Abstract
Background
Maintaining healthcare worker (HCW) work ability and continuing employment are critical issues. We examined the association of work ability with the previously unexamined work-environment factors of organizational support for safety (OSS) and surface-acting emotional labor (SaEL).
Methods
A self-administered survey covering work-environment features recruited employees of five U.S. public-sector healthcare facilities in 2018-2019 and 2021. Linear regression on the outcome of work ability, both cross-sectionally and longitudinally, focused on OSS and SaEL with additional independent variables.
Results
A total of 1,059 and 1,553 surveys were collected from predominantly female HCWs; 425 answered both questionnaires. Almost two thirds of the HCWs were either non-direct care or unlicensed HCWs. Work ability scores were lower in direct-care workers versus non-direct-care workers. In multivariable regression models, OSS was positively associated with work ability, and SaEL was negatively associated, cross-sectionally at baseline and follow-up. Job strain, back pain, and safety hazards were also negatively associated with work in both cross-sectional and longitudinal analyses, while social support and civility norms (combined) were positively associated. In the longitudinal analysis, only OSS positively predicted work ability, while SaEL had a borderline negative effect.
Conclusion
The finding that OSS predicts work ability is novel. Despite high turnover, there has been little examination of risk factors for low work ability among HCWs. The importance of this organizational predictor justifies organization-level interventions to improve workplace safety. Such actions could simultaneously address and mitigate the additional risk factors of SaEL, job strain, and back pain in order to improve work ability for all groups of HCWs.
Keywords: Organizational decision making, Organizational safety, Psychological safety, Safety management, Workplace safety
1. Introduction
While the quality of the healthcare work environment has been of high concern for a long time, experiences during the current COVID-19 pandemic demonstrated even more clearly the burden of job demands, lack of resources and emotional exhaustion faced by front-line healthcare workers (HCWs) [1]. Willingness to stay on the job suffers as a consequence and HCW retention is an increasingly urgent problem in many countries [2].
Work ability is a person's capacity to meet the physical, mental and interpersonal challenges at the workplace [3]. It could be considered the positive complement of work disability. Work ability is a useful explanatory lens for understanding the relationship between workers' health and their job status and function [4,5]. Health conditions that decrease work ability are also associated with increased all cause mortality [6] and disability [7]). Further, low work ability can lead to poor job outcomes such as low productivity [8], disengagement [[9], [10], [11]] job dissatisfaction, intention to leave the job, and early retirement [[10], [11], [12]]. In the healthcare sector, individuals' work ability is intertwined with their ability to provide care and the quality of care provided [13,14]. Large-scale empirical studies (e.g. [15]), research reviews [13] and meta-analyses [16] document the associations of work ability in HCWs with job attitudes, worker engagement, and intrinsic predictors such as long-term sickness, returning to work after sustaining an injury, and disability after a work injury [13].
Identifying environmental predictors of work ability that can be intervened upon is critical for supporting workers' physical and mental well-being and enabling people to continue working productively [16]. Job demands (emotional, mental and physical) can be risk factors for decreased work ability [16]. Work ability could potentially be improved by enhancing job resources such as coworker support, supervisor support, decision latitude, and control over work schedules [15,[17], [18], [19]].
Organizational safety practices are another resource. These affect mental and physical health, including acute traumatic injuries and musculoskeletal disorders, all of which can produce disability. Two qualitative studies showed that poor “safety culture practices” (lax enforcement of rules) were associated with lower work ability in blue collar workers/construction workers [20,21]. Brandt et al. reported that two items from the management commitment scale of the Nordic Occupational Safety Climate Questionnaire (NOSACQ-50) tool, predicted lower work ability in Danish blue-collar workers after 2 years [22].
We have previously identified organizational support for safety (OSS) as an important predictor of HCW burnout [23] and depression [24]. In view of these findings, organizational safety practices might also affect HCW work ability [5] but formal analysis of the relationship has been limited.
Healthcare is a service-oriented and emotionally taxing job that requires management of one's emotions. Surface-acting emotional labor (SaEL), or masking one's feelings to perform the job, is associated with adverse mental health in junior doctors, nurses, and nursing home workers [[25], [26], [27], [28]]. However, we have found no study examining its association with work ability. Given that SaEL is a predictor of burnout and depression [23,24], we considered it another plausible risk factor to examine.
As noted above, the consequences of low work ability in HCWs carry heavy repercussions. Hence, there is an urgent need to identify and address preventable risk factors like organizational commitment to safety and SaEL, alongside other antecedents of work ability, in a mixed HCW population.
1.1. Hypotheses
The primary hypothesis for this study was that OSS and SaEL were associated with work ability in staff of long-term healthcare facilities, cross-sectionally and longitudinally.
2. Methods
This study was part of a larger project, “Safety and Health through Integrated, Facilitated Teams” (SHIFT), carried out in five public sector healthcare facilities offering a mix of residential and outpatient services. The study sites had a high number of low-paid unlicensed allied personal and medical assistants, such as mental health workers. All facilities were unionized and had preexisting employee safety and health infrastructure. Participants completed a self-administered questionnaire during personal time and were compensated $10. The questionnaire included a combination of adapted versions of published validated instruments and new instruments developed for this study (Supplementary table 1). Further details about the study sites and methods have been reported [29].
Baseline surveys were conducted in 2018-2019 and follow-up surveys in 2021. For the present analyses, work ability was the study outcome, measured on each occasion as the mean of current physical, mental and interpersonal work ability [3].
