Examining the impact of perceived social support on mental workload in clinical nurses: the mediating role of positive coping style

This study aims to determine the status of nurses suffering from mental workload and to investigate relationship between mental workload and perceived social support, as well as the mediating effect of positive coping style for clinical nurses.

This was a cross-sectional study adhering to follow STROBE Statement (www.​Equator-network.org).

The data collected a total of 590 nurses from 3 tertiary hospitals in Chengdu between April and July 2023 with convenience sample. To be eligible, the nurses should meet the following inclusion criteria: (a) the certified registered nurses, (b) willing to participate in this study, (c) worked fulltime and directly involved in patient care. We excluded float nurses, nursing students, or additional nurses who were not on work due to illness, maternity leave. All of the participants in the survey worked in the emergency, intensive care, oncology, internal medicine, surgery, operating, pediatric, and psychiatry department. A mediation model was conducted, the sample size could follow the requirements at least 10 to 20 times the number of variable [25]. A total of 650 questionnaires were distributed, 590 questionnaires were valid, and the response rate of 90.8%.

The National Aeronautics and Space Administration Task Load Index (NASA-TLX) scale was used to measure the mental workload of the participants, which was developed by Hart and Staveland [26]. The Chinese version of NASA’s task load index scale was translated and culturally adapted by Liang to measure the level of nurses’ mental workload [27]. And the Chinese version of NASA’s task load index scale consists of 2 dimensions called workload perception and self-evaluation, where workload perception consists of 4 items (mental demand, physical demand, temporal demand and effort) and self-evaluation consists of 2 items (performance and frustration). The specific meanings of these items are as follows, mental demand, which indicates the process of required mental activity, such as thinking and calculating; physical demand, which signifies the need for physical activity, such as pushing and drawing; temporal demand defined as the time to complete the task; performance, which is an evaluation of satisfaction demonstrated in completing the task; effort, which indicates the level of effort an individual puts into completing a task; frustration, which is the discouragement and failure felt when doing a task. Each items have a score from 0 “low” to 20 “high” by a straight line, but performance item is reverse scored. Additionally, in this study, we used to calculate by adding the scores of six items rather than the weighted score, ranging from 0 to 100, and the higher scores indicating higher level of mental workload. The Cronbach’s α for Chinese version of NASA-TLX scale is 0.71. In this study, the Cronbach’s α of scale was 0.77. The model fitting index of confirmatory factor analysis (CFA) showed RMSEA = 0.045, GFI = 0.991, CFI = 0.993. The construct reliability (CR>0.6) and average variance extracted (AVE>0.5) of workload perception and self-evaluation were CR = 0.858, AVE = 0.603, Square root of AVE = 0.777 and CR = 0.730, AVE = 0.579, Square root of AVE = 0.761 (see Fig. 1 in Supplementary information). Further analysis of the discriminant validity found that the pairwise correlation between the factors was statistically significant and smaller than the corresponding square root of AVE, indicating that there was a certain degree of discriminant between the factors.

Social support was assessed by the Scale of Perceived Social Support (SPSS). it was developed by Zimet [15], and the Chinese version of SPSS translated by Jiang et al. with high reliability and validity [28]. The PSSS contains three dimensions, namely, support of family members, friends, and any significant others (colleague or leaders) and each dimension has 4 items, a total of 12 items. Among the SPSS, participants were asked to answer with a 7-point Likert scale, ranging from 1 “strongly disagree” to 7 “strongly agree”, and with higher scores indicating higher level of perceived social support. Additionally, the Cronbach’s α for the Chinese version of the scale is 0.80. In this study, the Cronbach’s α of scale was 0.97. The model fitting index of confirmatory factor analysis (CFA) showed RMSEA = 0.068, GFI = 0.947, CFI = 0.984. The construct reliability (CR>0.6) and average variance extracted (AVE>0.5) of family, friends and any significant others support were CR = 0.935, AVE = 0.784, Square root of AVE = 0.885, CR = 0.943, AVE = 0.810, Square root of AVE = 0.900 and CR = 0.942, AVE = 0.801, Square root of AVE = 0.895 (see Fig. 2 in Supplementary information). Further analysis of the discriminant validity found that the pairwise correlation between the factors was statistically significant and smaller than the corresponding square root of AVE, indicating that there was a certain degree of discriminant between the factors.

