Journal of Occupational

Introduction
Work is a broad concept and a poorly defined universal human activity. The satisfaction derived from one’s job could be described as an emotional response to the job conditions and a reaction to employees’ expectation of what the job should offer to them [1]. This construct has also been described as the degree of personal gratification an employee receives from their work and the pleasure or feeling of accomplishment derived from performing a job well [2]. Though definitions vary, in determining how an individual feels about their job, the interaction between factors intrinsic to the individual, factors intrinsic to the job, and factors extrinsic to both the individual and the job could be of considerable importance.
Employers are often preoccupied with how to make people happy at work. Happiness or job satisfaction at work is not an end per se, but it is a means that contributes to overall life satisfaction and employee mental health, thereby leading to organisational productivity. Thus, job satisfaction improves not only the quality of employees’ life, but also the quality and quantity of organizational productivity. For example, studies have reported satisfaction impacts among healthcare providers on patient satisfaction and the level of mortality at hospitals [3, 4]. Among all health measures, job satisfaction has been correlated with mental and psychological problems [5].
Several studies have demonstrated the relationship between job satisfaction and good mental health [6-9]. Although these studies have suggested the dependence of psychological health on job satisfaction, emerging evidence suggests that this relationship could be bidirectional, indicating that mental health could affect job satisfaction as well [10]. Since these studies have been mostly cross-sectional, the temporality of exposure could be difficult to ascertain. Research asserts that psychotherapy improved occupational outcomes and job satisfaction measures among depressed workers as clinical symptoms improved; thus, it must be stressed that an improvement in psychological health among workers could positively impact job satisfaction as well [10].
Healthcare workers are known to suffer more psychological distress than workers in other occupations [11]. This is consistent with population measures of psychological health, which showed around a third of doctors scored above the threshold of screening instruments [12]. This could be linked to the nature of clinical work and patients’ expectations. Work overload could also play a key role in exacerbating psychological distress among doctors in resource-poor settings with massive shortage of healthcare manpower. According to a Global Health Workforce Statistics report of the World Health Organisation (WHO), a recent increase occurred in brain drain among doctors from sub-Saharan Africa and mainly from Nigeria [13]. It is, therefore, necessary to investigate the job satisfaction level of medical doctors in Nigeria. This research aims to assess the level of job satisfaction among resident doctors in the selected teaching hospitals in Southern Nigeria and to determine the relationship between job satisfaction and psychological health.

Materials and Methods
This descriptive study was a web-based survey conducted among resident doctors in four selected teaching hospitals in Southern Nigeria. These hospitals included the University College Hospital ((UCH), Ibadan, the University of Calabar Teaching Hospital (UCTH), Calabar, the Ladoke Akintola University of Technology (LAUTECH), Ogbomosho, and the Lagos State University Teaching Hospital (LASUTH), Lagos. The hospitals were not selected randomly, but the selection was based on access to electronic mail addresses of resident doctors obtained with permission from the association executives of local resident doctors with the assurance of absolute confidentiality. The purpose of the survey was fully explained, and the potential participants were informed that the participation was voluntary. In addition, they were assured that the survey was totally anonymised, and that only aggregate data would be used for research purposes. Thus, the survey posed no potential harms to the participants as the data were well anonymised. This study was a full population survey, so it required no sampling or sample size calculation. All 643 resident doctors on the mailing lists were contacted via a link directing the participants to the survey questionnaire.
Although there are multidimensional instruments used to assess job satisfaction, the single-item global measure of job satisfaction gives some psychometric and practical advantages [14]. The single-item approach to job satisfaction measurement was adopted in this study by asking the respondents to rate their overall job satisfaction on a Likert scale of 1-5 with 1 = very dissatisfied, 2 = dissatisfied, 3 = neither satisfied nor dissatisfied, 4 = satisfied, and 5 = very satisfied. Thus, higher scores indicated greater job satisfaction.
The General Health Questionnaire (GHQ)-12 was utilized in this study to assess psychiatric morbidity in the study population. The GHQ-12 is a short and equally accurate version of the GHQ-60 developed by Goldberg for screening probable non-psychotic psychiatric morbidity in a population [15]. Two of the previously used methods for scoring the GHQ-12 were employed in this study, which included the most commonly used GHQ-scoring method (0-0-1-1) and the Likert scoring method (0-1-2-3). Sum scores of the GHQ-scoring method ranged from 0-12, while they ranged from 0-36 in the Likert scoring method. The Likert scoring method is less commonly used, but it offers some psychometric advantages by reducing data skewness [16]. For the GHQ-scoring method, a threshold cut-off of 2/3 was used to identify probable psychiatric cases, which was recommended as the most accurate method of screening [17, 18]. For the Likert scoring method, a threshold cut-off of 8/9 was adopted to identify probable psychiatric cases based on the scoring method [16]. The GHQ-12 has been previously validated in Nigeria with the sensitivity, specificity, and Cronbach’s α of 72%, 75%, and 0.70, respectively [19].
Data were managed using SPSS Statistics V.25.0 (IBM Corp. NY, United States). For the analysis of the responses from the GHQ-12, the software included data for only the respondents who answered all items in the GHQ when computing the GHQ and Likert scores. Thus, in estimating the relationship between job satisfaction and psychological health (the correlation and the chi-squared statistic), the number of the included samples was a little less than the total samples for estimating job satisfaction among the respondents. In addition, the reported findings for psychiatric caseness were related to the respondents who reported all items for the GHQ-12. The level of statistical significance was set at p < 0.05.
 
