To investigate the relationship between leisure-time physical activity and work-related repetitive strain injury (RSI), adjusted for sociodemographic characteristics and work-related physical and stress factors.
To investigate the relationship between leisure-time physical activity and work-related repetitive strain injury (RSI), adjusted for sociodemographic characteristics and work-related physical and stress factors.
The data source was the 2003 Canadian Community Health Survey, a national cross-sectional survey of 134,072 respondents. The analysis was limited to a sample of the survey population reporting full-time work during the past 12 months (n = 58,622). The outcome of interest was work-related RSI of the upper body. Multiple logistic regression was used to examine the relationship between the outcome and leisure-time physical activity level, adjusted for sociodemographic, health, and occupational characteristics. The potential effect of leisure-time physical activity with a high upper-body load was investigated in a secondary analysis.
The prevalence of upper-body work-related RSI was 5.9% in the Canadian population in 2003. An active lifestyle during leisure time was associated with a lower prevalence of work-related upper-body RSI (odds ratio 0.84, 99% confidence interval 0.75–0.95), after adjustment for work physical demands and other covariates. Female sex, obesity, smoking, age, work-related stress, and work physical demands were associated with RSI. In the secondary analysis, we did not find that participating in leisure-time activities with a high upper-body load was a risk factor for RSI.
Our study results indicate that being physically active during leisure time is associated with a decreased risk of upper-body occupational RSI, adding another potential health benefit to participation in leisure-time physical activity.
In 2002, Health Canada estimated the annual cost of musculoskeletal disease at more than $16 billion for medical care and lost wages, second only to cardiovascular disease (1). In 2001, ∼2.3 million Canadians, or ∼10% of the population ages 20 years or older reported a repetitive strain injury (RSI) (2), a type of musculoskeletal injury due to accumulation of microtrauma caused by repeated soft tissue overload. Most RSIs are work related, occur in the upper body, and affect individuals in their 30s and 40s, the prime working years. A recent review found that one-third of workers' compensation costs in private industry are due to RSI (3). In Canada, the prevalence of RSI has increased steadily over the past 10 years, with an ∼2% increase (500,000 additional cases) reported in 2001 compared with 1996 (2).
Sedentary lifestyles and occupations are common and are characterized by increased time in the seated position. This is in part due to an industrial shift from a resource-based economy to a service-based economy, with many more office, computer, and technology-based jobs. Lack of physical activity and increased sitting time may lead to muscle weakness and tightness in the upper body, which predisposes to RSI when occupational demands are repetitive or involve high loads. An active lifestyle may help to mitigate these muscle changes and protect against RSI. It is widely accepted that physical activity during leisure time has numerous health benefits and is associated with decreased risks of coronary heart disease, stroke, high blood pressure, cancer, and many other diseases (4), but it is not known if leisure-time physical activity decreases the risk of work-related RSI.
Many studies of RSI have been directed at specific jobs, have focused on either men or women, or have concentrated on the most severe cases such as carpal tunnel syndrome (5–8). Not many studies have been conducted on a population basis, and there are few, if any, reports of the effect of leisure-time physical activity on upper-body work-related RSI. Using cross-sectional data from the 2003 Canadian Community Health Survey (CCHS), a population-based national survey that provides detailed information on the type and amount of leisure-time physical activity, the present study examined the prevalence of RSIs and risk factors among Canadians working full time. The primary purpose was to investigate the potential benefits of leisure-time physical activity on the prevalence of work-related upper-body RSI, as part of a multifactor model of injury that adjusted for sociodemographic, physical work demands, and psychosocial (work stress) covariates and confounders. Because some leisure-time physical activities (e.g., tennis) place a high load on the upper body and have been associated with upper-body RSI (9), a secondary analysis was performed to adjust for the effect of these specific types of leisure-time physical activities.
