The association between team job crafting and work engagement among nurses: a prospective cohort study

doi:10.21203/rs.3.rs-2759914/v1

Background

Team-level job crafting has been put forward as a method to promote nurses' mental health. However, a longitudinal association is unclear. Therefore, the objective of this study was to investigate the association between team job crafting at baseline and work engagement, work performance, psychological distress, and intention to leave at three-month and six-month follow-ups among Japanese hospital nurses. Also, whether an increase in the team job crafting during 3 or 6 months was associated with an increase in the work engagement during 3 or 6 months of individual nurses was examined.

Methods

A multilevel prospective cohort study was conducted. Data were collected from nurses of five hospitals in Japan at baseline (T1) and follow-ups at 3-months (T2) and 6-months (T3). A total of 2,478 nurses were included. The team job crafting scale for nurses and its three subscales were measured for the independent variables. Ward-means were used as ward-level variables. The dependent variables were work engagement, work performance, psychological distress, and intention to leave. Hierarchical Linear Modeling (HLM) was used to examine the multilevel association.

Results

Data from 391 nurses nested in 30 wards were included in the analyses. The intraclass correlation coefficients (ICCs) at T1 were 0.02 for work engagement and 0.07 for team job crafting. The HLM revealed that ward-level team job crafting at T1 was not significantly associated with work engagement, work performance, psychological distress, and intention to leave at T2 or T3. The ward-level change (T3-T1) of “crafting for the task considering the team’s growth” (subscale for team job crafting) was significantly and positively associated with the change (T3-T1) in work engagement.

Conclusions

Ward-level team job crafting at baseline did not predict nurses' work engagement at the follow-ups. This finding contradicted previous cross-sectional and one-month follow-up studies in other countries. The impact of ward-level team job crafting on work engagement may be limited to more than a three-month follow-up.

Throughout the world, nurses have an important role in hospital care and are known to engage in stressful work. Poor working conditions for nurses, such as a high workload, professional conflicts, the emotional cost of caring, and shift work, have a negative impact on mental health (1). A recent study reported that one-third of nurses had depression and anxiety, and that half suffered from high emotional exhaustion and post-traumatic stress disorder (2). The prevalence of mental health problems was reported to be higher than those of other occupations (3). Improving a psychosocial work environment and mental health is necessary for nurses.

Recently, positive mental health has been a focus in the field of occupational health psychology (4). Work engagement, one of the most well-known positive mental health indicators, is defined as a positive, fulfilling, work-related state of mind characterized by vigor, dedication, and absorption (5). Previous studies have reported that work engagement was negatively related to symptoms of depression and anxiety (6), burnout (7), and psychological distress (8), and was positively related to job satisfaction and compassion satisfaction (7). In the nursing setting, work engagement is valuable not only to individual nurses but also to patients and hospitals (9, 10). For patient and hospital outcomes, work engagement was negatively associated with turnover intention from the hospital (7, 11) and positively associated with patient safety (12), work effectiveness (13), perceived care quality (14), and organizational commitment (15).

According to the Job Demands-Resources (JD-R) model, job resources are proposed as an important predictor of work engagement (16). Job resources were classified into three levels: task-level, interpersonal-level, and organizational-level (16). In nursing settings, interpersonal-level and organizational-level job resources are more emphasized to improve work engagement among nurses (9). Accordingly, improving such job resources could be effective.

Employees can improve their job resources by job crafting (17, 18). Job crafting was conceptualized in 2001 by Wrzesniewski and Dutton. It is defined as the physical and cognitive changes individuals make in the tasks or relational boundaries of their work (17). Job crafting involves three actions: task crafting, cognitive crafting, and relational crafting (17).

Recently, team-level job crafting, a new concept of job crafting, has been offered to increase nurses’ job resources (19). In many situations, job crafting occurs at the team level as well as at the individual level because workers in a workgroup typically experience common events, engage in the same work processes, and interact with each other (20). Leana et al. (2009) proposed "collaborative job crafting," which is a process of joint determination by a group of employees on how to alter their work to meet their shared objectives (20). Tims et al. (2013) also proposed team job crafting with the definition of "the extent to which team members combine efforts to increase structural and social job resources as well as challenging job demands and to decrease their hindering job demands," according to the JD-R model (21). There are three aspects relevant to nurses' team job crafting: crafting for the task considering the team's growth, crafting for members' respect and reflection of meaningful work, and crafting for smooth information sharing (19). For nurses who work in teams, team job crafting is expected to improve job resources more efficiently and thus is more likely to contribute to improved work engagement. While individual workers perform job crafting to fulfill their individual needs, team job crafting is accomplished by team members to achieve team goals or objectives (17, 19, 22). Thus, team job crafting is expected to improve interpersonal and organizational-level job resources.

A cross-sectional association between team job crafting and work engagement has been reported among general (non-healthcare) workers. For example, team job crafting was reported to be associated with individual work engagement, job satisfaction, work performance, organizational commitment, and quality of care (20, 21, 23, 24). For nurses, there were two previous studies. A cross-sectional study found a positive association between team job crafting and individual-level work engagement among medical professionals from a Chinese public hospital (25). Another study reported that baseline team job crafting was related to team-level work engagement at one-month follow-up among clinicians, including physicians and nurses in Vietnam (26). Therefore, team job crafting may also be associated with better work engagement among nurses.

However, previous studies investigated only cross-sectional or short-term (one-month) associations. The evidence of cross-sectional studies has the possibility of causal reversal between its association. While there was only one longitudinal study that clarified the association between team job crafting and work engagement, the dependent variable was team-level work engagement. Thus, the longitudinal relationship between ward-level team job crafting and individual-level work engagement has not been examined. Also, more than the one-month longitudinal relationship between team job crafting and work engagement is unknown.

The degree of association between team job crafting and work engagement may vary depending on the degree of the organizational climate. One aspect of the organizational climate is workplace social capital, which is described as the contextual psychosocial elements of a workplace characterized by interpersonal trust and reciprocity rules (27). Psychological safety could be another aspect of organizational climate, defined as the shared belief by team members that the team can safely take risks (28). It is argued that job crafting behavior improves positive affective well-being when the response from colleagues is supportive. Conversely, it improves negative affective well-being when the reaction from colleagues is hostile (29). Therefore, the team job crafting behavior would be more readily accepted by the team members in a workplace with a high organizational climate (28, 30). On the other hand, team job crafting would be less acceptable in a ward with an unaccommodating organizational climate because members may feel forced to do it, making it challenging to improve work engagement. Thus, team job crafting is more likely associated with work engagement in a ward with greater workplace social capital and psychological safety.

Additionally, team job crafting may be positively associated with work performance and negatively associated with psychological distress and intention to leave. Previous studies reported significant positive associations between team job crafting and job performance (20, 21, 24, 26). Also, previous studies reported significant negative associations between individual job crafting and psychological distress and intention to leave (31, 32). However, the longitudinal association between team job crafting and these outcomes was unknown.

This study aimed to investigate the longitudinal association between ward-level team job crafting at baseline and individual-level work engagement at three-month and six-month follow-ups among Japanese nurses. Also, the subgroup analyses investigated the difference in the association between ward-level team job crafting at baseline and individual-level work engagement at three-month and six-month follow-ups by organizational climate (i.e., high/low workplace social capital wards and high/low psychological safety wards). The associations between team job crafting at baseline and individual-level work performance, psychological distress, and intention to leave at three-month and six-month follow-ups were also examined as secondary outcomes.

In addition to examining the association between baseline team job crafting and work engagement in the follow-up surveys, the change-to-change relations among ward-level team job crafting and individual-level work engagement from the baseline to the follow-up survey by using scores of differences would also be helpful. A previous researcher demonstrated that investigating hypotheses in which changes in one construct are used to predict changes in another construct is a good way to find connections between dynamic processes (33). Therefore, We investigated whether an increase in the ward-level team job crafting score has a contextual effect with an increase in the work engagement scores of individual nurses nested in the ward.

Hypothesis

H1: Ward-level team job crafting (total score and each of three subscales) at baseline would be positively associated with individual-level work engagement at the three-month and six-month follow-ups.

H2: The degree of association between ward-level team job crafting (total score and each of three subscales) at baseline and individual-level work engagement at the three-month and six-month follow-ups would be greater in wards with higher degrees of workplace social capital and psychological safety at baseline.

H3: Ward-level team job crafting (total score and each of three subscales) at baseline would be positively associated with individual-level work performance and negatively associated with psychological distress and intention to leave at the three-month and six-month follow-ups.

H4: Ward-level change of team job crafting (total score and each of three subscales) from baseline to three-month and six-month follow-ups would be positively associated with an individual-level change of work engagement.

Study design

This was a prospective cohort study. Multilevel data were collected three times from hospital wards and nurses and midwives working at target hospitals. The first survey was conducted from December 2021-January 2022 (T1), the second survey was conducted in March 2022 (T2), and the third survey was conducted from June-July 2022 (T3). A self-report questionnaire assessed individual-level variables. The study protocol was registered at the UMIN Clinical Trials Registry (ID = UMIN000047810). The study was approved by The Research Ethics Committee of the Graduate School of Medicine/Faculty of Medicine at the University of Tokyo [No. 2018175NI-(3)]. This study has been reported according to the Strengthening the Reporting of Observational Studies in Epidemiology (34) (see Appendix 1).

Participants

Japanese hospitals having inpatient wards were approached using snowball sampling methods. The inclusion criteria of the wards were (a) wards in Japanese hospitals and (b) employed two or more nurses. Outpatient wards and clinics were excluded. The inclusion criteria of individuals were nurses and midwives working in the included hospital wards. Nurses and midwives working in outpatient wards and clinics were excluded.

Variables

Independent variables

Team job crafting was measured by the Japanese Team Job Crafting Scale (29). Internal consistency reliability and concept validity were determined to be acceptable. The scale includes 13 items (e.g., "The proficiency levels of newcomers are properly understood to help them grow", “positive events are shared among team members in order to reaffirm the value of our work,” and “a patient’s information is shared among team members even outside of meetings in order to provide better care for the patient as a team.”) measured on a five-point response scale (1 = not at all, 5 = extremely). Higher scores indicate higher team job crafting behavior in a ward. The total score, as well as each subscale score, was calculated by dividing the sum of item scores by the number of the items. The mean score of the three subscales (overall) was calculated as individual-level scores. The mean score of each of the three subscales (subscale 1, subscale 2, subscale 3) was also calculated. Ward-level scores were calculated by the ward-mean score (overall, subscale 1, subscale 2, subscale 3). Both individual- and ward-level scores were grand-mean centered on establishing a useful zero point (35).

Dependent variables

Work engagement

As the primary dependent variable, work engagement was assessed using the 9-item Japanese version of the Utrecht Work Engagement Scale (UWES) (36, 37). The UWES has nine items (e.g., "At my job, I feel strong and vigorous"). Every item was measured on a seven-point Likert scale ranging from 0 (Never) to 6 (Always), with higher scores meaning higher work engagement. Previously, the reliability and validity of the Japanese UWES were examined (37). Mean scores for each of the nine items were calculated.