Results
A total of 195 doctors responded to the survey out of the 643 email addresses contacted, which
amounted to a response rate of 30.3%. Eight questionnaires were unusable and did not contribute data to the survey; thus, responses from 187 participants were analysed. The respondents’ mean age was 36.5 ± 5.4, while about 60.9% aged between 30-39, 58.1% were male, and 61.0% worked at the University College Hospital, Ibadan (Table 1). About three-quarters (76.4%) of the respondents were married, more than two-thirds of them were senior registrars (69.8%), and 83.2% had at least six years of work experience since graduation from the medical school. In addition, the mean number of years of work experience since graduation was 9.8 ± 4.2. The most commonly represented specialties included surgery (20.0%), internal medicine (18.2%), community medicine (15.3%), and family medicine (14.7%).
The distribution of the job satisfaction scores approximated normality, indicating the use of parametric analyses. The mean of the overall job satisfaction score was 2.7 ± 1.1. Only 50 (26.7%) respondents expressed they were satisfied (45; 24.1%) or very satisfied (5; 2.7%) with their job. Other categories of the responses included neither satisfied nor dissatisfied (55; 29.4%), dissatisfied (57; 30.5%), and very dissatisfied (25; 13.4%).

The mean score of the GHQ-12 using the Likert scoring method was 14.4 ± 7.4, and it was 4.0 ± 3.5 using the GHQ scoring method. The psychiatric caseness assessed by GHQ scoring cut-off was 91/159 (57.2%). Psychiatric caseness assessed by Likert scoring cut-off was 122/159 (76.7%).

There was a similar moderate and significant negative correlation between job satisfaction scores and scores from both scoring methods, in which Pearson’s correlation coefficient r = -0.510, p = 0.01 (the Likert scoring method), r = -0.507, and p = 0.01 (the GHQ scoring method). The relationships between job satisfaction scores and GHQ scores were illustrated in scatter plots in Figs. 1 and 2. Accordingly, upon an increase in the respondents’ job satisfaction scores, the scores of the GHQ-12 decreased. In this respect, r2 was approximately 0.25, implying that about 25% of the variations in the GHQ-12 scores could be explained by the linear relationship between the GHQ-12 scores and the job satisfaction scores (Figs. 1 and 2). This was similarly demonstrated in the chi-squared analyses among the three categories of job satisfaction, including satisfied, neither satisfied nor dissatisfied, and dissatisfied as well as the two categories of probable psychiatric cases, i.e. cases and non-cases (Tables 2 and 3). There was also a statistically significant negative relationship between job satisfaction and probable psychiatric caseness. Using GHQ score cut-off, only about a third (31.7%) of the satisfied respondents were probable psychiatric cases compared to more than three-quarters (77.9%) of the respondents who were dissatisfied with their jobs (chi-square = 23.90, p < 0.001,     Table 2). Likewise, using Likert score cut-off, more than half (58.5%) of the satisfied respondents were probable psychiatric cases compared to almost all respondents (92.6%) who were dissatisfied with their jobs (chi-square = 19.00, p < 001, Table 3).