Data were drawn from Statistics Canada's 2003 CCHS, a cross-sectional survey that collects information related to the health status, health care utilization, and health determinants of Canadians approximately every 2 years. The CCHS was designed with a very large sample size (n = 134,072) to provide reliable estimates at the regional level (10). Data collection for cycle 2.1 began in January 2003 and continued for 11 months. The CCHS targets persons ages ≥12 years who are living in private dwellings in the 10 provinces and the 3 territories. Persons living on Indian reserves or crown lands, clientele of institutions, full-time members of the Canadian armed forces, and residents of certain remote regions are excluded from this survey. The CCHS represents ∼98% of the Canadian population ages ≥12. For the 2003 survey, there was an overall response rate of 80.7%. Details of the survey methodology have been published elsewhere (10). From the CCHS survey population of 134,072, a study sample of 58,622 full-time workers (over the previous year) was drawn for this study, including full-time workers ages 15–74.
The outcome of interest was the answer to the question, “In the past 12 months did you have any injuries due to repetitive strain which were serious enough to limit normal activities?” The outcome was further limited to RSIs in the upper body (neck, shoulder/upper arm, elbow/lower arm, wrist/hand) and those associated with work-related activities.
Leisure-time physical activity on the CCHS survey is measured using the Physical Activity Index (11), a standardized and validated index used in many health surveys (12–14). The index is derived from the average daily energy expended during leisure-time activities reported by the respondent in the past 3 months. It is calculated using the frequency and duration per session of activity and the established metabolic equivalent (MET) value of the activity (15). Because the primary focus of this study was to investigate active versus nonactive lifestyles, leisure-time physical activity was converted to a binary variable (active or nonactive). Based on common definitions adapted by the CCHS (10) and recommended by the Canadian Lifestyle and Fitness Research Institute (11), an active lifestyle was defined as accumulating an average of 3 METs per day from leisure-time physical activity.
Reports of the type of physical activities that respondents participated in during the past 3 months were grouped into 2 categories (high or low) based on frequent repetition or high load to the shoulder, elbow, and/or wrist/hand. Clinical opinion together with a review of the literature on sports biomechanics (9, 16) were used to categorize the activities by upper-body load as follows: high upper-body load (tennis, baseball, weight training, gardening, fishing, golf, bowling, hockey, volleyball, swimming, home exercise, skiing, cycling) and low upper-body load (walking, dancing, exercise class/aerobics, ice skating, rollerblading, jogging, soccer).
Individual characteristics included age (<30 years, 30–49 years, and ≥50 years), sex, body mass index (<30 kg/m2 and ≥30 kg/m2), usual daily or work habits during the past 3 months (usually sit, stand, or walk a lot; usually lift or carry light loads; do heavy work; or carry very heavy loads), marital status, education status (less than secondary school, secondary school graduate, at least some postsecondary education, or postsecondary graduate), and smoking status (daily smoker or not a daily smoker). Self-perceived work stress in the past 12 months was reported as high (most days extremely or quite a bit stressful) or low (most days a bit stressful, not very stressful, or not at all stressful).
The prevalences of RSI and leisure-time physical activity for the entire Canadian population and study sample of full-time workers were calculated and described by demographic, occupational, and socioeconomic characteristics. Crude odds ratios (ORs) were calculated for the relationships between RSI and leisure-time physical activity variables, as well as for other survey covariates. Potential collinearity and interaction between covariates were examined on a bivariate basis. Results from this analysis guided the selection of variables for regression modeling. Multiple logistic regression was used to examine if leisure-time physical activity was associated with a lower risk of work-related RSI, controlling for sex, age, level of work demands, smoking status, obesity, work stress, marital status, and education. A secondary analysis to investigate the effect of the type of leisure-time activity was performed whereby leisure-time physical activity was defined as high upper-body demands verses low upper-body demands. This was done to examine the potential association of upper-body work-related RSI with leisure-time physical activities that place repetitive or high load on the upper body.
