Construction and verification of a risk prediction model of psychological distress in psychiatric nurses

This study employed a cross-sectional survey design.

From August to September 2022, 812 psychiatric nurses were selected to participate in an anonymous survey via convenience sampling in China’s Shandong Province. According to the statistics of the Shandong Provincial Health Commission, by the end of 2021, 11,000 psychiatric nurses were registered in Shandong Province. According to the limited population sampling formula [28], the significance level α = 0.05, interval estimation confidence 1-α = 0.95, quantile k = 1.96, and P is usually set at 0.50 to obtain the most reliable sample size. The calculation formula for sample size is shown in Fig. 1. The calculation indicates that the number of samples to be investigated should be greater than 372, and 812 people are actually investigated. A stratified method of sampling was used to split seventeen prefecture-level cities in China’s Shandong Province into six categories by their geographical location: north, southwest, south, centre, provincial headquarters, and the Shandong peninsular area. The randomized number table method was used to choose one prefecture-level, higher education, and top mental hospital for each sector, with the study focusing on clinical nurses on the front lines at each institution. The inclusion criteria were as follows:(1)non-sex, marriage-restricted nurses between the ages of 18 and 60; (2) nurses with at least one year of experience working in mental health hospitals; (3) registered nurses with national nurse practitioner (NP) certification; (4) psychiatric nurses who operate in psychiatric institutes’ medical departments; and (5) psychiatric nurses who are on the job and voluntarily. Exclusion criteria include the following: (1) normal nurses at nonpsychiatric hospitals, including refresher and training nurses in mental institutions; (2) nurses who are on leave for more than 3 consecutive months; and (3) nurses who declined to engage in this study.

According to the sampling survey requirements, the nursing department of the surveyed unit is communicated with, and a person in charge of each unit is chosen. The researcher held a special meeting to introduce the questionnaire completion requirements and precautions to the person in charge of the surveyed unit. The questionnaire survey was conducted by electronic questionnaire on Questionnaire Star and included the following subquestionnaires, all of which were translated into Chinese and validated. The questionnaires used uniform guidelines and were answered anonymously to protect the subjects’ privacy. The time to complete the questionnaire was 15–25 min. A total of 904 questionnaires were filled out, excluding unqualified questionnaires with incomplete answers or logical errors, and 812 valid questionnaires were recovered, for an effective response rate of 89.82%.

All of the investigations in this study were authorized by the Shandong Mental Health Center’s Ethics Committee [2023] No. (37). All participants were informed of the purpose and procedures of the study and signed written informed consent prior to participation. Since no identifying information, such as the name and address, of any participant was collected, the privacy and anonymity of the participants were fully protected, and the data were aggregated and reported in summary form only.

We distributed 904 questionnaires and obtained 812 valid responses (response rate: 89.82%). The participants included 208 (25.62%) men and 604 (74.38%) women, aged

18–58 years (mean age = 32.69 ± 8.07. The number of working years ranged from 1 to 40 years, and the median number of working years M(P₂₅,P₇₅) was 6(3,15), as shown in Table 1.

In this study, 812 psychiatric nurses were divided into two groups according to the presence of psychological distress (K10 score ≥ 16). As shown in Table 2, there were mathematically significant differences between the two groups with respect to professional title, shift frequency, psychological capital (self-efficacy, hope, tenacity, and optimism dimensions and total score), job burnout (EE, DP, PA and total score), and the total score of organizational support and sleep quality (P < 0.05). (Table 2)

The 13 factors with statistical significance in the univariate analysis were taken as independent variables, and whether psychiatric nurses experienced psychological distress (K10 ≥ 16 points) was taken as the dependent variable for logistic regression analysis. The assignment table of the independent variables is shown in Table 3. The results show that the title of psychiatric nurse indicates the supervisor nurse, for every 1-point reduction in self-efficacy, for every 1-point increase in emotional exhaustion and personal accomplishment and for every 1-point increase in sleep quality. It is an independent risk factor for psychological distress in psychiatric nurses. The specific values are shown in Table 4.

The process of constructing this study’s psychological distress prediction model is shown in Fig. 2. The psychological distress diagram of psychiatric nurses was built based on the logistic regression findings, as shown in Fig. 3. This prediction model identifies the relevant score value of each item in the graph, adds the scores to create the total score, and calculates the likelihood of psychological distress of psychiatric nurses based on the score on the risk axis.