The primary variables were OSS and SaEL. OSS was computed from the management dimension of the NOSACQ-50 [30]. This scale has questions about: the presence of meaningful and practical safety routines at the workplace; employee influence on safety practices; employees’ participation in safety-policy-related decisions; collection of accurate information in accident investigations by the management; and judicious inquiry into accidents looking for causes, not guilty persons, when an accident occurs.
SaEL was measured with two items derived from [31].
For the other independent variables, psychosocial job strain was computed as the ratio of psychological job demands to decision latitude [32]; the sum of coworker and supervisor support [32]; civility norms in the workplace [33], and perceived presence of safety hazards (NIOSH 2008; Silver & Boiano, 2019). Back pain of at least moderate severity in the past 3 months was included because of the large literature on HCW turnover due to this condition (details of the questions and their sources are provided in Supplementary Table 1).
Age was dichotomized into younger (18-40 years) and older (≥40 years). Fourteen job titles were provided by the respondents and grouped into direct care of patients/residents or non-direct-care jobs such as dietary, housekeeping, facilities (maintenance) and security (Supplementary Table 2).
2.1. Statistical analysis
All analyses were done using SAS 9.4 [34]. Internal consistency of scales was measured using Cronbach's α coefficient. Work ability was compared between subgroups defined by gender, direct care/non-direct care jobs, and study site using t-test and one-way analysis of variance.
Spearman's correlation coefficients were computed among all pairs of variables because most of them had skewed distributions. Variables correlated with work ability at the level of p < 0.05 were retained for regression modeling. Regarding potential collinearity, no absolute correlations greater than 0.5 were observed. The variance inflation factor for all exposures was in the range of 1.02-1.5, suggesting no collinearity [35]. A combined score of coworker and supervisor support and civility norms (sum of the three ratings) was defined to create a comprehensive variable for the social landscape of the workplace and called the “social environment at work.”
Descriptive statistics were generated for demographic and work environment conditions in the baseline survey population, the follow-up population, and the group of workers who participated in both surveys (“overlap group”). Within the overlap group, paired t-tests assessed within person comparisons, i.e., whether these individuals' opinions of any work conditions differed on the two survey occasions. Two independent sample t-tests at both survey occasions assessed whether the individuals answering both surveys gave systematically different ratings than the respondents who were only surveyed once.
Linear regression analysis of work ability used forward stepwise model-fitting. Variables were retained if they had a p value ≤ 0.05 or if they changed the size of another coefficient by 15%. We fitted cross-sectional linear regression models for the baseline and final data sets, and longitudinal regression models with baseline exposures and the outcome at follow-up. Linear and longitudinal regression models were also stratified by direct care (yes/no).
3. Results
The questionnaire survey had 1059 participants at baseline (2018-2019) and 1553 at follow-up (2021); 428 individuals completed both baseline and follow-up surveys (“overlap” group). Participants were predominantly female in both surveys and more than one-half were direct care staff (Table 1). The mean ages for men were 48 years (baseline), and 47 years (follow-up), and for women 47 years (baseline) and 46 years (follow-up). Mean work ability scores were uniform among the sites. Mean work ability at baseline and at follow-up was lower in direct care workers (Table 1).
Table 1.
A: Characteristics of SHIFT study survey respondents employed in 5 public sector healthcare facilities. Baseline 2018-2019 (n=1059), follow-up 2021 (n=1553); B: Characteristics of SHIFT study survey respondents employed in 5 public sector healthcare facilities: n=428 survey respondents at both baseline and follow-up.
| Characteristics of SHIFT study survey respondents employed in 5 public sector healthcare facilities. Baseline 2018-2019 (n = 1059), follow-up 2021 (n = 1553). |
