Chronic fatigue denotes a persistent, sustained state integrating affective, cognitive and somatic components, whereas acute fatigue reflects transient incapacity that manifests immediately after task cessation (18–20). Inter-shift recovery refers to the degree to which acute fatigue dissipates between duty cycles, the persistent accumulation of fatigue can trigger a vicious cycle of chronic fatigue (18–20). Insufficient recovery compromises subsequent performance and elevates error probability (20).

The Occupational Fatigue Exhaustion/Recovery (OFER) model, developed and validated by Winwood et al. (18, 19), operationalizes these dynamics through three correlated sub-scales: chronic fatigue, acute fatigue and inter-shift recovery. The instrument was originally normed on hospital nurses but has been cross-validated in allied health professionals, including nursing managers (20, 21).

Complementarily, Effort–Reward Imbalance (ERI) model by Siegrist reveals that sustained discrepancy between high extrinsic effort and low extrinsic reward elicits sustained stress arousal, thereby precipitating adverse health outcomes (22, 23). Meta-analytic evidence connects ERI with depression, anxiety, metabolic dysregulation and, critically, occupational fatigue (23, 24).

ERI theory locates the psychosocial source of work-related ill health—effort-reward imbalance—whereas the OFER quantifies its expression, occupational fatigue. Their integration yields a multilevel risk architecture that identifies vulnerable employees and informs interventions.

China’s healthcare context is likely to amplify the fatigue trajectory outlined by the OFER and ERI models. Hospitals are simultaneously experiencing a rapid increase in patient acuity driven by an ageing population. Culturally, Chinese nursing managers tend to score high on collectivism and prioritize individual sacrifice for the sake of group harmony. This norm implicitly endorses the practice: many western Chinese nursing managers voluntarily remain on site after formal handover to ensure ward stability, effectively compressing inter-shift recovery time.

In this cross-sectional study, data collection was facilitated through an online survey. A non-probability convenience sampling method was employed, resulting in a sample composition of 24.6% deputy nurse managers, 57.3% nurse managers, 10.5% departmental nurse managers, and 7.6% directors of nursing departments. At the inception of the study, formal consultation was sought from the Chinese Nursing Association (CNA) to delineate the objectives and methodology. This association holds the preeminent authority concerning the representation of the nursing profession within China, with the majority of its members occupying roles such as directors of nursing departments across various regions of the country. Prior to data collection, the OFER and ERI scales were adapted for use in a managerial population. Content validity was assessed through an expert review by five specialists in occupational health and work-related stress, who rated the relevance and clarity of each item on a 4-point scale. The scale-level content validity index (S-CVI/Ave) reached 0.933 for the OFER and 0.913 for the ERI, both exceeding the 0.90 criterion for excellent content validity. Because the research team was based in western China, recruitment prioritized local individual CNA members, resulting in a sample dominated by nursing managers from that region. A total of 266 questionnaires were distributed through the CNA network. All participants held managerial positions in the nursing field. In accordance with ethical standards, all participants received an informational sheet prior to enrollment detailing the study’s purpose, anticipated duration, participant rights, and privacy protection policies.

Inclusion criteria were as follows: (1) Nurses in managerial positions; (2) Nurses with no language, reading, or other cognitive impairments were observed. Exclusion criteria were: (1) Nursing managers who did not formally hold a position of responsibility; (2) Nursing managers in further studies and visiting scholars; (3) Retired nursing managers (nursing managers who had retired and stepped down from their positions at the time of the survey); (4) Nursing managers on leave of absence (e.g., maternity, sick, paid/unpaid leave of absence) during the data collection period; (5) Those participating in other fatigue intervention studies. The sample size for this study was determined using the formula N ≥ 8(m) + 50 (25), where “m” denoted the number of independent variables (m = 15), and “N” represented the sample size. Taking into account the possibility of 20% missing data, the expected minimum sample size was 213. A total of 266 questionnaires were distributed, and 254 were returned. We excluded any response that met at least one of the following pre-specified exclusion criteria: (1) ≥ 20% missing data, (2) completion time < 3 min. Ultimately, 248 participants were included in the analysis (N = 248).

Data were collected from January 17 to March 16, 2023. The survey was conducted using an easily accessible and user-friendly online platform, Survey Star.

The survey instrument comprised two sections: one addressing the general information pertinent to the study and the other focusing on occupational fatigue and stress queries. The general information section was structured to gather crucial demographic and occupational data, encompassing variables such as gender (men and women), age (20–30 years, 31–40 years, 41–50 years and >50 years), marital status (unmarried, married and divorced or widowed), educational attainment (three-year college, undergraduate, and master’s degree or above), hospital location (eastern part, central part, and western part), hospital grade (primary hospital, secondary hospital, and tertiary hospital), years of experience (6–10 years, 11–20 years, and >20 years), position title (deputy nurse managers, nurse managers, departmental nurse managers, and directors of nursing departments), professional title (nurse practitioner, nurse-in-charge, associate professor of nursing and professor of nursing), department (Supplementary Table S1), working hours per week (≤40 h, 41-45 h, 46-50 h, and >50 h), and the frequency of night shifts per month (0 shift, 1–4 shifts, and ≥5 shifts).

