Introduction
An important and growing source of job stress for many workers is the challenge of managing one or more chronic medical conditions at work. Forty percent of US workers report persistent or recurrent musculoskeletal pain or other chronic physical health conditions that are currently limiting their ability to work [
1,
2]. With the increased age and obesity of the US workforce, the prevalence of chronic diseases is likely to increase [
3] and even younger workers are showing a higher prevalence of obesity and chronic health symptoms [
4,
5]. The most prevalent chronic conditions impacting working-age adults include low back pain, arthritis, migraine headaches, depression, diabetes, heart disease, and asthma [
6‐
8].
Chronic conditions can lead to unpredictable and fluctuating symptoms, daily oscillations in work tolerance and fatigue, increased injury risk, and complex self-care or treatment regimens [
9‐
11]. For workers with chronic conditions, conventional health and disability employer benefits may fail to satisfy day-to-day needs for job flexibility, leeway, and organizational support to cope effectively with periodic symptoms and task limitations. Over time, this can lead to increased job stress, dissatisfaction, turnover, and long-term disability [
12‐
18]. Persistent or intermittent symptoms and fluctuating dysfunction can challenge conventional workplace approaches to disability management and accommodation [
11,
19‐
21].
While legislative mandates exist for employers to support job accommodation in principle, workers with chronic medical conditions still report problems communicating their needs effectively, making allowable work style and workstation adjustments, dealing with discomfort, overcoming stigma and discrimination, and keeping up with productivity expectations [
22‐
24]. To overcome these problems, workers report a number of coping strategies: leveraging available job leeway and flexibility, careful planning and decision-making at work, obtaining sufficient job assistance and social support in and out of work, and communicating needs effectively and judiciously with peers and supervisors [
25,
26]. In addition to its toll on workers, chronic illness increases employer costs through reduced productivity, high turnover rates, absenteeism, and health care expense [
27‐
30]. To reduce these costs, more proactive strategies are needed for employers to help workers with episodic symptoms.
Besides objective work outcomes such as sickness absence and temporary disability leave, subjective self-report scales have been helpful to detect early perceptions of job-related limitations, disengagement, coping, and dissatisfaction among workers with chronic health conditions [
15,
23,
29]. One hypothesis from this literature is that workers with chronic health conditions are unable to recover from a high daily level of exhaustion after working hours, and this ultimately leads to job burnout over time [
23,
31]. Another theory is that workers slowly lose confidence in their abilities to perform routine tasks, to keep up with productivity demands, and to contribute meaningfully to their working teams [
15,
30,
32]. These feelings may be associated with decreased personal self-efficacy and reduced confidence to manage pain or solve health-related challenges at work [
32].
In the following study, we report the results of a randomized, controlled trial of a 5-session, worksite-based self-management program designed to address functional problems of workers with chronic health conditions (the
Manage at Work study) [
33]. We hypothesized that an employer-supported group intervention program adapted from principles of pain and illness self-management to the workplace context would improve work engagement and reduce work limitations.
Results
A total of 119 participants (98 female, 20 male, 1 unspecified gender) provided informed consent and were randomized to the intervention or control arms of the study. The most common occupations were administrative assistant (19%), manager/supervisor (17%), data analyst or research assistant (13%), medical assistant (12%), medical technologist (9%), nurse or nursing assistant (7%), and coding or billing specialist (7%). Other occupations included lab scientists, engineers, teachers, cashiers, assemblers, and counselors. The study population was comprised of mostly skilled workers in technical jobs with at least moderate levels of job security and opportunities for advancement.
