Manage at work: a randomized, controlled trial of a self-management group intervention to overcome workplace challenges associated with chronic physical health conditions

In the US, employers are required to provide reasonable accommodation for employees with disabling health conditions, and a job position must be held open for a sick worker from 3 months to several years, depending on the state of jurisdiction and whether the illness is deemed work-related. Intermittent health problems, with recurring flare-ups and/or short, periodic absences from work represent a special challenge for employers, and workers with chronic conditions are at greater risk for job loss in the US than in many other countries. While many large employers in the US offer employees programs in health promotion, health risk appraisal, and wellness coaching, few of these programs offer advice for managing chronic health conditions while also keeping up with work demands. Disability and insurance benefits for non-work related conditions can vary and depend on employee-paid private insurance coverage.

The proposed study design is a randomized, controlled trial (RCT) of an employer-sponsored psycho-educational group intervention program designed to improve workplace functioning among workers with chronic physical health symptoms (Clinical Trials Registry # NCT01978392). The study methodology involves recruiting employees with chronic conditions from five worksites, randomizing them to participate in a group intervention program or to a wait-list control arm, and assessing changes in baseline worker engagement and work limitations at 6- and 12-month follow-up. We hypothesize that group participation will improve health and disability outcomes by improving self-efficacy beliefs in several work-related domains (see Figure 1). A process evaluation will be conducted alongside the RCT. The study has been reviewed and approved by the Dartmouth College Committee for the Protection of Human Subjects (#24084).

The study population is workers who have at least one chronic physical health condition (>6 months) and are interested in exploring ways to deal effectively with health-related challenges associated with body aches, discomfort, and fatigue on the job. The recruitment procedure will include general workforce announcements (email, posters, flyers, employee newsletter) as well as individual notices to workers accessing wellness services or interacting with on-site employee healthcare providers. All inquiries about the project will be forwarded to a local project coordinator who will answer questions, obtain informed consent, administer the baseline survey, and schedule employees for workshop meetings.

Participants will be full-time workers (>20 hours per week) of age 18 years or older who self-report at least one chronic physical health condition (>6 months). All participants must read and speak in English. To prevent any unnecessary health disclosure, interested employees will not be required to divulge medical information about diagnoses and symptoms to the on-site project coordinator in order to qualify for the study; however, this background health information will be collected as part of the baseline survey after recruitment.

There are no diagnostic exclusions, but workers with an impending retirement or major job change (next 12 months) will be excluded, as will those who report no availability to attend group workshops before work, after work, or during lunch breaks. Workers who are currently out of work on a disability leave (>2 weeks) and workers unable to read and communicate in English will not be eligible for the study.

Randomization will occur at the individual level after workers provide written informed consent and complete the baseline survey. As recruitment of workers will occur over a period of time, a computer-generated block randomization schedule (with equal sized groups for every 20 participants) will be maintained by the project statistician and used to assign individuals to the intervention or control conditions, and separate randomization schedules will be maintained for each of the five participating employers. Allocation will be concealed and the randomization schedules will be kept confidential and secured by the project statistician, so group assignment will not be anticipated by those involved in patient recruitment. The research associate responsible for collecting and compiling follow-up survey data and the researchers performing data analyses will be kept blind to group assignment. The nature of the psycho-educational group intervention will make it impossible to blind study participants or group facilitators from randomization results.

Participants randomized to the intervention arm will be assigned to participate in five 2-hour group workshop sessions (or ten 1-hour sessions, if necessary to meet scheduling constraints) led by a specially trained facilitator (licensed psychologist or clinical social worker) and provided over a span of approximately 2–3 months. The sessions will be offered at the work site but not during working hours. The content of the intervention was developed from qualitative studies [30] and from a review of existing self-management intervention elements [29], then piloted and revised according to participant feedback. The intervention incorporates standard elements of existing evidence-based pain and illness self-management efforts but tailors key messages and discussion elements to workplace problems most relevant to workers with chronic physical health conditions. The theoretical basis for the intervention (Figure 1) is Bandura’s Self-Efficacy component of Social Cognitive Theory, which suggests that “the self-assurance with which people approach and manage difficult tasks determines whether they make good or poor use of their capabilities” (p. 35) [34]. Each session is focused on different self-management strategies, with each session containing a mix of facilitator presentation, group discussion, case illustrations, role play, completion of in-session self-assessments and activities, and brief homework assignments. Participants are encouraged to support each other through communications outside of the scheduled meetings, though this interaction is voluntary. Approximately equal time is allocated to the topics of improving comfort, modifying work, communicating effectively, applying systematic problem-solving strategies, and dealing with negative thoughts and emotions (see Table 1).

