To conduct a longitudinal randomized controlled trial comparing a work place ergonomic intervention versus a control (i.e., provision of written educational materials) for persons with rheumatoid arthritis (RA) or osteoarthritis (OA) regarding employment, physical, and symptom outcomes.
Eighty-nine participants (mean age 50.54 years, 87% women, 38% with RA, 62% with OA) were randomized into the intervention or control group. Seventy-five participants completed baseline, 12-month, and 24-month evaluations. Outcome measures included Arthritis Impact Measurement Scales 2 (AIMS2) physical, symptom (i.e., pain), and role scores (i.e., impact of arthritis on employment); Job Satisfaction Survey (JSS); and Brief Symptom Inventory Global Severity Index. The study design was a 2-factor (treatment and time) with repeated measures on 1 factor (time) design used with baseline as a covariate for 12- and 24-month data.
Between-group analyses indicated differences at 24 months for the AIMS2 role score (P < 0.03), with the intervention group reporting less arthritis-related impact on their work. Within-group analyses indicated significant improvements for the intervention group in AIMS2 change scores for physical functioning and symptom variables at 12 months (P < 0.04 and P < 0.01, respectively) and 24 months (P < 0.01 and P < 0.01, respectively). Job satisfaction (JSS) decreased at 12 months for both the intervention (P < 0.01) and control groups (P < 0.01), and at 24 months for the control group (P < 0.01).
An ergonomic work place intervention (versus a control) is associated with decreased arthritis-related work difficulties over 2 years for individuals with OA and RA, as well as improvements in physical functioning and pain.
Arthritis has been identified as the leading cause of disability and the second most frequently occurring chronic condition in the US (1), as well as the third leading cause of work disability (2). Work limitations due to arthritis are reported by 30% of those with the disease, or 1 in 20 working-age adults (3). US residents with arthritis who are working report that, on average, they miss 1 day of work every 2 weeks due to arthritis-related impairments (4). A number of longitudinal studies have reported a work attrition rate of 26–33% following an initial diagnosis of rheumatoid arthritis (RA), and that 50% of individuals with RA are projected to be unable to work 10 years after the initial diagnosis (5–10).
Arthritis-related work limitations also impact personal and societal financial outcomes. In 2003, the overall costs to the economy related to arthritis and other rheumatic conditions were estimated to be $128 billion, or 1.2% of the US gross national product (11). With the anticipated increase in an aging population, arthritis and associated work limitations are projected to increase significantly in the future (3, 12).
Risk factors and job variables leading to work limitations or job cessation for persons with arthritis have been described (2). Several studies (13–15) identified increasing age, nonprofessional positions, less education, and number of days of work loss as negative indicators for maintaining employment. Numerous studies (16–20) indicated pain and severity of functional disability as limiting duration and type of employment, although other studies indicated specific symptoms such as pain, joint function, and laboratory findings as nonpredictors of work limitations and cessation (5). Other factors associated with vocational success for persons with arthritis include support of coworkers, supervisors, and family members; perceived high job status; positive attitudes toward the job; workstation modifications; satisfactory transportation; and self-employment (13, 15–16). More common and effective accommodations reported were the ability to alter work hours, flexible scheduling, the ability to control the pace of work tasks, participation in decision making, fewer work place limitations (e.g., sitting and standing long periods, lifting, reaching, and repetitive tasks), the ability to modify tasks and workstations, assistance from others, medications, and changing jobs (21, 22). Gignac (23) also identified anticipatory coping behaviors and problem-solving strategies as successful measures to sustain employment for persons with arthritis.
Although the adverse effect of arthritis on employment has been well documented, there is limited research on the efficacy of different interventions to promote positive employment outcomes. A review (24) of 6 uncontrolled studies of vocational rehabilitation (VR) programs for persons with rheumatic diseases concluded that there was a trend for a positive but limited effect of VR programs on the ability of persons with arthritis to maintain employment or return to work. Another review of 9 studies of job retention for persons with chronic diseases also reported limited effectiveness of programs focused on work accommodations and self-efficacy (25). Regardless of the benefit of specific VR programs for persons with arthritis, persons with arthritis infrequently use VR services (16, 26, 27).