Participants endorsed a median of 3 chronic health conditions (range 1–11). The most frequent chronic conditions were as follows: back or neck problems (85%), hand/arm problems (61%), leg or feet problems (55%), migraine or severe headaches (43%), vision problems (31%), stomach or bowel disorders (24%), asthma, bronchitis, or emphysema (22%), mental health disorders (17%), hearing problems (12%), cardiovascular disease (5%), severe skin disorders (4%), and diabetes (4%). Participants responded to either employer-based (92%) or community-based (8%) announcements about the study. The most prevalent comorbidity was to report a chronic musculoskeletal pain problem affecting more than one body part. Most participants (83.2%) were working in a healthcare setting. Participants reported missing from 0 to 70 workdays (M = 5.7, SD = 12.1) due to chronic health conditions over the previous six months.
Table
1 shows demographic characteristics by study groups, and these data reflect a mostly White, college educated, middle-aged cohort of workers with significant industry experience and job tenure. Despite random assignment to the intervention and control arms of the study, there was a trend for the control group to be older (mean age = 47.9 versus 44.2) and there was a statistically significant group difference on job tenure, with the control group having more workers with > 5 years (61% versus 42%). Because these two variables have implications for both work and health outcome variables, they were retained as covariates in the subsequent multivariate group comparisons of study outcomes. There were no statistically significant differences or notable trends between the control and intervention groups on any other demographic variables.
Table 1
Baseline demographic characteristics of study participants by randomized group assignment
Age | 46.0 (12.7) | 44.2 (12.0) | 47.9 (13.2) | t = 1.624 | 0.107 |
Chronic health conditions | 3.7 (1.9) | 3.7 (2.1) | 3.6 (1.7) | t = 0.208 | 0.836 |
Children/dependents at home | | | χ2 = 1.66 | 0.799 |
Yes | 45 (37.8) | 22 (36.7) | 23 (39.0) | | |
No | 74 (62.2) | 38 (63.3) | 38 (64.4) | | |
Gender | | | | χ2 = 3.16 | 0.206 |
Male | 20 (16.8) | 13 (21.7) | 7 (11.9) | | |
Female | 98 (82.4) | 46 (76.7) | 52 (88.1) | | |
Not specified | 1 (0.8) | 1 (1.7) | 0 (0.0) | | |
Race | | | | χ2 = 0.17 | 0.918 |
Asian | 2 (1.7) | 1 (1.7) | 1 (1.7) | | |
Black | 7 (5.9) | 3 (5.0) | 4 (6.8) | | |
White | 108 (90.8) | 55 (91.7) | 53 (89.8) | | |
Not reported | 2 (1.7) | 1 (1.7) | 1 (1.7) | | |
Ethnicity | | | | χ2 = 0.00 | 1.000 |
Hispanic | 4 (3.4) | 2 (3.3) | 2 (3.3) | | |
Non-Hispanic | 112 (94.1) | 56 (93.3) | 56 (94.9) | | |
Not reported | 3 (2.5) | 2 (3.3) | 1 (1.7) | | |
Marital status | | | | χ2 = 2.21 | 0.531 |
Never married | 25 (21.0) | 12 (20.0) | 13 (22.0) | | |
Married/partnered | 67 (56.3) | 35 (58.3) | 32 (54.2) | | |
Divorced/separated | 25 (21.0) | 13 (21.7) | 12 (20.3) | | |
Widowed | 2 (1.7) | 0 (0.0) | 2 (3.4) | | |
Annual income | | | | χ2 = 5.00 | 0.840 |
