New nurses’ practice environment, job stress, and patient safety attitudes: a cross-sectional study based on the job demands-resources model

The study is based on the JD-R model [9], which groups job characteristics into job demands and resources. This model has been commonly applied in nursing practice and explores the role of work participation in nursing practice [10]. According to this theoretical model, various job demands and resources can interact and jointly affect organizational outcomes. Patient safety attitudes are a crucial organizational outcome in clinical settings [11]. In light of this, we employed this model to investigate the effects of various work factors of new nurses on patient safety.

Job demands were defined as: "physical, psychological, social or organizational aspects of the job that require sustained physical and/or psychological (cognitive and emotional) effort or skills and are therefore associated with certain physiological and/or psychological costs ([9], p312])." Research has shown that job-related stress in nursing is associated with adverse effects on both physical [12] and mental [13] health of nurses. Furthermore, increased stress levels lead to lower quality of care and negative nursing outcomes [14]. In this study, new nurses' job stress is considered the job demand. Job resources were defined as: "functional in achieving work goals; reduce job demands and the associated physiological and psychological costs; or stimulate personal growth, learning, and development ([9], p312])." High levels of job resources will significantly improve motivation and organizational outcomes, and we use a practice environment as a job resource because a positive practice environment reduces work-related stress [15] and improves nursing outcomes. In this study, we proposed a theoretical model based on the JD-R and analyzed the links between the practice environment, job stress, and patient safety attitudes among new nurses (Fig. 1).

This study used a cross-sectional design and followed the STROBE guideline for cross-sectional studies [30].

Participants were selected from seven general tertiary hospitals across six cities in Anhui, China. Inclusion criteria were: (1) possession of a Chinese nursing certificate; (2) work experience within a range of 1 to 2 years, and (3) currently full-time working nurses. The study excluded individuals: (1) not engaged in clinical nursing during the research period, such as those on maternity or sick leave; and (2) refuse to sign informed consent form.

Klien [31] recommended a minimum sample size of 200 for stable estimates when testing SEM. According to the SEM estimation parameters calculation method by Reykov and Marcoulides [32], the model is expected to have 35 estimated parameters. Empirically, the sample size is calculated as ten times the estimated parameters [33], resulting in a sample size of 350. Considering a possible 10% non-response rate and potential invalid questionnaire, we used a convenience sampling method to recruit a total of 400 new nurses.

The data were collected from April 2022 to August 2022. The trained investigators traveled to the hospitals where the participants worked. Before distributing the questionnaires, the participants were informed of the study's objective, process, and ethical principles followed by signing the informed consent. Four hundred questionnaires were distributed, 390 were collected, and 20 invalid questionnaires were excluded due to incompletion. A total of 370 participants completed the questionnaires with a 92.5% response rate.

The study has been approved by the ethical committee of the College of Nursing of Wannan Medical College (LL-2022BH02); all methods were performed in accordance with the Helsinki guidelines [41].

Statistical analysis was conducted utilizing SPSS (Statistical Product and Service Solutions, version 23). Frequencies, percentages, means, and standard deviations were used to describe the distribution of participants' sociodemographic data, job stress, practice environment, and patient safety attitudes. The Shapiro–Wilk test was used to assess the normal distribution of variables. As the variables were skewed distribution, the Spearman correlation coefficient was employed to analyze correlations. To enhance the understanding of the results, we use R (version 4.3.1) to visually represent the correlations between the variable dimensions. Given a skewed distribution, Smart PLS (version 3) was preferred to analyze the structural equation model SEM and test the theoretical hypothesis model.

For assessing reflective models, factor loading (> 0.708), Cronbach's α (> 0.9), and composite reliability (CR > 0.9) were adopted [42, 43]. Convergent validity was assessed using the average variance extracted (AVE), which should be higher than 0.5, and discriminant validity was used the heterotrait-monotrait (HTMT) ratio of the correlations, which should be less than 0.85 for the conceptually different constructs [44].