Characteristics of SHIFT study survey respondents employed in 5 public sector healthcare facilities: n = 428 survey respondents at both baseline and follow-up |
||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| Baseline n (%) | Follow-up n (%) | Baseline Work ability: mean (SD) |
Follow-up work ability: mean (SD) | Baseline work ability: median (q-q3) | Follow-up work ability: median (q1-q3) | Baseline n (%) | Baseline Work ability: mean (SD) |
Follow-up work ability: mean (SD) | Baseline work ability: median (q1-q3) | Follow-up work ability: median (q1-q3) | |
| Work ability: | 8.78 (1.43) | 8.60 (1.58) | 9 (2) | 9 (2) | 8.84 (1.28) | 8.42 (1.58) | 9 (1.66) | 9 (2) | |||
| Age | |||||||||||
| 18-<40 yr | 279 (26%) | 460 (30%) | 8.76 (1.36) | 8.59 (1.63) | 9 (2) | 9 (2) | 102 (24.%) | 8.64 (1.57) | 8.55 (1.44) | 9 (2) | 9 (2.33) |
| 40-70 yr | 682 (64%) | 982 (63%) | 8.79 (1.46) | 8.61 (1.57) | 9 (1.67) | 9 (2) | 281 (66%) | 8.94 (1.145) | 8.38 (1.59) | 9.33 (1.66) | 9 (2) |
| Missing | 98 (9%) | 111 (7%) | 45 (11) | ||||||||
| Gender | |||||||||||
| Male | 366 (34%) | 585 (38%) | 8.88 (1.39) | 8.61 (1.65) | 9 .33 (1.66) | 9 (2) | 149 (35%) | 8.97 (1.18) | 8.43 (1.58) | 9.33 (1.66) | 9 (1.66) |
| Female | 689 (65%) | 956 (62.%) | 8.75 (1.41) | 8.62 (1.55) | 9 (2) | 9 (2) | 276 (65%) | 8.83 (1.22) | 8.42 (1.58) | 9 (2) | 9 (2.33) |
| Site | |||||||||||
| VA Hospital | 311 (29%) | 493 (32%) | 8.59 (1.5) | 8.58 (1.49) | 9 (1.66) | 9 (2) | 88 (21%) | 8.84 (1.33) | 8.54 (1.41) | 9 (1.66) | 9 (1.66) |
| Soldiers' home | 194 (18%) | 197 (13%) | 8.67 (1.60) | 8.58 (1.67) | 9 (2) | 9 (2) | 82 (19%) | 8.84 (1.41) | 8.28 (1.71) | 9 (1) | 9 (1.66) |
| Soldiers' home | 138 (13%) | 116 (7%) | 8.86 (1.50) | 8.61 (1.60) | 9.33 (2) | 9 (2) | 45 (10%) | 8.92 (1.13) | 8.46 (1.47) | 9 (1.83) | 8.66 (2.33) |
| Mental health facility (1) | 300 (28%) | 594 (38%) | 8.77 (1.29) | 8.55 (1.72) | 9 (2) | 9 (2) | 156 (36%) | 8.80 (1.27) | 8.42 (1.65) | 9.16 (2) | 9 (2.33) |
| Mental health facility (2) | 116 (11%) | 153 (10%) | 8.94 (1.25) | 8.90 (1.33) | 9 .33 (1.66) | 9.33 (1.66) | 57 (13%) | 8.89 (1.21) | 8.37 (1.58) | 9.33 (1.66) | 8.66 (2.33) |
| Job type | |||||||||||
| Non-direct care | 478 (45%) | 678 (44%) | 8.91 (1.41) | 8.64 (1.63) | 9.33 (1.66) | 9 (2) | 195 (46%) | 8.99 (1.31) | 8.56 (1.53) | 9.33 (1.67) | 9 (2) |
| Direct care | 581 (55%) | 875 (56%) | 8.67 (1.44)∗ | 8.59 (1.50) | 9 (2) | 9 (2) | 233 (54%) | 8.73 (1.26)∗ | 8.28 (1.62) | 9 (2) | 8.66 (2.33) |
Note: ∗ = p < 0.05; SD, standard deviation. The p values from Wilcox scores, except for Kruskal–Wallis test for Site. “Transgender” category, transgender n = 3 removed for Wilcoxon test on gender.
Assessments of baseline work environment conditions did not differ between the overlap group (n = 428) and the remainder of the baseline survey participants (Table 2). Similarly, ratings by the overlap group at follow-up did not differ much from those of the remainder of the follow-up participants, except that work ability scores were slightly higher and SaEL (the adjustment of outward emotional expressions to meet professional requirements without changing underlying feelings) was slightly lower in the group only surveyed at follow-up. Mean SaEL values were higher and work ability scores were lower at follow-up. Thus, the overlap subgroup appeared generally representative of the larger cohort on both survey occasions.
Table 2.
Comparisons of work environment features in SHIFT all-employee survey respondents: Two-sample t-tests for baseline (f0) and follow-up (f2) groups versus the overlapping subjects; paired t-test for comparisons within overlapping subjects (N = 428)
| Variable name | Scale range | Baseline survey (f0) values: 2-sample t-test (428 vs 631) |
p | Follow-up survey (f2) values: 2-sample t-test (428 vs. 1131) |
p | Paired t-tests: f0 vs. f2 within overlap group |
||
|---|---|---|---|---|---|---|---|---|
| Overlap (n = 428) mean (SD) | Baseline only (631) mean (SD) | Overlap (n = 428) mean (SD) | Follow-up only (n = 1131) mean (SD) | p | ||||
| Work ability | (higher is good) | 8.84 (1.28) | 8.74 (1.52) | 0.24 | 8.42 (1.58) | 8.67 (1.60) | 0.0052 | <0.0001 |
| Physical work ability | (higher is good) | 9.09 (1.30) | 8.93 (1.62) | 0.07 | 8.71 (1.62) | 9.00 (1.58) | 0.0021 | <.0001 |
| Mental work ability | (higher is good) | 8.70 (1.56) | 8.63 (1.77) | 0.49 | 8.29 (1.89) | 8.51 (1.93) | 0.04 | <.0001 |
| Interpersonal work ability | (higher is good) | 8.73 (1.63) | 8.63 (1.79) | 0.39 | 8.25 (2.00) | 8.51 (1.94) | 0.02 | <.0001 |
