Assessment of mental health status among Chinese nursing staff in the intensive care unit: a network analysis

Abstract

Backgrounds:

Nursing is the key group to provide healthcare services, and it is easy for nursing staff to develop mental health problems.

Aims:

The study aimed to evaluate prevalence of psychological symptoms in nurses working in an intensive care unit (ICU) and the inter-relationship of associations of psychological symptoms using network analysis.

Methods:

This study is a cross-sectional design study. The Chinese version of the Symptom Check List-90 (SCL-90) was used to measure the psychological status of ICU nurses. The network structure of psychological symptoms was characterised, and indices of ‘Expected influence’ were used to identify symptoms central to the network. Network stability was examined using a case-dropping bootstrap procedure.

Results:

Multiple logistic regression analysis found those who had worked more than 15 years were less likely to experience positive psychological symptoms, whereas nurses working in emergency ICU and other ICUs, nurses working in departments with over 16 beds were more likely to develop psychological symptoms. In addition, ‘Anxiety’, ‘Mental degeneration’ and ‘Depression’ were central symptoms in the network.

Conclusions:

ICU nurses reported a high level of psychological symptoms, which may affect the quality of their work and worsen public health problems.

Introduction

Mental disorder is a common health problem in all walks of life, which is associated with numerous negative consequences, such as increased morbidity, causing poor performance, low productivity and reduced quality of life, and it has aroused widespread concerns in society (Chen et al., 2021; Gautham et al., 2020). Nurses are playing a crucial role in taking care of patients, and it is a stressful profession. Thus, they are at risk of developing mental health problems (Charkhat Gorgich et al., 2017; Xin et al., 2019). Mental health problems can influence nurses’ work performance, health, job satisfaction and intention to leave (Aiken et al., 2001; Cheung and Yip, 2015; Ramírez-Elvira et al., 2021).

The intensive care unit (ICU) is a fast-paced, demanding and tension-charged environment. Nurses working in the ICU are predisposed to workplace stress owing to direct and indirect exposure to traumatic and critical events, excessive workload (Cavalheiro et al., 2008; Saini et al., 2011), high patient care demands (Bakker et al., 2005; Padilha et al., 2008), long shift work (Jensen et al., 2018), close contact with death and severe illness (Giannis et al., 2021) and even medical violence (Dafny and Beccaria, 2020; Rayan and Baker, 2019). All of these reasons may lead to a heightened stress response combined with negative emotions (Rayan and Baker, 2019). Meta-analysis indicated that the prevalence for high emotional exhaustion was 31% (95% CI, 8–59%), for high depersonalisation was 18% (95% CI, 8–30%) and for low personal accomplishment was 46% (95% CI, 20–74%) among nurses working in the ICU (Ramírez-Elvira et al., 2021). To lower the risk and alleviate the negative outcomes caused by mental health problems of nurses working in the ICU, it is important to understand the prevalence and correlates of mental health problems.

A network approach is a novel approach where mental health problems are viewed more as a system where symptoms form interacting networks among themselves (Borsboom, 2017). Recently, network analysis was applied on patients with anorexia nervosa and it was found that depression, anxiety and interpersonal sensitivity (SCL-90) were the highest centrality symptoms among the network analysis in this population (Solmi et al., 2019). However, findings based on a sample from patients with anorexia nervosa do not apply to nurses, especially nurses in ICU. To date, no network analysis studies have been published on psychological symptoms in nurses working in ICU in China. These gaps in the psychological symptoms in nurses working in ICU literature provided us with the impetus to evaluate overall psychological symptoms in nurses working in ICU and the inter-relationship of associations of psychological symptoms using network analysis.

Methods

Study design and participants

This cross-sectional study was conducted in December 2018 to assess the mental health status of nursing staff working in ICU. To enhance sample representativeness, frontline registered nurses working in ICU from each province in China (except for Tibet) were invited to participate in this study. Those who were on leave and nursing students were excluded. An anonymous online survey tool called ‘Wenjuanxing’ was used to collect data. The study protocol was approved by the Clinical Ethics Committee of the First Hospital of Hebei Medical University.