Work performance

The validated Japanese short version of the WHO Health and Work Performance Questionnaire (WHO-HPQ) was used to evaluate work performance (38). The scale consists of one item that rates an individual's total job performance over the preceding month ranging from 0 (Worst) to 10 (Best). Higher scores mean higher work performance.

Psychological distress

Psychological distress was assessed with the Japanese version of the K6 questionnaire, which consists of six items and asks respondents how frequently they had experienced psychological distress symptoms in the previous 30 days (39). The performance of the scale is essentially identical (39). On a five-point Likert scale, responses range from 0 (None of the time) to 4 (All of the time), with higher scores indicating higher psychological distress. A total score (0 to 24) was calculated.

Intention to leave

Intention to leave was measured with the four-items of the Japanese version of the intention to leave scale (e.g., "I consider my decision to work for this employer as an obvious mistake") (40, 41). The participants responded to each item on a five-point Likert scale ranging from 1 (I disagree completely) to 5 (I agree completely). Higher scores indicate higher intention to leave. The mean score of the items was calculated.

Confounding variables

Job stressors, workplace social support, and demographic variables were measured as confounding variables. For the multilevel analysis, these variables were treated as individual-level and grand-mean-centered.

Job demands, job control, supervisor/coworker support

Job demands, job control, and supervisor/coworker support were measured by items from the Brief Job Stress Questionnaire, which is widely used to assess psychosocial factors at work in Japan (42, 43). All items were scored on a four-point Likert scale ranging from 1 (Not at all) to 4 (Definitely), with higher scores indicating higher job demands, job control, and supervisor/coworker support. The total score was calculated by dividing the sum of item scores by the number of items in the present study. Test-retest reliability and validity were reported to be at adequate levels (43).

Effort-Reward Imbalance

Effort-reward imbalance was measured by the validated Japanese short version of the effort-reward imbalance (ERI) questionnaire (44, 45). The effort scale consists of three items related to work pressure and immersion, and the reward scale consists of seven items related to financial, esteem-related, and organizational rewards. The items were rated on a 4-point Likert scale (1 = strongly disagree, 4 = strongly agree). Higher scores mean higher effort and reward. Sum scores for effort and reward were calculated, and higher scores indicate more effort and reward, respectively. A high ERI ratio reflects a lack of reciprocity between efforts and rewards (high cost/low gain). A study provided that the cut-off point of 1.4 in the Japanese short version of the ERI is almost equivalent to the cut-off point of 1.0 in the English version (46). Therefore, we used the cut-off point of 1.4 and created a dichotomy variable (> 1.4 and \(\le\)1.4).

Demographic variables

Gender (men, women, other, no response), age (20–29, 30–39, 40–49, 50–59, 60-), educational status (professional school, university/graduate school, other), marital status (not married, married, no response), and professional experience (years) (1–4, 5–9, 10–19, 20–29, 30-) were measured. Adjusting for confounding variables in the analyses, "no response" selected for the questions about gender and marital status was included in the "no response" category.

Variables for the subgroup analyses

Workplace social capital

Workplace social capital was measured by the Japanese version of the Workplace Social Capital Scale (47, 48). The Japanese version was preliminarily tested and proved to have an acceptable level of internal consistency reliability and construct validity (48). The scale includes eight items (e.g., "We have a 'we are together' attitude") measured on a five-point Likert scale, and the mean score was calculated. Higher scores indicate higher workplace social capital. The ward-level score was calculated by the ward mean of the individual score and grand-mean centered.

Psychological safety

Psychological safety was measured by the Japanese version of the Psychological safety scale (28, 49). Internal consistency reliability and validity were reported at an acceptable level. The scale consists of seven items measured on a seven-point Likert scale, and the mean score was calculated. Higher scores mean higher psychological safety. The ward-level score was calculated by the ward mean of the individual score and grand-mean centered.

Sample size calculation

The target sample size was calculated by accounting for intraclass correlations (ICC) of the organizational outcomes (50). Sample sizes should be multiplied by the design effect (1+[m − 1] ρ), where m is the average cluster size, and ρ is ICC. In a previous study, ICC for work engagement among Japanese nurses was 0.09, and the number of clusters was set to 22 (51). The effect size of team job crafting was estimated at 0.20 based on a validation study of the team job crafting scale (19). Based on these estimations, 2277 participants from 103 wards were needed at an alpha error rate of 0.05 and a beta error rate of 0.20 using G*Power version 3.1.9.6 (52).

Statistical analysis

In the analyses, wards with one or fewer respondents were additionally excluded because ward-level team job crafting, calculated as an average of ward members' ratings, might not be stable and it is more likely to be affected by individual characteristics of the ward members (53).

First, descriptive statistics for the variables and the ICCs for team job crafting, work engagement, performance, psychological distress, intention to leave, workplace social capital, and psychological safety were calculated. The design effect was also calculated to assess a need for multilevel framework data analysis. Then, HLM was conducted to investigate multilevel relationships between team job crafting at T1 and dependent variables at T2 and T3 to test the first, second, and third hypotheses.

Multilevel regression modeling techniques (54, 55) were employed to investigate the hypotheses. The data were conceptualized as a two-level model, consisting of an individual at the first level and a ward at the second level. Four steps of multilevel analyses at T2 and T3 were conducted. The first step of the analysis was to determine the variance at individual and ward levels without explanatory variables (null model: Model 0). This model contains random groups and random variation within groups. Then, individual-level and ward-level team job crafting were entered into the null model (crude model: Model 1), the T1 dependent variable was entered into Model 1 (T1 dependent variable adjusted model: Model 2), and demographic variables and work-related variables were entered into Model 2 (demographic and work-related variables adjusted model: Model 3). The overall team job crafting score were included as independent variables for Model 1, Model 2, and Model 3. The score of three team job crafting subscales were included only in Model 3. Akaike Informal Criteria (AIC) was calculated to check the best fit model. The equation for the adjusted model (Model 3) could be explained as follows.

Level 1 (individual-level)

Y (work engagement at T2 and T3) _ij = β _0j + β_1j* (team job crafting)_ij+β_2j* (T1 work engagement)_ij+ β_3j* (gender)_ij+ β_4j* (age)_ij+ β_5j* (education)_ij+ β_6j* (marital status)_ij+ β_7j* (job demand)_ij+ β_8j* (job control)_ij+ β_9j* (supervisor support)_ij+ β_10j* (coworker support)_ij+ β_11j* (effort-reward imbalance)_ij+ e_ij

Level 2 (ward-level)

β_0j = γ₀₀ + γ₀₁* (team job crafting)_0j + u_0j

Note

Y_ij means the score of work engagement of individual nurse i who attends wards j. e and u are assumed to be error terms.

To test the fourth hypothesis, HLM were also conducted. The ward-level change in team job crafting from T1 to T2 and from T1 to T3 was put into the model as independent variables. Each score of three team job crafting subscales as well as total score were included the analyses. The individual-level change in work engagement from T1 to T2 and from T1 to T3 was put into the model as dependent variables.

Nurses with no data on wards have been excluded from the analyses. Using the multiple imputation approach for parameter estimation (56, 57), the wards and nurses with partially missing values, except for the ward data, or who dropped out of the study during follow-up were included in the evaluated model. An independent variable, dependent variables, and confounding variables were used for the imputation. A total of 100 simulated completed datasets were generated by predictive mean matching. R version 4.2.1 with mice package for the multiple imputations was used for each analysis (58).

Subgroup analyses

First subgroup analyses were conducted separately for wards with high scores (higher than or equal to the median) and low scores (less than the median) of workplace social capital and psychological safety at baseline. Although previous studies used a quartile point for cut-off for workplace social capital (59, 60) and psychological safety (61), this study used the median score because of the smaller sample size.

Changes to protocol

Two changes were made to the protocol registered at the UMIN Clinical Trials Registry (UMIN-CTR) (ID = UMIN000047810). First, the original plan was only for analysis that investigates the association between baseline team job crafting and work engagement at three-month and six-month follow-ups. However, because the survey timing crossed the fiscal year and the pandemic of COVID-19 infection, a decrease in team job crafting was expected. Thus, an analysis of change-to-change relations of team job crafting and work engagement was added. Second, in the protocol, a quartile point is used for cut-off for workplace social capital and psychological safety. However, this study used the median score for the cut-off point because of the smaller sample size for wards.

Participants

Figure 1 shows the study's participation flow chart of the wards and nurses. A total of twelve hospitals were invited to participate in the study, and five hospitals agreed to participate. Of these, 2478 nurses nested in 99 wards were included. A total of 460 nurses nested in 36 wards completed the T1 survey (response rate of individual nurses = 18.6%). Of 460 nurses, 63 nurses were excluded due to missing ward variables. Also, six nurses in six wards with only one respondent were additionally excluded. The average numbers of nurses who completed the T1 survey from each ward were 24 (Min: two respondents and Max: 36 respondents in a ward). At the T2 survey, 29 wards representatives and 279 nurses completed the survey (response rate of individual nurses = 60.7%). Seven wards and 181 nurses refused or dropped out from T1 to T2. At the T3 survey, 27 wards representatives and 208 nurses completed the survey (response rate of individual nurses = 45.2%). Nine wards and 252 nurses refused or dropped out from T1 to T3. Nurses who responded to the T1 survey were included in the analyses even if they dropped out at the T2 or T3 survey. The actual collected and analyzed data was 391 participants from 30 wards, and the post hoc statistical power of analysis (1 - β) would be 0.20.

Participant characteristics

Table 1 shows the characteristics of the 30 wards and 391 nurses at T1. Among the wards, 14 (46.7%) included 2–9 respondents, 8 (26.7%) included 10–19 respondents, 7 (23.3%) included 20–29 respondents, and 1 (3.3%) included 30–39 respondents. Types of clinical departments of wards were internal medicine wards (4), surgical wards (4), psychiatric wards (13), pediatric wards (2), and mixed wards or other (7). Of the nurses, most (78.5%) were female.

Table 2 presents the means and SDs of the independent variables at baseline and dependent variables at baseline, 3-month, and 6-month follow-up. Table 3 shows the intraclass correlation coefficients (ICCs) for the study variables and individual-level and ward-level correlations among variables. ICCs were calculated to estimate the proportion of variance explained at each level of the studies (62). The results showed that 95% of the variance in T2 work engagement and 99% in T3 work engagement is explained by variables from individuals, and 5% of the variance in T2 and 1% in T3 thru work engagement from wards. The design effect was 1.60, which could not exceed 2.0. On the other hand, ICCs for independent (i.e., team job crafting and each domain of team job crafting) ranged from 0.03 to 0.12, indicating that 3–12% of the variances were explained by these ward-level variables. Additionally, ICCs for the change in team job crafting scores between T1 and T2 ranged from 0.01 to 0.04, and between T1 and T3 ranged from 0.01 to 0.03. The change in work engagement scores between T1 and T2 was 0.15, and between T1 and T3 was 0.05.