For statistical analysis, the frequency weights provided by Statistics Canada for CCHS 2.1 were used to generate population point estimates that adjusted for the differing probabilities that individuals were selected in the sampling strategy. To estimate the precision around the survey estimates, a model-based estimation of variance was used (17, 18). Frequency weights were rescaled to produce a standardized probability weight with an average value of 1 and a sum equal to the sample size. This standardized weight variable was calculated by dividing the survey frequency weight for each individual by the mean weight value for the analysis data set. The probability weights were used in the regression analyses for estimation of standard errors and coefficients of variation for the study sample. This method takes into account the unequal probabilities of selection in calculating variances. If the assumed model is true for the observed units, then it can be shown that the model-based variance should be close to design-based variance for large samples and large populations (18). However, because this method does not take into account the stratification and clustering of the sample's design, the variance estimates may be an underestimate (10, 18). Because this study was based on a very large sample size even after re-basing the weight to the sample size, a stringent cut point of P = 0.01 was used for univariable and multivariable analysis. Statistical analyses were performed using SPSS software, version 13.0 (19).
The prevalence of RSI in the overall CCHS survey population from all causes in the past 12 months was ∼2.8 million Canadians, or 10.9%. The prevalence of RSIs that were due to work was 1.3 million, approximately half of all RSIs. Upper-body injuries accounted for 63.3% of all RSIs in full-time workers; the remaining 36% involved the spine and lower-body region. Of upper-body RSIs, the most common areas affected were the wrist/hand (39%), shoulder (29%), and elbow (26%).
Crude ORs provided evidence that female sex, physical work demands, work-related stress, obesity, daily smoking, ages 30–49 years, and being previously married (divorced/separated/widowed) were significantly associated with upper-body work-related RSI (Table 1). Being physically active was protective against RSI (crude OR 0.76, 99% confidence interval [99% CI] 0.68–0.85), as was being over 50 years of age. There was no association between upper-body work-related RSI and high upper-body load from leisure-time physical activities. Adjusted ORs obtained from multiple logistic regression are presented in Table 1. The strength of association for RSI and being physically active was lower after adjustment for covariates (adjusted OR 0.84, 99% CI 0.75–0.95) but remained significant. Other ORs were similar between the univariable and multivariable models. Education level remained nonsignificant, but was retained in the final model to provide estimates adjusted for a measure of socioeconomic status. There was no observed association or change in model effects when high upper-body load from leisure-time physical activity was added to the model during our secondary analysis.
|Prevalence (%)||Unadjusted OR (99% CI)||Adjusted OR (99% CI)†|
|Female||42.3||1.50 (1.37–1.64)‡||1.68 (1.52–1.85)‡|
|30–49||62.9||1.37 (1.15–1.63)‡||1.27 (1.03–1.57)‡|
|≥50||12.3||0.74 (0.64–0.86)‡||0.74 (0.64–0.88)‡|
|Leisure activity level|
|Active||25.1||0.76 (0.68–0.85)‡||0.84 (0.75–0.95)‡|
|Usual work demands|
|Lift heavy loads||26.5||2.33 (2.04–2.67)‡||2.80 (2.42–3.24)‡|
|Lift light loads||23.0||1.90 (1.65–2.19)‡||2.31 (1.99–2.70)‡|
|Sitting||10.7||1.60 (1.39–1.85)‡||1.70 (1.47–1.98)‡|
|High||32.6||1.53 (1.40–1.68)‡||1.44 (1.30–1.58)‡|
|Yes||21.5||1.41 (1.27–1.56)‡||1.29 (1.15–1.44)‡|
|Yes||15.8||1.27 (1.13–1.43)‡||1.25 (1.11–1.42)‡|
The results from this study confirm that RSI is a prevalent condition in Canada, affecting 10.9% of the population (2.8 million individuals). The results also indicate that the prevalence of RSI continues to increase in the Canadian population (up from 8% in 1997 and 10% in 2001 ) and that work-related RSI is a major contributor to the burden of disease (approximately half of the prevalence in this study sample and in previous surveys ). The results of this study confirm relationships between RSI and female sex (20, 21), obesity (2, 21), smoking (2, 21), age (30–40 years) (2, 21), and self-perceived work stress (22, 23).