The calibrate function was used to create a calibration curve to assess the model’s goodness of fit. As shown in Fig. 4, the slope of the curve was close to one. The model’s prediction effect was tested via receiver operating characteristic (ROC) curves. As shown in Fig. 5, the area under the ROC curve was 0.916 (P < 0.001), the highest approximate entry index was 0.705, the ideal critical value was 0.610, the sensitivity was 0.894, and the specificity was 0.811.

Nurses in charge are vulnerable to psychological anguish, which could be because this category of nurse forms the backbone of clinical work and is subjected to severe professional demands by family, teaching, scientific research, and family pressure [35].

The results of the single-factor analysis revealed that shift frequency was an influential factor of psychological distress but not a predictor of psychological distress in psychiatric nurses in the multivariate analysis. This may be because shift frequency changes nurses’ work and rest schedules; however, it is not an independent factor causing psychological distress.

Psychological anguish is more prevalent among psychiatric nurses who have low self-efficacy. In this study, a single-factor analysis revealed statistically significant differences in the level of organizational support between psychiatric nurses with and without psychological distress. However, in the regression analysis, this variable was not included in the regression equation, suggesting that organizational support was not an independent predictor of psychological distress in psychiatric nurses. In the present research, single-factor analysis revealed statistically significant differences in the overall score of psychological capital and each of its four components—self-efficacy, hope, resilience, and optimism—between psychiatric nurses with and without psychological distress. Regression studies revealed that self-efficacy predicted the probability of psychological discomfort. These findings demonstrate that nurses’ psychological capital is closely connected to the prevalence of psychological discomfort; this could be because psychological capital is a good psychological state demonstrated by people who are growing and developing. Psychological capital is an essential positive psychology resource and an individual’s basic psychological characteristic that may successfully enhance individual progress [13]. Improving an individual’s self-confidence and optimistic attitude can effectively reduce the distress of negative emotions, such as anxiety and depression [36], and reduce the individual’s stress response [37]. Individuals have a high level of psychological capital, allowing them to face stress with a positive attitude, better perceive the respect, care and attention of the organization, and, thus, have more self-confidence and be able to respond positively when encountering problems and setbacks. Therefore, it stimulates the positive and optimistic mentality of nurses. This positive emotion brings more positive energy to the individual, reducing the psychological distress of anxiety, depression and other negative emotions caused by stressful events. Self-efficacy was a predictor of psychological distress in psychiatric nurses. Self-efficacy is the ability of people to motivate themselves, recognize their own resources and take the necessary actions to behave in a certain way in a given environment [38]. When facing challenging work, individuals with high self-efficacy are confident that they will succeed through hard work. When encountering crises and obstacles, confident people are usually more optimistic, allowing them to face positive setbacks and difficulties with solid confidence. Therefore, in the career training of psychiatric nurses, group training programs to improve self-efficacy should be added, and more motivational language and measures should be provided at work to improve nurses’ self-confidence.

Emotional exhaustion (EE) is severely reduced. Increased emotional exhaustion (EE) is more prone to psychological distress in psychiatric nurses, which could be due to the nature of the job. Patients with mental problems, particularly those with severe mental disorders, may encounter violent events, suicide, self-injury, escape, and other unexpected events as a result of diminished mental health and mental illness [27]. Psychiatric nurses also experience high rates of workplace violence [39]. The institutions where the subjects of this study are located are psychiatric hospitals in Shandong Province, and most patients admitted have severe mental disorders, with higher risks and greater pressure on nurses [29]. Psychiatric nurses are often in a state of high emotional tension and are prone to emotional exhaustion [31].