Longitudinal association between team job crafting and work engagement

Tables 4 − 1 and 4 − 2 is the main results of HLM on work engagement at T2 and T3, respectively. Almost 95% of the variance was situated at the individual level, larger than the ward level (5%). The analyses revealed that ward-level team job crafting at T1 have no significant relationships on work engagement at T2 (γ₀₁ = -0.20 [SE = 0.28], p = 0.492) and T3 (γ₀₁ = -0.26 [SE = 0.31], p = 0.402). No significant association was found between any subscale of ward-level team job crafting at T1 and work engagement at T2 and T3.

Subgroup analysis

The first subgroup analyses showed no significant association between the total score and any subscale of ward-level team job crafting at T1 and work engagement at T2 and T3 in the higher or lower degree of workplace social capital subgroup and the higher or lower degree of psychological safety subgroup at T1. The results are shown in Supplementary Tables (Tables S1-1 to S2-4).

Secondary dependent variables results

The results of HLM indicated no significant association between total score and any subscale of ward-level team job crafting at T1 and work performance, psychological distress, and intention to leave at T2 and T3 (Tables 5 − 1 and 5 − 2, Tables 6 − 1 and 6 − 2, and Tables 7 − 1 and 7 − 2, respectively).

Associations between changes in team job crafting and changes in work engagement

Additional analyses were conducted to examine the association between the ward-level change of team job crafting (T2 – T1 and T3 – T1, respectively) and the individual-level change of work engagement (T2 – T1 and T3 – T1, respectively) (see Tables 8 − 1 and 8 − 2). The ward-level change in the total score of team job crafting from T1 to T2 and from T1 to T3 was not significantly associated with the individual-level change in work engagement. The ward-level change in crafting for the task considering the team's growth from T1 to T3, which is one of the subscales of team job crafting, was significantly and positively associated with the individual-level change in work engagement from T1 to T3 (Coefficient = 0.59 [SE = 0.29], p = 0.044). The individual-level changes in the total team job crafting score, crafting for the task considering the team's growth, and crafting for smooth information sharing of individual-level team job crafting from T1 to T2 were significantly and positively associated with the change of work engagement from T1 to T2. The individual-level changes in the total score and each subscale of individual-level team job crafting from T1 to T3 was also significantly and positively associated with the change in work engagement from T1 to T3.

Contrary to the hypotheses, this longitudinal study found no significant multilevel association between ward-level team job crafting at baseline and individual-level work engagement at three-month or six-month follow-ups among nurses. The result was same for the association between each subscale of ward-level team job crafting and work engagement. Subgroup analyses showed that there was no significant difference in the longitudinal relationship between the total and any subscale score of ward-level team job crafting and work engagement in the degree of workplace social capital and psychological safety. Ward-level team job crafting had no significant longitudinal associations with work performance, psychological distress, or intention to leave. The additional analyses for testing the difference found that ward-level change of "crafting for the task considering the team's growth," one of the subscales of team job crafting, from T1 to T3 was significantly and positively associated with an individual-level change of work engagement from T1 to T3. The change in the individual-level team job crafting was significantly and positively related to the change in work engagement from T1 to T2 and from T1 to T3, respectively.

The main hypothesis of the association between ward-level team job crafting at T1 and individual-level work engagement at T2 and T3 was not supported. These results were inconsistent with previous cross-sectional and longitudinal studies (21, 23–26). One possible explanation may be the difference in the follow-up period. A previous study investigated its longitudinal association with a one-month follow-up (26), whereas the current study investigated three-month and six-month follow-ups. Our results suggested that the ward-level behavior of team job crafting might have a synchronous association with work engagement, while the association could not exist over three-months follow-up. According to a meta-analysis that examined the longitudinal association between individual-level job crafting and work engagement, longitudinal observational studies with more than three-month follow-ups found no significant association (63). Job crafting involves a temporary requirement to achieve short-term work goals, such as enhancing a particular work process and making decisions by team members on which job resources are needed to mobilize (21, 64, 65). As a result, the benefit of team job crafting is likely to be temporary.

Considering cross-sectional evidence in previous research, the association between team job crafting and work engagement found in prior studies may have been causally reversed, at least in part. A concept paper suggested that engaged workers are more likely to increase their job resources, that is to say, are more likely to craft their jobs (66). They also said that engaged workers tend to try to be actively involved in organizational matters (66). Therefore, team job crafting might be observed in a workplace where team members' work engagement is already high in the previous studies.

The difference in the systems and culture of Japanese hospitals may also have influenced the results because the previous studies among nurses were conducted in China and Vietnam. For example, Japan has the most hospital beds and a low nurse-to-patient ratio (67). Due to a higher workload, Japanese nurses work more involuntary overtime than in other countries (68–70). It is possible that nurses who participated in this study were so occupied with their daily work that doing team job crafting did not improve their work engagement, but made them feel like they were forced to do it by other team members.

Another possible reason for the null result may be that other non-negligible confounding variables could have distorted the results. First, the impact of COVID-19 on work engagement might be much larger. In Japan, the sixth wave of the COVID-19 pandemic occurred from January to April 2022 (during the T2 survey), and the seventh wave occurred from June to September (during the T3 survey) (71). During the pandemic's peak, nurses perceived a high emotional work and a heavy workload (72), and the COVID-19 crisis was negatively related to work engagement among front-line health workers (73). Thus, the spread of COVID-19 may have resulted in lower work engagement, regardless of the baseline level of team job crafting. The second possible confounding factor might be organizational change. At Japanese hospitals, many nurses often retire, and new nurses are hired from the end of March to the beginning of April, resulting in major changes in the composition of ward members. Since the appearance of a negative assessment of changes in the work environment was reported to have a negative impact on the psychological well-being of nurses (74), it may lead to low work engagement.

The ICCs for team job crafting were low (0.03 to 0.12), a result that was inconsistent with a previous study (19). A possible explanation is that the low response rate of the current study might cause the low ICC of team job crafting. If participants with relatively unengaged team job crafting behavior did not respond to the questionnaire, team job crafting scores between the wards would have stayed the same, and within-group similarities could not have been detected. Other possibilities include that the team job crafting scale could not capture the ward-level variance of team job crafting. Another is that the team job crafting may not be a collective construct, while it is supposed to be so in the proposed theory (20, 21). If so, even if they are in the same wards, individual nurses would have different perspectives on the degree of crafting their jobs. This might require a reconsideration of the concept of team job crafting.

Moreover, the ICCs for work engagement (0.01 to 0.05) and the design effect (1.60) were low. Previous studies among healthcare workers, including nurses, have also found ICCs ranging from 0.05 to 0.09, indicating the variance of work engagement is less likely to be shared by the group and more likely to be explained at the individual level (51, 75). If work engagement is a variable provided by individual-level factors, and there are few group-level effects on work engagement, the model of the hypothesis assuming group-level team job crafting association may have low plausibility. Accordingly, when investigating the multi-level impact on work engagement among healthcare workers, it is necessary to consider that the contextual effect of work engagement may be small.

The second hypothesis proposing that the degree of association between ward-level team job crafting at T1 and individual-level work engagement at T2 and T3 was greater in the wards with higher degrees of workplace social capital and psychological safety was not supported. The present study found no association between ward-level team job crafting and work engagement in either subgroup classified by the degree of organizational climate. Although team job crafting behavior is more likely to occur when working in a better organizational climate (30, 76–78) the degree of organizational climate may not make a difference in understanding and accepting team job crafting behaviors. It may not be involved in the longitudinal relationship between team job crafting and work engagement.

The third hypothesis regarding a positive association between ward-level team job crafting at T1 with work performance, psychological distress, and intention to leave at T2 and T3 was not supported. Similar to the reasons why H1 was not supported, the effects of team job crafting may not maintain for a long period to improve work performance and decrease psychological distress and intention to leave among nurses.

The fourth hypothesis for a positive association between the ward-level change of team job crafting scores (from T1 to T2 and T3) with the individual-level change of work engagement scores (from T1 to T2 and T3) was partially supported. An increase in "crafting for the task considering the team's growth" through six months was associated with work engagement increases through six months. While ward-level team job crafting behavior at a single point in time does not predict future work engagement, an increased degree of crafting for their growth may predict an increase in individual's work engagement. This subscale represents a concept similar to collaborative job crafting and another theory that focuses on task crafting and increasing structural job resources (19). The fact that increasing a task-focused type of team job crafting is positively associated with increased work engagement is consistent with the results of previous studies (21, 23, 24, 26). Also, an individual-level increase in team job crafting was significantly and positively related to the increase in work engagement from T1 to T2 and from T1 to T3, respectively. This may indicate that the perception by individual nurses of an increase in the extent to which team job crafting is practiced in their wards would improve their work engagement.

Limitation

There are several limitations to this study. First, the number of participants and wards included in the analysis met only about 20% of the expected sample size, and response rates varied by ward. The low sample size may cause very low statistical power in the analyses (1-β = 0.20) and lead to type 1 and type 2 errors. Multilevel modeling may also not be possible due to the small number of participating wards. Additionally, the low response rate of the participants (18.6%) and the low follow-up rates (60.7% for the T2 survey and 45.2% for the T3 survey, respectively) can cause selection bias. Second, all individual- and ward-level variables were measured using self-report questionnaires. Measured values could contain information bias and measurement errors. Third, the results may be distorted by confounding variables that this study could not consider, such as other types of job stressors and environmental determinants outside the workplace. Fourth, limited causal inferences could be made when testing the fourth hypothesis, as there was no temporal separation between the independent and dependent variables. Finally, because the samples were extracted from a restricted area in Japan by snowball sampling, the generalizability of the results is restricted. High-quality investigations, such as the random sampling method, must be conducted to solve these limitations.

Implication

The current study investigated the multilevel longitudinal associations between ward-level team job crafting and individual-level work engagement and other work-related outcomes among nurses. The results showed no significant longitudinal association. While team job crafting has been put forward as a way to promote work engagement, this study implies that there is no long-term association, and the results of previous studies may be causally reversed. However, the ward-level increase in the crafting for the task considering team growth was associated with an individual-level increase in work engagement. Encouraging task crafting for growth may contribute to increasing nurses' work engagement. Future studies should clarify under what conditions team job crafting can be better related to work engagement by using a high-quality design.

This study investigated the association between ward-level team job crafting and individual-level work engagement among Japanese nurses using a multilevel longitudinal design. The current study found that ward-level team job crafting at baseline was not associated with positive individual outcomes such as work engagement and work performance and negative individual outcomes such as psychological distress and intention to leave at three-month and six-month follow-ups. The ward-level increase in the crafting for a task considering team growth may predict an individual-level increase in work engagement. The impact of ward-level team job crafting on work engagement may be limited to a longer-term follow-up.

