Practice environment is the single most important aspect influencing the well-being of nurses (Aiken et al. 2012; Coetzee et al. 2013). Within the practice environment, the characteristic of manager’s ability, leadership and support for nurses was the most influential on job satisfaction (Lambrou et al. 2014); so, also it was found that nurse managers influenced the retention of RNs (Mokoka et al. 2010:8). However, the precise relationship between leadership, job satisfaction and intentions to leave among RNs working in hospitals in the North West (NW) and Free State (FS) provinces remains unclear.

The main research questions in this study were:

A quantitative, descriptive cross-sectional survey design was used to investigate the relationship between leadership, job satisfaction and RNs’ intention to leave hospitals in the FS and NW provinces. A cross-sectional survey design allowed the researcher to examine RNs’ experiences of leadership, job satisfaction and intention to leave simultaneously at one point of time (Burns & Grove 2009:695).

This research focused on private and public hospitals in the FS and NW provinces. The population for this research study included all RNs in the medical and surgical wards of Level 1 three public hospitals in the FS province, Level 1 two hospitals in the NW province and three private hospitals from each of the selected private hospital groups in the FS and NW provinces. RNs in adult medical and surgical units participated in this study. An all-inclusive sample was used for RNs in medical and surgical units. A total of 680 questionnaires were distributed in the hospitals, and 204 RNs completed the questionnaire, amounting to a response rate of 33.3%. The total sample comprised 204 RNs (n = 204) from the FS and NW provinces, of whom 100 RNs worked in public hospitals and 104 RNs worked in private hospitals.

Inclusive criteria implied that RNs participating in this study had to be:

Data collection followed two processes to accommodate the dual (private and public) health-care system in South Africa. Data were collected using the self-administered questionnaire. In the private hospitals, a hospital employee (nurse) was appointed by the management of the hospital to manage data collection under supervision of a project manager. The project manager orientated the fieldworker for the RN4CAST programme and trained the fieldworker in the distribution and collection of the questionnaires. The fieldworker delivered the questionnaires to all nurses in the selected wards. Nurses returned the completed questionnaires to the fieldworker within one week. In the public hospitals, the project team delivered the questionnaires to all RNs in the selected wards on the morning of data collection and collected the completed surveys approximately six hours later.

The self-completion questionnaire was divided into four sections:

Section A enquired about your job, focusing on the practice environment of the RNs, and included the Practice Environment Scale of the Nurse Work Index (PES-NWI), questions related to job satisfaction, intention to leave and the Maslach Burnout Inventory (MBI).

Section B focused on quality and safety. In this section, RNs were asked to respond to issues related to safety and quality of care to patients delivered in their specific units.

Section C enquired about the nurses’ most recent shift at work in hospital. This section focused on questions related to the work schedules of RNs, nurse-to-patient ratios and details about the RNs’ most recent shifts.

Section D asked questions about ‘you the participant’. In this section, the demographic characteristics of RNs were explored, including questions related to age, gender and level of education (Sermeus et al. 2011).

In this study, the researcher only used the Nurse Manager’s Ability, Leadership and Support of Nurses subscales of the PES-NWI to measure leadership, overall job satisfaction and intention to leave (from section A of the questionnaire), and demographic data (from section D of the questionnaire).

Both descriptive (means, standard deviations and percentages) and inferential statistics (Spearman’s rank order correlations, t-tests, Cronbach’s alpha and statistical significance) were used to analyse the data. In the RN4CAST programme, the raw information was entered and captured by computer software EpiData by the North-West University’s (NWU’s) Statistical Consultation Services (SCS). The data were analysed using the computer software statistical programme for social sciences (SPSS) version 16.0. (SPSS Inc, 2009). An analysis of the data was done in collaboration with a statistician from SCS.

The leadership subscale of the revised PES-NWI consists of four items, namely: 1) supervisors support nurses; 2) a nurse manager is a good manager and a leader; 3) a nurse manager gives praise and recognition for a job well done; and 4) a nurse manager supports nurses in decision-making, even if it entails conflict with physicians. In international studies, Cronbach’s alpha coefficient for this subscale ranged between 0.63 and 0.84 (Bruyneel et al. 2009:205). In the national RN4CAST studies, the Cronbach’s alpha coefficient for this subscale was 0.86 (Klopper et al. 2012:690). In this study, the Cronbach’s alpha coefficient was 0.71, which means this could be accepted as being a reliable subscale to measure the variable of leadership.

