, the high rates of attrition were linked to profession (doctor) and duration of employment (≤2 years). These individual factors, in addition to age, gender and marital status have been previously associated with attrition in low-and middle-income countries [1]. For high levels of doctor attrition in rural settings, factors such as high work load and burn out, poor living and working conditions, low salaries and limited career growth opportunities have been implicated [1,11,14].
Further, we hypothesize that the specific risk of losing doctors and individuals who have been at the hospital for a short period may be linked to the two national programs. First, the Human Resources for Health program expanded residency training opportunities for general practitioners in the same year we conducted this study [9].
While this opportunity for further training should ultimately improve the number and quality of clinicians available, it may have also contributed to the loss of doctors leaving rural areas for training in the urban settings, as similarly identified in the 2016 mid-term review of Rwanda's third Health Sector Strategic Planning [15]. The introduction of secondary referral hospitals in each of Rwanda's five provinces where graduate specialists from the Human Resources for Health program will be deployed to decongest tertiary referral hospitals, may improve access to this higher level care in rural areas [11]. In addition to decentralization of the specialists, strategies such as recruiting students of rural origin, including rural health needs in the training curricula and conducting training in rural facilities to increase awareness and concern for rural health challenges [1,2,5,12] could improve placement and retention in the rural areas post Human Resources for Health training.
Secondly, while the compulsory service program posts general practitioners to rural district hospitals for two years of service, without interventions to encourage retention, these doctors may leave rural settings after service completion. A paper reviewing the impact of compulsory service programs globally noted that while compulsory service programs increase physician presence in rural areas in about 70 countries implementing the program, high attrition rates of physicians after service completion was a major challenge [16]. The shortage of doctors in Rwanda that leads to their high demand in urban areas and private sector, poor working conditions, and lack of opportunities for other sources of income such as in private practice make rural areas less desirable to health workers [1,13,14]. Recurrent attrition after compulsory service may affect the quality of care delivered, continuity of chronic care by breaking up patient-provider relationships, and the morale of remaining staff due to increased workload [10,16]. We suggest linking compulsory service program with additional rural attraction and retention strategies such as deploying health care workers of rural origin to their home districts during compulsory service program, improving living conditions and offering professional development opportunities in the rural areas [1,2,5,8].
There are a number of limitations to our study. First, since we only included one rural district hospital that receives support from Partners In Health, our study may not be generalizable to the rest of