Ethics approval and consent to participate

The study was approved by The Research Ethics Committee of the Graduate School of Medicine/Faculty of Medicine at the University of Tokyo [No. 2018175NI-(3)]. All methods were carried out in accordance with relevant guidelines and regulations. Informed consent was obtained from all subjects.

Consent for publication

Not applicable

Availability of data and materials

The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.

Competing interests

The authors AS, KI, and NK are employed at the Department of Digital Mental Health, an endowment department supported with an unrestricted grant from 15 enterprises (https://dmh.m.u-tokyo.ac.jp/c), outside the submitted work.

MI, KW, US, HA, YK, and YM have no conflicts of interest to disclose.

Funding

This work was supported by JST SPRING, Grant Number JPMJSP2108.

Authors' contributions

MI, AS, KW, KI, and NK set up the research question. MI, US, and YM prepared the questionnaire, and MI and AS recruited the participants according to the inclusion criteria. MI and HA analyzed the participants’ data, and MI, AS, KW, KI, YK, YM, and NK interpreted the data regarding the team job crafting and work engagement. MI was a major contributor to writing the manuscript. All authors read and approved the final manuscript.

Acknowledgments

Not applicable

McVicar A. Workplace stress in nursing: a literature review. J Adv Nurs. 2003;44(6):633–42.
Havaei F, MacPhee M, McLeod CB, Ma A, Gear A, Sorensen C. A provincial study of nurses' psychological health and safety in British Columbia, Canada. 2020.
Kessler RC, Akiskal HS, Ames M, Birnbaum H, Greenberg P, . A RM, et al. Prevalence and effects of mood disorders on work performance in a nationally representative sample of US workers. Am J Psychiat. 2006;163(9):1561-8.
Schaufeli WB, Bakker AB. Job demands, job resources, and their relationship with burnout and engagement: a multi-sample study. J Organ Behav. 2004;25(3):293-315.
Wilmar B. Schaufeli MS, Vicente González-romá & Arnold B. Bakker. The measurement of engagement and burnout: a two sample confirmatory factor analytic approach. 2002.
Innstrand ST, Langballe EM, Falkum E. A longitudinal study of the relationship between work engagement and symptoms of anxiety and depression. Stress Health. 2012;28(1):1-10.
Sawatzky JA, Enns CL. Exploring the key predictors of retention in emergency nurses. J Nurs Manag. 2012;20(5):696-707.
Gómez-Salgado J, Domínguez-Salas S, Romero-Martín M, Romero A, Coronado-Vázquez V, Ruiz-Frutos C. Work engagement and psychological distress of health professionals during the COVID-19 pandemic. J Nurs Manag. 2021;29(5):1016-25.
Keyko K, Cummings GG, Yonge O, Wong CA. Work engagement in professional nursing practice: A systematic review. Int J Nurs Stud. 2016;61:142-64.
Garcia-Sierra R, Fernandez-Castro J, Martinez-Zaragoza F. Work engagement in nursing: an integrative review of the literature. J Nurs Manag. 2016;24(2):E101-11.
De Simone S, Planta A, Cicotto G. The role of job satisfaction, work engagement, self-efficacy and agentic capacities on nurses' turnover intention and patient satisfaction. Appl Nurs Res. 2018;39:130-40.
Brooks Carthon JM, Hatfield L, Plover C, Dierkes A, Davis L, Hedgeland T, et al. Association of Nurse Engagement and Nurse Staffing on Patient Safety. J Nurs Care Qual. 2019;34(1):40-6.
Spence Laschinger HK, Wilk P, Cho J, Greco P. Empowerment, engagement and perceived effectiveness in nursing work environments: does experience matter? J Nurs Manag. 2009;17(5):636-46.
Wong CA, Spence Laschinger HK, Cummings GG. Authentic leadership and nurses' voice behaviour and perceptions of care quality. J Nurs Manag. 2010;18(8):889-900.
Botha E, Beukes I. Organisational commitment, work engagement and meaning of work of nursing staff in hospitals : original research. SA J Ind Psychol. 2013;39(2):1-10.
Bakker AB, Demerouti E. The Job Demands-Resources model: state of the art. J Manage Psychol. 2007;22(3):309-28.
Wrzesniewski A, Dutton JE. Crafting a Job: Revisioning Employees as Active Crafters of Their Work. Acad Manage Rev. 2001;26(2):179-201.
Demerouti E. Multiple Levels in Job Demands-Resources Theory: Implications for Employee Well-being and Performance. Handbook of well-being. 2018.
Iida M, Watanabe K, Imamura K, Sakuraya A, Asaoka H, Sato N, et al. Development and validation of the Japanese version of the team job crafting scale for nurses. Res Nurs Health. 2021;44(2):329-43.
Leana C, Appelbaum E, Shevchuk I. Work Process and Quality of Care in Early Childhood Education: The Role of Job Crafting. Acad Manage J. 2009;52(6):1169-92.
Tims M, Bakker AB, Derks D, van Rhenen W. Job Crafting at the Team and Individual Level:Implications for Work Engagement and Performance. Group Organ Manage. 2013;38(4):427-54.
Tims M, Twemlow M, Fong CYM. A state-of-the-art overview of job-crafting research: current trends and future research directions. Career Dev Int. 2021;27(1):54-78.
Llorente-Alonso M, Topa G. Individual Crafting, Collaborative Crafting, and Job Satisfaction: The Mediator Role of Engagement. Revista de psicología del trabajo y de las organizaciones. 2019;35(3):217-26.
McClelland GP, Leach DJ, Clegg CW, McGowan I. Collaborative crafting in call centre teams. J Occup Organ Psych. 2014;87(3):464-86.
Hu Q, Schaufeli WB, Taris TW, Shimazu A, Dollard M, Taris L, et al. Resource crafting: Is it really ‘resource’ crafting—or just crafting? Front Psychol. 2019;10:90-614.
Luu TT. Collective job crafting and team service recovery performance: a moderated mediation mechanism. Mark Intell Plan. 2017;35(5):641-56.
Oksanen T, Suzuki E, Takao S, Vahtera J, Kivimäki M. Workplace social capital and health. Global perspectives on social capital and health: Springer; 2013. p. 23-63.
Edmondson A. Psychological safety and learning behavior in work teams. Admin Sci Quart. 1999;44(2):350-83.
Tims M, Parker SK. How coworkers attribute, react to, and shape job crafting. Organ Psychol Rev. 2020;10(1):29-54.
Hu Q, Dollard MF, Taris TW. Organizational context matters: Psychosocial safety climate as a precursor to team and individual motivational functioning. Safety Sci. 2022;145:105524.
Sakuraya A, Shimazu A, Eguchi H, Kamiyama K, Hara Y, Namba K, et al. Job crafting, work engagement, and psychological distress among Japanese employees: a cross-sectional study. Biopsychosoc Med. 2017;11:6.
Oprea B, Păduraru L, Iliescu D. Job crafting and intent to leave: the mediating role of meaningful work and engagement. J Career Dev. 2022;49(1):188-201.
Henk CM, Castro-Schilo L. Preliminary Detection of Relations Among Dynamic Processes With Two-Occasion Data. Structural equation modeling. A Multidisciplinary Journal. 2016;23(2):180-93.
Vandenbroucke JP, von Elm E, Altman DG, Gøtzsche PC, Mulrow CD, Pocock SJ, et al. Strengthening the Reporting of Observational Studies in Epidemiology (STROBE): Explanation and Elaboration. Epidemiology (Cambridge, Mass). 2007;18(6):805-35.
Enders CK, Tofighi D. Centering predictor variables in cross-sectional multilevel models: a new look at an old issue. Psychol Methods. 2007;12(2):121-38.
Schaufeli WB, Bakker AB, Salanova M. The measurement of work engagement with a short questionnaire: A cross-national study. Educ Psychol Meas. 2006;66(4):701-16.
Shimazu A, Schaufeli W, Kosugi S, Suzuki A, Nashiwa H, Kato A, et al. Work engagement in Japan: validation of the Japanese version of the Utrecht Work Engagement Scale. Applied Psychology. 2008;57(3):510-23.
Kawakami N, Inoue A, Tsuchiya M, Watanabe K, Imamura K, Iida M, et al. Construct validity and test-retest reliability of the World Mental Health Japan version of the World Health Organization Health and Performance Questionnaire Short Version: a preliminary study. Ind Health. 2020:2019-0090.
Furukawa TA, Kawakami N, Saitoh M, Ono Y, Nakane Y, Nakamura Y, et al. The performance of the Japanese version of the K6 and K10 in the World Mental Health Survey Japan. Int J Methods Psychiatr Res. 2008;17(3):152-8.
Tsuno K, Kawakami N, Shimazu A, Shimada K, Inoue A, Leiter MP. Workplace incivility in Japan: Reliability and validity of the Japanese version of the modified Work Incivility Scale. J Occup Health. 2017:16-0196-OA.
Geurts S, Schaufeli W, De Jonge J. Burnout and intention to leave among mental health-care professionals: A social psychological approach. J Soc Clin Psychol. 1998;17(3):341.
Inoue A, Kawakami N, Shimomitsu T, Tsutsumi A, Haratani T, Yoshikawa T, et al. Development of a short questionnaire to measure an extended set of job demands, job resources, and positive health outcomes: the new brief job stress questionnaire. Ind Health. 2014:2013-0185.
Inoue A, Kawakami N, Shimomitsu T, Tsutsumi A, Haratani T, Yoshikawa T, et al. Development of a short version of the new brief job stress questionnaire. Ind Health. 2014.
Siegrist J, Wege N, Pühlhofer F, Wahrendorf M. A short generic measure of work stress in the era of globalization: effort–reward imbalance. Int Arch Occu Env Hea. 2009;82(8):1005-13.
Tsutsumi A, Kawanami S, Horie S. Effort-reward imbalance and depression among private practice physicians. Int Arch Occu Env Hea. 2012;85(2):153-61.
Kurioka S, Inoue A, Tsutsumi A. Optimum cut-off point of the Japanese short version of the effort-reward imbalance questionnaire. J Occup Health. 2013:12-0235-OA.
Kouvonen A, Kivimäki M, Vahtera J, Oksanen T, Elovainio M, Cox T, et al. Psychometric evaluation of a short measure of social capital at work. BMC public health. 2006;6(1):1-10.
Odagiri Y, Ohya Y, Inoue S. Nihongoban syokuiki so-syaru kyapitaru sitsumonshi no shinraisei to datousei no kentou [Reliability and validity of the Japanese version of the Workplace Social Capital Scale]. Sangyo Eiseigaku Zasshi. 2010;52:631.
Jyunichi M, Kei F. Fusagu syokuba ni Yumoa wo—syokuba yumoa ga sinritekianzensei wo kaishite souzousei ni oyobosu eikyou—[Humor in a blocked workplace—The effect of workplace humor on creativity via psychological safety]. Nihon Shinrigakkai Taikai Happyou Ronbunsyu Nihon Shinrigakkai dai 83 kai Taikai. 2019.
Campbell MK, Elbourne DR, Altman DG. CONSORT statement: extension to cluster randomised trials. BMJ. 2004;328(7441):702-8.
Kunie K, Kawakami N, Shimazu A, Yonekura Y, Miyamoto Y. The relationship between work engagement and psychological distress of hospital nurses and the perceived communication behaviors of their nurse managers: A cross-sectional survey. Int J Nurs Stud. 2017;71:115-24.
Faul F, Erdfelder E, Buchner A, Lang A-G. Statistical power analyses using G* Power 3.1: Tests for correlation and regression analyses. Behav Res Methods. 2009;41(4):1149-60.
Fujita S, Kawakami N, Ando E, Inoue A, Tsuno K, Kurioka S, et al. The association of workplace social capital with work engagement of employees in health care settings. J Occup Environ Med. 2016;58(3):265-71.
Bryk AS, Raudenbush SW. Hierarchical linear models: Applications and data analysis methods: Sage Publications, Inc; 1992.
Snijders TA, Bosker RJ. Multilevel analysis: An introduction to basic and advanced multilevel modeling: sage; 2011.
Madley-Dowd P, Hughes R, Tilling K, Heron J. The proportion of missing data should not be used to guide decisions on multiple imputation. J Clin Epidemiol. 2019;110:63-73.
Sterne JA, White IR, Carlin JB, Spratt M, Royston P, Kenward MG, et al. Multiple imputation for missing data in epidemiological and clinical research: potential and pitfalls. BMJ. 2009;338.
R Core Team R. R: A language and environment for statistical computing. R foundation for statistical computing Vienna, Austria; 2018.
Kouvonen A, Oksanen T, Vahtera J, Stafford M, Wilkinson R, Schneider J, et al. Low workplace social capital as a predictor of depression - The Finnish public sector study. Am J Epidemiol. 2008;167(10):1143-51.
Gao J, Weaver SR, Dai J, Jia Y, Liu X, Jin K, et al. Workplace Social Capital and Mental Health among Chinese Employees: A Multi-Level, Cross-Sectional Study. PloS one. 2014;9(1):e85005-e.
Ihsaan M, Yousaf J, Sohaib M, Majeed A. Measuring the levels of learning in organizations: A comparative study of food sector of Pakistan. International Journal of Organizational Learning and Change. 2013;1(1):67-70.
Hox JJ, Moerbeek M, Van de Schoot R. Multilevel analysis: Techniques and applications: Routledge; 2017.
Frederick DE, VanderWeele TJ. Longitudinal meta-analysis of job crafting shows positive association with work engagement. Cogent Psychol 2020;7(1):1746733.
Tims M, Bakker AB, Derks D. Development and validation of the job crafting scale. J Vocat Behav. 2012;80(1):173-86.
Harju LK, Hakanen JJ, Schaufeli WB. Can job crafting reduce job boredom and increase work engagement?: A three-year cross-lagged panel study. J Vocat Behav. 2016;95-96(95/96):11-20.
Bakker AB. Engagement and “job crafting”: Engaged employees create their own great place to work. Handbook of employee engagement: Edward Elgar Publishing; 2010.
Health equipment: Organization for Economic Co-operation and Development. 2021. Available from: https://www.oecd-ilibrary.org/social-issues-migration-health/hospital-beds/indicator/english_0191328e-en [Accessed 20th October 2022]
Kikuchi Y, Nakaya M, Ikeda M, Okuzumi S, Takeda M, Nishi M. Relationship between job stress, temperament and depressive symptoms in female nurses. Int J Occup Med Env. 2014;27(3):426-34.
Watanabe M, Yamauchi K. Subtypes of overtime work and nurses’ fatigue, mental status, and work engagement: A latent class analysis of Japanese hospital nurses. J Adv Nurs. 2019;75(10):2122-32.
Bacon DR, Stewart KA. Results of the 2017 AORN salary and compensation survey. AORN J. 2017;106(6):476-93.
Ministry of Health LaW. Kokunai no hasseijyokyo nado. 2022 Available from: https://www.mhlw.go.jp/stf/covid-19/kokunainohasseijoukyou.html [Accessed 20th October 2022]
Giménez-Espert MDC, Prado-Gascó V, Soto-Rubio A. Psychosocial Risks, Work Engagement, and Job Satisfaction of Nurses During COVID-19 Pandemic. Front Public Health. 2020;8:566896.
Liu D, Chen Y, Li N. Tackling the Negative Impact of COVID-19 on Work Engagement and Taking Charge: A Multi-Study Investigation of Frontline Health Workers. J Appl Psychol. 2021;106(2):185-98.
Verhaeghe R, Vlerick P, Gemmel P, Maele GV, Backer GD. Impact of recurrent changes in the work environment on nurses’ psychological well-being and sickness absence. J Adv Nurs. 2006;56(6):646-56.
Huhtala M, Geurts S, Mauno S, Feldt T. Intensified job demands in healthcare and their consequences for employee well‐being and patient satisfaction: A multilevel approach. J Adv Nurs. 2021;77(9):3718-32.
Mansour S, Tremblay D-G. Psychosocial safety climate as resource passageways to alleviate work-family conflict: A study in the health sector in Quebec. Pers Rev. 2018.
Mäkikangas A, Bakker AB, Schaufeli WB. Antecedents of daily team job crafting. Eur J Work Organ Psy. 2017;26(3):421-33.
Chen C-Y. Does work engagement mediate the influence of job resourcefulness on job crafting? Int J Contemp Hosp M. 2019;31(4):1684-701.