The assessment of occupational fatigue was conducted utilizing the Occupational Fatigue Exhaustion/Recovery Scale (OFER) (18–20). Chronic fatigue is defined as a complex psychological, physiological, and emotional state that encompasses multiple components, including a depressive mood (18–20). Acute fatigue is a relative state of “incapacity” resulting directly from fatigue caused by previous activities, which is most evident after work (18–20). Inter-shift recovery refers to the extent to which perceived acute work-related fatigue dissipates between consecutive work shifts. Insufficient inter-shift recovery may lead to persistent fatigue at the start of the next shift, thereby affecting work efficiency and quality (18–20). The OFER consists of three dimensions, each comprising five items. Using a 7-point Likert scale with scores ranging from 0 to 6, corresponding to “strongly disagree” responses to “strongly agree,” and items 9, 10, 11, 13, and 15 were reverse-scored (18–20). Each dimension was assessed by calculating a score derived from dividing the total number of items by 30 and multiplying by 100. The resulting scores were categorized as follows: 0–25 indicated low levels, 26–50 indicated mid-low levels, 51–75 indicated mid-high levels, and 76–100 indicated high levels of occupational fatigue (20, 21). A higher score corresponded to an elevated level of chronic fatigue, acute fatigue, or inter-shift recovery. Fang et al. evaluated the Chinese version of the OFER scale among nurses in mainland China and verified its robust psychometric properties (20). Their study reported Cronbach’s alpha values of 0.83, 0.85, and 0.86 for chronic fatigue, acute fatigue, and inter-shift recovery subscales, respectively (20). In the present study, we utilized the Chinese version of the OFER scale, and the Cronbach’s alpha values were determined to be 0.89, 0.87, and 0.78, respectively.

The Effort-Reward Imbalance (ERI) scale (22, 23, 26), is extensively utilized to assess perceived work-related stress, particularly among nurses. This scale comprises 23 items distributed across three dimensions: effort, reward, and overcommitment. Responses to the items in the effort and reward domains are rated on a scale from 1 to 5, where a score of 1 signifies no stressful experience, and a score of 5 signifies a very high stress level. The overcommitment domain items are rated on a scale from 1 to 4, with a score of 4 indicating complete agreement with the statement. The total scores range from 6 to 30 for the effort domain, 11 to 55 for the reward domain, and 6 to 24 for the overcommitment domain. Items 7 through 17 are reverse-scored. Notably, work-related stress is a pivotal component of the ERI model, which is frequently employed to evaluate the impact of work-related stress on individuals’ physical and mental health by examining the imbalance between input and output (22, 23, 26). Work-related stress can be quantified using the Effort-Reward Ratio (ERR) and overcommitment. Effort-Reward Ratio (ERR) is calculated by dividing the effort score by the reward score, which is then multiplied by a correction factor of 0.5454 to account for the number of items assessed (26). A value of 0 suggests a positive working environment characterized by a manageable workload and greater profits, while a value exceeding 1.0 signifies a substantial discrepancy between effort and performance (26). The Chinese adaptation of the ERI scale has exhibited strong psychometric properties, as evidenced in prior research (26). In the present study, Cronbach’s alpha coefficients for the effort, reward, and overcommitment subscales were reported as 0.90, 0.89, and 0.77, respectively.

The data analysis was conducted utilizing SPSS software, version 27.0. The demographic and occupational variables were categorical data. Some demographic variables (marital status, education, department, and night shifts per month) were dichotomized at the median value of the sample; no statistically significant differences were observed before or after dichotomization (all p > 0.05). Data from the OFER scale and the ERI scale exhibited non-normal distribution. Based on predefined criteria (detailed as OFER scale brief) (20, 21), we categorized OFER scores of chronic fatigue, acute fatigue, and inter-shift recovery into four groups (e.g., chronic fatigue levels for low, mid-low, mid-high, and high). Comparative analysis between demographic or occupational variables and the OFER scale was performed using the Chi-Square. Categorical associations were first examined with Pearson’s χ² test. When more than 20% of cells had expected counts below 5, the Fisher–Freeman–Halton (FFH) exact test was used. If the complete enumeration required for the FFH test exceeded the default 5-min time limit in SPSS, we replaced it with the Monte Carlo exact method based on 10,000 random tables; 99% confidence intervals for the p-values are reported. We assessed the correlation between OFER and ERI via Spearman correlation analysis. After checking the proportional-odds assumption, we fitted the appropriate ordinal or multiple logistic regression model and included variables significant at p < 0.05 to estimate their associations with occupational fatigue.

All procedures performed in studies involving human participants were in accordance with the ethical standards of the Ethics Committee of West China Second Hospital of Sichuan University (Ethics No.: 2020114) and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards. Informed consent was obtained from all individual participants included in the study.

