Background/Purpose

Although physical activity is crucial for managing symptoms and comorbidities in rheumatoid arthritis (RA), the majority of the RA population does not reach recommended health-enhancing physical activity levels.  Longitudinal studies are scarce and more knowledge is needed about predictors of different patterns of physical activity over time in large RA samples. The aim was to, within a large well-defined sample of persons with rheumatoid arthritis, identify and describe groups demonstrating different trajectories of physical activity over two years and to identify baseline predictors for being in each specific group.

Methods

This longitudinal study used data from the Swedish Quality Registries and a questionnaire covering sociodemographic, disease-related, psychosocial and physical activity variables. Physical activity was assessed by the International Physical Activity Questionnaire as total weekly hours of vigorously intense activity, moderately intense activity and walking. A total of 2752 participants responding to at least two out of three questionnaires at baseline, 12 months and 24 months were aged 18-75 years, diagnosed with RA according to the ACR criteria and independent in daily living (Stanford Health Assessment Questionnaire ≥ 2.0). K-means cluster analysis was used to identify three trajectories of hours of weekly physical activity. Multinomial logistic regression was used to identify predictors of trajectory membership.  Multiple imputation was used to impute missing data in potential predictors, which included sociodemographic, disease-related, psychosocial and physical activity variables.

Results

We identified three patterns of physical activity: The “Stable High” group performed on average 25 hours of physical activity per week (n=272), the “Decreasing” group went from 22 to eight hours per week (n=564) and the “Stable Low” performed three hours of physical activity per week (n=1916). Predictors of being in the “Stable High” group versus the other two groups were male gender and already established health-enhancing physical activity at baseline. Increased age predicted being in the “Decreasing” group versus the other two groups. Predictors of being in the “Stable Low” group versus the other two groups were female gender, more activity limitation, lower self-efficacy for exercise and not having established health-enhancing physical activity at baseline. For detailed results, see Table 1.

Conclusion

The results indicate that physical activity levels are still low in the RA population. Different trajectories of physical activity over time exist and are possible to identify, with each trajectory being associated with a unique set of predictors. Modifiable predictors to improve physical activity include activity limitation and self-efficacy for exercise.