$10,000–$19,999 | 3 | 1 | 2 | | |
$20,000–$29,999 | 14 | 8 | 6 | | |
$30,000–$39,999 | 27 | 12 | 15 | | |
$40,000–$49,999 | 27 | 15 | 12 | | |
$50,000–$59,999 | 11 | 7 | 4 | | |
$60,000–$69,999 | 10 | 5 | 5 | | |
$70,000–$79,999 | 13 | 5 | 8 | | |
$80,000–$89,999 | 5 | 2 | 3 | | |
$90,000–$99,999 | 4 | 1 | 3 | | |
$100,000 or over | 4 | 3 | 1 | | |
(Missing) | 1 | 1 | 0 | | |
Highest education | | | | χ2 = 3.69 | 0.450 |
< 12 years | 1 (0.8) | 0 (0.0) | 1 (1.7) | | |
High school | 9 (7.6) | 6 (10.0) | 3 (5.1) | | |
Some college | 54 (45.4) | 28 (46.7) | 26 (44.1) | | |
Bachelor’s degree | 27 (22.7) | 15 (25.0) | 12 (20.3) | | |
Post-bachelor’s | 28 (23.5) | 11 (18.3) | 17 (28.8) | | |
With current employer | | | | χ2 = 14.17 | 0.015 |
0–6 months | 8 (6.7) | 4 (6.7) | 4 (6.8) | | |
6–12 months | 8 (6.7) | 8 (13.3) | 0 (0.0) | | |
1–2 years | 8 (6.7) | 6 (10.0) | 2 (3.4) | | |
2–5 years | 34 (28.6) | 17 (28.3) | 17 (28.8) | | |
5–10 years | 29 (24.4) | 9 (15.0) | 20 (33.9) | | |
> 10 years | 32 (26.9) | 16 (26.7) | 16 (27.1) | | |
With current industry | | | | χ2 = 3.67 | 0.599 |
0–6 months | 2 (1.7) | 0 (0.0) | 2 (3.4) | | |
6–12 months | 5 (4.2) | 4 (6.7) | 1 (1.7) | | |
1–2 years | 10 (8.4) | 5 (8.3) | 5 (8.5) | | |
2–5 years | 24 (20.2) | 14 (23.3) | 10 (16.9) | | |
5–10 years | 26 (21.8) | 12 (20.0) | 14 (23.7) | | |
> 10 years | 52 (43.7) | 25 (41.7) | 27 (45.8) | | |
Working > 1 employer | | | | χ2 = 0.01 | 0.973 |
No | 105 (88.2) | 53 (88.3) | 52 (88.1) | | |
Yes | 14 (11.8) | 7 (11.7) | 7 (11.9) | | |
Attendance records maintained by group facilitators reflected a moderate to high level of engagement in the intervention group, with 81% of participants completing at least half of the workplace self-management sessions. Subjective ratings of satisfaction with the intervention program were also high (mean overall satisfaction rating of 8.8 out of 10). Participants reported no adverse events related to study participation.
The means and standard deviations for baseline and follow-up measures are shown in Table
2. Of the original sample (n = 119), 98 participants (82%) completed the 6-month survey. All tests of statistically significant changes in groups are controlled for the baseline covariates of age and job tenure. Both groups showed increasing work limitations over the 6-month period, and there was no significant group x time interaction (p > 0.05) that would suggest the intervention program had attenuated this effect. On work engagement, however, the intervention group improved slightly, while the control group experienced a decline over the same 6-month period (interaction test,
p < 0.05). There was also a statistically significant improvement in turnover intent for the intervention group versus the control group (interaction test,
p < 0.05). Secondary measures of work self-efficacy, work fatigue, and job satisfaction showed a trend for the intervention group to improve in contrast with the control group, but these differences did not reach statistical significance.