Table 1. Baseline characteristics of wards and nurses at hospitals which participated in the study (Nj = 30, Ni = 391)

Ward-level variables (Nj = 30)	N (%)	Mean (SD)	Min-Max
Respondents in a ward
2-9	14 (46.7)
10-19	8 (26.7)
20-29	7 (23.3)
30-39	1 (3.3)
Clinical department of wards
Internal medicine wards	4 (13.3)
Surgical wards	4 (13.3)
Psychiatric wards	13 (43.3)
Pediatric wards	2 (6.7)
Mixed wards, other	7 (23.3)
Individual-level variables (Ni = 391)
Gender
Men	82 (21.0)
Women	307 (78.5)
Other	1 (0.3)
No response	0 (0.0)
Missing	1 (0.3)
Age		36.6 (11.8)	21-70
20-29	146 (37.3)
30-39	89 (22.8)
40-49	86 (22.0)
50-59	49 (12.5)
60-	16 (4.2)
Missing	5 (1.3)
Education
Professional school	339 (86.7)
University/Graduate school	49 (12.5)
Other	2 (0.5)
Missing	1 (0.3)
Marital status
Not married	205 (52.4)
Married	180 (46.0)
No response	3 (0.8)
Missing	3 (0.8)
Professional experience (years)
1-4	88 (22.5)
5-9	73 (18.7)
10-19	95 (24.3)
20-29	63 (16.1)
30-	35 (9.0)
Missing	37 (9.5)

Nj = the number of wards; Ni = the number of individuals.

Table 2. Means, standard deviations of independent variables, variables for subgroup analyses, and dependent variables (Nj= 30, Ni=391)

		Baseline (T1)			3-month follow-up (T2)			6-month follow-up (T3)
Individual-level variables	Range	Ni	Mean	SD	Ni	Mean	SD	Ni	Mean	SD
Team job crafting (total)	1-5	379	3.20	0.64	267	3.20	0.68	198	3.17	0.61
Crafting for task	1-5	382	3.34	0.66	268	3.28	0.70	200	3.29	0.59
Crafting for respect	1-5	389	2.96	0.77	270	3.04	1.00	200	2.96	0.75
Crafting for information share	1-5	387	3.31	0.72	271	3.31	0.69	198	3.26	0.69
Work engagement	0-6	384	2.48	1.00	268	2.53	1.01	201	2.42	1.13
Work performance	0-11	382	5.07	1.76	268	5.38	1.76	200	4.90	1.93
Psychological distress	0-24	388	5.30	4.89	269	5.13	4.98	202	5.55	5.07
Intention to leave	1-4	388	2.80	0.52	270	2.80	0.52	203	2.80	0.56
Workplace social capital	1-4	387	2.95	0.58	-	-	-	-	-	-
Psychological safety	1-7	385	4.44	0.88	-	-	-	-	-	-
Ward-level variables	Range	Nj	Mean	SD	Nj	Mean	SD	Nj	Mean	SD
Team job crafting (total)	1-5	30	3.20	0.24	26	3.20	0.24	25	3.13	0.26
Crafting for task	1-5	30	3.33	0.21	26	3.28	0.25	25	3.27	0.28
Crafting for respect	1-5	30	2.96	0.32	26	3.03	0.34	25	2.91	0.31
Crafting for information share	1-5	30	3.31	0.26	26	3.20	0.24	25	3.22	0.29
Workplace social capital	1-5	30	2.95	0.30	-	-	-	-	-	-
Psychological safety	1-7	30	4.44	0.37	-	-	-	-	-	-

Nj = the number of wards; Ni = the number of individuals.

Table 3. Intra class correlations (ICCs) for each variable and multilevel correlations among ward-level variables (Nj= 30, Ni=391)

ICC [95%CI]

1.

2.

3.

4.

5.

6.

7.

8.

9.

10.

11.

12.

13.

14.

15.

16.

17.

18.