Unfortunately, there have been few controlled studies of interventions to improve the employment outcomes of persons with arthritis. The majority of these intervention studies focused on workers at risk to develop work-related problems, workers with periods of being unable to work (28–30), or general interventions offered at training sites rather than specific interventions related to job-specific issues at their specific work place. Two interventions (31, 32) provided at individuals' existing work places indicated an increase in full-time work, a decrease in lost days, and an improvement in the use of self-management techniques and work modifications. Another intervention study (33) involved “job raising” programs with 10 weeks of general instruction regarding self-confidence and knowledge of “job accommodations.” However, the use of volunteers to manage this program was described as “problematic.” Another study (34) demonstrated that specific VR interventions were associated with 55% fewer job disruptions than a control group. Since individuals with arthritis are highly susceptible to developing musculoskeletal disorders (35), the application of ergonomic interventions for prevention of arthritis-related disability in the work place is warranted (36). Even though current evidence for the effectiveness of ergonomic work place interventions is inconsistent and reviews indicate that positive results may be attributed to design flaws (37–39), the Centers for Disease Control and Prevention (35) recommends ergonomics for the prevention of work-related musculoskeletal disorders for workers with arthritis.
The aim of this study was to evaluate the long-term (i.e., 12 and 24 months) effects of a work place ergonomic intervention for persons with RA or osteoarthritis (OA). It was hypothesized that more positive outcomes would occur in the intervention group in terms of employment (i.e., impact of arthritis on work performance and job satisfaction), physical functioning, pain, and psychological well-being.
Significance & Innovations
This article is significant in that an individualized, 5-hour, ergonomic intervention comprised of an assessment, work plan, and 1 telephone followup with an occupational therapist led to small and statistically significant improvements in arthritis-related difficulties in work performance, compared to a control group that received written information only.
The worksite-based ergonomic intervention evaluated in this study targeted individuals with osteoarthritis and rheumatoid arthritis who are still working, and contributes to a new model for state vocational rehabilitation services to consider, i.e., preventing work loss instead of treating work disability.
SUBJECTS AND METHODS
This repeated-measures randomized controlled trial evaluated a work place ergonomic intervention compared to a control group (i.e., provision of only written educational information).
Eighty-nine participants were recruited over a 3.5-year period throughout Missouri from regional arthritis centers, newspaper and radio advertisements, payroll inserts, employer health fairs, and rheumatology clinics. Potential participants were screened through telephone interviews to determine their interest and eligibility. Inclusion criteria were a diagnosis of RA or OA confirmed by a physician, competitive full-time or part-time employment, age between 18 and 61 years, and no other significant medical (e.g., traumatic brain injury, cancer) or psychiatric histories (e.g., schizophrenia, bipolar disorder). Approval was received from the University of Missouri Health Science Institutional Review Board and participants provided signed informed consents. Ninety-one percent of the participants were employed full time and 9% were employed part time (i.e., 20–32 hours/week). Given the limited number of participants, all full- and part-time participants were included in the analyses.
Figure 1 shows the Consort diagram indicating the differing number of participants in each group who completed measures at the baseline, 12-month, and 24-month data collection points. Table 1 lists the demographic characteristics, arthritis-specific variables, and employment characteristics of the sample. Financial data were obtained regarding hourly, weekly, monthly, or annual income, although all income data were expressed as annual income. The Occupational Information Network (O*NET) (40) was used to obtain information regarding type of employment.
Table 1. Baseline characteristics of the study groups*
Study sample (n = 89)
Intervention group (n = 48)
Control group (n = 41)
Values are the number (percentage) unless otherwise indicated. O*Net = Occupational Information Network.
Maximum range 51 years includes 1 participant with a diagnosis of juvenile rheumatoid arthritis.
Upon enrollment and completion of baseline data collection, participants were randomized to an intervention (n = 48) or control group (n = 41) using a table of random numbers. Data were collected via mail at baseline and postenrollment following intervention at 3, 6, 12, and 24 months. Screening, enrollment, group assignment, and data collection were completed by a grant project director. For the purposes of this article, only baseline, 12-month, and 24-month data are reported.
Treatment group (work place ergonomic intervention).