The sample size was calculated using the formula N = Za²P(1 – P)/d², in which a = 0.05 and Za = 1.96, and the estimated acceptable margin of error for proportion d was 0.05. The prevalence of mental health problems among nurses working in ICU was estimated to be 31% based on a previous study (Ramírez-Elvira et al., 2021). Assuming that 10% of those invited would refuse participation in this study, at least 362 participants had to be included.

Measures

Two questionnaires were used in this study. A self-designed data collection form was used to collect the basic sociodemographic characteristics, and the Chinese version of the Symptom Check List-90 (SCL-90) was used to measure the psychological status of ICU nurses. The SCL-90 is a self-report scale to assess mental health or psychopathological symptoms, with 90 items covering 9 domains, including somatization, obsessive symptoms, interpersonal relation, depression, anxiety, hostility, terrifying, bigoted and mental degeneration (Roskin and Dasberg, 1983; Yu et al., 2020). Each item was scored from 1 (none) to 5 (extremely), with a higher score reflecting higher frequency and intensity of symptoms. As recommended previously (Chen et al., 2020; Du et al., 2020), a SCL-90 total score of ⩾160 was considered as ‘having positive psychological symptoms’.

Statistical analyses

The data analyses were performed by SPSS 22.0. Comparisons between nurses with and without positive psychological symptoms concerning basic demographic characteristics were conducted by chi-squared tests. Multiple logistic regression analyses examined sociodemographic correlates of psychological symptoms. The correlation between two variables was measured using the Spearman correlation test. Positive psychological symptoms were entered as the dependent variable, whereas the variables that significantly differed in univariate analyses were entered as independent variables. The significance level was set at 0.1 (two-tailed).

Network analysis

The network was computed using the R software (R Core Team, 2020). We computed the polychoric correlations of items of psychological symptoms to investigate the edges of the network and estimated the graphical Gaussian model (GGM), a popular network model, with the graphic least absolute shrinkage and selection operator and extended Bayesian information criterion model using the R package ‘qgraph’ (Epskamp et al., 2018). GGM is a pairwise Markov random field model used for interval or ordinal data. The edges in the GGM can be interpreted as partial correlation coefficients. The network was visualised using the ‘qgraph’ package, where thicker edges represented stronger relationship between nodes. We also estimated the centrality index expected influence (EI) of nodes to determine which symptoms are more central (influential) in the network (Beard et al., 2016). To further assess the accuracy of the index, we used the bootstrap method in the ‘bootnet’ package to investigate the stability of central index-EI using the correlation stability coefficient (CS-coefficient). We set the cut-off for the CS-coefficient at 0.25 for all the indices because evidence showed that when centralities do not differ from one another, the CS-coefficient is mostly below 0.25 (Epskamp et al., 2018).

Result

Overall, 2280 nurses were invited to participate in the survey, and 2029 completed the survey, yielding a response rate of 89%. A total of 2029 Chinese ICU nurses participated in this study. Of the participants, 55.1% (95% CI: 53–57%) of them had positive psychological symptoms. As shown in Table 1, nurses were mainly from General ICU (71.8%), and most of them have been working in ICU for more than 5 years (56.4%). In addition, 87.8% of them were female, and 72.1% of them had bachelor’s degrees or above. The majority of the nurses were married (61.3%) and had children (51.0%). Over 90% of the nurses had night shifts (96.7%) and worked more than 40 hours per week (92.3%).

Table 1.

Comparison between ICU nurses with and without positive psychological symptoms concerning basic demographic characteristics in 2018 (n = 2029).

