Understanding the factors influencing health-worker employment decisions in South Africa

The study utilized cross-sectional data collected in 2009 in KwaZulu-Natal, South Africa. The sample of HWs was selected purposively from three public hospitals, two private hospitals and one NGO hospital in the province, and consisted of professional nurses, staff nurses and nursing assistants.

Questionnaires were administered to HWs in urban and peri-urban areas. The researchers used a proportional sampling method in sourcing participants from each employment type (sector), meaning that the number of participants from a given sector was proportionate to the number of HWs working in that sector. The largest number of HWs were selected from the public sector (n = 430), followed by the NGO (n = 133) and private (n = 131) sectors. A total of 694 HWs participated in the study.

The survey instrument was based upon the Health Worker Incentives Survey Immpact Toolkit[19]. Questions included: background (demographic) information, cadre type, job and workplace perceptions, assessment of job satisfaction as well as the desire to change their current employer and or location.

Data analysis was conducted using the Statistical Package for the Social Sciences (SPSS Inc., version 18.0, Chicago, IL, USA). The primary unit of analysis in this study was the type of health-care employer (public, private or NGO). Bivariate analysis was used to compare responses from participants in each of the employer types, with both descriptive and inferential statistical methods used. Chi-square tests were conducted to detect levels of significance in demographic and categorical variables (e.g. sex, gender, level of education, employer type), while the one-way analysis of variance (F test) was carried out to test for levels of significance in continuous (numerical) variables, such as age. Participants rated various aspects of their working environment on a 5-point Likert-type rating scale (with 1 being low and 5 being high), and the F test was used to test for differences in ratings between participants from the three employer types.

Ethical approval was obtained from the University of KwaZulu-Natal Humanities Research Ethics Committee (Ethical Approval Number: HSS/0703/07). All participants involved in the study were informed about the nature of the study and what their participation would entail. Electronic and hard copies of the data are stored under lock and key at the Health Economics and HIV and AIDS Research Division office, and are only accessible by the researchers involved in the study who are bound by confidentiality agreements.

Table 1 shows that 90% (n = 620) of the total sample were female – this was consistent across employer types. No significant difference in age was observed between the three employer types, with the mean age being 36.96 years (standard deviation [SD] = 10.38) (F = 1.650, P = 0.193). The public sector (37%, n = 121) had significantly more professional nurses than the private (23%, n = 27) and NGO (24%, n = 22) sectors (χ ² = 14.764, P < 0.01). This is consistent with the national profile[17]. Significantly more personnel from the private sector (46%, n = 60) compared with the public (31%, n = 131) and NGO (29%, n = 39) sectors indicated that they were married and living with their spouse (χ ² = 21.960, P < 0.05). HWs in the public (51%, n = 207) and NGO (48%, n = 60) sectors were more likely than HWs in the private sector (21%, n = 26) to have served their current employer for five years or more (χ ² = 152.490, P < 0.001). The majority of the sample either resided in a private dwelling that was rented (37%, n = 247), or in a privately owned dwelling (32%, n = 216) that they were in the process of purchasing. For 75% of the sample, HW income was either the only income in the home or it made up at least 75% of total household income.

Participants were asked to rate various aspects of the conditions in which they work on a 5-point rating scale, with 1 being low and 5 being high. When asked to compare the level of their workload with that of colleagues in other sectors, workers in the public sector (mean = 4.55, SD = 0.88) rated their workload significantly higher than workers in the NGO (mean = 4.33, SD = 0.97) and private (mean = 3.81, SD = 1.20) sectors (F = 23.869, P < 0.001). Personnel from the public (mean = 4.16, SD = 1.03) and NGO (mean = 4.14, SD = 0.92) sectors reported significantly more stress at work compared with workers in the private sector (mean = 3.73, SD = 1.21) (F = 8.131, P < 0.001). In general, workers who felt more stressed at work were more likely to consider migrating (F = 5.251, P < 0.01). Also, workers at public facilities (mean = 2.09, SD = 1.25) rated the level of remuneration as significantly lower than workers in the NGO (mean = 2.23, SD = 1.24) and private (mean = 2.61, SD = 1.06) sectors (F = 7.868, P < 0.001). Moreover, personnel from the public sector (mean = 2.27, SD = 1.19) rated the standard of their working premises as much lower than workers from the NGO (mean = 3.24, SD = 0.98) and private (mean = 3.89, SD = 1.12) sectors (F = 106.806, P < 0.001). In addition, public HWs (mean = 2.11, SD = 0.94) rated the adequacy of available human resources available as significantly lower than the NGO (mean = 2.32, SD = 1.02) and private (mean = 3.04, SD = 1.00) sectors rated theirs (F = 43.474, P < 0.001). Finally, public HWs (68%, n = 280) were significantly less likely to have received any in-service training since being employed when compared with workers from the private (83%, n = 104) and NGO (87%, n = 115) sectors (χ ² = 27.631, P < 0.001). A similar, but more acute trend was observed when they were asked whether they received any such training in the past 12 months (Table 2).

HWs in the public sector especially experience poorer working conditions, as indicated by participants’ self-reports on stress, workload, level of remuneration, standard of work premises, level of human resources and frequency of in-service training. Our findings are supported by others, confirming that several salient push and pull factors play a role in HW plans and decisions to migrate[17, 20, 21].