The intervention group received individual work place assessments, resulting in a work plan to improve arthritis-related vocational difficulties. Two work place–based sessions of ∼2.5 hours were provided by an occupational therapist with a background in arthritis care and ergonomics. Participants in the intervention group were responsible for obtaining employer approval for the onsite visits. The initial evaluation included an interview determining specific job tasks and arthritis-related difficulties, as well as observation and photographing of job task performance. Job routines and essential job functions were identified, the impact of arthritis on job performance and functioning was assessed, and an ergonomic assessment of the work environment was completed. An individual intervention plan was then developed and reviewed at the second session. At the participant's discretion, the supervisor participated in this session. The plan provided a summary of the job assessment, photographs identifying areas of concern, and specific recommendations to better manage arthritis-related impairments in the work setting. Typical recommendations consisted of methods to establish routines and workflow, including body mechanics and exercises, suggested workstation and equipment modifications or changes, and person-specific recommendations (e.g., orthoses, lifestyle changes, etc.). There was a followup phone call within a month to determine if modifications to the work plan were desired and, if necessary, a session for assisting with purchase or installation of modifications or equipment.
Additionally, a resource manual was provided with guides for self-management of arthritis and possible ergonomic interventions in the work setting. The manual was comprised of Arthritis Foundation education materials and information about arthritis, including local resources related to work accommodations, work-related resources and regulations, job-related risks, and job-specific guides and modifications.
Control group (written educational materials).
Control group participants were contacted via telephone by the occupational therapist and provided the same resource manual that was provided for the intervention group. All of the participants were encouraged to investigate state VR services as a resource, although only 3 qualified for services, given they were employed at the time during which they participated in the study (and 1 was receiving VR services at the time of enrollment).
All outcome measures were collected using self-administered questionnaires that assessed outcome in terms of employment satisfaction (Job Satisfaction Survey [JSS] [41–43]), impact of arthritis on work performance (Arthritis Impact Measurement Scales [AIMS2] role score [44, 45]), physical functioning (AIMS2 physical functioning score [44, 45]), pain (AIMS2 symptom score [44, 45]), and psychological well-being (Brief Symptom Inventory [BSI] ).
The JSS assessed attitudes related to job satisfaction in 9 areas: pay, promotion, supervision, fringe benefits, contingent rewards, operating procedures, coworkers, nature of work, and communication (41–43). The 36 questions are structured according to a 1–6 Likert scale ranking (ranging from disagree very much to agree very much), and a total score is computed from all items. Internal consistency (range 0.60–0.91), test–retest reliability (0.71 for the total scale), and adequate discriminant and convergent validities have been reported (43). Higher scores represent greater job satisfaction (range 36–216).
AIMS2 (44, 45) measured the health status of persons with arthritis, satisfaction with their functioning, perceptions of the impact of arthritis on their current health and functioning, and priorities for areas of improvement in health. It is composed of 78 items with 12 subscales of 4–6 items per section and 10 demographic items. Items rank performance, attitude, or feelings using a 5-point Guttman scale. Subscales are combined into a 5-component health status model, including measures of physical functioning, affect, pain symptoms, social interaction, and impact of arthritis on employment.
The AIMS2 physical component score assessed the impact of arthritis in 6 areas of physical functioning: mobility, walking/bending, hand/finger functioning, arm functioning, self-care, and ability to engage in household tasks. The physical component score is the average of these 6 scores. The AIMS2 symptom component score, composed of 5 items, is a measure of general pain associated with arthritis (e.g., severity, frequency, duration, and impact on sleep). The AIMS2 role component score measured the perceived degree of work impairment caused by arthritis and is based on 4 items that assess the impact of arthritis on the ability to work (e.g., frequency of days of work affected, duration of time able to work, impact on accuracy of work, and frequency of need to change the manner in which work was completed).
Lower scores represented better health status. AIMS2 reliability and validity have been well documented (44, 45). Internal consistency via Cronbach's alpha for the 12 sections ranged from 0.72–0.91 for participants with RA and from 0.74–0.96 for participants with OA. Test–retest intraclass correlation coefficients ranged from 0.78–0.94. The instrument has also been shown to demonstrate sensitivity to change, especially in terms of physical functioning and pain symptoms (44).