| Organizational support for safety | 1-4 (higher is good) | 2.74 (0.53) | 2.76 (0.52) | 0.57 | 2.7 (0.54) | 2.82 (0.55) | 0.0001 | 0.09 |
| Safety hazards | 1-4 (higher is bad) | 2.35 (0.58) | 2.31 (0.58) | 0.29 | 2.24 (0.57) | 2.23 (0.60) | 0.82 | 0.0007 |
| Emotional labor | 1-4 (higher is bad) | 2.57 (0.99) | 2.54 (0.99) | 0.61 | 2.67 (0.97) | 2.51 (1.01) | 0.0064 | 0.05 |
| Job strain | (higher is bad) | 0.28 (0.10) | 0.29 (0.10) | 0.25 | 0.28 (0.11) | 0.288 (0.11) | 0.97 | 0.82 |
| Psychological demands | 2-8 (higher is bad) | 4.7 (1.27) | 4.8 (1.32) | 0.15 | 4.73 (1.35) | 4.67 (1.34) | 0.48 | 0.94 |
| Decision latitude | (higher is good) | 17.23 (2.90) | 17.16 (3.00) | 0.7 | 17.07 (2.95) | 16.91. (2.91) | 0.31 | 0.42 |
| Physical demands | 5-20 (higher is bad) | 10.68 (3.63) | 10.51 (3.9) | 0.49 | 10.57 (3.54) | 10.36 (3.55) | 0.29 | 0.48 |
| Civility norms | 1-4 (higher is good) | 2.72 (0.72) | 2.78 (0.71) | 0.19 | 2.75 (0.70) | 2.76 (0.74) | 0.75 | 0.42 |
| Coworker support | 2-8 (higher is good) | 5.72 (1.58) | 5.59 (1.55) | 0.18 | 5.69 (1.50) | 5.82 (1.58) | 0.14 | 0.79 |
| Supervisor support | 2-8 (higher is good) | 5.85 (1.19) | 5.81 (1.23) | 0.6 | 5.81 (1.24) | 5.89 (1.24) | 0.22 | 0.3 |
| Social support | 4-16 (higher is good) | 11.57 (2.31) | 11.39 (2.33) | 0.20 | 11.54 (2.31) | 11.73 (2.44) | 0.16 | 0.58 |
| Social environment | (higher is good) | 4.76 (0.91) | 4.72 (0.93) | 0.50 | 4.77 (0.91) | 4.83 (0.98) | 0.26 | 0.88 |
SD, standard deviation.
The independent variables were correlated with each other in the range of -0.52 to 0.51 at baseline (Supplementary Table 3) and with similar ranges in the later survey. Work ability was most strongly associated with OSS and the presence of safety hazards. Organizational support for safety was strongly correlated with safety hazards, SaEL, coworker support, and psychosocial job strain. Direct care staff had higher ratings of most physical and psychosocial demands than those in other jobs.
3.1. Regression analyses
In multivariable regression models, perceived OSS (positive effect) and SaEL (negative effect) were associated both cross-sectionally and longitudinally with work ability (Table 3), although the effect of SaEL was weak in the overlap group cross-sectionally. In addition, job strain, safety hazards, and back pain were negatively associated with work ability in all three cross-sectional models, while the social environment was positively associated cross-sectionally and longitudinally. Each of these six variables had the same direction in the crude and the demographically adjusted models for job characteristics. None of the interactions between pairs of work exposures was statistically significant.
Table 3.
Coefficients for linear regression models of work ability in relation to work environment characteristics: Employees of 5 public sector healthcare facilities. Baseline 2018-2019 (N = 1059), follow-up 2021 (N = 1553), and overlapped (N = 428) (adjusted for age, gender, job type, and site)
| Baseline: Crude models | Follow-up: Crude models | Baseline: Cross-sectional multivariable model | Follow-up: cross-sectional multivariable model | Longitudinal multivariable model | |
|---|---|---|---|---|---|
| OSS | 0.73 (<0.0001) | 0.85 (<0.0001) | 0.26 (0.0179) | 0.31 (0.0017) | 0.43 (0.02) |
| SaEL | -0.28 (<0.0001) | -0.50 (<0.0001) | -0.08 (0.08) | -0.31 (<0.0001) | -0.18 (0.05) |
| Social environment | 0.49 (<0.0001) | 0.42 (<0.0001) | 0.23 (0.0002) | 0.050 (0.40) | 0.02 (0.86) |
| Job strain | -4.08. (<0.0001) | -3.56 (<0.0001) | -2.09 (<0.0001) | -1.50 (0.0009) | – |
| Back pain | -0.42 (<0.0001) | -0.40 (<0.0001) | -0.29 (0.0028) | -0.18 (0.051) | – |
| Safety hazards | -0.58 (<0.0001) | -0.75 (<0.0001) | -0.08 (0.37) | -0.21 (0.014) | – |
| Jobs used | 795 | 1222 | 344 | ||
| r | 0.16 | 0.16 | 0.03 |
OSS, organizational support for safety; SaEL, surface-acting emotional labor.
3.2. Stratified regression analysis
After stratification by direct care (yes/no), OSS had a consistently positive association with work ability in all subsets of the data except for non-direct care workers longitudinally (Table 4). At baseline, the coefficient was nearly 75% larger in non-direct-care workers than in direct-care workers. However, at follow-up (cross-sectionally) and in the longitudinal model, the difference was in the opposite direction, i.e., positive associations were observed only for direct care workers.
Table 4.