Variable	Total (n = 2029)		SCL-90 total score ⩾ 160 (n = 1117)		SCL-90 total score < 160 (n = 912)		Univariate analysis
	N	%	N	%	N	%	χ2	df	p
Gender							0.231	1	0.631
Female sex	1781	87.8	797	87.4	984	88.1	−	−	−
Male	248	12.2	115	12.6	133	11.9
Education							2.351	1	0.125
College	566	27.9	239	26.2	327	29.3	−	−	−
Bachelor or above	1463	72.1	673	73.8	790	70.7
Marriage							0.332	1	0.564
Yes	1243	61.3	565	62.0	678	60.7	−	−	−
No	786	38.7	347	38.0	439	39.3
Children							0.267	1	0.605
Yes	1035	51.0	471	51.6	564	50.5	−	−	−
No	994	49.0	441	48.4	553	49.5
Ranking							6.675	2	0.036
Junior nurse	1524	75.1	667	73.1	857	76.7	−	−	−
Intermediate nurse	448	22.1	211	231	237	21.2
Senior nurse	57	2.8	34	3.7	23	2.1
Years of working							10.043	3	0.018
0–5	904	44.6	380	41.7	524	46.9	−	−	−
6–10	720	35.5	330	36.2	390	34.9
11–15	216	10.6	116	12.7	100	9.0
>15	189	9.3	86	9.4	103	9.2
Department							6.603	2	0.037
General ICU	1456	71.8	639	70.1	817	73.1	−	−	−
Emergency ICU	97	4.8	37	4.1	60	5.4
Other ICU a	476	23.5	236	25.9	240	21.5
Number of beds							33.431	3	<0.0001
3–10	304	15.0	154	16.9	150	13.4	−	−	−
11–15	408	20.1	148	16.2	260	23.3
16–20	598	29.5	241	26.4	357	32.0
>20	719	35.4	369	40.5	350	31.3
Adding bed situation							0.595	3	0.898
Always	209	10.3	154	16.9	150	13.4	−	−	−
Often	428	21.1	148	16.2	260	23.3
Occasionally	707	34.8	241	26.4	357	32.0
Never	685	33.8	369	40.5	350	31.3
Frequency of night shift							8.576	4	0.073
2–3 days	681	33.6	333	36.5	348	31.2	−	−	−
4–5 days	1019	50.2	430	47.1	589	52.7
6 days	186	9.2	83	9.1	103	9.2
>6 days	76	3.7	32	3.5	44	3.9
Not fixed	67	3.3	34	3.7	33	3.0
Working hours per week							4.070	2	0.131
<40	156	7.7	64	7.0	92	8.2	−	−	−
40–60	1431	70.5	632	69.3	799	71.5
>60	442	21.8	216	23.7	226	20.2

Bolded values are p < 0.05.

ICU: intensive care unit.

^aNeurosurgery ICU, Cardiology ICU, Thoracic surgery ICU, Surgical comprehensive ICU, Comprehensive internal medicine ICU, Respiratory ICU and Paediatric ICU.

There is a significant correlation between ranking and years of working (r = 0.662, p < 0.001); therefore, ranking was not included in the multiple logistic regression. Multiple logistic regression analysis found those who had worked more than 15 years (p = 0.01; OR = 0.56, 95% CI: 0.36–0.87) were less likely to have experienced positive psychological symptoms, compared with those who worked less than 5 years. And nurses working in emergency ICU (p = 0.03; OR = 1.27, 95% CI: 1.02–1.57) and other ICU (p = 0.05; OR = 1.59, 95% CI: 1.01–2.50) were more prone to have positive psychological symptoms, compared with nurses working in comprehensive ICU. In addition, nurses working in the department with beds between 16 and 20 (p < 0.0001; OR = 1.82, 95% CI: 1.42–2.34) and more than 20 (p < 0.0001; OR = 1.56, 95% CI: 1.25–1.94) were more likely to develop positive psychological symptoms compared with ICUs with beds between 3 and 10, as shown in Table 2.

Table 2.

Demographic characteristics independently associated with nurses with and without positive psychological symptoms in 2018 (n = 2029).