The BSI (46) was administered to assess the psychological well-being of persons with OA and RA, given the psychological difficulties that are often associated with arthritis (47) and their potential impact on employment. The BSI evaluated 9 symptom areas of psychological distress (e.g., depression, anxiety, etc.). This study used only the Global Severity Index (GSI), a summary score that is used as a general measure of psychological well-being. Item responses require a 5-point Likert scale ranking of the amount of distress for each item (where 0 = not at all and 4 = extremely). The raw scores are the summation of item scores divided by the number of items in each of the specific areas. Adequate internal consistency (range 0.71–0.85), test–retest reliability (range 0.68–0.91), and convergent, discriminant, and construct validity have been reported (46). Lower GSI scores indicate less severity of symptoms. The published mean ± SD score for the GSI for a nonclinical sample is 0.31 ± 0.30.
A 2-factor (treatment and time) with repeated measures on 1 factor (time) design used the baseline data points as a covariate for the data analysis for the 12- and 24-month followup. Measures of other data points are not discussed here. Summary statistics such as means, SDs, medians, and frequencies were computed for each variable. To compare differences between the groups for continuous variables at baseline, Wilcoxon's rank sum test was used. To compare differences between groups for categorical variables, a chi-square test was employed.
The nonparametric analysis of covariance was performed for between-group comparison of 12-month and 24-month data due to lack of normality for the data. The baseline values for the AIMS2 physical and symptom components were used as covariates in the analysis. The analysis was conducted by first regressing the value at each post–time point on baseline value and 2 extra covariates, and then comparing the residual from the regression between the 2 groups using Wilcoxon's 2-sample test. The results were summarized for each post–time point. Data were analyzed using SAS, version 9.2. For within-group analysis, the signed rank test was used for comparison between time points in each group.
For the statistical analysis, the AIMS2 scales were recoded and normalized for scores to be reported in a 0–10 range (where lower scores represent good health status). The AIMS2 user's guide (45) emphasizes that there are not established criteria for assessing change and suggests determining a percentage of change desired or evaluating change through use of effect sizes.
Given the fact that the focus of the study was on improving employment outcomes, the AIMS2 role score was used to determine the sample size required in order to draw meaningful conclusions. Previous research using the AIMS2 role scale revealed that a nonmedical intervention yielded an effect size of 0.62 (48). Therefore, to detect an effect size of 0.62, a sample size of 42 participants per group was required. The alpha level was set at 0.05 with a power of 0.80.
Recruitment and retention.
Eighty-nine individuals were initially enrolled and completed baseline data. Eighty participants completed 12-month data and 75 participants completed 24-month data. Wilcoxon's rank sum test indicated that there were no significant differences in the outcome variables between the group of 75 who completed both the 12- and 24-month data sets and the 14 noncompleters of 24-month data. Nonetheless, given that only 75 of the original 89 participants completed the 24-month data, and that the power analysis suggested the need for 42 participants per group, the data need to be interpreted with caution.
Comparison of outcome variables between the intervention and control groups at baseline indicated significant differences between groups in terms of arthritis-related physical symptoms (AIMS2 physical score, P < 0.03) and pain (AIMS2 symptom score, P < 0.05), with the control group reporting less physical symptoms and pain (Table 2).
At 12 months there were no significant differences between the intervention and control groups on scores for the JSS, GSI, and AIMS2 physical, symptom, and role components. At 24 months there was a significant difference between the groups for the AIMS2 role score (P < 0.03), with the intervention group reporting less arthritis-related impact on vocational functioning than the control group. There were no other differences between the groups at 24 months for the 4 other outcome measures.
To measure the treatment effects within each group, change scores were calculated between baseline and 12- and 24-month data collection points (Table 3). For the intervention group, the results showed that physical symptoms (i.e., AIMS2 physical score) significantly improved at both 12- (P < 0.04) and 24-month followup (P < 0.01), pain (i.e., AIMS2 symptom score) also improved at both 12- (P < 0.01) and 24-month followup (P < 0.01), and job satisfaction (i.e., JSS) significantly declined at 12 months (P < 0.01), but not at 24 months (P < 0.08).
The sample size for each variable for each group at each time point is the same as those listed in Table 2. JSS = Job Satisfaction Survey; AIMS2 = Arthritis Impact Measurement Scales 2; BSI = Brief Symptom Inventory; GSI = Global Severity Index.