1. Team job crafting (T1)

0.07

[0.02-0.17]

1.00

0.87*

0.91*

0.90*

0.57*

0.43*

0.41*

0.32*

0.30*

0.12*

0.24*

0.23*

-0.30*

-0.26*

-0.28*

-0.37*

-0.29*

-0.32*

2. Crafting for task (T1)

0.03

[-0.01-0.12]

0.73*

1.00

0.68*

0.67*

0.46*

0.34*

0.35*

0.29*

0.28*

0.12*

0.22*

0.23*

-0.25*

-0.22*

-0.24*

-0.31*

-0.23*

-.028*

3. Crafting for respect (T1)

0.12

[0.05-0.23]

0.94*

0.56*

1.00

0.73*

0.51*

0.39*

0.41*

0.34*

0.29*

0.08

0.20*

0.22*

-0.25*

-0.22*

-0.21*

-0.36*

-0.30*

-.031*

4. Crafting for information (T1)

0.07

[0.01-0.17]

0.91*

0.48*

0.82*

1.00

0.55*

0.42*

0.37*

0.24*

0.23*

0.13*

0.22*

0.14*

-0.30*

-0.24*

-0.27*

-0.33*

-0.26*

-.026*

5. Workplace social capital (T1)

0.21

[0.12-0.37]

0.74*

0.54*

0.68*

0.69*

1.00

0.51*

0.40*

0.34*

0.37*

0.11*

0.11

0.17*

-0.29*

-0.22*

-0.28*

-0.42*

-0.31*

-0.35*

6. Psychological safety (T1)

0.12

[0.05-0.23]

0.74*

0.47*

0.66*

0.78*

0.79*

1.00

0.32*

0.25*

0.32*

0.25*

0.15*

0.34*

-0.39*

-0.33*

-0.40*

-0.33*

-0.37*

7. Work engagement (T1)

0.02

[-0.02-0.09]

0.25*

0.36*

0.24*

0.10

0.21*

0.14*

1.00

0.83*

0.79*

0.43*

0.44*

0.45*

-0.31*

-0.21*

-0.37*

-0.51*

-0.47*

-0.49*

8. Work engagement (T2)

0.05

[-0.01-0.15]

0.10

0.08

0.15*

0.04

0.16*

0.22*

0.62*

1.00

0.82*

0.39*

0.47*

-0.27*

-0.23*

-0.41*

-0.44*

-0.53*

-0.50*

9. Work engagement (T3)

0.01

[-0.05-0.12]

0.29*

0.12*

0.34*

0.28*

0.30*

0.27*

0.64*

0.65*

1.00

0.39*

0.44*

0.59*

-0.36*

-0.33*

-0.47*

-0.44*

-0.52*

10. Work performance (T1)

0.003

[-0.03-0.07]

0.09

0.05

0.01

0.20*

0.12*

0.22*

0.36*

0.14*

0.29*

1.00

0.51*

0.54*

-0.37*

-0.25*

-0.36*

-0.14*

-0.19*

11. Work performance (T2)

0.04

[-0.02-0.14]

0.02

-0.01

0.09

-0.04

0.12*

0.04

0.34*

0.46*

0.42*

0.14*

1.00

0.51*

-0.43*

-0.34*

-0.45*

-0.21*

-0.24*

12. Work performance (T3)

0.01

[-0.05-0.12]

-0.01

-0.08

0.02

0.00

-0.04

0.01

0.49*

0.35*

0.63*

0.49*

0.39*

1.00

-0.38*

-0.34*

-0.43*

-0.27*

-0.31*

-0.26*

13. Psychological distress (T1)

0.03

[-0.01-0.10]

-0.16*

-0.09

-0.16*

-0.13*

-0.37*

-0.53*

-0.55*

-0.43*

-0.35*

-0.40*

-0.48*

1.00

0.73*

0.63*

0.34*

0.27*

14. Psychological distress (T2)

0.06

[-0.002-0.18]

-0.11*

0.10

-0.17*

0.03

-0.25*

-0.38*

-0.45*

-0.49*

-0.29*

-0.42*

-0.44*

0.76*

1.00

0.67*

0.25*

0.22*

15. Psychological distress (T3)

-0.05

[-0.08-0.03]

-0.27*

0.05

-0.32*

-0.35*

-0.10

-0.35*

-0.34*

-0.23*

-0.59*

-0.40*

-0.36*

-0.69*

0.56*

0.76*

1.00

0.34*

0.36*

0.45*

16. Intention to leave (T1)

0.14

[0.06-0.27]

-0.40*

-0.32*

-0.34*

-0.39*

-0.47*

-0.44*

-0.57*

-0.36*

-0.55*

-0.40*

-0.08

-0.16*

0.41*

0.28*

0.32*

1.00

0.64*

0.63*

17. Intention to leave (T2)

0.21

[0.10-0.38]

-0.39*

-0.14*

-0.40*

-0.43*

-0.53*

-0.57*

-0.36*

-0.50*

-0.37*

-0.05

-0.12*

0.35*

0.34*

0.83*

1.00

0.61*

18. Intention to leave (T3)

0.20

[0.08-0.38]

-0.38*

-0.13*

-0.38*

-0.46*

-0.52*

-0.51*

-0.09

-0.19*

-0.42*

-0.34*

-0.22*

-0.16*

0.10*

0.30*

0.40*

0.54*

0.69*

1.00

Upper triangular matrix indicates individual-level correlations, and lower triangular matrix indicates ward-level matrix.

Nj = the number of wards; Ni = the number of individuals. ICC = intra-class correlation coefficient interval. *p<0.05

Table 4-1. Multilevel association between individual- and ward-level team job crafting and work engagement at T2 (Nj= 30, Ni= 391)

	Model 0 (Null model)		Model 1 (Crude model)		Model 2 (adjusted^a)		Model 3 (adjusted^b)
Fixed effects	Coefficient (SE)	p value	Coefficient (SE)	p value	Coefficient (SE)	p value	Coefficient (SE)	p value
Intercept	2.51 (0.327)	<0.001	2.50 (0.32)	<0.001	2.51 (0.32)	<0.001	2.52 (0.32)	<0.001
Individual-level	Coefficient (SE)	p value	Coefficient (SE)	p value	Coefficient (SE)	p value	Coefficient (SE)	p value
Team job crafting			0.470 (0.12)	<0.001	-0.05 (0.09)	0.590	-0.10 (0.09)	0.266
Crafting for task
Crafting for respect
Crafting for information
Ward-level	Coefficient (SE)	p value	Coefficient (SE)	p value	Coefficient (SE)	p value	Coefficient (SE)	p value
Team job crafting			-0.32 (0.34)	0.358	-0.16 (0.30)	0.586	-0.20 (0.28)	0.492
Crafting for task
Crafting for respect
Crafting for information
Random effects	Coefficient		Coefficient		Coefficient		Coefficient
Intercept	0.05		0.06		0.10		0.10
Residual variance	0.94		0.94		0.90		0.90
AIC	1123.738		1053.022		715.433		743.486

	Model 3a (adjusted^b)		Model 3b (adjusted^b)		Model 3c (adjusted^b)
Fixed effects	Coefficient (SE)	p value	Coefficient (SE)	p value	Coefficient (SE)	p value
Intercept	2.52 (0.32)	<0.001	2.52 (0.33)	<0.001	2.52 (0.32)	<0.001
Individual-level	Coefficient (SE)	p value	Coefficient (SE)	p value	Coefficient (SE)	p value
Team job crafting
Crafting for task	-0.05 (0.10)	0.599
Crafting for respect			-0.03 (0.07)	0.645
Crafting for information					-0.13 (0.08)	0.075
Ward-level	Coefficient (SE)	p value	Coefficient (SE)	p value	Coefficient (SE)	p value
Team job crafting
Crafting for task	-0.38 (0.31)	0.216
Crafting for respect			-0.10 (0.22)	0.667
Crafting for information					-0.08 (0.32)	0.783
Random effects	Coefficient		Coefficient		Coefficient
Intercept	0.10		0.10		0.10
Residual variance	0.90		0.90		0.90

^aAdjusted by T1 work engagement. ^bAdjusted by age, gender, educational status, marital status, nurse experiences, T1 work engagement, job demand, job control, supervisor support, coworker support, effort-reward imbalance. Nj = the number of wards; Ni = the number of individuals.

	Model 0 (Null model)		Model 1 (Crude model)		Model 2 (adjusted^a)		Model 3 (adjusted^b)
Fixed effects	Coefficient (SE)	p value	Coefficient (SE)	p value	Coefficient (SE)	p value	Coefficient (SE)	p value
Intercept	2.38 (0.37)	<0.001	2.38 (0.36)	<0.001	2.39 (0.36)	<0.001	2.39 (0.36)	<0.001
Individual-level	Coefficient (SE)	p value	Coefficient (SE)	p value	Coefficient (SE)	p value	Coefficient (SE)	p value
Team job crafting			0.52 (0.13)	<0.001	-0.004 (0.100)	0.966	-0.10 (0.12)	0.405
Crafting for task
Crafting for respect
Crafting for information
Ward-level	Coefficient (SE)	p value	Coefficient (SE)	p value	Coefficient (SE)	p value	Coefficient (SE)	p value
Team job crafting			-0.32 (0.36)	0.377	-0.15 (0.31)	0.633	-0.26 (0.31)	0.402
Crafting for task
Crafting for respect
Crafting for information
Random effects	Coefficient		Coefficient		Coefficient		Coefficient
Intercept	0.04		0.05		0.07		0.06
Residual variance	0.96		0.95		0.93		0.94
AIC	1200.156		1128.861		904.821		927.033

Table 4-2. Multilevel association between individual- and ward-level team job crafting and work engagement at T3. (Nj= 30, Ni= 391)

	Model 3a (adjusted^b)		Model 3b (adjusted^b)		Model 3c (adjusted^b)
Fixed effects	Coefficient (SE)	p value	Coefficient (SE)	p value	Coefficient (SE)	p value
Intercept	2.39 (0.36)	<0.001	2.39 (0.36)	<0.001	2.39 (0.36)	<0.001
Individual-level	Coefficient (SE)	p value	Coefficient (SE)	p value	Coefficient (SE)	p value
Team job crafting
Crafting for task	0.000 (0.10)	0.999
Crafting for respect			-0.06 (0.09)	0.491
Crafting for information					-0.15 (0.10)	0.158
Ward-level	Coefficient (SE)	p value	Coefficient (SE)	p value	Coefficient (SE)	p value
Team job crafting
Crafting for task	-0.57 (0.34)	0.098
Crafting for respect			-0.07 (0.23)	0.757
Crafting for information					-0.09 (0.29)	0.771
Random effects	Coefficient		Coefficient		Coefficient
Intercept	0.05		0.06		0.06
Residual variance	0.95		0.94		0.94

^aAdjusted by T1 work engagement. ^bAdjusted by age, gender, educational status, marital status, nurse experiences, T1 work engagement, job demand, job control, supervisor support, coworker support, effort-reward imbalance. Nj = the number of wards; Ni = the number of individuals.

