Prevalence and determinants of depression, anxiety and stress among psychiatric nurses in Ghana: a cross-sectional study

The study was a cross-sectional study conducted in three public psychiatric hospitals: Accra Psychiatric Hospital, Ankaful Psychiatric Hospital, and Pantang Hospital between March 2020 and May 2021. Accra Psychiatric Hospital is located in Adabraka in the Greater Accra Region of Ghana. The hospital was built in 1904 and commissioned in 1906 with a bed capacity of 200 patients. Currently, the hospital has a bed capacity of 600 patients. It is the oldest and most populated psychiatric hospital in Ghana [26]. Ankaful Psychiatric Hospital was built outside the nation’s capital after constructing the Accra Psychiatric Hospital. It was established in 1965 with a bed capacity of 500, but currently, the hospital has a total bed capacity of 311 [27]. The Pantang Hospital is the largest psychiatric hospital commissioned in 1975. It is a 500-bed facility located close to the Pantang village, 1.6 km off the Accra-Aburi Road and 25 km from Accra. The hospital is used as a training centre for nursing and medical students all over the country and beyond. The hospital also operates a polyclinic, maternity, child welfare clinic and an eye clinic that serves the general population [28].

The population for this study included nurses who work in the three public psychiatric hospitals in Ghana. The study population comprised 993 nurses working in three public psychiatric hospitals in Ghana. This involved 462 nurses from Accra Psychiatric Hospital, 210 nurses from Ankaful Psychiatric Hospital, and 321 from Pantang Hospital. A probability proportionate to the size and simple random sampling technique were used to select participants.

The sample size was determined by applying Miller and Brewer’s formula [29]. It states that at 95% confidence level; \(\mathrm{n}=\frac{N}{\left(1+N\ (a)2\right)}\). Where, n- desired sample size, N- target population, a- level of statistical significance of 0.05, 1-is a constant.

Therefore, the sample size, \(\mathrm{n}=\frac{993}{\left(1+993(0.05)2\right)}=285.139=285.\)

The calculated sample size of 285 was increased by 10% to 314 to ensure that samples were not lost during data collection and cleaning and increased the statistical power [30].

The sample size for each hospital was determined from the calculated sample size of 314, proportionate to the population of each hospital.

The sample size for Accra Psychiatric Hospital (N = 462)

The sample size for Ankaful Psychiatric Hospital (N = 210)

Sample size for Pantang Hospital (N = 321)

In all, 146 nurses were recruited from the Accra Psychiatric hospital, 102 from the Pantang hospital and 66 from the Ankaful psychiatric hospital.

After the determination of the sample size for each hospital, a simple random sampling procedure was used to obtain the required number of respondents from each hospital. The sample frame of the population for each hospital was determined by getting a list of all professional nurses working in that hospital from the nursing administration. A random number generator was used to generate random numbers and registered the name in the sample frame corresponding to the numbers to constitute the sample for that particular hospital. This was continued until the required number was met.

Three research assistants with bachelor’s degree in mental health nursing were recruited and trained to assist in the data collection process. The training involved ensuring proper administration of the questionnaire and ethical procedures including measures to ensure respondent’s confidentiality and anonymity during data collection. The training lasted for 2 h. The data were collected by the first author and research assistants at the three hospitals. The respondents filled in the self-administered questionnaires and returned them within 1 week. Three hundred and eleven questionnaires were returned (99.0% response rate) and entered into the Statistical Package for Social Sciences for analysis.

Data were collected using a self-administered questionnaire. The instrument comprised 58 items which were divided into four sections; the first section focused on the socio-demographic characteristics of the respondents (e.g., “Gender”, “Age”, “Marital status”, “Educational status”, “Name of the hospital”, “Number of years spent in the nursing profession “, “Department/Ward”, “Monthly income”, and “Religion”).