For the control group, job satisfaction declined at both 12- (P < 0.01) and 24-month followup (P < 0.01). There were no significant changes within the control group on any of the other outcome measures.
This study indicated that the work place ergonomic intervention was effective in decreasing arthritis-related difficulties in the work place over a 2-year period when compared to the control group. Within-group comparisons also indicated that the work place intervention group improved in physical functioning and pain over 24 months, although the control group did not. These findings suggest that work place ergonomic interventions offered in the work place are effective in improving the employment, physical functioning, and pain outcomes of persons with OA and RA who are employed and only beginning to develop arthritis-related difficulties.
Between-group comparisons indicated that the work place intervention led to significant improvements in only the AIMS2 role component score at 24 months (but not job satisfaction, physical functioning, pain, or psychological well-being). In general, the intervention was associated with improvements in a person's lost or abbreviated work days, perceived quality of work performance, and necessity of altering work secondary to arthritis-related problems. Table 2 indicates that the AIMS2 role score showed general stability from baseline to 12 months with improvement from 12 to 24 months for the intervention group (i.e., from 1.92 to 1.89 to 1.49), whereas for the control group it showed general stability from baseline to 12 months with a decline from 12 to 24 months (i.e., from 1.73 to 1.65 to 2.16). These results show promise for further ergonomic interventions to determine their benefit in extending work life (i.e., beyond 2 years).
The results were also encouraging in that within-group analyses indicated that the work place intervention was associated with statistically significant improvements in both physical functioning and pain at 12 and 24 months. These findings for the intervention group are consistent with previous studies that noted that when ergonomic changes were introduced workers were 2.6 times less likely to become work disabled (13), and that individuals with ankylosing spondylitis were 2.5 times less likely to become work disabled when workstations were adjusted (49).
Of practical importance, the results further suggest that work place ergonomic interventions should be provided to persons with arthritis while they are still working, in contrast to the current policies of state VR programs that typically provide services only after individuals are no longer able to work. The financial feasibility of delivering individualized work place assessments is not immediately apparent, although such a simple intervention (i.e., a 2.5-hour assessment, a 2.5-hour intervention, and a followup phone call by an occupational therapist) is likely to be very cost effective (i.e., disability payments, decreased work productivity) compared to traditional VR models (i.e., provision of services after disabilities develop).
Unexpectedly, the job satisfaction of both the intervention and control groups declined after 12 months, as well as at 24 months for the control group. The reason for the decline is difficult to determine, but it is noted that all JSS scores for both groups are within the normal limits according to the test manual (i.e., the generally satisfied range: 144–216) (41–43).
The results of the study are limited because the sample included primarily middle-aged women with high levels of education who were employed mainly in administrative positions. Specifically, 61% of the participants' jobs were considered sedentary, low physical demand, and in office/management/teaching settings, limiting the study's generalizability. Further similar research is recommended for persons in more traditional manual labor positions that involve more physically demanding tasks.
Another limitation relates to the relatively mild degree of work place–related disability reported by the sample (i.e., initial mean AIMS2 role scores of 1.92 of 10 and 1.73 of 10 for the intervention and control groups, respectively). Such low scores suggest a decreased chance to detect meaningful improvements. As a result, although the changes in AIMS2 role scores between the groups were statistically significant, the clinical significance of this finding is uncertain.
The results are also limited, given that the sample consisted of individuals with a median duration of arthritis of only 5 years, that the participants self-selected for the study, which suggests that they had a belief that employers would be supportive of such efforts (21), and that the control group had significantly better functioning at baseline.
Future research may benefit from extending the time during which the benefits of the work place intervention can be evaluated (e.g., 5–10 years) and the use of objective measures of vocational outcomes that provide more specific information about vocational functioning than the AIMS2 (e.g., number of hours worked, measures of productivity, etc.). Future research will benefit from determining if the physical improvements noted in this study are related to the ergonomic intervention or other rehabilitation therapies (e.g., occupational therapy, physical therapy). Collaboration with employers and state VR agencies is also suggested.
All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be published. Dr. Johnstone had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Study conception and design. Baldwin, Johnstone, Hewett, Sharp.
Acquisition of data. Baldwin, Johnstone.
Analysis and interpretation of data. Baldwin, Johnstone, Ge, Hewett, Smith, Sharp.