Table 5-1. Multilevel association between individual-level and ward-level team job crafting and work performance at T2

	Model 0 (Null model)		Model 1 (Crude model)		Model 2 (adjusted^a)		Model 3 (adjusted^b)
Fixed effects	Coefficient (SE)	p value	Coefficient (SE)	p value	Coefficient (SE)	p value	Coefficient (SE)	p value
Intercept	5.30 (0.48)	<0.001	5.30 (0.48)	<0.001	5.29 (0.48)	<0.001	5.29 (0.48)	<0.001
Individual-level	Coefficient (SE)	p value
Team job crafting			0.60 (0.20)	<0.001	0.44 (0.18)	0.015	0.56 (0.20)	0.006
Crafting for task
Crafting for respect
Crafting for information
Ward-level	Coefficient (SE)	p value	Coefficient (SE)	p value	Coefficient (SE)	p value	Coefficient (SE)	p value
Team job crafting			-0.29 (0.59)	0.625	-0.28 (0.57)	0.627	-0.09 (0.57)	0.869
Crafting for task
Crafting for respect
Crafting for information
Random effects	Coefficient		Coefficient		Coefficient		Coefficient
Intercept	0.03		0.03		0.05		0.05
Residual variance	0.97		0.97		0.95		0.95
AIC	1544.696		1480.671		1407.723		1424.262

	Model 3a (adjusted^b)		Model 3b (adjusted^b)		Model 3c (adjusted^b)
Fixed effects	Coefficient (SE)	p value	Coefficient (SE)	p value	Coefficient (SE)	p value
Intercept	5.29 (0.49)	<0.001	5.29 (0.48)	<0.001	5.29 (0.49)	<0.001
Individual-level	Coefficient (SE)	p value	Coefficient (SE)	p value	Coefficient (SE)	p value
Team job crafting
Crafting for task	0.44 (0.19)	0.022
Crafting for respect			0.37 (0.16)	0.023
Crafting for information					0.42 (0.17)	0.016
Ward-level	Coefficient (SE)	p value	Coefficient (SE)	p value	Coefficient (SE)	p value
Team job crafting
Crafting for task	-0.30 (0.60)	0.611
Crafting for respect			0.03 (0.42)	0.948
Crafting for information					-0.20 (0.56)	0.724
Random effects	Coefficient		Coefficient		Coefficient
Intercept	0.05		0.04		0.05
Residual variance	0.95		0.96		0.95

^aAdjusted by T1 work performance. ^bAdjusted by age, gender, educational status, marital status, nurse experiences, T1 work performance, job demand, job control, supervisor support, coworker support, effort-reward imbalance. Nj = the number of wards; Ni = the number of individuals.

Table 5-2. Multilevel association between individual-level and ward-level team job crafting and work performance at T3 (Nj= 30, Ni= 391)

	Model 0 (Null model)		Model 1 (Crude model)		Model 2 (adjusted^a)		Model 3 (adjusted^b)
Fixed effects	Coefficient (SE)	p value	Coefficient (SE)	p value	Coefficient (SE)	p value	Coefficient (SE)	p value
Intercept	4.86 (0.69)	<0.001	4.88 (0.69)	<0.001	4.86 (0.69)	<0.001	4.86 (0.69)	<0.001
Individual-level	Coefficient (SE)	p value
Team job crafting			0.53 (0.23)	0.021	0.36 (0.21)	0.088	0.28 (0.25)	0.270
Crafting for task
Crafting for respect
Crafting for information
Ward-level	Coefficient (SE)	p value	Coefficient (SE)	p value	Coefficient (SE)	p value	Coefficient (SE)	p value
Team job crafting			-0.73 (0.70)	0.301	-0.77 (0.66)	0.247	-0.76 (0.65)	0.244
Crafting for task
Crafting for respect
Crafting for information
Random effects	Coefficient		Coefficient		Coefficient		Coefficient
Intercept	0.04		0.05		0.06		0.05
Residual variance	0.96		0.95		0.94		0.94
AIC	1602.785		1549.46		1467.047		1466.555

	Model 3a (adjusted^b)		Model 3b (adjusted^b)		Model 3c (adjusted^b)
Fixed effects	Coefficient (SE)	p value	Coefficient (SE)	p value	Coefficient (SE)	p value
Intercept	4.87 (0.68)	<0.001	4.85 (0.69)	<0.001	4.86 (0.69)	<0.001
Individual-level	Coefficient (SE)	p value	Coefficient (SE)	p value	Coefficient (SE)	p value
Team job crafting
Crafting for task	0.34 (0.22)	0.126
Crafting for respect			0.29 (0.20)	0.145
Crafting for information					-0.02 (0.23)	0.922
Ward-level	Coefficient (SE)	p value	Coefficient (SE)	p value	Coefficient (SE)	p value
Team job crafting
Crafting for task	-1.02 (0.72)	0.157
Crafting for respect			-0.53 (0.50)	0.296
Crafting for information					-0.55 (0.61)	0.365
Random effects	Coefficient		Coefficient		Coefficient
Intercept	0.05		0.05		0.05
Residual variance	0.95		0.94		0.95

^aAdjusted by T1 work performance. ^bAdjusted by age, gender, educational status, marital status, nurse experiences, T1 work performance, job demand, job control, supervisor support, coworker support, effort-reward imbalance. Nj = the number of wards; Ni = the number of individuals.

	Model 0 (Null model)		Model 1 (Crude model)		Model 2 (adjusted^a)		Model 3 (adjusted^b)
Fixed effects	Coefficient (SE)	p value	Coefficient (SE)	p value	Coefficient (SE)	p value	Coefficient (SE)	p value
Intercept	5.74 (1.56)	0.002	5.79 (1.54)	0.001	5.64 (1.49)	0.003	5.63 (1.47)	0.003
Individual-level	Coefficient (SE)	p value
Team job crafting			-1.94 (0.54)	<0.001	-0.23 (0.43)	0.587	-0.27 (0.46)	0.567
Crafting for task
Crafting for respect
Crafting for information
Ward-level	Coefficient (SE)	p value	Coefficient (SE)	p value	Coefficient (SE)	p value	Coefficient (SE)	p value
Team job crafting			1.05 (1.69)	0.536	0.45 (1.42)	0.753	0.03 (1.33)	0.982
Crafting for task
Crafting for respect
Crafting for information
Random effects	Coefficient		Coefficient		Coefficient		Coefficient
Intercept	0.05		0.04		0.04		0.03
Residual variance	0.95		0.95		0.96		0.97
AIC	2351.54		2260.225		2024.213		2020.849

Table 6-1. Multilevel association between individual-level and ward-level team job crafting and psychological distress at T2 (Nj= 30, Ni= 391)

	Model 3a (adjusted^b)		Model 3b (adjusted^b)		Model 3c (adjusted^b)
Fixed effects	Coefficient (SE)	p value	Coefficient (SE)	p value	Coefficient (SE)	p value
Intercept	5.62 (1.48)	0.003	5.62 (1.48)	0.003	5.62 (1.48)	0.003
Individual-level	Coefficient (SE)	p value	Coefficient (SE)	p value	Coefficient (SE)	p value
Team job crafting
Crafting for task	-0.38 (0.46)	0.413
Crafting for respect			-0.19 (0.37)	0.613
Crafting for information					-0.04 (0.40)	0.922
Ward-level	Coefficient (SE)	p value	Coefficient (SE)	p value	Coefficient (SE)	p value
Team job crafting
Crafting for task	0.90 (1.42)	0.526
Crafting for respect			-0.18 (0.99)	0.857
Crafting for information					-0.24 (1.29)	0.854
Random effects	Coefficient		Coefficient		Coefficient
Intercept	0.03		0.03		0.03
Residual variance	0.97		0.97		0.97

^aAdjusted by T1 psychological distress. ^bAdjusted by age, gender, educational status, marital status, nurse experiences, T1 psychological distress, job demand, job control, supervisor support, coworker support, effort-reward imbalance. Nj = the number of wards; Ni = the number of individuals.

Table 6-2. Multilevel association between individual-level and ward-level team job crafting and psychological distress at T3 (Nj= 30, Ni= 391)

	Model 0 (Null model)		Model 1 (Crude model)		Model 2 (adjusted^a)		Model 3 (adjusted^b)
Fixed effects	Coefficient (SE)	p value	Coefficient (SE)	p value	Coefficient (SE)	p value	Coefficient (SE)	p value
Intercept	5.65 (1.03)	<0.001	5.70 (1.02)	<0.001	5.60 (1.01)	<0.001	5.61 (1.01)	<0.001
Individual-level	Coefficient (SE)	p value
Team job crafting			-1.86 (0.60)	0.003	-0.44 (0.55)	0.429	-0.37 (0.62)	0.548
Crafting for task
Crafting for respect
Crafting for information
Ward-level	Coefficient (SE)	p value	Coefficient (SE)	p value	Coefficient (SE)	p value	Coefficient (SE)	p value
Team job crafting			0.95 (1.64)	0.563	0.49 (1.54)	0.749	1.03 (1.55)	0.507
Crafting for task
Crafting for respect
Crafting for information
Random effects	Coefficient		Coefficient		Coefficient		Coefficient
Intercept	0.02		0.02		0.04		0.03
Residual variance	0.98		0.98		0.97		0.98
AIC	2361.626		2274.19		2142.758		2137.773

	Model 3a (adjusted^b)		Model 3b (adjusted^b)		Model 3c (adjusted^b)
Fixed effects	Coefficient (SE)	p value	Coefficient (SE)	p value	Coefficient (SE)	p value
Intercept	5.60 (1.00)	<0.001	5.59 (1.03)	<0.001	5.60 (1.02)	<0.001
Individual-level	Coefficient (SE)	p value	Coefficient (SE)	p value	Coefficient (SE)	p value
Team job crafting
Crafting for task	-0.52 (0.57)	0.365
Crafting for respect			-0.03 (0.47)	0.941
Crafting for information					-0.32 (0.53)	0.544
Ward-level	Coefficient (SE)	p value	Coefficient (SE)	p value	Coefficient (SE)	p value
Team job crafting
Crafting for task	2.13 (1.72)	0.217
Crafting for respect			0.29 (1.13)	0.796
Crafting for information					0.68 (1.41)	0.628
Random effects	Coefficient		Coefficient		Coefficient
Intercept	0.02		0.03		0.03
Residual variance	0.98		0.97		0.97

^aAdjusted by T1 psychological distress. ^bAdjusted by age, gender, educational status, marital status, nurse experiences, T1 psychological distress, job demand, job control, supervisor support, coworker support, effort-reward imbalance. Nj = the number of wards; Ni = the number of individuals.

Table 7-1. Multilevel association between individual-level and ward-level team job crafting and intention to leave at T2 (Nj= 30, Ni= 391)

	Model 0 (Null model)		Model 1 (Crude model)		Model 2 (adjusted^a)		Model 3 (adjusted^b)
Fixed effects	Coefficient (SE)	p value	Coefficient (SE)	p value	Coefficient (SE)	p value	Coefficient (SE)	p value
Intercept	2.82 (0.12)	<0.001	2.83 (0.12)	<0.001	2.83 (0.11)	<0.001	2.83 (0.11)	<0.001
Individual-level	Coefficient (SE)	p value
Team job crafting			-0.22 (0.05)	<0.001	-0.04 (0.05)	0.330	-0.02 (0.05)	0.696
Crafting for task
Crafting for respect
Crafting for information
Ward-level	Coefficient (SE)	p value	Coefficient (SE)	p value	Coefficient (SE)	p value	Coefficient (SE)	p value
Team job crafting			-0.04 (0.19)	0.832	-0.02 (0.14)	0.917	0.03 (0.14)	0.855
Crafting for task
Crafting for respect
Crafting for information
Random effects	Coefficient		Coefficient		Coefficient		Coefficient
Intercept	0.18		0.18		0.10		0.08
Residual variance	0.82		0.82		0.90		0.92
AIC	547.858		506.801		330.471		382.962

	Model 3a (adjusted^b)		Model 3b (adjusted^b)		Model 3c (adjusted^b)
Fixed effects	Coefficient (SE)	p value	Coefficient (SE)	p value	Coefficient (SE)	p value
Intercept	2.83 (0.11)	<0.001	2.83 (0.11)	<0.001	2.83 (0.11)	<0.001
Individual-level	Coefficient (SE)	p value	Coefficient (SE)	p value	Coefficient (SE)	p value
Team job crafting
Crafting for task	-0.01 (0.05)	0.827
Crafting for respect			-0.02 (0.04)	0.548
Crafting for information					-0.01 (0.04)	0.764
Ward-level	Coefficient (SE)	p value	Coefficient (SE)	p value	Coefficient (SE)	p value
Team job crafting
Crafting for task	0.17 (0.15)	0.238
Crafting for respect			-0.01 (0.10)	0.935
Crafting for information					-0.01 (0.13)	0.955
Random effects	Coefficient		Coefficient		Coefficient
Intercept	0.08		0.07		0.07
Residual variance	0.92		0.93		0.93

^aAdjusted by T1 intention to leave. ^bAdjusted by age, gender, educational status, marital status, nurse experiences, T1 intention to leave, job demand, job control, supervisor support, coworker support, effort-reward imbalance. Nj = the number of wards; Ni = the number of individuals.