Coefficients for linear regression models (stratified by direct-care vs. non-direct-care jobs) of work ability in relation to work environment characteristics: Employees of 5 public sector healthcare facilities baseline 2018-2019 (N = 1059), follow-up 2021 (N = 1553), and overlapped (N = 428). (Adjusted for age, gender, job type, and site)
| Baseline cross-sectional model |
Follow-up cross-sectional stratified model |
Longitudinal stratified model |
|||||||
|---|---|---|---|---|---|---|---|---|---|
| Unstratified | Dcare = 0 | Dcare = 1 | Unstratified | Dcare = 0 | Dcare = 1 | Unstratified | Dcare = 0 | Dcare = 1 | |
| OSS | 0.26 (0.0179) | 0.41 (0.015) | 0.19 (0.19) | 0.31 (0.0018) | 0.19 (0.22) | 0.41 (0.0010) | 0.50 (0.017) | -0.31 (0.43) | 0.56 (0.03) |
| Emotional labor | -0.08 (0.08) | -0.05 (0.46) | -0.13 (0.05) | -0.31 (<0.0001) | -0.42 (<0.0001) | -0.24 (<0.0001) | -0.18 (0.04) | -0.15 (0.29) | -0.21 (0.1) |
| Social environment | 0.23 (0.0002) | 0.21 (0.02) | 0.22 (0.00063) | 0.05 (0.35) | 0.07 (0.35) | -0.006 (0.93) | 0.06 (0.57) | 0.14 (0.43) | 0.01 (0.95) |
| Job strain | -2.09 (<0.0001) | -2.28 (0.0008) | -1.96 (0.0082) | -1.50 (0.0011) | 0.57 (0.40) | -2.98 (<0.0001) | -0.05 (0.95) | 0.14 (0.91) | -0.23 (0.86) |
| Back pain | -0.29 (0.0028) | -0.21 (0.13) | -0.32 (0.015) | -0.18 (0.05) | -0.19 (0.17) | -0.20 (0.10) | -0.05 (0.78) | 0.14 (0.61) | -0.06 (0.83) |
| Safety hazards | -0.08 (0.37) | -0.19 (0.17) | -0.02 (0.0082) | -0.21 (0.01) | --0.38 (0.008) | --0.13 (0.22) | -0.11 (0.53) | -0.10 (0.74) | 0.28 (0.21) |
| direct care | -0.14 (0.14) | -0.003 (0.97) | -0.05 (0.79) | ||||||
| aic | 2641 | 1207 | 1445 | 4321 | 1989 | 2325 | 1222 | 550 | 595 |
| Jobs used | 795 | 366 | 429 | 1222 | 548 | 674 | 306 | 144 | 162 |
OSS, organizational support for safety.
For the other independent variables, some effects were inconsistent between direct care and non-direct-care workers (Table 4). Surface-acting emotional labor at baseline had a negative effect only for direct care workers. However, at follow-up, the associations were more robust and 55% larger for non-direct care workers. There was a negative trend in the longitudinal models; neither coefficient was statistically significant, but this sample was about one-fourth of the size.
Job strain had a negative effect at follow-up and longitudinally for direct-care workers only. Safety hazards generally had a weak negative association. Back pain was consistently associated with lower work ability across the stratified analyses except for non-direct-care workers in the longitudinal analysis.
4. Discussion
This study of mixed licensed and unlicensed HCWs showed that work ability (combined physical, interpersonal, and psychological dimensions) was higher both cross-sectionally and prospectively with better organizational commitment to safety, after adjusting for other covariates. Surface-acting emotional labor had a consistently negative association but was weaker in two of the three multivariable models. To the best of our knowledge, the associations of work ability with OSS and surface-acting emotional labor are new findings for healthcare workers. Job strain also had a consistent and strong negative effect. Other associations with safety hazards, back pain, and the social environment were also in the expected directions, although they were less robust in the multivariable models.
Several elements of safety climate (from the NOSACQ tool) were previously found to be associated with lower physical and mental work ability after 2 years in Danish blue-collar workers [22]. Our results add to those findings with a methodologically stronger approach by using the complete nine-item scale for perceived organizational commitment to safety as a predictor in HCWs.
The odds of having safe work experiences during the pandemic in HCWs were predicted by prepandemic OSS [36]. The present finding that OSS was the single most important predictor of physical, mental, and interpersonal work ability over four years, at least in these direct-care workers, allows occupational health and safety policymakers a tangible point of intervention.
Direct-care workers reported lower mean work ability than non-direct-care workers, but both groups had highly skewed distributions of this outcome with low variance, and they experienced a similar decreased in work ability during the pandemic. We cannot easily explain why OSS predicted work ability only in the direct-care providers but not in the other group. Perhaps OSS is not as important to workers who do not provide direct care, or perhaps there was simply not enough variance in the outcome in this group. Even though we did not have statistical power to identify predictors of work ability in non-direct care workers, it is possible that OSS would still be relevant to their work ability as well. Since any changes in OSS would likely be made at an organizational level, they could affect the work environment of both direct care and non-direct care workers. Keeping this in mind, the finding that OSS predicted work ability in the longitudinal models is even more important.
According to the Conservation of Resources (COR) Theory [37], employees are motivated to retain and protect workplace resources that are most salient to their roles. It is possible that organizational safety climates serve as key resources for direct care staff, impacting their well-being and work ability. For non-direct care providers, organizational safety may be perceived as less directly relevant to their core work resources, thus contributing to the non-significant association observed in our analysis.
Organizational practices regarding safety should be a resource for employees to mitigate the effect of job demands [38]. Managerial safety practices, such as rule-oriented and participative leadership [21,39] have been associated with positive safety outcomes in other industries. Safe patient-handling programs in nursing homes resulted in reduced back pain, reduced compensation claims, and eventually turnover [40]. Therefore, management commitment to safety can have broader implications beyond injury prevention. The association of back pain with work ability in this material suggests that workplace safety can enhance both the person's performance and their confidence about remaining on the job.