Low psychiatric nurses’ PA scores may be due to the high-risk and high-relapse features of individuals who have mental illnesses [32]. The relapse and repeated hospitalization of patients with mental disorders, the low cure rate for patients with severe mental disorders and the continuous decline in social functions [40] reduce the value experience of psychiatric nurses, frustrating their self-confidence and resulting in a diminished sense of work achievement. Moreover, most psychiatric nursing work involves basic, medication, psychological, and rehabilitation nursing, among other types [35]. Psychological nursing skills, requiring strength-based nursing care for psychiatric nurses, can lead to a sense of accomplishment for nurses [24, 41]. However, improving psychological nursing skills requires some organizational support, including professional training, practice and case supervision, to increase personal professional skills and recognition from patients, prompting a sense of value for psychiatric nurses. Without sufficient organizational support and intrinsic motivation for personal growth, the year-after-year performance of traditional daily psychiatric care, such as primary care, will cause psychiatric nurses to lack a sense of self-worth. Moreover, the utilization of professional skills and advanced equipment to rescue critically ill patients involve the urgent challenge of saving lives but also provides nurses an intuitive experience of treatment and saving lives for, increasing their sense of self-worth [42]. However, these complex first aid tasks for critically ill patients are rarely used in psychiatric departments. According to Xia Lei [40], the recognition rate of low personal achievement among psychiatric nurses is as high as 61.9%, which is greater than that among psychiatrists (47.8%) and psychotherapists (40.0%).

Psychiatric nurses who sleep poorly are more prone to experience psychological disturbances. According to this study, psychiatric nurses with greater sleep quality scores are more likely to experience psychological discomfort. This may be due to the bidirectional interaction connection between sleep quality and psychological suffering, such as anxiety and depression. Sleep deprivation activates the body’s stress response. If a person remains in this state for an extended period of time, he or she can develop psychological disorders such as depression and anxiety. This emotional and psychological distress prolongs fast-wave sleep and aggravates sleep disorders. This type of psychological distress and sleep disorders has interdependent effects and ultimately leads to the deterioration of an individual’s physical and psychological state [24]. In this study, single-factor analysis revealed statistically significant variations in the frequency of nurses’ shifts between psychiatric nurses with and without psychological distress. However, this variable was not included in the equation of the regression analysis, suggesting that the frequency of nurses’ shifts may have some effect on sleep, even though it was not an independent predictor of psychological distress in psychiatric nurses. Research has shown that the frequency of nurses’ shifts is positively correlated with sleep latency, sleep efficiency and subjective sleep quality [41]. The biological rhythm of nurses on shifts is disrupted, affecting body functions, leading to insomnia or lethargy, and resulting in worse sleep quality [42]. The prevalence of shift work sleep disorder (SWSD) among psychiatric nurses in China is 37.5% [43]. The nursing profession determines the work pattern of shifts. Currently, healthy sleep queues for nurses have been established in Japan [44] and Norway [45]. This finding also suggests that shifts may cause the psychological distress of nurses through its impact on the quality of sleep.

We constructed a nomogram for psychological distress prediction for psychiatric nurses, as shown in Fig. 2. This study uses the operating characteristic curve of the research object to test the prediction effect of the model. When the region under the contour of the curve is 0.5– to 0.7, the prediction impact is poor; when it is 0.7– to 0.9, the prediction effect is medium; and when it is > 0.9, the forecasting effect is high [46]. In this study, the model’s circumference of the receiver operating characteristic curve is 0.916, the highest Youden index value is 0.699, its ideal critical value is 0.610, its sensitivity is 0.894, and its specificity is 0.811, indicating it has a strong prediction impact. A calibration curve was created to assess the model’s quality of fit. The slope of the curve is close to one, indicating that the model accurately predicts the risk of emotional distress among nurses working in psychiatry. The prediction model developed has a high degree of discrimination and calibration. The nomogram-based psychological distress prediction model for psychiatric nurses visualizes complicated equations and estimates the likelihood of psychological distress occurrence among psychiatric nurses on the basis of the nomogram model score, which is simple and straightforward to apply.

First, this was a multicentre cross-sectional study with regional representation. However, only clinical nurses in tertiary hospitals in Shandong Province were sampled, which does not fully represent the psychological distress level of nurses in Shandong Provincial Psychiatric Specialized Hospitals. Therefore, future studies in other parts of China are needed to verify the results of this study. Second, this study employed a survey for self-assessment, which is relatively subjective. Finally, this study only used a nomogram, a model prediction method, for visualization. Although the model prediction fits well, it does not compare multiple methods. Using various artificial intelligence prediction methods for statistical analysis in later research and selecting optimal prediction models are recommended.