Variable	B	Odds ratio	95% CI	p-Value
Years of working
0–5	−	−	−	−
6–10	−0.227	0.797	0.652–0.974	0.027
11–15	−0.530	0.589	0.434–0.799	0.001
>15	−0.159	0.853	0.608–1.197	0.359
Department
Comprehensive ICU	−	−	−	−
Emergency ICU	0.220	1.246	0.812–1.912	0.315
Other ICU	−0.233	0.792	0.640–0.979	0.031
Number of beds
3–10	−	−	−	−
11–15	0.614	1.848	1.360–2.510	<0.001
16–20	0.470	1.601	1.207–2.123	0.001
>20	0.008	1.008	0.767–1.325	0.954
Frequency of night shift
2–3 days	−	−	−	−
4–5 days	0.272	1.313	1.076–1.602	0.007
6 days	0.152	1.164	0.834–1.625	0.372
>6 days	0.433	1.541	0.945–2.515	0.083
Not fixed	0.008	1.008	0.587–1.731	0.977

Bolded values are p < 0.05.

ICU: intensive care unit; CI: confidence interval.

Network structure

Figure 1 presents the network structure of psychological symptoms. The predictability of symptoms is shown as ring-shaped pie charts. The mean predictability was 0.83, indicating that on average 83% of each node’s variance could be accounted for by other nodes. The model showed that the edge between D (‘Depression’) and O (‘Obsessive’) (average edge weight = 0.291) was the strongest positive edge in the network, followed by the edge between B (‘Bigoted’) and M (‘Mental degeneration’) (average edge weight = 0.290), and the edge between I (‘Interpersonal’) and D (‘Depression’) (average edge weight = 0.281).

Figure 1.

Network structure of psychological symptoms in Chinese nursing staff in intensive care units in 2018, demonstrated the node centrality of psychological symptoms.

In terms of EI, A (‘Anxiety’) had the highest EI centrality, followed by M (‘Mental degeneration’) and D (‘Depression’) in the whole network, indicating that these three symptoms were important and central for understanding psychological symptoms among ICU nurses. In contrast, several other symptoms were marginal, such as H (‘Hostility’), S (‘Somatization’) and Ter (‘Terrifying’) (Figure 1).

For stability of the network, the centrality of EI had an excellent level of stability (i.e. CS-coefficient = 0.75), meaning 75% of the sample could be dropped and the structure of the network would not significantly change (Figure 2). The results of the bootstrap 95% CI for edges, and bootstrapped differences tests for edge weight are shown in Supplemental Figure S1. The bootstrap difference test showed that most comparisons between edge weights were statistically significant (Supplemental Figure S1).

Figure 2.

The stability of centrality and bridge centrality indices using case-dropping bootstrap in Chinese nursing staff in intensive care units in 2018 demonstrated the network structure is stable.

Discussion

This study was the first to examine the prevalence of positive psychological symptoms of ICU nurses and associations between psychological symptoms using network analysis. In the network of psychological symptoms of nurses working in ICU, Anxiety, Mental degeneration and Depression were the most influential symptoms.

In this study, the prevalence of psychological symptoms was 55.1% (95% CI: 53–57%). There are several reasons that lead to high positive psychological symptoms of ICU nurses, including the accelerated departure of nurses from their profession due to dissatisfaction with the working environment (e.g. ICU experience deaths of patients, abusive behaviour towards ICU nurses by patients’ families; Aiken et al., 2001; Ramírez-Elvira et al., 2021) and work-related stresses involving patients under their direct care (e.g. addressing end-of-life issues, performing cardiopulmonary resuscitation, involvement in post-mortem care and prolonging life by artificial support; Mealer et al., 2007).