The second section focused on the prevalence of depression. Beck’s Depression Inventory (BDI) was used to assess depression. The BDI is a 21-item self-reporting scale on a 4-point scale: 0- (Never- do not apply to me), 1- (Sometimes- applied to me to some degree), 2- (Often- applied to me to a considerable degree), 3- (Almost always- applied to me very much). Examples of some of the items are “I am so sad and unhappy that I can’t stand it”, “I feel I am a complete failure as a person”, and “I blame myself for everything bad that happens”. It has a minimum score of 0 and a maximum score of 63. The score of 0–9 indicates (minimal depression), 10–16 (mild depression), 17–29 (moderate depression) and 30–63 (severe depression) [31]. A cut-off score of 9 indicates a normal depression level. In this study, Cronbach’s alpha reliability coefficient for BDI was 0.91, demonstrating good internal consistency.

Anxiety was assessed using Beck’s Anxiety Inventory (BAI) in the third section. It is a 21-item self-report scale that is used to evaluate anxiety symptoms among adults on a 4-point Likert scale, which ranges from 0- (Not at all), 1- (Mildly- but it didn’t bother me much), 2- (Moderately- it wasn’t pleasant at times), 3- (Severely- it bothered me a lot). Some example items are “Wobbliness in legs”, “Unable to relax”, “Fear of worst happening”, and “Hands trembling”. A score of 0–7 indicates (low anxiety), 8–15 (mild anxiety), 16–25 (moderate anxiety) and 26–63 (severe anxiety). A cut-off score of 7 indicates a normal anxiety level [32]. The Cronbach’s alpha reliability coefficient for BAI was 0.91, indicating good internal consistency.

The fourth section of the instrument determined the prevalence of stress using the Perceived Stress Scale-10 (PSS). PSS is a 10-item scale used to assess respondents’ perception of stressful experiences over the previous month. Items on the scale are rated on a 5-point Likert scale which ranges from 0- (Never), 1- (Almost Never), 2- (Sometimes), 3- (Fairly Often), and 4- (Very Often). Some example items are “Upset because of something that happened unexpectedly?”, “You could not cope with all the things that you had to do?” and “Things were going your way?”. The scores were calculated after reversing the positive item’s score. It has a minimum score of 0 and a maximum score of 40; a high score indicates greater stress. Six out of the ten items of the PSS-10 are considered negative (1,2,3,4,5,6), and the remaining four are positive (7,8,9,10). A cut-off score of 13 indicates normal stress levels. A score of 0–13 represent (low stress), 14–26 (moderate stress) and 27–40 (high perceived stress) [33]. The Cronbach’s alpha reliability coefficient for PSS in the current study was 0.74, demonstrating good internal consistency.

Although there is no information regarding the validation of BDI, BAI and PSS in Ghana [34], these scales have been widely used in the Ghanaian setting [35‐40] and found useful and applicable [41]. A cross-cultural study that involved Ghana, indicated that there was no significant difference in Becks; Depression Inventory scores among different countries [42]. Pre-testing of the instrument was conducted using 40 respondents at public hospitals with Mental Health Units within the Cape Coast Metropolis and Komenda-Edina-Eguafo-Abirem Municipality in Ghana to improve the validity and reliability of the instrument.

Table 1 shows that 60.1% of the respondents were female, and 69.1% were aged 25–34 years. More than half (59.5%) of the respondents were diploma holders. The results further revealed that 28.3% of the respondents had spent 4 to 6 years in the nursing profession, and 50.2% of them earned between 1500 to 2000 Ghana cedis (the equivalent of 300–400 US dollars) monthly.

Table 2 shows that 15.1% of the study respondents experienced mild depression whilst 1.9% experienced severe depression. Additionally, it was found that 19.0 and 5.7% of the respondents experienced mild and moderate anxiety, respectively. Furthermore, it was observed that 42.8% of the respondents experienced moderate stress.