Nurses’ overall job satisfaction was measured with a single item that ranged from 1 (very dissatisfied) to 4 (very satisfied). Published reliability coefficients for single-item overall job satisfaction ranged from 0.70 to 0.80 (Sermeus et al. 2011:4; Wanous, Reichers & Michael 1997). Nine specific aspects of job satisfaction (work schedule flexibility, opportunities for advancement, independence at work, professional status, wages, educational opportunities, annual leave, sick leave and study leave) were measured that ranged from 1 (very dissatisfied) to 4 (very satisfied). In addition, intent to leave was measured on a 2-point scale as ‘yes’ or ‘no’.

The validity of the questionnaire was determined by an international pilot study done in Belgian hospitals during 2009. The predictive validity of the instrument used in the International Hospital Outcome Study (IHOS) was used in preparing the nurse survey questionnaire that was used in the RN4CAST project. The sample of 179 nurses had response rates that ranged from 67% to 79% across hospitals. Therefore, the IHOS Nurse Survey Questionnaire, used in the RN4CAST project, was a strong and psychometrically sound instrument (Bruyneel et al. 2009:202). The RN4CAST has been validated for allowing measurement, evaluation and comparison of nursing work environment factors that influence the work force (Bruyneel et al. 2009:203).

The demographic information of the participants is presented in Table 1.

Table 1 shows that out of 204 RNs, 94% (n = 193) were females and 3.9% (n = 8) were males. This correlates with the SANC statistics indicating that FS had 7009 female RNs and 995 male RNs, while the NW province had 8145 female RNs and 1091 male RNs (SANC 2014). The majority of RNs (74%; n = 158) had diplomas in nursing. Of the 204 RNs, 34.7% (n = 67) were 40–49 years old. Of the RNs, 22.1% (n = 45) worked for 0–5 years, while 17.7% (n = 36) had worked for 6–10 years; 35.8% (n = 73) had worked for less than five years in the same hospital, and only 1% (n = 2) of RNs had worked for more than 36 years.

Out of the 204 RNs, 71.0% (n = 142) agreed that their supervisors supported nurses, as shown in Table 2. Most respondents (75.7%; n = 153) agreed that their nurse managers were good leaders and good managers, and (65.8%; n = 131) agreed that their nurse manager will support the nurses’ decisions. However, with regard to nurse managers giving praise and recognition for a job well done, 55% (n = 112) disagreed with this item.

Leadership in medical and surgical units in the FS and NW provinces was reportedly effective, except that nurse managers did not praise nurses and did not recognise a job well done. However, the mean for the entire leadership subscale was 2.89 (SD = 0.63), which indicates that the requisite features of this subscale were present in the practice environments (Lake 2002).

Table 3 indicates that more respondents were satisfied 70.5% (n = 141) with their current jobs compared with the respondents who were dissatisfied 29.5% (n = 59). These findings were in line with national data, indicating that 32.2% of South African medical and surgical RNs experienced dissatisfaction with their jobs (Coetzee et al. 2013:162). Positive findings confirm that most participants (82.4%; n = 164) were satisfied with their independence at work and 81.4% (n = 162) were satisfied with their work schedule flexibility. Dissatisfaction was reported based on wages (50%; n = 101), opportunities for advancement (40.1%; n = 79) and educational opportunities (34.1%; n = 67).

Most respondents were satisfied with their annual leave (79.4%; n = 158) and with their sick leave (71.6%; n = 141), but 40.9% (n = 81) were dissatisfied with their study leave.

Most respondents (81.3%; n = 161) were satisfied, while 18.9% (n = 37) were dissatisfied with nursing as a career of choice. This finding might explain the high incidence of job dissatisfaction because some respondents might have experienced job dissatisfaction because they did not enjoy nursing.

As many as 46.1% (n = 94) of the respondents indicated that, as a result of job dissatisfaction, they intended leaving their jobs in the current hospitals within the next year. Only 29.5% (n = 59) of the respondents were dissatisfied with their jobs and 18.1% (n = 37) were dissatisfied with their career choices. However, 46.1% (n = 94) of the respondents intended leaving their current employment positions within the next year as a result of job dissatisfaction.

Respondents, who intended to leave their places of employment within the next year as a result of job dissatisfaction, were asked what type of work they would seek. This item was measured on a three-point scale: ‘Nursing in another hospital, nursing, but not in a hospital or a non-nursing job’. Only 28% (n = 26) of these respondents would continue nursing in a hospital, 49.5% (n = 47) would continue nursing but not in a hospital setting and 22.4% (n = 21) would seek non-nursing employment. The finding indicates that 46.1% (n = 94) of the respondents intended leaving their current hospital jobs within the next year as a result of job dissatisfaction and only 28% (n = 26) would continue to work in hospitals.

Another question in relation to respondents’ intentions to leave their current jobs were asked, namely: ’If you were looking for another job, how easy do you think it would be for you to find an acceptable job in nursing?’ This item was measured on a four-point scale ranging from 1 (very difficult) to 4 (very easy). Most respondents (72.5%; n = 148) felt that it would be easy to find an acceptable job in nursing, while 22% (n = 45) felt that it would be difficult.

