Going Mobile: Measuring Physical Activity in Arthritis Research

physexerciseIn this article I’m going to give a brief and mostly non-technical review of three recent research trials that seem to confirm that measured physical activity is now being recognised to be an important outcome measure in patients with arthritis and Ankylosing Spondylitis. The wearable technology to make these measurements possible is now available and their scientific validation has been established, particularly in the field of sports science. In sports science parlance ‘physical activity’ is defined precisely in terms of energy expenditure so that activity levels can be compared across the population. This outcome relates more to the duration and intensity of exercise carried out rather than the specifics of movement that might be impaired by arthritis. The focus here is on the impact of exercise on fitness and health rather than on flexibility and mobility.

Hernandez-Hernandez studied a group of 50 RA patients compared to 50 matched healthy controls using a questionnaire of physical activity called the IPAQ and a tri-axial RT6 accelerometer. The authors stated that they wished to study the correlation between disease activity and physical activity. Those in the study wore the accelerometer device continuously for 5 days and their activity was broken down into light, moderate and vigorous activity using standard cut points. ‘MVPA’ was defined as the amount of time spent in moderate or vigorous physical activity. Disease activity was monitored using the DAS28 and disability assessed using the HAQ. They also recorded fatigue using the FACIT questionnaire and quality of life using the standard SF-36 questionnaire. Results: Although there was a weak correlation with the IPAQ, the self-reported questionnaire did not closely match the accelerometer data. The authors found that accelerometry was able to capture physical activity accurately & that recorded levels of moderate and vigorous physical activity (MVPA) in RA patients was significantly lower than that of healthy controls. The MVPA correlated with the HAQ but not with disease activity. Thirty of the RA patients had a second assessment carried out 6 months after the initial assessment. In this subset they found that a significant change in DAS28 (>1.2) correlated negatively with a change in MVPA.

  • This is possibly the first paper to show a convincing negative correlation between DAS28 and monitored physical activity in the patient’s natural setting.
  • This is also one of the first papers to use the IPAQ in RA patients and to demonstrate a difference in physical activity between RA patients and controls. At the moment it looks as if studies should use both a physical activity questionnaire and a tool to measure activity. Even though 5 days of activity were measured, this is still only a snapshot in time for a patient with RA.


Swinnen et al studied 40 patients with Ankylosing Spondylitis using a ‘SenseWear Pro 3’ armband device for monitoring daily activity. The patients wore the armband all day every day for 5 consecutive days including three week days and two days at the weekend. They worked out energy expenditure (METS) and compared it to values expected from age matched controls. They also examined the relationship between the BASDAI and the actual mobility measured. They did find that AS patients had a lower level of physical activity than controls (especially in the vigorous or very vigorous activity ranges) but did not find that the BASDAI correlated with the degree of difference in activity compared to controls.

  • The SenseWear Pro 3 armband has been validated as a tool capable of measuring energy expenditure fairly accurately. However, the location of the sensor may not be optimal for measuring mobility restriction in AS.
  • The two groups were fairly well matched, but the Healthy Control group had a significantly higher proportion of patients in employment than the AS group. Whilst this is not surprising, it might have been better to only include people in employment in a study like this. Although they attempted to control for this effect, the low numbers in the study may have made detailed analysis difficult.


    • This is one of the first papers to measure daily mobility accurately in patients with AS and to attempt to relate this to the BASDAI. A previous study by Plasqui was limited by a very small group size and the use of a less accurate pedometer measurement tool.
    • The authors have employed a well-considered methodology for making sense of the level of physical activity measured by these devices and they have demonstrated that patients can wear them for >90% of the 24 hour periods studied.
    • The study did not report on variations in BASDAI/mobility over time which probably would have been more informative.

stepActivityThe third article in my brief review is by Prioreschi et al in Rheumatology (May 2014). This was a small study studying the effect of DMARD therapy in 18 RA patients using accelerometry and compared results against a matched control group. They also recorded the duration of morning stiffness and more typical CDAI and HAQ score assessments. The patients wore Actical accelerometers for 2 consecutive weeks. Significant improvements in CDAI, HAQ, CRP, and duration of morning stiffness were found after 3 months therapy, and there was also an improvement in the level of moderate physical activity at the end of the study.

  • This is probably the first study to split the day into segments for comparing functional activity: they found that RA patients were less active in the morning and early afternoon, but that this difference from controls disappeared after DMARD therapy.
  • This is one of the first studies to relate the duration of stiffness against the measured activity levels. The results give objective confirmation that activity levels are related to the symptom of stiffness
  • The minimum required % of the day spent wearing the accelerometer was not specified in the paper.
  • The decrease in CRP was found to be related to the increase in moderate activity by multivariate regression analysis.
  • The authors have shown that this methodology can be used as a sensitive outcome measure for detecting changes in habitual physical activity
  • The authors commented that most of the participants were obese: the reliability of hip/belt worn accelerometers can be reduced in this group. However, the finding that DMARD therapy actually leads to increased moderate intensity physical activity in this group is even more likely to lead to health improvements in general.

Overall, I am really encouraged by these studies. I think we are going to learn a lot about our patients over the next few years as we get down to studying the impact of our treatments on their daily activity. I can see this sort of data (‘observations of daily living‘) being incorporated into the medical record and used to provide virtual ‘nudges’ to encourage patients to keep up their exercise programmes. Of course it will only be of benefit if it is done carefully & part of a rehab programme that the patient really wants to follow.