Working years have been negatively associated with the prevalence of psychological symptoms in ICU nurses, which is consistent with some studies that focused on nurses in Alexandria (Arafa et al., 2003) and nurses in the Turkish Republic of Northern Cyprus (Özgür et al., 2011). Both of the studies revealed that fewer years of working experience was found to be a significant predictor of psychological ill health among nurses. Another study also found that nurses in Hong Kong with clinical experience of less than 10 years were more likely to report stress (Cheung and Yip, 2015). This study revealed that the frequency of night shift was related to being more likely to experience positive psychological symptoms, which may be related to high levels of fatigue, stress and burnout among the nurses working in ICU (Chen et al., 2021).

As younger nurses are expected to have fewer years of working experience, longer rotating night shifts and spend more time to be familiar with their work as compared with their counterparts, it is expected that this variable was significantly associated with psychological ill-health (Cheung and Yip, 2015). Emergency ICU nurses are routinely confronted with severe injuries, death and suicide and are also frequently exposed to verbal and physical aggression (Bennett et al., 2005; McFarlane and Bryant, 2007), which is consistent with our study that emergency ICU nurses had higher psychological symptoms than their corresponding figures. A previous study reported that nurses working in emergency departments experienced work negative problems during working. For example, 75% of emergency nurses were confronted with aggressive behaviour, compared with 43% of internal medicine wards, 23% of surgical wards and 9.9% of paediatric wards and gynaecology (Adriaenssens et al., 2012). Furthermore, verbal aggression was reported with high prevalence rate for nurses worked in emergency compared to the nurse worked in another department. For example, verbal aggression over the same period was reported by 75% of emergency nurses, compared with 39% of them in internal medicine wards, 25% of them in surgical wards and 12.4% of them in paediatric wards and gynaecology (Winstanley and Whittington, 2004). The number of beds is related to the workload of the nurses in ICU. The workload is affected by stress due to several interrelated factors, including the competence of the nurse, the allocated time and effort for the care of the patients, and the intensity and complexity of anticipated care services. A previous study reported that workload caused feelings of personal insufficiency, insensitivity, unhappiness, hopelessness and intolerance (Turan and Ançel, 2020). Therefore, it is observed that workload has increased the likelihood of developing mental health problems such as depressive symptoms (Gong et al., 2014), anxiety, sleep disturbance, emotional burnout and somatic complaints (Arrogante and Aparício-Zaldivar, 2017; Jaradat et al., 2016; Shen et al., 2016).

The symptom ‘Anxiety’ was a prominent central symptom in the network of Chinese nurses in ICUs as indicated by EI. In general, the prevalence of anxiety in ICU nurses is higher than in other nursing specialties (Adriaenssens et al., 2012). Chinese nurses are known to work in stressful environments. This of course involves characteristics of the occupation, such as concern for occupational hazards, infection or injury, fear of potential medical malpractice and related consequences, physical and mental exhaustion due to intensive care of emergency patients and seeing patients die and suffer. In addition, the nursing model in Chinese hospitals has changed rapidly in recent years. China’s medical services have also undergone earth-shaking reforms, breaking away from the previous market-oriented service system and emphasising the public welfare of state-owned hospitals. All the circumstances and changes make nursing practice more demanding and challenging (Cheng et al., 2000; Weyers et al., 2006).

Our network psychological symptoms indicated that ‘Mental deterioration’ was the second central symptom in this network. As China’s population grows, nurses take on an increasing and complex workload, forcing them to take on a greater stressful emotional burden while feeling lower levels of job satisfaction and reward (Li et al., 2018). All these negative job-related factors may worsen ’nurses’ physical health, leading to burnout and a higher risk of mental deterioration (Deng et al., 2020; Gao et al., 2012; Oyama and Fukahori, 2015). Therefore, assessment and identification of high-risk populations with poor mental health or mental deterioration, and subsequent intervention and treatment before onset, should be considered key to preventing mental illness.

