Leisure time physical activity and sickness absenteeism; a prospective study

Abstract

Aim To establish whether workers with frequent leisure time physical activities are at higher or lower risk of sickness absence compared to inactive workers.

Method Self reported and company recorded sickness absence data were collected during 18 months of follow-up for 8902 workers. Frequency of leisure time physical activities was queried at baseline.

Results Overall, we found that workers active in their leisure time twice or more each week reported significantly less sickness absence compared to inactive workers (14.8 versus 19.5 days/year), mainly due to a decrease in sick leave because of musculoskeletal disorders.

Conclusion Demotivating sports participation by making workers liable for workdays lost due to sporting injuries might be counter-productive in decreasing absenteeism and its related costs. Promoting worker participation in sport might lead to reduced absenteeism.

Introduction

The balance between the health benefits and burden of leisure time sports in terms of injury risk and subsequent use of healthcare resources, workers absenteeism and commercial costs is important and has been discussed in national parliaments [1,2]. Absenteeism resulting from leisure time physical activities is often easily identified and may lead to the worker being ‘blamed’. In contrast, the positive health effects of leisure time physical activity are much more diffuse and not recognizable as such. Surprisingly, however, scientific data on the overall effect of leisure time physical activity on worker absenteeism are limited. Although some studies reported a beneficial effect of leisure time physical activity on long-term sickness absence [3] or absenteeism due to musculoskeletal disorders [4], none reported on the overall effects on sick leave. The aim of this study was to determine the relationship between leisure time physical activity and sickness absence.

Methods

For the current analysis, data from a prospective cohort study were used, which has been described elsewhere [5]. In short, in this study, workers from 45 companies were followed for 54 months from May 1998. Information on the frequency of leisure time physical activities [6] and sickness absenteeism information provided by study participants and their employers was used. Workers were categorized as being ‘inactive’, indicating leisure time physical activity once a week or never, or being ‘active’ two or more times a week. Two independent sources of sickness absence data were gathered. Firstly, the self-reported duration, frequency and reason for sickness absence between 8 and 12 months after baseline. The reason for sickness absence was coded by a trained coder using the Dutch version of the International Classification of Primary Care (ICPC) [7]. Secondly, employer sickness absence records during 1 year starting 6 months after the baseline measurement. Logistic regression was used to analyse differences in sickness absence risk, Poisson regression to analyse difference in duration of sickness absence (in calendar days), between active and inactive workers. Due to the complex logistics involved in sending out the questionnaires and gathering the company sickness absence data we could not exactly match the time windows of both data collection methods.

Results

A total of 12 140 workers were eligible for inclusion. We excluded workers who reported having a chronic condition at baseline (n = 2397), were pregnant or laid off during the observation period (n = 360) and who reported to be on sickness absence at baseline (n = 481). This left 8902 workers for analysis. Of the workers, 31% were categorized as inactive and 69% as active. Self-reported sickness absence data were available for 6837 (77%) and employer sickness absence data for 6375 workers (72%). Reasons for non-availability of sickness absence data were changed employer (n = 460) and inability of the employer to provide data (n = 2067).

The results for the self-reported duration, frequency and number of sickness absence days in total and per ICPC group and the company provided sickness absence records are shown in Table 1. Overall, for the self-reported data, active workers reported significantly less frequently sickness absence (OR: 0.87; 95% CI: 0.78–0.97), especially for musculoskeletal disorders of the spinal column (OR: 0.62; 95% CI: 0.45–0.84), compared to the inactive workers. Also the duration of sickness absence was, on average, 1 day shorter for active workers. This was mainly due to a shorter duration of diseases of the digestive tract. The overall number of self-reported sickness absence days was significantly lower (19%) in the active workers, with less absent days for most disease categories, except for psychological problems (no difference) and social problems (more absent days).

The number of sickness absence days as provided by the employer showed only a slightly lower sickness absence risk for the active workers. However, the mean duration of the sickness absence periods was 23% shorter in the active compared to the inactive workers (P < 0.001). Also, the annual number of sickness absence days per worker was lower in the active compared to the inactive workers (respectively, 14.8 and 19.5 days a year). Adjustment for age, gender and educational level, did not change the results for any of the reported relations (specific data not shown).

Discussion

In summary, we found that workers who reported being physically active more than once a week reported significantly less sickness absence. So far, this relationship has only been investigated elsewhere in detail in relation to musculoskeletal disorders [4], with findings similar to ours. For psychological problems, also an important cause of sickness absence, we found a small positive effect, although not significant. Exercise is an accepted method of treating patients with depression [8]. The positive effects are explained through the positive influence of physical exercise on both psychosocial and physiological health [9,10]. Only for sickness absence due to social problems was a small but significant effect found in favour of inactive workers. One could speculate that for workers with social problems, sport might be a way of release. Although the study is of a longitudinal design, some limitations remain. Possibly, workers' health problems simultaneously reduce the ability of the worker to participate in sports and lead to increased sick leave. Although, by excluding all workers reporting a chronic condition and workers at sick leave at baseline, we tried to minimize this bias. Also some differences were found between the self reported and company recorded sickness absence data, especially in the mean duration and total number of sick leave days. This is probably caused by the difference in observation periods (4 months versus 1 year) which excludes long-term sickness absence periods for the self-reported data. Calculating the costs due to lost productive workdays (using Dutch labour cost data [11]) being leisure time inactive was associated with a decreased productivity of €488 per worker per year.

We conclude that being physically active in leisure time is related to lower absenteeism rates. Leisure time physical activity should therefore not be discouraged by making workers liable for workdays lost work due to leisure time sporting injury. In our view, it is more important to keep people, and workers in particular, physically active.

Conflicts of interest

None declared.

References