Table 2
Changes in work outcomes by group assignment
Primary outcome measures | | | | | | | |
Work engagement | 4.53 (0.99) | 4.79 (1.36) | + 0.22 | 5.10 (1.01) | 4.68 (1.32) | − 0.36 | 7.88 | 0.006 |
Work limitation | 7.66 (4.80) | 8.15 (5.16) | + 0.10 | 6.79 (4.90) | 7.77 (5.51) | + 0.19 | 0.16 | 0.686 |
Secondary outcome measures | | | | | | | |
Job satisfaction | 6.62 (1.98) | 6.43 (2.27) | − 0.09 | 6.80 (1.59) | 6.40 (2.17) | − 0.21 | 0.20 | 0.660 |
Work fatigue | 2.88 (1.14) | 2.67 (1.29) | − 0.17 | 2.89 (1.34) | 2.80 (1.32) | − 0.07 | 0.45 | 0.505 |
Work self-efficacy | 3.97 (0.97) | 4.12 (1.13) | + 0.14 | 4.05 (1.10) | 4.01 (1.20) | − 0.03 | 1.23 | 0.271 |
Days lost to illness | 4.57 (9.91) | 2.82 (5.66) | − 0.22 | 7.88 (15.64) | 8.25 (14.45) | + 0.02 | 0.26 | 0.615 |
Turnover intent | 2.21 (1.10) | 2.12 (1.14) | − 0.08 | 2.03 (1.04) | 2.41 (1.26) | + 0.33 | 4.05 | 0.047 |
On the 6-month recall of days lost to illness there was a marked reduction in the intervention group (from 4.6 to 2.8) versus the control group (from 7.9 to 8.3), but this difference in means was not statistically significant due to the positive skew of these data (i.e., only a few workers reported a high number of days lost to ill health). However, a post-hoc analysis of the number of workers reporting more than 10 days of lost time (McNemar test, p < 0.05) showed a statistically significant worsening of sick days in the control group (those with > 10 days increased from 5 to 10 individuals) but not in the intervention group (those with > 10 days decreased from 4 individuals to 1 individual).
Discussion
This study presents a novel employer-sponsored strategy to improve the working conditions for employees with chronic health problems using a group psycho-educational format. The conceptual framework for the intervention was based around concepts of self-efficacy and health self-management, but targeted to address challenges of communication, pacing, problem-solving, and job modification unique to the workplace setting. This registered randomized clinical trial (Clinicaltrials.gov: NCT01978392) provided a strong methodological framework to evaluate benefits of the MANAGE AT WORK intervention program. Though the project was ended prematurely due to an unexpected closing of the sponsoring research institute, preliminary results indicate an improvement in workplace engagement while other health and work outcomes are approaching statistically significant gains when compared with a wait-list control group.
This study adds to the growing literature on the struggles of aging or ill workers to stay on the job despite recurrent or episodic symptoms, but workplace intervention strategies for this population have been sparse in the occupational rehabilitation literature [
44]. While much of the existing research has focused on quantifying health effects on work productivity, other research has been helpful to illustrate the more subjective struggles of workers to maintain social relations, negotiate regarding accommodations, develop a realistic understanding of personal capabilities, and to demonstrate an appropriate level of assertiveness and knowledge to overcome organizational barriers [
46,
47]. Employers may find it difficult to provide a supportive work environment for workers with chronic health conditions when symptoms are intermittent and when there may be no explicit request for accommodation. As reported in previous studies of chronic medical conditions among working populations [
48], many participants reported multiple comorbid health conditions, often involving multiple body parts and systems. The most prevalent problems related to widespread musculoskeletal pain.
One broad indication from the study is the relative success of recruiting at-risk employees, making meetings convenient and safe for participants, fostering discussion, and encouraging workplace self-management strategies without any reported incidents of negative interactions or adversarial relations at work. Thus, it may be possible to expand on workers’ efforts to obtain needed assistance and manage their workloads responsibly without disrupting employers and co-workers. One limitation of this approach is its reliance on workers to make or request adaptations in the workplace. The communication module of the training was built around the idea of promoting assertive communication about health, but this was balanced by the realities of working relationships, organizational culture, and workplace codes of conduct. Similarly, the job modification module was designed to make better use of existing leeway and flexibility using a problem-solving approach, while considering potential organizational or task-related barriers and constraints. Some workplace accommodation requests require more formal medico-legal processes, but the focus of the intervention was on informal types of job modification and assistance shown to be key for this working population with more episodic and transient problems [
23].