Our network analysis reported that ‘Depression’ is one of the central symptoms. Due to fast-paced, demanding and tension-charged ICU environments, it may cause negative mental health outcomes such as depression among nurses in ICUs (Huang et al., 2022). Chronic exposure to depressive symptoms may lead to individuals having a high risk of developing future episodes of a major depressive disorder and greater chronic morbidity (Huang et al., 2022). Additionally, high working stress, burnout and the impairment of recovery from heavy work stress are the causes of depressive symptoms (Vandevala et al., 2017). Educational strategies targeted at healthcare professionals and innovative methods of organising and providing healthcare intervention need to be proposed to improve the recognition and management of depression in nursing in hospitals (Cabana et al., 2002; Katon et al., 2001).

The strengths of this study included the multicentre study design and use of advanced network analysis statistics designed to identify central symptoms of psychological symptoms in ICU nurses. However, several methodological limitations should also be acknowledged. Firstly, causal relations between sociodemographic information and psychological symptoms could not be determined due to the cross-sectional study design. Secondly, certain risk and/or protective factors that may also influence psychological symptoms of ICU nurses were not examined, such as the balance between work and family, and relationships with family members.

Conclusion

In conclusion, ICU nurses reported a high level of psychological symptoms, which may affect the quality of their work and worsen public health problems. Central symptoms (e.g. ‘Anxiety’, ‘Mental degeneration’ and ‘Depression’) identified in this study could be targeted in interventions and preventive measures for psychological symptoms of nurses working in ICU. Training, psychological interventions and judicious use of resources may all promote the mental health of nurses working in ICU. Further studies are needed to explore the specific influencing factors and interventions that promote mental health.

Key points for policy, practice and/or research.

• ICU nurses reported a high level of psychological symptoms, which may affect the quality of their work and worsen public health problems.

• Central symptoms (e.g. ‘Anxiety’, ‘Mental degeneration’ and ‘Depression’) identified in this study could be targeted in interventions and preventive measures for psychological symptoms of nurses working in ICU.

• Training, psychological interventions and judicious use of resources may all promote the mental health of the nurses working in ICU.

• Using the results of a central symptoms based network analysis for setting treatment or intervention targets for nursing staff has the potential to reduce the risk of mental health problems.

Biography

Liang-Nan Zeng got her PhD degree in the University of Macau and now works as a clinical nurse in the Department of Nursing, at Chengdu Fifth People’s Hospital.

Hong Cai is a third-year PhD student in the University of Macau, she got her Bachelor’s and Master’s degrees at Jilin University, School of Nursing.

Fei Gao is an engineer working in Hebei General Hospital in China.

Bi Guan works as a nurse leader in the Department of Nursing, at Chengdu Fifth People’s Hospital. Her research areas include mental health problems and evidence-based clinical practice research.

Wen-Jin Chen is a doctor from Xuanwu Hospital of Capital Medical University in China. Specialising in the treatment of neurocritical patients. His research interests include evidence-based clinical practice and epidemiology in mental health.

Wei He is a doctor from the Beijing Tongren Hospital and expertise in the treatment of neurocritical patients.

Tang-Ming Peng is a doctor from the Chengdu Fifth People’s Hospital, mainly engaged in the treatment of patients with cerebrovascular disease.

Xiao-Peng Li is a medical professional with experience in treating neurocritical patients from the First Affiliated Hospital of Xinjiang Medical University. He is interested in the mental health of occupational professionals.

Yan Li is involved in nursing management and is a member of the Chengdu Fifth People’s Hospital.

Sui Sum Kung works as a doctor in the Division of Neurosurgery, Department of Surgery, Kaohsiung Medical University Hospital, Kaohsiung Medical University.

Xiao-Meng Wang’s research areas include evidence-based medicine, epidemiology of mental health and the clinical research of mental health problems.

Wei Liu worked in EICU in Xuzhou Central Hospital for more than ten years and has a high interest in mental health problems of nurses and doctors working in high stress units.

Di Zhao works in Hebei Province Hospital of Chinese Medicine, Affiliated Hospital of Hebei University of Chinese Medicine. His research areas include CBT intervention and depression.

Shu Yan works in Chengdu Fifth People’s Hospital. His research areas include mental health problems and evidence-based clinical practice research.