Our study supports a large body of literature that links physical demands at work such as repetition and awkward postures with an increased risk of RSI (24). Our study was not designed to investigate the specific effects of work-related physical demands on the risk of RSI, but we were able to adjust for physical demands at work using a more global measure based on self-report of 4 categories of usual activity or work activity during the past 3 months. We interpreted survey responses as work physical demands because the study population was limited to active full-time workers. Although there can be challenges with self-reported physical demands, it is important to note that we were able to adjust for work physical demands while investigating the effect of leisure-time physical activity in a multivariable model. It is possible that work demands during the past 3 months are not reflective of work experiences during the past 12 months or beyond (cumulative work history), but we believe that workers who usually work in high-demand jobs would tend to do so consistently (e.g., construction, warehousing) and might eventually rotate out of these positions to less physically demanding jobs with more seniority. As a result, there may be some misclassification within the past reporting year, but this would result in an underestimation of the effect.
An active lifestyle in leisure time was associated with a lower prevalence of work-related upper-body RSI. This relationship persisted after adjustment for work physical demands, a strong predictor of RSI, and other covariates. This finding provides evidence for a hypothesis that an active lifestyle outside of work may protect against work-related RSI, adding another potential health benefit to leisure-time physical-activity participation. This is consistent with a population cohort study of new cases of RSI using 1994–1995 to 2000–2001 survey data (20) that reported a protective effect of leisure-time activity on the risk of work-related RSI (OR 0.88), although the confidence interval included 1 (95% CI 0.30–2.58). Leisure-time physical activity may play an important role in facilitating necessary mechanical and metabolic processes for the musculoskeletal system that counter the repetitive or sedentary effects of many jobs, thereby improving musculoskeletal health. Leisure-time physical activity may improve musculoskeletal health in sedentary (e.g., sitting) workers by offsetting muscle weakness and tightness in the upper body. For physically active workers (e.g., repetitive lifting), leisure-time physical activity may provide better balance of movement and muscle activity than the repetitive loads often encountered in the work place. Workers who predominantly stand or walk at work (variety of movement without lifting) had the lowest levels of work-related RSI of the 4 physical work demand categories.
Previous studies that have investigated RSI (all causes) and leisure-time physical activity have found that an active lifestyle is a risk factor for RSI (2, 25). This may be largely accounted for by the inclusion of RSI caused by sport and leisure activity (as well as work-related RSI) in the analysis. These other causes of RSI were purposely excluded in our analysis to look specifically at work-related RSI.
One potential concern with higher levels of leisure-time physical activity is that it may place increased load on the musculoskeletal system, potentially contributing to RSI. This potential association was investigated by analyzing whether leisure-time physical activities that placed frequent high loads on the upper body (e.g., baseball, tennis) were associated with RSI. We did not find a relationship between RSI and physical activities with high upper-body loads, either in the bivariate or adjusted analysis. This indicates that choice of leisure-time physical activity (at the levels and categorization investigated in this study) is not a potential risk factor for occupational RSI. Our criterion for frequency of a particular type of upper-body activity was participating at least 12 times over the previous 3-month period (1 time per week on average). Although our analysis indicates that this is a safe frequency of activity with regard to RSI, we may learn more by investigating a higher frequency of activities with a higher upper-body demand in future studies.