Table 3 presents which socio-demographic characteristics of respondents predict depression. The model comprised six predictors, namely; gender, age, department, education status, hospital and income. The criterion variable was the depression levels of nurses working in psychiatric hospitals. The model fitting information for the above-specified model was significant, p = .008. The goodness of fit indices revealed that the data fit the model, p = .719. The Nagelkekerke pseudo-R-square value was .540 indicating that the socio-demographic variables of the respondents accounted for 54% of the variances in depression levels of mental health nurses. The educational level of respondents was found to be significantly associated with depression. Particularly, diploma and bachelor’s degree holders were more likely than master’s holders to be minimally depressed relative to severely depressed. It was found that the department (unit/ward) of the respondents was associated with depression. Those working in the Administration and Out-Patient Department (OPD) were less likely than those in Chronic Wards to be minimally depressed comparable to severely depressed. The income of nurses was positively and significantly associated with levels of depression. Nurses with income less than 1500 cedis equivalent to 300 United States dollars were more likely than those taking 3000+ cedis equivalent to 600 United States dollars to experience less depression relative to severely depressed.

Table 4 presents the analysis of which socio-demographic variables of respondents predict anxiety. The model comprised six predictors, namely; gender, age, department, education status, hospital and income. The criterion variable was the anxiety levels of nurses working in psychiatric hospitals. The model fitting information for the above-specified model was significant, p = .026. The goodness of fit indices revealed that the data fit the model, p = .059. The Nagelkekerke pseudo-R-square value was .380 indicating that the socio-demographic variables of the respondents accounted for 38% of the variances in anxiety levels of mental health nurses. The findings suggest that a higher level of education was associated with lower anxiety levels. This is to say that nurses who have diplomas and degrees were more likely to be anxious than those with master’s degrees. Nurses working at the Ankaful Psychiatric Hospital and Accra Psychiatric Hospital appeared to have higher anxiety levels than Pantang Hospital. This study indicates that the respondents’ age was associated with anxiety. Younger nurses were more likely to suffer from severe anxiety than older nurses. Furthermore, the results also found that male nurses were more likely than female nurses to have low pressure relative to severe anxiety.

Results from Table 5 present which respondents’ socio-demographic characteristics predict their level of stress. The model comprised six predictors, namely; gender, age, department, education status, hospital and income. The criterion variable was the stress levels of staff working in psychiatric hospitals. The model fitting information for the above-specified model was significant, p = .043. The goodness of fit indices revealed that the data fit the model, p = .107. The Nagelkekerke pseudo-R-square value was .610 indicating that the socio-demographic variables of the respondents accounted for 61% of the variances in stress levels of mental health nurses. There appeared to be a negative association between educational level and stress. Nurses with higher academic levels were more likely to experience lower stress levels and vice versa. Additionally, it was found that nurses in the Administration and OPD were more likely to be highly stressed than those in the Acute and Chronic Ward. However, those in the Chronic Ward were more likely to have higher stress levels than those in the Acute Ward. It was observed that the income of nurses was negatively and significantly associated with stress levels. Thus, nurses who take a higher income were likely to experience low-stress levels than those with relatively more minor income. The result generally indicates that older nurses were likely to experience low stress levels and younger nurses were more likely to have high stress levels.

Significant strengths of this study are that the sample size was large and representative of the population of psychiatric nurses in three public psychiatric hospitals in Ghana. Also, the structured questionnaire allowed the nurses to independently respond to the items on the questionnaire, which ensured increased reliability and validity. Our study used three different standardised psychometric scales to assess the prevalence of depression, anxiety and stress, which adds to the validity of the findings. However, this study did not include a comparison to conclude regarding depression, anxiety and stress levels in psychiatric nurses. The study was quantitative and cross-sectional in nature and was unable to explore participants’ experiences. A mixed methodology that includes some qualitative data may also have an added perspective to the study findings.

The study’s findings suggested that psychiatric nurses experience depression and anxiety and moderate stress level. In addition, the results showed that a higher level of education was associated with lower levels of anxiety and stress among psychiatric nurses. This highlights the need for hospital management to encourage further education and provide Continuing Professional Development (CPD) programmes on the management of depression, anxiety and stress by nurses to enhance patient care and improve their quality of life. Further education will have added value which may lead to improved salaries or income. The findings further indicate that psychiatric nurses who worked at the OPD were more likely to experience a high level of depression and stress than those in the Acute and Chronic wards. Therefore, stakeholders in mental health such as the Mental Health Authority need to design preventive strategies to reduce the risk of depression and stress at the OPD.