In our study, the association of SaEL with work ability in HCWs shows how the effort of managing and suppressing ones' emotions impairs job performance. Suh and Punnett have demonstrated how SaEL predicts depression longitudinally, showing that SaEL is an important predictor of one's mental health. Primary prevention of SaEL is complicated due to the intrinsic interpersonal nature of healthcare work; however, organizations can take steps to make sure that employees are heard and feel protected when they are in a confrontational situation. This would be similar to some of the practices assessed by the OSS scale. More interventional studies are needed to evaluate the effectiveness of such efforts.
It is the responsibility of the organization to reduce this strain by implementing safety practices that are both consistent with the official messaging and effective in protecting workers.
Individual-oriented programs which seek to bolster HCW well-being, resilience, and peer support through behavioral interventions have met with limited success in improving mental health [41].
4.1. Limitations and strengths
This study utilized self-reported data on both exposures and the outcome; hence, common-method bias may exist. Individuals with reduced work ability might be more likely to exaggerate their levels of job demands and other stressors, leading to differential misclassification of exposures, i.e., information bias, and reverse causality in the cross-sectional analyses. Over one thousand employees were surveyed on each occasion, but the number for whom there were longitudinal data was somewhat limited. Further, the results might not be generalizable throughout the healthcare workforce since the study population consisted of largely of unionized employees from public-sector agencies providing chronic and residential care services.
This study was different from many prior studies of HCW work ability in that it had a moderate sample size, involved multiple sites, and included both clinical and nonclinical employees. Furthermore, we surveyed a wide range of occupational features and examined factors not often studied in relationship to work ability, specifically, OSS, surface-acting emotional labor, and back pain; others have studied the impact of injuries on variously defined work ability cross-sectionally. Measured constructs and scales were adapted from well-reviewed tools, studied for internal consistency, and found to be reliable.
4.2. Conclusions and future implications
The present findings support more organization-based interventions, such as improving safety policies and demonstrating commitment to safety in practice and improving the social environment at work by increasing coworker and supervisor support and civility. In our longitudinal results, a one-unit increase in OSS is associated with more than a half-unit gain in work ability, highlighting the practical significance of safety support interventions for direct care employees. These results indicate that, particularly for those with direct-care jobs, organizational safety support is likely to be impactful in real-world settings. Interventions in this area could potentially drive meaningful improvements in workforce capacity and well-being. With reference to the specific elements assessed by the OSS items, our findings support incorporating employee input into safety policies; accident reporting without scapegoating workers; and practicing participatory ergonomics by encouraging employees to take part in decision-making. Further, managerial skills and leadership training could be improved.
It is of great import for the management of healthcare institutions to invest in organizational policies promoting improved managerial commitment to employee safety. This would foster the safety culture of an organization-a cardinal resource for the betterment of worker well-being, thus leading to the building and sustaining of a healthy and resilient workforce. In the postpandemic work environment for HCWs, countries and organizations developing better safety policies and ensuring their implementation would improve the workers' perception of safety; these organizations would also have better work ability as opposed to those with a laissez-faire attitude toward worker safety.
CRediT authorship contribution statement
Sundus Siddique: Writing – review & editing, Writing – original draft, Methodology, Investigation, Formal analysis, Data curation, Conceptualization. Rebecca Gore: Writing – review & editing, Validation, Software, Methodology, Investigation, Formal analysis, Data curation. Tomas Hemmingsson: Writing – review & editing, Validation, Supervision. Laura Punnett: Writing – review & editing, Validation, Supervision, Software, Resources, Project administration, Investigation, Funding acquisition.
Ethics statement
This study was approved by the Institutional Review Board at University of Massachusetts Lowell (No. 16-131-PUN-XPD).
This research was supported by the U.S. National Institute for Occupational Safety and Health (NIOSH), Grant Number U19 OH008857 and U19 OH012299. The contents of this publication are solely the responsibility of the authors and do not necessarily represent the official views of NIOSH.
The authors declare that there is no potential conflict of interest in the research, authorship, funding and/or publication of this article.
Availability of data and materials
Data collection forms are available from the research team at: cphnew@uml.edu. Data files are not publicly available due to privacy guarantees to study participants, many of whom could be identified from their roles and demographic characteristics even in de-identified data sets.
Patients or the public
Patients or the public were not involved in the design, or conduct, or reporting, or dissemination plans of this particular study. However the broader SHIFT study was designed using participatory ergonomics. In which HCWs gave feedback based on the Total Worker health principals.
Important contribution to policy
Our study has the novel finding that organizational support for safety (OSS) predicts work ability in HCWs. These results indicate the need for organization level changes to improve workplace safety. These changes could cause simultaneously alleviate other job demands like SaEL, job strain and back pain for all groups of HCWs.
Funding
This research was supported by the U.S. National Institute for Occupational Safety and Health (NIOSH), Grant Number U19 OH008857 and U19 OH012299. The contents of this publication are solely the responsibility of the authors and do not necessarily represent the official views of NIOSH.The authors declare that there is no potential conflict of interest in the research, authorship, funding and/or publication of this article.
Conflicts of Interest
No competing interest.
Footnotes
Supplementary data to this article can be found online at https://doi.org/10.1016/j.shaw.2026.07.003.
Contributor Information
Sundus Siddique, Email: dr_sundus_siddique@outlook.com.
Rebecca Gore, Email: Rebecca_Gore@uml.edu.
Tomas Hemmingsson, Email: tomas.hemmingsson@su.se.
Laura Punnett, Email: Laura_Punnett@uml.edu.