In our study, the scores on the three dimensions of the OFER scale (chronic fatigue, acute fatigue, and inter-shift recovery) were found to be non-normally distributed. We summarized them using frequencies and percentages (Supplementary Table S2). Similarly, the three dimensions of the ERI scale (Effort, Reward, and Overcommitment) also exhibited non-normal distributions. These dimensions were characterized using medians and quartiles (Supplementary Table S3). As illustrated, most nursing managers reported experiencing mid-low to mid-high levels of chronic fatigue, mid-high to high levels of acute fatigue, and mid-low to mid-high levels of inter-shift recovery. Concurrently, the median scores for effort and overcommitment were lower than those for the reward score. Furthermore, 64.92% of the ERRs were lower than value 1.

Table 1 presents the distinct characteristics of nursing managers categorized by varying levels of chronic fatigue, acute fatigue, and inter-shift recovery. Significant differences were identified in hospital grade across chronic fatigue levels (p < 0.05). Furthermore, significant differences were noted in hospital grade, professional title, position title, and weekly working hours across varying levels of acute fatigue (p < 0.05). Similarly, significant differences were observed in hospital location and weekly working hours among different levels of inter-shift recovery (p < 0.05).

Table 2 illustrates the positive correlation between chronic fatigue levels and effort and overcommitment scores (p < 0.05). Similarly, acute fatigue levels also exhibit a positive correlation with effort and overcommitment scores (p < 0.05). In contrast, inter-shift recovery demonstrates a negative correlation with both effort and overcommitment scores (p < 0.05). Although effort score and overcommitment score were moderately inter-correlated, auxiliary ordinary least squares regression indicated no problematic multicollinearity (tolerance = 0.61, variance inflation factor = 1.65).

Regression analysis examined the relationship between participants’ characteristics, work stress and levels of chronic fatigue, acute fatigue, and inter-shift recovery. In this analysis, chronic fatigue, acute fatigue, and inter-shift recovery levels were designated dependent variables, and the participants’ characteristics and work stress variables with statistical significance (p < 0.05) in the Chi-Square test or Spearman correlation analysis were classified as independent variables.

Table 3 presents the ordinal regression model for the level of chronic fatigue, χ² (4) = 156.20 (p < 0.001); no concerning multicollinearity was detected among the variables, with all variance inflation factors (VIFs) ranging from 1.17 to 1.67. Prior to model fitting, the proportional odds assumption was evaluated using the parallel-lines test, χ² (8) = 8.68 (p = 0.37), indicating that the assumption was not violated and the ordinal logistic model was therefore appropriate. The analysis revealed that the effort score (OR = 1.23, 95% CI 1.16–1.31) and overcommitment score (OR = 1.32, 95% CI 1.19–1.46) were significantly associated with chronic fatigue levels. Compared with tertiary hospitals, primary hospitals (OR = 5.11, 95% CI 1.21–21.31) were also associated with higher chronic fatigue. The Cox & Snell R² of the model was 0.47, suggesting that these variables account for approximately 47% of the variation in chronic fatigue levels among nursing managers.

Table 4 presents the multiple logistic regression model for acute fatigue. Since the overall parallel-lines test was significant χ² (26) = 56.88 (p < 0.001), we employed multiple logistic regression. The model for acute fatigue, χ² (39) = 193.82 (p < 0.001); no concerning multicollinearity was detected among the variables, with all variance inflation factors (VIFs) ranging from 1.16 to 1.82. Using high acute fatigue as the reference, higher effort scores were associated with lower odds of low (OR = 0.77, 95% CI 0.66–0.90), mid-low (OR = 0.73, 95% CI 0.64–0.83) and mid-high acute fatigue (OR = 0.84, 95% CI 0.76–0.93), while higher overcommitment scores showed the same downward gradient across these three levels. A ≤ 40 h work-week increased the likelihood than a >50 h of mid-low (OR = 20.00, 95% CI 1.45–275.85) and mid-high fatigue (OR = 15.21, 95% CI 1.46–158.62). Primary hospital and deputy/departmental nurse manager positions, relative to tertiary hospitals and directors of nursing positions, exhibited directionally varied yet consistently protective associations (OR 0.02–0.31, 95% CIs below 1). The Cox & Snell R² was 0.54, suggesting that these variables collectively account for approximately 54% of the variables in acute fatigue levels among nursing managers.

Table 5 presents the ordinal regression model for inter-shift recovery, χ² (7) = 100.54 (p < 0.001); no concerning multicollinearity was detected among the variables, with all variance inflation factors (VIFs) ranging from 1.03 to 1.93. Prior to model fitting, the proportional odds assumption was evaluated using the parallel-lines test, χ² (14) = 18.15, p = 0.20, indicating that the assumption was not violated and the ordinal logistic model was therefore appropriate. The analysis revealed that the effort scores (OR = 0.86, 95% CI 0.81–0.92), overcommitment scores (OR = 0.84, 95% CI 0.76–0.93), and working hours per week were significantly associated with inter-shift recovery levels. The Cox & Snell R² was 0.33, suggesting that these variables collectively account for approximately 33% of the variation in inter-shift recovery levels among nursing managers.