The Simplified Coping Style Questionnaire (SCSQ) was revised by Xie to be suitable with high reliability and validity in China populations [21]. The SCSQ was extracted into the positive coping style subscale and the negative coping style subscale. The simplified coping style questionnaire-Positive Coping Style (SCSQ-PCS) has 12 items including divert attention, seek support, adopt problem-solving methods, suppress negative emotions and participating activities. Additionally, participants were asked to answer with a 4-point Likert scale, ranging from 0 “never used” to 3 “always used”. The higher positive coping scores indicated positive coping style. In this study, the Cronbach’s α of scale was 0.74. The model fitting index of confirmatory factor analysis (CFA) showed RMSEA = 0.048, GFI = 0.967, CFI = 0.984. The construct reliability (CR>0.6) and average variance extracted (AVE>0.5) of positive coping style were CR = 0.945, AVE = 0.590, Square root of AVE = 0.768 (see Fig. 3 in Supplementary information). Further analysis of the discriminant validity found that the pairwise correlation between the factors was statistically significant and smaller than the corresponding square root of AVE, indicating that there was a certain degree of discriminant between the factors.

All data were collected electronic questionnaire through online questionnaire service platform in China (Questionnaire Star). Before the survey, we obtained the consent and support of the nursing management department of each hospital. With the assistance and cooperation of nursing supervisors in each department, we distributed electronic questionnaires through the Questionnaire Star platform to clinical nurses who met the inclusion criteria. At the same time, the process of collecting questionnaires fully followed to the principles of informed consent and confidentiality. In addition, the instructions on the first page of the questionnaire included guidelines for informed consent, requirements for completing the questionnaire, and instructions on data privacy protection, ensuring that informed consent and voluntary participation of each participant in this study. In addition, each hospital has trained surveyors responsible for the data collection process of the corresponding hospital, including the questionnaire filling instructions, how to successfully submit the questionnaire, and regularly log in to the Questionnaire Star platform for data monitoring. To ensure the quality of the questionnaire, surveyors can log in to the Questionnaire Star platform to set questionnaire management measures, such as limiting each IP address to submit questionnaire only once and requiring all questions to be answered before successful submission. After data collection, two researchers checked the quality of questionnaire and eliminated invalid questionnaires. The criteria of invalid questionnaires were following: (a) apparently regular and repeated responses to the questionnaire, (b) questionnaires with conflicting answers (the working years are not consistent with the age), (c) repeated submission of questionnaires, (d) questionnaires with excessively long or short completion times.

Data were exported from Questionnaire Star to Statistical Package for Social Sciences (SPSS) software version 26.0 and double-checked by two researchers. And missing data were eliminated at the first checked the quality of data collection and delete outliers and extremes values. In addition, the absolute values of maximum skewness and maximum kurtosis of variables range from 0.065 to 0.583, so the normal distribution of data were guaranteed. The descriptive analysis was presented with demographic characteristics and work-related information by percentage. Similarly, the descriptive analysis of mental workload, perceived social support and positive coping style were determined with means and standard deviations (SD). T-test and analysis of variance (ANOVA) were used to analyze the influence of demographic factors on mental workload. Furthermore, multiple linear regression and Pearson’s correlation analysis was conducted to evaluate the relationship among mental workload, perceived social support and positive coping style. Finally, the AMOS 26.0 software was used to examine the mediation model path between three variable relationships. In the study, the likelihood-ratio chi-square/degree of freedom ratio (χ²/df<3), goodness-of-fit index (GFI>0.9), adjusted goodness of fit (AGFI>0.9), comparative fit index (CFI> 0.9), Tucker-Lewis index (TLI>0.9), normed fit index (NFI>0.9), Incremental Fit Index (IFI>0.9), root mean square error of approximation (RMSEA<0.08) were used to judge the rationality of the model. In addition, all p ≤ 0.05 were supposed to be significant with a two-tailed test and the mediational effect is significant if the CI is 95% and the bias-corrected lower and upper limits with not including zero through a bootstrapped sample of 5000 [29].