The correlation matrix in Table 4 presents the leadership subscale of the PES-NWI-revised in relation to overall job satisfaction and the nine aspects of job satisfaction. The p-value presents the statistical significance of the variables, whereas the correlation coefficient presents the magnitude, direction and strength of the relationship between the variables.

There was a large positive relationship between leadership and overall job satisfaction of RNs in their current hospitals (r = 0.47; p = 0.000). Furthermore, all nine aspects of job satisfaction showed a medium positive relationship (r = 0.33 – 0.44; p = 0.000).

Overall job satisfaction had a large positive relationship (r = 0.46; p = 0.000) with opportunities for advancement and a medium positive relationship (r = 0.32–0.44; p = 0.000) with all other job-related aspects.

The relationship between nurses’ intention to leave and the variables, leadership and job satisfaction, is visually presented by means of a 2×4 cross-tabulation. The reason for this is because intention to leave is measured on a two-point scale (nominal data) while leadership and job satisfaction are measured on a four-point scale (ordinal data).

In Table 5, the relationship between leadership items and intention to leave is presented with percentages, frequency (in brackets) and statistical significance.

There is a small practical significant difference (d = 0.19) between the intention to leave and a supervisor who supported nurses. Table 5 indicates that RNs who intended to leave their current hospitals were more likely to disagree (37.4%; n = 34) that supervisors supported nurses, compared with RNs who intended to stay (22.4%; n = 24).

There is a medium practical significant difference (d = 0.26) between intention to leave and a nurse manager who was reportedly a good manager and leader. Table 5 demonstrates that RNs who intended leaving their current hospitals were more likely to disagree (32.7%; n = 30) that their nurse managers were good managers and leaders, compared with RNs who intended to stay (17.6%; n = 19).

There was a medium practical significant difference (d = 0.38) between intention to leave, and praise and recognition for a job well done. Table 5 shows RNs who intended to leave were more likely to disagree (63.0%; n = 58) that they received praise and recognition for a job well done than RNs who intended to stay (49.1%; n = 53).

There is a medium practical significant difference (d = 0.27) between intention to leave and a nurse manager who backed up the nurses’ decision-making, even if there was conflict with physicians. RNs who intended to leave were more likely to disagree (46.7%; n = 43) that their nurse manager supported them in decision-making, even if the conflict was with the physician, compared with RNs who intended to stay (23.6%; n = 25).

Table 6 portrays the relationship between RNs’ intention to leave and overall job satisfaction.

There is a large practical significant relationship between RNs’ intention to leave and their overall job satisfaction (d = 0.50).

This is presented in Table 6 which shows that RNs who intended to leave their current hospitals were more dissatisfied (50.5%; n = 47) with their current jobs than RNs who intended to stay (11.3%; n = 12).

There was a medium practical significant difference between intention to leave and opportunities for advancement (d = 0.41). Table 6 shows that RNs who intended to leave their current hospitals were more dissatisfied (59.8%; n = 55) with the opportunities for advancement compared with RNs (23.1%; n = 24) who had no intention to leave their current hospitals.

There was a medium practical significant difference between intention to leave and independence at work (d = 0.39). RNs who intended to leave their current hospitals were more dissatisfied (31.2%; n = 29). with independence at work than RNs who intended to stay (5.7%; n = 6).

There was a medium practical significant difference between intention to leave and wages (d = 0.43). Table 6 indicates that RNs who intended to leave their current hospitals were more dissatisfied (67%; n = 63) with wages compared with RNs who intended to stay (34.6%; n = 37) at their current hospital.

There was a medium practical significant difference (d = 0.33) between intention to leave and educational opportunities. Table 6 shows that RNs who intended to leave their current hospitals were more dissatisfied (50%; n = 46) with educational opportunities compared with RNs who intended to stay (20.2%; n = 21).

There was a medium practical significant difference (d = 0.31) between intention to leave and satisfaction with annual leaves. Table 6 shows that RNs who intended to leave their current hospitals were more dissatisfied (31.8%; n = 29) with annual leaves compared with RNs who intended to stay (11.2%; n = 12).

There was a medium practical significant difference between intention to leave and sick leave, evidenced by a d-value of 0.40. As shown in Table 6, RNs who intended to leave were more dissatisfied (41.6%; n = 37) with sick leaves compared with RNs who intended to stay (17.8%; n = 19) at their current hospitals.

There was a medium practical significant difference (d = 0.36) between intention to leave and satisfaction with study leave. Table 6 shows that RNs who intended to leave their current hospitals were more dissatisfied (55.4%; n = 51) with study leaves compared with RNs who intended to stay (28.6%; n = 30).