Participants randomized to the control arm will receive no intervention during the 12-month period of study participation. However, after completing their 12-month follow-up, participants in the control group will be invited to attend a full-day Saturday workshop. The intent of the full-day workshop will be to provide the same self-management information as in the intervention arm, but on a delayed basis and in a more condensed format. The full-day workshop will be led by the same trained facilitators as the intervention group and participants will receive all of the same materials.

As part of each assessment, participants will be asked to report the frequency and type of medical and rehabilitation services used over the previous 6 months at both follow-up assessments.

A process evaluation will be conducted alongside the randomized controlled trial to gain insight into the feasibility of this intervention for more widespread dissemination and to detect any particular study challenges that might be identified from more qualitative assessments. The process evaluation will be based on the RE-AIM framework, which consists of five dimensions: Reach, Efficacy/effectiveness, Adoption, Implementation, and Maintenance [56]. The goals of the process evaluation will be to assess the reach of the program (at both the employer and participant level), to identify any dose–response relationships between levels of participation and outcomes, to evaluate adherence to the study protocol, to assess the satisfaction and experiences of workers and facilitators, and to identify facilitators and barriers for future implementation. Information will be collected by quantitative as well as qualitative methods. In the final follow-up survey, participants will be asked whether they might be willing to participate in an in-depth interview that would include the topics shown above. Since it is not feasible to interview all stakeholders, purposeful sampling will be applied based on relevant characteristics of the target population in combination with the answers to the questionnaire.

The estimated effect size for the primary outcome measure (Work Limitations Questionnaire) is based on the size of effect obtained in a prior intervention trial focusing on workers with depression [38]. In that study, the effect size for various WLQ subscales ranged from 0.51 to 0.87 standard deviation units, suggesting a medium effect size, f = 0.25. With a target recruitment of 300 volunteers (60 from each organization), an assumed attrition rate of 20 %, and an alpha level of 0.05, the statistical power to detect a medium effect size (f = .25) on the primary outcome measure is 0.96 without consideration of the nested design or the possible need for covariates when making pre-post group comparisons. However, given the possible complication of non-equivalent treatment sites and the inclusion of several covariates in the final group comparison [57, 58], a power estimate in the range of 0.80 – 0.85 would be a more conservative estimate based on various scenarios simulated in the PASS 11 Power Analysis & Sample Size Software [59].

After providing informed consent, participants will complete a baseline survey and then be assigned to the intervention or control arms using a block randomization schedule (blocked in groups of 20) maintained privately by the project statistician. Six and 12 months after initial recruitment, participants will be mailed a follow-up survey (or provided access to an on-line survey) to provide for re-assessment of primary and secondary outcome measures. Non-responders will be sent two additional reminders, and all participants will receive a payment of $50 for completing each of the 3 research survey assessments (baseline, 6 months, 12 months).

The primary analytic strategy will be to compare the intervention and control groups on changes in outcome measures at 6- and 12-month follow-up using a multilevel linear mixed model that will take into account the employer and the 3 repeated measurements, and also allow for missing data on either the 6- or 12-month assessment. The two primary outcome measures will be work limitations (total work productivity index from the Work Limitations Questionnaire) and work engagement (total score from the Utrecht Work Engagement Scale). Because of the nesting of participants within 5 different employers, a multilevel analysis will be necessary with employer treated as a fixed factor. This will help to account for any systematic differences in the working populations, job demands, and policies and practices within these organizations. We also anticipate the inclusion of at least 3 individual-level covariates (e.g., age, gender, and number of chronic conditions) in the principal analysis if there are relevant group differences at baseline.

A number of background variables will be assessed to check that the randomization has yielded equivalent comparison groups at baseline. Background variables will include: age, gender, level of education, income, number and type of chronic conditions, blue-collar versus white-collar jobs, job stress, hours of work per week, shift work, number of household dependents, and job and industry tenure.