Appendix A. Supplementary data
The following is the Supplementary data to this article:
References
- 1.Rudolph Cort W.Z.H. “The COVID-19 generation”: a cautionary note. J Work Aging Retirement. 2020;6(3):139–145. doi: 10.1093/workar/waaa009. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.De Vries N., Boone A., Godderis L., Bouman J., Szemik S., Matranga D., De Winter P. Vol. 60. 2023. (The race to retain healthcare workers: a systematic review on factors that impact retention of nurses and physicians in hospitals). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Ilmarinen J., Tuomi K., Eskelinen L., Nygård C.-H., Huuhtanen P., Klockars M. Summary and recommendations of a project involving cross-sectional and follow-up studies on the aging worker in Finnish municipal occupations (1981—1985) J Scandinavian J Work Environ Health. 1991:135–141. [PubMed] [Google Scholar]
- 4.Tuomi K., Ilmarinen J., Eskelinen L., Järvinen E., Toikkanen J., Klockars M. Prevalence and incidence rates of diseases and work ability in different work categories of municipal occupations. J Scandinavian J Work Environ Health. 1991:67–74. [PubMed] [Google Scholar]
- 5.Rudolph Cort W., Hannes Zacher. Age inclusive human resource practices, age diversity climate, and work ability: Exploring between-and within-person indirect effects. J Work Aging Retirement. 2021;7(4):387–403. [Google Scholar]
- 6.Seitsamo J., Martikainen R. Work ability and all cause mortality: a 25-year longitudinal study among Finnish municipal workers. J Promot Work Ability Towards Product Ageing. 2008:101–104. [Google Scholar]
- 7.Alavinia S.M., De Boer A., Van Duivenbooden J., Frings-Dresen M., Burdorf A. Determinants of work ability and its predictive value for disability. Occupat Med. 2009;59(1):32–37. doi: 10.1093/occmed/kqn148. [DOI] [PubMed] [Google Scholar]
- 8.Vänni K., Virtanen P., Luukkaala T., Nygård C.-H. Relationship between perceived work ability and productivity loss. Int J Occup Saf Ergon. 2012;18(3):299–309. doi: 10.1080/10803548.2012.11076946. [DOI] [PubMed] [Google Scholar]
- 9.Kerksieck P., Ballouz T., Haile S.R., Schumacher C., Lacy J., Domenghino A. 2023. Post COVID-19 condition, work ability and occupational changes in a population-based cohort; p. 31. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Sundstrup E., Jakobsen M.D., Mortensen O.S., Andersen L.L. Joint association of multimorbidity and work ability with risk of long-term sickness absence: a prospective cohort study with register follow-up. J Scandinavian J Work Environ Health. 2017:146–154. doi: 10.5271/sjweh.3620. [DOI] [PubMed] [Google Scholar]
- 11.Palmlöf L., Skillgate E., Talbäck M., Josephson M., Vingård E., Holm L.W. Poor work ability increases sickness absence over 10 years. Occupat Med. 2019;69(5):359–365. doi: 10.1093/occmed/kqz083. [DOI] [PubMed] [Google Scholar]
- 12.Seitsamo J., Martikainen R. CRC Press; 2008. Work ability and all cause mortality: a 25-year longitudinal study among Finnish municipal workers. Promotion of work ability towards productive aging; pp. 111–114. [Google Scholar]
- 13.Cadiz D., Brady G., Truxillo D., Zaniboni S., editors. Biannual age in the workplace conference. 2019. The deleterious effects of age discrimination through perceptions of work ability. St Gallen, Switzerland. [Google Scholar]
- 14.Cadiz D.M., Brady G., Rineer J.R., Truxillo D.M. A review and synthesis of the work ability literature. J Work. Aging Retirement. 2019;5(1):114–138. [Google Scholar]
- 15.McGonagle A.K., Fisher G.G., Barnes-Farrell J.L., Grosch J.W. Individual and work factors related to perceived work ability and labor force outcomes. J Journal of Applied Psychology. 2015;100(2):376. doi: 10.1037/a0037974. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Brady G.M., Truxillo D.M., Cadiz D.M., Rineer J.R., Caughlin D.E., Bodner T. Opening the black box: examining the nomological network of work ability and its role in organizational research. J Appl Psychol. 2020;105(6):637. doi: 10.1037/apl0000454. [DOI] [PubMed] [Google Scholar]
- 17.Rudolph Cort W.M.A.K. Exploring age-conditional effects in the emotional labor–perceived work ability linkage: a daily diary study. J Work. Aging Retirement. 2019;5(2):163–174. [Google Scholar]
- 18.Bakker A.B., Demerouti E., Euwema M.C. Job resources buffer the impact of job demands on burnout. J Occup Health Psychol. 2005;10(2):170. doi: 10.1037/1076-8998.10.2.170. [DOI] [PubMed] [Google Scholar]
- 19.Mayerl H., Stolz E., Waxenegger A., Rásky É., Freidl W. The role of personal and job resources in the relationship between psychosocial job demands, mental strain, and health problems. J Front Psychol. 2016;7:1214. doi: 10.3389/fpsyg.2016.01214. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Choudhry R.M., Fang D., Mohamed S. The nature of safety culture: a survey of the state-of-the-art. J Safety Sci. 2007;45(10):993–1012. [Google Scholar]
- 21.Ajslev J.Z.N., Lund H.L., Møller J.L., Persson R., Andersen L.L. Habituating pain: questioning pain and physical strain as inextricable conditions in the construction industry. J Nordic J Work Life Stud. 2013;3(3):195–218. [Google Scholar]