Table 7-2. Multilevel association between individual-level and ward-level team job crafting and intention to leave at T3 (Nj= 30, Ni= 391)

	Model 3a (adjusted^b)		Model 3b (adjusted^b)		Model 3c (adjusted^b)
Fixed effects	Coefficient (SE)	p value	Coefficient (SE)	p value	Coefficient (SE)	p value
Intercept	2.82 (0.13)	<0.001	2.81 (0.13)	<0.001	2.82 (0.13)	<0.001
Individual-level	Coefficient (SE)	p value	Coefficient (SE)	p value	Coefficient (SE)	p value
Team job crafting
Crafting for task	-0.03 (0.06)	0.619
Crafting for respect			-0.06 (0.05)	0.242
Crafting for information					-0.04 (0.05)	0.410
Ward-level	Coefficient (SE)	p value	Coefficient (SE)	p value	Coefficient (SE)	p value
Team job crafting
Crafting for task	0.19 (0.19)	0.331
Crafting for respect			0.07 (0.14)	0.641
Crafting for information					-0.07 (0.17)	0.670
Random effects	Coefficient		Coefficient		Coefficient
Intercept	0.09		0.11		0.09
Residual variance	0.90		0.88		0.90

	Model 0 (Null model)		Model 1 (Crude model)		Model 2 (adjusted^a)		Model 3 (adjusted^b)
Fixed effects	Coefficient (SE)	p value	Coefficient (SE)	p value	Coefficient (SE)	p value	Coefficient (SE)	p value
Intercept	2.79 (0.14)	<0.001	2.80 (0.14)	<0.001	2.81 (0.13)	<0.001	2.81 (0.13)	<0.001
Individual-level	Coefficient (SE)	p value
Team job crafting			-0.27 (0.07)	<0.001	-0.01 (0.06)	0.115	-0.06 (0.06)	0.363
Crafting for task
Crafting for respect
Crafting for information
Ward-level	Coefficient (SE)	p value	Coefficient (SE)	p value	Coefficient (SE)	p value	Coefficient (SE)	p value
Team job crafting			-0.04 (0.22)	0.847	0.01 (0.18)	0.966	0.05 (0.18)	0.789
Crafting for task
Crafting for respect
Crafting for information
Random effects	Coefficient		Coefficient		Coefficient		Coefficient
Intercept	0.19		0.19		0.13		0.11
Residual variance	0.81		0.81		0.86		0.88
AIC	621.234		575.536		425.462		482.440

^aAdjusted by T1 intention to leave. ^bAdjusted by age, gender, educational status, marital status, nurse experiences, T1 intention to leave, job demand, job control, supervisor support, coworker support, effort-reward imbalance. Nj = the number of wards; Ni = the number of individuals.

Table 8-1. Multilevel association between change of team job crafting (T2-T1) and change of work engagement (T2-T1) (Nj= 23, Ni= 248)

	Model 0 (Null model)		Model 1 (Crude model)		Model 2 (adjusted^a)
Fixed effects	Coefficient (SE)	p value	Coefficient (SE)	p value	Coefficient (SE)	p value
Intercept	0.00 (0.07)	0.954	-0.00 (0.06)	0.979	-0.02 (0.07)	0.775
Individual-level changes	Coefficient (SE)	p value	Coefficient (SE)	p value	Coefficient (SE)	p value
Team job crafting (T2-T1)			0.25 (0.08)	0.001	0.25 (0.08)	0.001
Crafting for task (T2-T1)
Crafting for respect (T2-T1)
Crafting for information(T2-T1)
Ward-level changes	Coefficient (SE)	p value	Coefficient (SE)	p value	Coefficient (SE)	p value
Team job crafting (T2-T1)			0.56 (0.36)	0.117	0.60 (0.37)	0.107
Crafting for task (T2-T1)
Crafting for respect (T2-T1)
Crafting for information(T2-T1)
Random effects	Coefficient		Coefficient		Coefficient
Intercept (Ward)	0.17		0.16		0.17
Residual variance	0.83		0.84		0.83
AIC	443.746		435.6811		486.779

	Model 2a (adjusted^a)		Model 2b (adjusted^a)		Model 2c (adjusted^a)
Fixed effects	Coefficient (SE)	p value	Coefficient (SE)	p value	Coefficient (SE)	p value
Intercept	-0.02 (0.07)	0.749	-0.01 (0.07)	0.844	-0.02 (0.07)	0.818
Individual-level changes	Coefficient (SE)	p value	Coefficient (SE)	p value	Coefficient (SE)	p value
Team job crafting (T2-T1)
Crafting for task (T2-T1)	0.25 (0.07)	<0.001
Crafting for respect (T2-T1)			0.06 (0.04)	0.183
Crafting for information(T2-1)					0.21 (0.06)	0.001
Ward-level changes	Coefficient (SE)	p value	Coefficient (SE)	p value	Coefficient (SE)	p value
Team job crafting (T2-T1)
Crafting for task (T2-T1)	0.20 (0.39)	0.605
Crafting for respect(T2-T1)			0.46 (0.27)	0.090
Crafting for information(T2-T1)					0.49 (0.30)	0.108
Random effects	Coefficient		Coefficient		Coefficient
Intercept (Ward)	0.21		0.18		0.16
Residual variance	0.79		0.82		0.84

^aAdjusted by age, gender, educational status, marital status, nurse experiences, job demand, job control, supervisor support, coworker support, effort-reward imbalance. Nj = the number of wards; Ni = the number of individuals.

Table 8-2. Multilevel association between change of team job crafting (T3-T1) and change of work engagement (T3-T1) (Nj= 21, Ni= 175)

	Model 0 (Null model)		Model 1 (Crude model)		Model 2 (adjusted^a)
Fixed effects	Coefficient (SE)	p value	Coefficient (SE)	p value	Coefficient (SE)	p value
Intercept	-0.04 (0.06)	0.503	-0.05 (0.05)	0.288	-0.06 (0.05)	0.238
Individual-level changes	Coefficient (SE)	p value	Coefficient (SE)	p value	Coefficient (SE)	p value
Team job crafting (T3-T1)			0.51 (0.11)	<0.001	0.55 (0.11)	<0.001
Crafting for task (T3-T1)
Crafting for respect (T3-T1)
Crafting for information(T3-T1)
Ward-level changes	Coefficient (SE)	p value	Coefficient (SE)	p value	Coefficient (SE)	p value
Team job crafting (T3-T1)			0.40 (0.31)	0.190	0.40 (0.31)	0.204
Crafting for task (T3-T1)
Crafting for respect (T3-T1)
Crafting for information(T3-T1)
Random effects	Coefficient		Coefficient		Coefficient
Intercept (Ward)	0.04		0.01		0.00
Residual variance	0.96		1.00		1.00
AIC	372.744		349.278		391.148

	Model 2a (adjusted^a)		Model 2b (adjusted^a)		Model 2c (adjusted^a)
Fixed effects	Coefficient (SE)	p value	Coefficient (SE)	p value	Coefficient (SE)	p value
Intercept	-0.05 (0.05)	0.316	-0.07 (0.05)	0.204	-0.06 (0.06)	0.531
Individual-level changes	Coefficient (SE)	p value	Coefficient (SE)	p value	Coefficient (SE)	p value
Team job crafting (T3-T1)
Crafting for task (T3-T1)	0.34 (0.10)	<0.001
Crafting for respect (T3-T1)			0.29 (0.10)	0.003
Crafting for information(T3-1)					0.41 (0.09)	<0.001
Ward-level changes	Coefficient (SE)	p value	Coefficient (SE)	p value	Coefficient (SE)	p value
Team job crafting (T3-T1)
Crafting for task (T3-T1)	0.59 (0.29)	0.044
Crafting for respect(T3-T1)			0.40 (0.26)	0.123
Crafting for information(T3-T1)					0.10 (0.30)	0.748
Random effects	Coefficient		Coefficient		Coefficient
Intercept (Ward)	0.00		0.00		0.04
Residual variance	1.00		1.00		0.96

^aAdjusted by age, gender, educational status, marital status, nurse experiences, job demand, job control, supervisor support, coworker support, effort-reward imbalance. Nj = the number of wards; Ni = the number of individuals.

Competing interest reported. The authors AS, KI, and NK are employed at the Department of Digital Mental Health, an endowment department supported with an unrestricted grant from 15 enterprises (https://dmh.m.u-tokyo.ac.jp/c), outside the submitted work. MI, KW, US, HA, YK, and YM have no conflicts of interest to disclose.

The association between team job crafting and work engagement among nurses: a prospective cohort study

Status:

Journal Publication

Version 1

Abstract

Background

Methods

Results

Conclusions

Figures

Background

Hypothesis

Methods

Study design

Participants

Variables

Independent variables

Dependent variables

Work engagement

Work performance

Psychological distress

Intention to leave

Confounding variables

Job demands, job control, supervisor/coworker support

Effort-Reward Imbalance

Demographic variables

Variables for the subgroup analyses

Workplace social capital

Psychological safety

Sample size calculation

Statistical analysis

β0j = γ00 + γ01* (team job crafting)0j + u0j

Subgroup analyses

Changes to protocol

Results

Participants

Participant characteristics

Longitudinal association between team job crafting and work engagement

Subgroup analysis

Secondary dependent variables results

Associations between changes in team job crafting and changes in work engagement

Discussion

Limitation

Implication

Conclusions

Declarations

References

Tables

Additional Declarations

Supplementary Files

Status:

Journal Publication

Version 1

β_0j = γ₀₀ + γ₀₁* (team job crafting)_0j + u_0j