Monitoring physical activity after a cardiovascular event: What is ‘fit’ for purpose?

In the September issue of European Journal of Preventive Cardiology, Lavie and colleagues discussed how physical activity across the lifespan is a cost-effective strategy that must be implemented globally to improve overall health, functional capability, longevity and quality of life.1 At the same time, they acknowledge the challenges with encouraging people to become and to remain physically active. In addition to the known benefits of being and staying physically active and reducing sedentary time for disease prevention,2–4 physical activity is a critical component in the care and recovery of patients following a cardiovascular event, such as stroke.5,6 Exercise therapy can have positive effects on depressive symptoms, emotional well-being, cognition, executive functioning and memory, balance, mobility, bone health, walking ability, upper extremity muscle strength, functional capacity, the ability to perform activities of daily living, and quality of life.6,7 The efficacy of exercise interventions in stroke rehabilitation is frequently determined using objective physical activity measurement instruments, such as accelerometers or pedometers. However, with so many instruments for measuring physical activity now available, it is often difficult to assess which is the best to use and for what purpose. For example, some may be interested in measuring gait,8 balance,9 or arm movement,10 while others may be interested in measuring energy expenditure.11 In the December issue of European Journal of Preventive Cardiology, Mandigout and colleagues12 present a detailed assessment of four different devices (Armband, Actigraph GT3X, Actical, and pedometer) to determine which is best suited to measure energy expenditure during activities of daily living in patients within the first six months post-stroke. The design of this study is important for a number of reasons. First, the position of a device on the body may influence its accuracy and sensitivity to record particular types of activity. Studies on upper limb and lower limb sensor placement suggest both may be useful in this patient group,13,14 although there are limited data on the effect of sensor position on the measurement of activities of daily living.15 Mandigout et al. assessed accelerometers located at the ankle, hip, wrist and arm (comparing both the plegic and non-plegic sides), in addition to pedometers worn at the hip and neck, so that participants wore 14 devices in total while performing the test. Second, selecting this early phase in stroke recovery is important. Guidelines suggest early exercise intervention can reduce the negative effects of bedrest and inactivity, while taking advantage of the heightened neuroplasticity present in the early post-stroke period.6 Lastly, the replication of activities of daily living, including walking and moving from sitting to standing, as the test protocol for patients wearing the different devices enhances the extrapolation of results to the daily living conditions of stroke survivors. Many stroke patients experience limitations in activities of daily living following a stroke, with reduced independence associated with lower long-term quality of life,16 and increased later life care requirements.17 The energy expenditure results from the simulated activities of daily living test protocol are presented only as a whole and not for different activities individually. However, for individuals, health professionals and/or researchers using these devices as part of an intervention to increase overall energy expenditure it may not be important to see how accurately energy expenditure is estimated for each activity. A further strength of this study is the comparison with energy expenditure data from a portable indirect calorimeter worn by patients throughout the test protocol. Inevitably, some of the devices and locations performed better than others in predicting energy expenditure across the range of simulated activities of daily living. The overall conclusion from the study was that all devices underestimate energy expenditure and, therefore, are not accurate for use in this early post-stroke population. There are a number of factors that could influence this finding. Patients using walking aids were not excluded from the study and the use of aids has been shown to influence energy expenditure prediction from wearable activity sensors.18 Exclusion of this patient group would clearly limit the real-world application of the findings. Comparison of the devices themselves is somewhat problematic. Many have proprietary algorithms based on recordings from one or more sensors, often in addition to the movement sensors in the devices. How these data are used to contribute to the energy expenditure estimation and the assumptions that underlie this are often not available to researchers. As such, we often operate with something akin to a black box where the contents are unknown. Access to the inner workings of these devices allows researchers to validate their use in specific populations, for example, the modification of the Actical activity counts thresholds for stroke patients.11 We need to understand the assumptions underlying the sensor output as, even under controlled laboratory conditions using a shaker plate, different sensors provide different results.19 The comparison with indirect calorimetry may also be problematic. There is some suggestion that indirect calorimeter estimation of energy expenditure may reflect a cost of movement post-stroke that may or may not be directly proportional to the task.20 Additionally, energy expenditure forms only one dimension of physical activity that may be important. Interventions may also be focused on increasing overall movement, balance, improving gait, improving cardiorespiratory fitness, targeting upper body strength or simply reducing sedentary behaviour.21 For these outcomes, a different criterion measure is needed to assess the validity of the measurement instrument. Previous reviews have shown that accelerometers are a valid tool for measuring physical activity in people after stroke,22 but suggest that both responsiveness and predictive value be assessed prior to using these devices. The relative change in activity over the period of any post-stroke physical activity intervention is still useful to provide awareness and feedback to patients and encourage their daily activity. In this situation, the absolute accuracy of devices may be less of a concern. Guidelines recommend that exercise programmes should be at least four weeks in duration to influence depressive symptoms, but to achieve the most favourable overall outcomes, structured aerobic exercise is recommended for at least three days per week for at least eight weeks, with lighter forms of physical activity promoted on the other days of the week.6 Research suggests that the benefits of exercise interventions on daily functioning and quality of life may reduce if the activity is not sustained.23 The use of activity monitors to support health professionals when giving feedback to post-stroke patients has shown some improvement in patients’ walking activity.24 However, even in post-stroke patients, where activity can translate to immediate short-term and longer term benefits in functional capacity, health and quality of life, research suggest that physical inactivity is prevalent.7 It seems that we have yet to find the most effective way to support individuals to become more active and to sustain this practice in everyday life. In addition to physical outcomes, there is evidence that physical activity interventions provide stroke survivors with a way to engage in valued activities, and support continuity with their pre-stroke sense of self.25 Understanding the patient’s ‘purpose’ in engaging with interventions may enhance intervention efficacy. Multidimensional assessment of wearable sensors, allowing patients to see progress in their movement, balance and gait, for example, as well as the more traditional methods of assessing physical activity such as duration, energy expenditure and intensity may also support more effective interventions. When selecting our measurement instrument and intervention, ultimately do we not want both the patient and the device to be fit for purpose? The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article. The author(s) received no financial support for the research, authorship, and/or publication of this article.