HWs in the three sectors differed significantly as to whether or not they were considering alternative employment (χ ² = 22.925, P < 0.001). HWs at NGOs (73%, n = 91) featured most prominently, followed by workers in the public sector (59%, n = 230), while less than one-half (47%, n = 56) of workers in the private sector indicated that they were considering moving. Age significantly predicted the desire to leave, with those seeking alternative employment (mean = 35.23, SD = 8.94) being significantly younger than those who were not (mean = 41.93, SD = 11.91) (F = 19.544, P < 0.001). Nursing cadre (χ ² = 3.267, P = 0.514), sex (χ ² = 14.725, P = 0.142) and marital status (χ ² = 18.138, P = 0.112) did not significantly predict their desire to leave/stay. Several factors play a role in influencing the flow of HWs away from the public sector. As is evident from the results, there are various push factors associated with this sector, and a number of pull factors associated with the private sector (Table 3).

Although few studies have explicitly set out to determine the variation of push and pull factors across age groups, it is interesting to note that young HWs between the ages of 20 and 29 were more likely to cite deteriorating working conditions as a reason for wanting to leave South Africa, while older respondents (<60 years old) were less likely to identify heavy workloads as a reason for wanting to emigrate[22]. This is also the case in most developing countries where public-sector health facilities in general are characterized by poor infrastructure, management problems, unequal distribution of resources and high numbers of people seeking treatment, accentuated by the high prevalence of HIV and AIDS in most populations on the continent, and in South Africa in particular[17]. Elsewhere it has been argued that while countries in east and southern Africa may not be in a position to provide financial incentives, there are alternatives that have shown positive results in retaining HWs within their respective public health sectors[23]. Lesotho, Mozambique, Malawi and Tanzania have provided housing to their HWs whilst the provision of transport, childcare, food and employee support centres have all elicited positive HW retention outcomes[23].

Within South Africa, the need to address the problem of both internal and external movement of HWs led to the development and implementation of the Occupational Specific Dispensation for HWs in the public sector to improve their conditions of service and remuneration. Since 2007 when the Occupational Specific Dispensation strategy was developed, the translation of its objectives into practice has evidently not been satisfactory. In 2011 the Department of Health estimated an annual attrition rate of 25%, which excludes an additional 6% for retirement, death and change of profession[2]. This means that every year 25% of the potential workforce (not just new graduates) move from their current position, out of the South African public health sector[2].

The South African Department of Health Strategy 2010/11–2012/13, in priority area five, called for the re-opening of nursing schools and colleges along with the recruitment of HWs from countries with an excess of these professionals[24]. The Strategy went further in placing emphasis on examining the role of community HWs in the public health-care system[24]. The Human Resources for Health (HRH) Strategy for South Africa 2012–2016 further emphasizes the need to strengthen and professionalize the management of human resources and prioritize health workforce needs[2]. To realize this goal, the government set time horizons for the short (1 to 3 years), medium (3 to 5 years) and long (10 to 20 years) term, during which they plan on achieving specific objectives[2]. In expressing part of their strategic direction for the near to distant future, government has committed itself to the scaling up and revitalization of education, training and research, and to strengthening and professionalizing the management of human resources and prioritizing health workforce needs. Based on the findings of this study, as well as other studies[11‐13], these goals, if realized, could go some way in curbing the flow of HWs away from the public sector and to other countries. Based on these findings it will serve to attract others into the profession, and perhaps into the public sector. However, often documents propose effective strategies to achieve goals only for it to be discovered later that these goals were not realized due to the failure of authorities and stakeholders to effectively execute or implement these strategies.

To minimize HW attrition and the resultant negative effects, innovative efforts are required to address the causes of HWs’ dissatisfaction and to further identify the nonfinancial factors that influence HWs’ choices, especially given the inability to increase remuneration within a constrained fiscus. The South African Strategic Plan for HIV, tuberculosis and sexually transmitted infections 2012 to 2016 does call for the need to explore ‘innovative financing’ as a mechanism for raising additional resources for the public health care system[25]. Fryatt suggests a number of avenues that South Africa could explore to raise funds and whilst these resources will not solely be allocated to HRH, they could be used to fund efficacious attraction and retention strategies[25]. The government of Malawi responded to their HW crisis by increasing salaries by 50%, improving working conditions, re-enrolment of retired HWs and investment in training for HRH, all of which yielded positive results[26]. Ghana, for their innovative education and training strategies, were recognized with an award from the Health Worker Migration Policy Council[26]. Limited resources, inadequate education and career opportunities along with weak management systems leads to a shortage of HWs[26]. Further research is needed in those countries that have successfully addressed these issues. South Africa, through policy, has set out ways to address the HRH shortage, but implementation is key to ensure there are adequate staffing levels across cadres and in traditionally under-resourced areas.

The study is limited by its use of a cross-sectional design that entailed the gathering of data at only one point in time from nurses in selected health facilities in one province of South Africa. The use of proportional sampling does not allow us to generalize these findings to other types of health professionals. Other categories of health professionals, such as pharmacists, physiotherapists, dentists, occupational therapists and psychologists, who are crucial to the delivery of health-care services, were not included. The generalization of these findings to other nurses across the public, private and NGO sectors is further limited by the fact that the nurses participating in this study do not necessary represent nurses across the country working in these sectors.