The structural models and determining that the sample predictive power of the model is explained by the coefficient of determination R², when the model does not have collinearity issues (variance inflation factor, VIF < 5) [45]. R² ranges from 0 to 1, with higher values indicating greater explanatory power. Henseler and colleagues state that R² values of 0.75, 0.50, and 0.25 can be considered significant, moderate, and weak [46]. The predictive accuracy of the model was explained using the blindfolding-based cross-validated redundancy measure Q², which should have a value greater than 0 to indicate the predictive accuracy of this structural equation model. As a rule of thumb, Q² values above 0, 0.25, and 0.50 indicate small, medium, and large predictive accuracies for PLS pathway models [42]. The global fit of the PLS modeling was explained using the goodness-of-fit (GoF), \(\mathrm{GoF}=\sqrt{\overline{\mathrm{Communality}\ast\overline{\mathrm R^2}}}\), with 0.1, 0.25, and 0.36 being the small, medium, and large values of the global fit of the model, respectively [47].

Finally, we used the bootstrapping method with 5,000 resamples to evaluate the significance of the path coefficient and mediation effect (t-values should be higher than 1.96 or smaller than -1.96 for the two-tailed test).

Three hundred and seventy participants' socio-demographic characteristics are shown in Table 1. The average age of the new nurses was 25.43 ± 1.56 years. The majority (61.6%) was 23 ~ 26 years old, female (92.7%), 63.5% possessed advanced diploma. About one third (31.6%) had less than 6 months of work experience, and 33.2% had more than 18 months. Majority of them (40.5%)worked in internal medicine.

Descriptive statistics of practice environment, job stress and patient safety attitudes are shown in Table 2. The skewness and kurtosis of the three scores ranged from -0.582 ~ 0.770 and -0.259 ~ 1.471 respectively, combined with the results of the Shapiro–Wilk test, p values were all < 0.001. It was concluded that the scale scores did not follow a normal distribution.

The average score of practice environment was 100.25 ± 13.18, and the mean score for job stress was 68.92 ± 16.76. The average patient safety attitudes score was 126.99 ± 14.39, indicating that the participants' patient safety attitudes score was at a moderate level, and the 'perceptions of management' was highest. The Cronbach's α of the three variables ranged from 0.911 to 0.970, and the Cronbach's α of each dimension ranged from 0.682 to 0.944, indicating that these scales have good reliability in this study.

Table 3 shows the results of the Spearman correlation analysis between the main variables. There was a positive correlation between practice environment and patient safety attitudes (r = 0.660, p < 0.001). Job stress was significantly negatively associated with both the practice environment (r = -0.346, p < 0.001) and patient safety attitudes (r = -0.388, p < 0.001).

The detailed results of the correlation between the variables and their dimensions are depicted in Fig. 2. The results were much the same as the relationship between the major variables. The stress recognition dimension of the patient safety attitudes scale has a significant positive correlation with the dimensions of job stress and has no significant correlation with the dimensions of practice environment and patient safety attitudes.

The results of our study revealed that the level of patient safety attitudes of new nurses was at a moderate level. The findings are similar to those of Li YZ et al. [48], Zhang, F et al. [49], and Ünver S, et al. [50]. These outcomes show that new nurses generally have more positive attitudes towards patient safety. Among the six dimensions, the findings showed that the highest mean score was for "perceptions of management", while the lowest score was for "stress recognition". This finding is inconsistent with that of Zhang, F et al. [49], and Ünver S, et al. [50]. Their studies reported that the highest score dimension for patient safety attitudes was the teamwork climate. Some scholars consider teamwork to be an essential factor in improving patient safety, and nurses with positive attitudes towards patient safety were more likely to cooperate and contribute to teamwork [51]. The reason for this inconsistency could be the differences in the study populations. Our study participants are new nurses undergoing standardized training and rotating between different departments in the hospital. In this case, they require consistent management and training. Therefore, perceptions of management are extremely important in shaping new nurses' attitudes towards patient safety. When management prioritizes patient safety, it encourages new nurses to enhance their attitudes toward patient safety. However, a couple of studies [52‐54] have shown that the stress recognition dimension received the lowest score. This suggests that new nurses may have difficulty identifying stressful situations they come across in the workplace. The reason for this may be that usually, new nurses are assigned to do the relatively easy caring work, but not to care for critically ill patients. Besides, there are usually mechanisms set up in the unit to deal with emergencies, so they believe that even in high-pressure situations, working performance is not significantly affected [50]. Based on this finding, it is recommended that clinical managers should act as mentors to guide new nurses in identifying and managing the stresses they encounter.