- 22.Brandt M., Sundstrup E., Andersen L.L., Wilstrup N.M., Ajslev J.Z. Safety climate as a predictor of work ability problems in blue-collar workers: prospective cohort study. BMJ Open. 2021;11(3) doi: 10.1136/bmjopen-2020-040885. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Siddique S., Gore R., Zhang Y., Punnett L. Emotional exhaustion in healthcare workers: moving beyond coping skills to improve organizational conditions. J Occup Environ Med. 2024;10:1097. doi: 10.1097/JOM.0000000000003063. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Siddique S.G.R., Zhang Y., Punnett L., SHIFT Research Team Impact of surface-acting emotional labor on depression in healthcare workers: the role of emotional exhaustion as a mediator. Work Stress Health Conf (WSH Conference) Nov 2021. 2021 [Google Scholar]
- 25.Rogers M.E., Creed P.A., Searle J. Emotional labour, training stress, burnout, and depressive symptoms in junior doctors. J Vocation Educ Train. 2014;66(2):232–248. [Google Scholar]
- 26.Aung N., Tewogbola P. The impact of emotional labor on the health in the workplace: a narrative review of literature from 2013–2018. AIMS Public Health. 2019;6(3):268. doi: 10.3934/publichealth.2019.3.268. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Yoon S.L., Kim J.H. Job-related stress, emotional labor, and depressive symptoms among Korean nurses. J Nurs Scholar. 2013;45(2):169–176. doi: 10.1111/jnu.12018. [DOI] [PubMed] [Google Scholar]
- 28.Suh C., Punnett L. Surface-acting emotional labor predicts depressive symptoms among health care workers over a 2-year prospective study. Arch Occup Environ Health. 2021;94:367–375. doi: 10.1007/s00420-020-01585-8. [DOI] [PubMed] [Google Scholar]
- 29.Punnett L., Nobrega S., Zhang Y., Rice S., Gore R., Kurowski A. Safety and Health through Integrated, Facilitated Teams (SHIFT): stepped-wedge protocol for prospective, mixed-methods evaluation of the Healthy Workplace Participatory Program. BMC Public Health. 2020;1463 doi: 10.1186/s12889-020-09551-2. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Kines P., Lappalainen J., Mikkelsen K.L., Olsen E., Pousette A., Tharaldsen J., Tómasson K., Törner M. Nordic Safety Climate Questionnaire (NOSACQ-50): a new tool for diagnosing occupational safety climate. J Int J Indus Ergonom. 2011;41(6):634–646. [Google Scholar]
- 31.Brotheridge C.M., Grandey A.A. Emotional labor and burnout: comparing two perspectives of “people work”. J Vocat Behav. 2002;60(1):17–39. [Google Scholar]
- 32.Karasek Jr RA. Job demands, job decision latitude, and mental strain: implications for job redesign. J Admin Sci Quart. 1979:285–308. [Google Scholar]
- 33.Walsh B.M., Magley V.J., Reeves D.W., Davies-Schrils K.A., Marmet M.D., Gallus J.A. Assessing workgroup norms for civility: the development of the Civility Norms Questionnaire-Brief. J Bus Psychol. 2012;27:407–420. [Google Scholar]
- 34.Cary N. SAS Institute Inc; USA: 2013. Base SAS 9.4 procedures guide: statistical procedures. [Google Scholar]
- 35.Vatcheva K.P., Lee M., McCormick J.B., Rahbar M.H. Multicollinearity in regression analyses conducted in epidemiologic studies. J Epidemiol. 2016;6(2) doi: 10.4172/2161-1165.1000227. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Siddique Sundus RS., Mansi Bhardwaj, Rebecca Gore, Haylee Coupal, Laura Punnett. Health care organization policies for employee safety and COVID-19 pandemic response: a mixed-methods study. J Occup Environ Med. 2023;65(1):1. doi: 10.1097/JOM.0000000000002741. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Hobfoll S.E. Conservation of resources: a new attempt at conceptualizing stress. Am Psychol. 1989;44(3):513. doi: 10.1037//0003-066x.44.3.513. [DOI] [PubMed] [Google Scholar]
- 38.Reiman T., Oedewald P. Measuring maintenance culture and maintenance core task with CULTURE-questionnaire––a case study in the power industry. J Safety Sci. 2004;42(9):859–889. [Google Scholar]
- 39.Grill M., Pousette A., Nielsen K., Grytnes R., Törner M. Safety leadership at construction sites: the importance of rule-oriented and participative leadership. Scand J Work Environ Health. 2017:375–384. doi: 10.5271/sjweh.3650. [DOI] [PubMed] [Google Scholar]
- 40.Lahiri S., Latif S., Punnett L. An economic analysis of a safe resident handling program in nursing homes. Am J Indus Med. 2013;56(4):469–478. doi: 10.1002/ajim.22139. the ProCare Research Team [DOI] [PMC free article] [PubMed] [Google Scholar]
- 41.Akinnusotu O., Bhatti A., Doubeni C.A., Williams M. Supporting mental health and psychological resilience among the health care workforce: gaps in the evidence and urgency for action. J Annal Family Med. 2023;21(Suppl. 2):S100–S102. doi: 10.1370/afm.2933. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
Data collection forms are available from the research team at: cphnew@uml.edu. Data files are not publicly available due to privacy guarantees to study participants, many of whom could be identified from their roles and demographic characteristics even in de-identified data sets.