Although our study provides evidence of an association between leisure-time physical activity and reduced risk of upper-body work-related RSI, the cross-sectional design precludes the determination of a cause and effect relationship. An alternative hypothesis of reverse causality must be considered: individuals reporting upper-body RSI may be less likely to be active outside of work. This may be particularly true for activities with high upper-body demands and may explain why there was no observed relationship between high upper-body activities and risk of RSI. Considering prevalent cases of RSI, individuals with any prior symptoms may purposely avoid these types of activities. However, because most upper-body RSIs involve the hand and wrist and leisure-time physical activities include any type of activity (including walking), RSI would not necessarily limit all leisure-time physical activity and participation in some form of activity may be protective for musculoskeletal health.
The prevalence of an RSI in the last 12 months was based on self-reported information. It was not known if the RSIs had actually been diagnosed by a health care professional. However, it has been consistently shown that self-report of musculoskeletal injuries is valid in population surveys and epidemiologic research, including occupational musculoskeletal injuries (26) and chronic musculoskeletal disease such as arthritis (27–30). Some research has demonstrated that when individuals become more aware of RSI, they are more likely to report it (31, 32). Therefore, the CCHS may overestimate the prevalence of RSI, compared with studies that use more stringent definitions or clinical measures. Furthermore, RSI is a broad diagnostic category and is referred to in the literature by various other names including work-related musculoskeletal disorder, cumulative strain disorder, and nonspecific work-related upper-limb disorder (33, 34). Even if the exact diagnosis is incorrect, the presence of work-related upper-body symptoms still provides evidence of musculoskeletal injury that affects the upper limbs and occurs in occupational settings (34).
Leisure-time physical activity levels during the past 3 months were also self-reported, with potential for recall and social desirability bias, which may result in misclassification. However, it has been repeatedly shown that self-reports in questionnaires on physical activity are both practical and valid in epidemiologic studies (35–37). Furthermore, there is no reason to believe that misclassification would be differential, i.e., related to the study outcome. The occupational data available in this cycle of the CCHS did not include specific job titles or duties. This limitation was overcome in some measure by information on the daily work physical demands (sit, stand, lift light loads, lift heavy loads). Although many studies have used occupational title in analysis, use of physical demands may be a more accurate indicator of mechanical load and occupational risk on the musculoskeletal system. The upper-body loads during leisure-time physical activity are variable in a number of sports and activities (e.g., weight training). This may have resulted in misclassification that underestimated potential associations.
The data from this study are cross-sectional, making direction of any association difficult to ascertain. However, this study was based on a large comprehensive population health sample representative of the Canadian population. Previous studies of work-related RSI have tended to focus on specific occupational or industrial groups. Information on a wide and diverse range of recreational parameters as well as on demographic, work, socioeconomic, and health variables was available. This allowed for an in-depth analysis and for the inclusion of known confounders in examining the primary research question. The estimates of the standard errors and confidence intervals may not fully take into account the sophisticated survey sampling design and may be underestimates of the variance around the estimates, resulting in statistically significant conclusions. Future research should include a prospective investigation of the relationship between physical activities and RSI, including a more detailed investigation of the role of upper-body load during physical activities.
Our study results indicate that being physically active in leisure time may protect against work-related RSI. Work-related RSI is a common problem that has significant costs in terms of lost wages and productivity, medical care, and quality of life. It is a growing problem in Canada that may be associated with an increase in sedentary occupations and lifestyles. Work-related RSI has significant public health implications and is important to workers and workers' compensation boards. It is important to identify lifestyle factors that may protect against RSI in order to inform employers, compensation boards, and the working population. Identification of safe physical activities may reduce RSI and provide multiple other health benefits.
Mr. Ratzlaff 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 design. Ratzlaff, Koehoorn.
Acquisition of data. Ratzlaff, Koehoorn.
Analysis and interpretation of data. Ratzlaff, Koehoorn.
Manuscript preparation. Ratzlaff, Gillies, Koehoorn.
Statistical analysis. Ratzlaff, Koehoorn.
The authors thank Dr. Bob Hogg for guidance with data acquisition and analysis and providing comments on the draft versions of this article.