Introduction
Stroke is a leading cause of disability worldwide and one third of people who suffer from a stroke are left with permanent deficits impacting on their function (Carroll et al., Reference Carroll, Courtney-Long, Stevens, Sloan, Lullo, Visser and Brown2014). Physical impairments and activity limitations resulting from stroke may contribute to an inactive lifestyle. Current research suggests that people after stroke are not meeting the recommended levels of physical activity (English, Manns, Tucak, & Bernhardt, Reference English, Manns, Tucak and Bernhardt2014). A recent meta-analysis found that people after stroke walked only 50% of the recommended number of steps per day for healthy older adults, and they were significantly less active than people with other chronic illnesses (Field, Gebruers, Shanmuga Sundaram, Nicholson, & Mead, Reference Field, Gebruers, Shanmuga Sundaram, Nicholson and Mead2013).
Physical activity can reduce detrimental secondary changes after stroke such as loss of skeletal muscle (English, McLennan, Thoirs, Coates, & Bernhardt, Reference English, McLennan, Thoirs, Coates and Bernhardt2010; Scherbakov & Doehner, Reference Scherbakov and Doehner2011), change in muscle structure (Addison, Marcus, Lastayo, & Ryan, Reference Addison, Marcus, Lastayo and Ryan2014; Hafer-Macko, Ryan, Ivey, & Macko, Reference Hafer-Macko, Ryan, Ivey and Macko2008), loss of bone mass (Borschmann, Reference Borschmann2011; Borschmann, Pang, Bernhardt, & Iuliano-Burns, Reference Borschmann, Pang, Bernhardt and Iuliano-Burns2012) and reduced cardiorespiratory fitness (Billinger, Coughenour, MacKay-Lyons, & Ivey, Reference Billinger, Coughenour, MacKay-Lyons and Ivey2011; Pang, Charlesworth, Lau, & Chung, Reference Pang, Charlesworth, Lau and Chung2013; Smith, Saunders, & Mead, Reference Smith, Saunders and Mead2012). Physical activity has also been shown to modify stroke risk factors by improving glucose regulation, lipid metabolism, insulin sensitivity, blood pressure regulation and cellular function (Endres et al., Reference Endres, Gertz, Lindauer, Katchanov, Schultze, Schröck and Laufs2003; Ivey, Ryan, Hafer-Macko, Goldberg, & Macko, Reference Ivey, Ryan, Hafer-Macko, Goldberg and Macko2007).
In order to achieve these health benefits, international stroke guidelines recommend that people after stroke should participate in moderate intensity exercise for 20 to 60 minutes per session (or three 10 to 15 minutes exercise bouts throughout the day), at least three times a week (Billinger et al., Reference Billinger, Arena, Bernhardt, Eng, Franklin, Johnson and Roth2014). This activity dose is lower than current physical activity recommendations for heathy adults, which recommend 20 minutes of vigorous-intensity activity at least three days a week and strength training at least two days a week (Haskell et al., Reference Haskell, Lee, Pate, Powell, Blair, Franklin and Bauman2007). Both sets of guidelines advocate that for best health results, structured exercise combined with an increase in participation in daily household, occupation and social roles is needed (Billinger et al., Reference Billinger, Arena, Bernhardt, Eng, Franklin, Johnson and Roth2014; Haskell et al., Reference Haskell, Lee, Pate, Powell, Blair, Franklin and Bauman2007). The mismatch between physical activity guidelines and reported low activity levels of people after stroke highlights an urgent healthcare need to implement targeted interventions to promote physical activity. It is, therefore, important to understand how to support people after stroke to uptake and maintain physical activity behaviours.
Wearable sensors and Mobile Health (mHealth) technology are a potential low-cost and readily accessible platform for assessing and promoting physical activity. Wearable sensors can be defined as body worn sensors, such as accelerometers, sensors that are embedded in clothes or worn as accessories (Bonato, Reference Bonato2009). mHealth refers to all health-related interventions that are delivered using mobile devices such as mobile phones, smart phones and tablets (Dicianno et al., Reference Dicianno, Parmanto, Fairman, Crytzer, Yu, Pramana and Petrazzi2015). Smartphone technology is a component of mHealth. These technologies have the potential to facilitate behaviour change by incorporating techniques such as knowledge of results and social support (for example, via sharing of results on social media) (Yang, Maher, & Conroy, Reference Yang, Maher and Conroy2015). However, the feasibility and applicability of these technologies for people after stroke is unclear, and the validity of activity-related data derived from them is often poor (Bort-Roig, Gilson, Puig-Ribera, Contreras, & Trost, Reference Bort-Roig, Gilson, Puig-Ribera, Contreras and Trost2014; Morris, MacGillivray, & Mcfarlane, Reference Morris, MacGillivray and Mcfarlane2014). This narrative review provides an overview of the validity, feasibility and design of the current devices available to measure physical activity and also discusses the potential for advances in technology to promote physical activity in people after stroke.