One feasibility concern that was evident from our preliminary exchanges with employers was the need to maintain the privacy of participating workers. This problem was addressed in the study methodology in several ways. First, a worker will not be required to disclose the nature of his or her health problem in order to qualify for study inclusion or as part of group discussions (however, this information will be collected as part of the confidential research questionnaire). While this may lead to a more heterogeneous sampling, we felt it was more important to preserve the workers’ rights to safeguard personal health information. Second, the intervention itself will be scheduled during lunchtime hours or after work, when there will be no need to coordinate work absences with a regular supervisor. Third, the informed consent and study enrollment process will be conducted through the employers’ Employee Assistance Programs (EAPs) or equivalent institutions that are accustomed to dealing with sensitive employee information and where safeguards are already in place to protect the confidentiality of workers. With these methodological enhancements in place, we expect that workers will be able to volunteer without the risk of workplace stigma or embarrassment.

One methodological dilemma was the choice of an appropriate control group. In a recent review on standard patient self-management programs in 19 randomized controlled trials, the experimental condition showed improved outcomes over care as usual, education leaflets, or waiting-list control groups [28]. While a more rigorous “attention control” condition might provide the best guard against a Hawthorne or similar effect (e.g., the control group attending discussion meetings on another topic), this has not been the standard in studies of self-help interventions, as self-care interventions are not commonly perceived by participants as desirable or comforting in the same vein as massage therapy, supportive psychotherapy, or other hands-on or empathic treatments for pain. Thus, we believe that individuals randomized to the wait-list control arm will be unlikely to experience a high level of dejection and disappointment that would represent a serious bias in the measurement of outcomes at follow-up months later.

While our initial intent was to focus on workers with chronic pain only, we were swayed toward a broader enrollment of workers for the following reasons. First, there is existing evidence that self-management interventions are relevant and effective for a broad range of physical health conditions, not just chronic pain [19, 21, 22, 60]. Second, we observed no problems delivering the 10-hour intervention program to a pilot group of community volunteers with a high level of diagnostic heterogeneity. Third, employers expressed concern that specifying “chronic pain” in the description of the program might discourage workers from participating due to potential stigma and embarrassment. Though announcements and advertisements for the program will specify “chronic or recurrent physical symptoms”, there will be no effort to screen potential candidates based on illness or diagnosis.

Measures were chosen for the study in order to address potential concerns of both workers and their employers. For example, the WLQ provides an opportunity to assess intervention benefits in terms of improved work productivity, but the UWES is more focused on the level of psychological attachment workers feel for their jobs. Similarly, measures of sickness absence, turnover intent, etc. are more directed to employer concerns, while measures of self-efficacy and well-being may be of more importance to workers. The intervention itself also strikes a balance between the need for productivity and the need for worksite wellness. Like all SM interventions, the instructional and participatory elements are based on social-cognitive theory and are designed to boost perceptions of mastery and self-efficacy in the workplace.

Though we developed no definitive criteria for the inclusion of employers in the study, the researchers did consider the issue of worksite readiness. Given the nature of the intervention, it would seem ineffective to provide workers with information about workplace self-management without supportive employer policies and practices that would enable communication and problem solving. Thus, employers with a poor wellness culture or adversarial labor-management relations may not be appropriate to host the study. In actuality, benefits of the group intervention approach might be stronger if partnered with a matching organizational effort to improve practices (e.g., supervisor training, participatory ergonomic approaches, etc.) but this was beyond the scope of the current study. Future studies might adopt a more organizational framework, but privacy issues and organizational status quo represent significant hurdles.

One considerable strength of the study is the recruitment and participation of employees at the workplace. This should generate a more representative sampling of affected workers compared with other forms of recruitment (e.g., through medical clinics or patient lists) or when participation in group meetings requires off-site travel. Also, by conducting the study in actual workplaces, the study should provide useful information about feasibility of implementation as an employer-sponsored health program. The randomized, controlled design of the study also provides a strong basis for investigating effectiveness of the intervention.

This study will be relevant for workers with chronic physical health conditions and for all employers. However, this approach may be especially relevant in occupational settings with an aging workforce who face concerns of a large portion of their workforce experiencing chronic conditions, and where some normative level of leeway and decision-making autonomy can be afforded to workers with regard to work style and the organization and prioritization of work tasks. The group intervention strategy in this study may also depend on employer policies and procedures that support worker self-management efforts and have a strong health and wellness culture as a foundation. This study will give some insight into the effectiveness of self-management intervention strategies to reduce disability and improve worker well-being. Results of the study will become available in 2015.