The two primary outcome measures were work engagement and work limitations. Work engagement was the outcome measure that showed the greatest responsiveness to the intervention program. This construct was introduced in recent years to capture a more qualitative element of job satisfaction and is defined as “…a positive, fulfilling, work-related state of mind that is characterized by vigor, dedication, and absorption” [
38]. For workers with chronic or episodic health problems, negative feelings about work may take the form of burn-out, exhaustion, feeling less capable and isolated, and experiencing a general disengagement from work that is well captured by this psychological construct. Our self-management intervention may address these beliefs by identifying more options for positive change, finding ways to communicate needs effectively to managers and co-workers, and building a sense of mastery to overcome future challenges. As such, the intervention may have led workers to feel more interested and active in their day-to-day work and less absorbed by growing health problems and feelings of helplessness.
The other primary outcome measure for the study was the Work Limitations Questionnaire (WLQ). In contrast to work engagement, the WLQ did not show improvement in response to the intervention in our preliminary analyses. This measure has been very instrumental in research and evaluation to quantify the work productivity losses attributable to various medical conditions [
9,
15,
35], but it has only begun to be used as an individual-level metric for evaluating intervention effectiveness. The intervention failed to improve participants’ perceived ability to meet time pressures or perform physically demanding aspects of their work relative to the control group. One possible explanation is that the WLQ was not sufficiently responsive to short-term intervention, as the measure has shown statistically significant improvements in only one published trial (for depression) [
49,
50]. Another possibility is that workers with chronic health concerns are less concerned about their raw productivity output than about their psychological well-being. The uptake of future workplace interventions for workers with chronic health conditions may depend on showing improved well-being while also enhancing productivity.
An important practical implication of the study is that employers can provide health coaching or other educational or counseling programs to workers with chronic or episodic illness in the same vein as other types of workplace health promotion (e.g., smoking cessation or weight loss programs). In this example, groups were facilitated by Employee Assistance Program (EAP) counselors, who were specially trained to deliver the intervention program. An advantage of this approach was that EAP counselors had both a strong mental health training background but also understood issues of organizational culture, workplace relationships, and disability employer regulations. A second implication of our results is that workers did experience some increased attachment to their jobs as a product of a facilitated group intervention addressing health-related challenges. Providing more options for dealing with workplace challenges and a greater sense of mastery may improve the dedication and commitment of workers with chronic health conditions.
Our study is not without limitations. The premature cessation of funding on the project led to a smaller sample size (119 vs. 300) and a reduced statistical power to show significant differences between groups. In terms of the intervention itself, an organizational component may have reinforced or strengthened the self-management and coping efforts of workers. Using cluster randomized designs may help to reduce the potential for cross-contamination between intervention and control groups in future studies. Most participants reported multiple chronic conditions, so the effect of overlapping comorbidities was difficult to assess. A shared effort that places equal responsibility on workers and supervisors to solve problems would place less onus on workers and more emphasis on organizational factors that can be changed.
Participating organizations may have been more apt than the average employer to allow workers to communicate with managers and peers about health concerns without any threat of retaliation or punitive treatment, though this is certainly an area of concern for more widespread implementation. The study sample included only a few occupational and industry groups, and effectiveness of the intervention may vary by occupation and industry. Future studies might combine this type of program with other organizational efforts (e.g., supervisor training, communications from management, ergonomic evaluations, peer coaches) to respond to the problem more completely. Another future direction would be to develop readiness assessment tools for employers to participate in such programs targeting workers with chronic health conditions.
In conclusion, this study demonstrates the initial feasibility of offering an employer-sponsored psycho-educational program to workers with chronic physical health conditions, and participation in this program resulted in improved levels of work engagement. Workers with chronic health conditions represent a fast-growing subset of the workforce in the US and elsewhere, and novel intervention strategies involving employer partnerships are sorely needed. We conclude that a group psycho-educational program is a feasible and effective way to improve worker engagement for those workers with chronic health conditions. Further efforts to develop and test employer-sponsored programs of this type should be pursued in different jurisdictions, contexts, and medical conditions to understand feasibility and effectiveness for more widespread dissemination.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.