This study also found that new nurses' patient safety attitudes were positively correlated with their practice environment and were negatively correlated with their job stress, which is consistent with the findings of existing studies [55‐58]. It is worth noting that the stress perception dimension of the safety attitudes questionnaire was positively correlated with job stress and not significantly correlated with practice environment. This result is unusual because it differs from the other dimensions of the Safety Attitudes Questionnaire, which may be due to the fact that when job stress increases, new nurses have to expend more energy dealing with these stressors, which tends to produce higher perceptions of stress.

In our study, the practice environment had a direct positive effect on patient safety attitudes, which is similar to that of Al Ma'mari Q et al. [55]. They also found that the practice environment predicted nurses' patient safety attitudes and that there was a relationship between a good work environment and the quality of nursing care and improved patient safety outcomes [58]. We concluded that a positive nursing practice environment promotes, among other factors, the involvement of new nurses in quality management, improved collaboration between healthcare providers and the provision of adequate resources. These conditions can be used as working resources to improve attitudes to patient safety and thus provide safer care for patients. It suggests that nurse leadership is a key factor in building and sustaining a healthy work environment. Effective nurse leadership is a precursor to a healthy work environment [59]. Therefore, nurse managers should understand and improve the status quo of the clinical practice environment in different departments and improve their nursing leadership to promote safe care.

Our results indicated that job stress had a direct negative effect on new nurses' patient safety attitudes, which congruent with the findings of Yalçın Akgül G et al. [56], whose studies found that a statistically significant negative correlation between organizational stress and patient safety attitudes. Job stress is a major cause impacting the physical and mental health of healthcare workers as well as in the quality of health services [60], and working in high-pressure environments can increase the risk of care errors and adverse events [61]. Therefore, job stress is one of obstacles in shaping patient safety attitudes, and nurse managers need to identify new nurses' job stress in a timely manner and provide effective stress management interventions to reduce new nurses' job stress and promote a safe nursing environment.

Our findings also support the hypothesis of the practice environment has an indirect positive effect on patient safety attitudes through mediation of job stress. Several studies [55, 62, 63] have shown that many factors, such as burnout, fatigue, workload, working conditions and workplace violence, affect nurses' job stress and reduce their attitudes to patient safety. These factors can lead to adverse events by increasing nurses' stress, and with job stress serving as a mediating factor. The study's results demonstrated that the practice environment indirectly affects the safety attitudes of new nurses by reducing job stress. According to the stress and coping theory [64], stress arises from the interplay between an individual and their external environment when the individual's coping resources are insufficient to manage external environmental threats. Therefore, improving the practice environment of new nurses can help to reduce their job stress and thus improve patient safety attitudes.

The present study has several strengths and limitations. The first strength is that it employed the JD-R model to examine how practice environment and job stress to the patient safety attitudes of new nurses. The second strength is multicentered study with a variety of samples' characteristics, which makes the results more representative and applicability, and the third strength is the model was tested using scientific methods to assess model fit. The study's limitations include: (1) the sample size was satisfied for testing SEM based on experts' recommendation [31], but 370 samples were not large enough to make the study results more applicable to other regions or countries. (2) Some other variables within the JD-R model had not been included in this study, which may influence the comprehensiveness of J-DR model regarding new nurses' practice environment, job stress, and patient safety attitudes, and (3) we did not consider the effect of covariates on SEM, so the effect sizes in the results should be referred with caution.