The total of 590 questionnaires and the descriptive analysis were described all characteristics information of participants. It was observed that most of the nurses were female (85.25%), married (55.42%) and with children (51.86%); the mean score of age was 31.41 ± 8.55 years. The major education level of nurses was bachelor degree (64.75%). Moreover, nurses (61.36%) attended the psychological training (Table 1).

The mental workload score of 590 participants was 68.86 ± 17.53, where the score of the dimensions were workload perception (48.11 ± 14.27) and self-evaluation (20.75 ± 6.18). Perceived social support score was 59.08 ± 15.51. Positive coping style were21.69 ± 8.32 (Table 2). Moreover, nurses who not attended psychological training reported higher mental workload (p<0.05) (Table 1).

The results showed that the higher the perceived social support, the lower the mental workload (r= -0.340, p < 0.01). Similarly, the more inclined to adopt positive coping style, the lower the mental workload (r= -0.348, p < 0.01). In addition, with the increase of perceived social support, positive coping style will also increase (r = 0.572, p < 0.01) (Table 2).

Multiple linear regression models were used to explore the impact of variables with statistical significance from the univariate analyses (psychological training), perceived social support and positive coping style on mental workload. The collinearity diagnosis indicated no multicollinearity among the covariates. The results showed that the multiple linear regression model explained 16.6% of the total variation of mental workload, indicating that psychological training (β=-0.143, p<0.010), positive coping style (β=-0.250, p<0.010) and perceived social support (β=-0.209, p<0.010) can negatively predict mental workload (Table 3).

Our study further explored the relationship between mental workload, perceived social support and positive coping style by structural equation model, as shown in Fig. 1.

In addition, the mediator of positive coping style was examined by 5000 bootstrap analyses with 95% confidence intervals. All the fit indices were appropriated as follow: χ²/df = 1.818, GFI = 0.993, AGFI = 0.979, TLI = 0.995, NFI = 0.995, CFI = 0.998, RMSEA = 0.037. It indicated that the fit indices well. Also, perceived social support had a significant direct effect of on mental workload (β= -0.296, p < 0.001), account for 64.8% of the total effect. And the direct effect of positive coping style on mental workload (β= -0.282, p< 0.001). The above data indicated that hypothesis of H1 was valid. Furthermore, the indirect mediating effects of perceived social support → positive coping style → mental workload was − 0.161 (p< 0.001) with a 35.2% of total effect value. Consequently, the result suggested that positive coping style partially mediates the relationship between perceived social support and mental workload, and the hypothesis of H2 was valid. See above data (Table 4; Fig. 2).

It is worth noting that this study has some important implications. The theoretical significance of this study is to expand and verify the understanding of human-based dynamics model of mental workload, which believes that resource supply and individual characteristics are the key variables that affect mental workload. Our results showed that positive coping style play a mediating role between perceived social support and mental workload, indicating that improving perceived social support and individual tendency to adopt positive coping style can reduce the level of mental workload. This study has expanded the empirical research results of the application of human-based dynamics model of mental workload in clinical nurses, and enriched the localization research of the model in Chinese nurses.

There are some limitations in this study. Firstly, although variables of the cross-sectional design of this study were established using structural equation model on the basis of hypothesized relationships according to the literature review, we cannot draw causal explanations at a clearly defined time frame. Therefore, the further longitudinal design should be considered. Secondly, the convenience sampling method was employed, and the sample consisted exclusively of nurses from general tertiary hospitals in China. This sampling approach may not fully capture the characteristics of the target population, as it excludes other occupational groups and regions. Additionally, the findings of this study are primarily relevant to clinical nurses in a specific hospital, limiting the generalizability of the results to nurses in other medical institutions or from diverse cultural backgrounds. Therefore, future research should utilize stratified sampling to include a broader range of regions and more varied populations, thereby improving the generalizability of the findings. Lastly, future studies could examine other factors influencing mental workload, such as organizational culture, leadership styles, or workload distribution. Longitudinal research designs could also be employed to explore causal relationships between these variables over time.