Wearable Sensors
Pedometers
Pedometers are small, lightweight devices that count steps taken during walking and other locomotion activities. These devices are generally inexpensive and easy to use. Pedometers typically contain a spring-suspended, electric lever arm or a magnetic reed proximity switch that moves up and down with the vertical displacement of the hip (Schneider, Crouter, Lukajic, & Bassett, Reference Schneider, Crouter, Lukajic and Bassett2003). More expensive devices contain a horizontal beam and a piezoelectric crystal similar to accelerometers (Schneider et al., Reference Schneider, Crouter, Lukajic and Bassett2003). Pedometers are most commonly worn at the waist or hip.
Research in healthy populations have found the Yamax Digi-Walker to be one of the most affordable and accurate pedometers (Bassett Jr et al., Reference Bassett, Ainsworth, Leggett, Mathien, Main, Hunter and Duncan1996; Schneider et al., Reference Schneider, Crouter, Lukajic and Bassett2003). However, this does not translate well to the stroke population, with only modest validity when compared with observed step counts (ICC=0.58) (Elsworth et al., Reference Elsworth, Dawes, Winward, Howells, Collett, Dennis and Wade2009). Pedometer accuracy has been found to decrease at slower walking speeds (<0.9m/s) and with asymmetrical gait patterns (Berlin, Storti, & Brach, Reference Berlin, Storti and Brach2006; Fulk et al., Reference Fulk, Combs, Danks, Nirider, Raja and Reisman2014; Motl, McAuley, Snook, & Scott, Reference Motl, McAuley, Snook and Scott2005). This inaccuracy could be due to the mechanism of the pedometer or the analysis algorithms not being sensitive enough, to detect the smaller hip displacement that occurs at slower walking speeds, leading to the pedometer underestimating step counts (Melanson et al., Reference Melanson, Knoll, Bell, Donahoo, Hill, Nysse and Levine2004). One study did find that an ankle-worn pedometer was more accurate in evaluating step counts at slow speeds than pedometers worn at the waist or hip, however, pedometers still significantly undercount steps when compared to accelerometers (Karabulut, Crouter, & Bassett Jr, Reference Karabulut, Crouter and Bassett2005).
Accelerometers
Accelerometers are body worn sensors that utilise piezoelectric sensors to detect acceleration of movements. Uniaxial, biaxial and triaxial accelerometers evaluate movement in one, two or three planes respectively, and can provide total step count, activity counts, step rate and postural transition data (Chen & Bassett, Reference Chen and Bassett2005). There are several brands of accelerometers that have been used in research to measure physical activity, and the accelerometers also differ by site of wear such as wrist, arm, hip and ankle (Murphy, Reference Murphy2009).
Arm and Wrist Mounted
In recent years there has been a dramatic increase in the number of wrist-mounted activity trackers designed for the general population. Activity trackers are aesthetically pleasing, unobtrusive and often provide either real-time results or have user friendly mobile application interfaces. These wrist-worn trackers contain triaxial accelerometers, and in some models, other sensors such as altimeters or heart rate monitors. They are able to record step activity, calories burnt and activity patterns. However, the proprietary algorithms are intended for use in healthy adults and may not be sensitive enough to detect movement at low walking speeds. The Fitbit Ultra® (a small device that can be worn on a waist belt) and the Nike+ Fuelband® (wristband) were validated against observed step counts in patients after stroke or traumatic brain injury (Fulk et al., Reference Fulk, Combs, Danks, Nirider, Raja and Reisman2014). Fitbit Ultra® demonstrated moderate agreement with observed step counts (ICC = 0.73) but the Nike+ Fuelband® was found to be inaccurate (ICC = 0.20) for use in this population (Fulk et al., Reference Fulk, Combs, Danks, Nirider, Raja and Reisman2014). As is often the case with commercially available products, the hip worn Fitbit Ultra® is no longer available and has been replaced by activity wristbands. The validity of the latest versions of wrist worn activity trackers are not known in the stroke population, but it is possible that their sensitivity to slow or asymmetrical walking may be limited by the proprietary algorithms that determine step counts programmed for healthy adults.
Hip Mounted
The hip or waist is usually the preferred site for accelerometers to measure energy expenditure. Hip mounted accelerometers, such as the Caltrac and TriTrac RT3, have been validated in the healthy population with strong correlation to energy expenditure as measured by doubly labelled water (gold standard measure of physical activity) (Pambianco, Wing, & Robertson, Reference Pambianco, Wing and Robertson1990; Rothney, Schaefer, Neumann, Choi, & Chen, Reference Rothney, Schaefer, Neumann, Choi and Chen2008). However, both the Caltrac and TriTrac RT3 were found to demonstrate poor to moderate test-retest reliability (r = 0.44 and ICC = 0.68 respectively) in people after stroke (Haeuber, Shaughnessy, Forrester, Coleman, & Macko, Reference Haeuber, Shaughnessy, Forrester, Coleman and Macko2004; Hale, Pal, & Becker, Reference Hale, Pal and Becker2008). Generally, hip mounted accelerometers, similar to hip mounted pedometers, do not account for the gait asymmetries in people after stroke and thus it may not be the most accurate site for activity measurement.
Ankle Mounted
A recent systematic review evaluating measurement devices found the StepWatch Activity Monitor (SAM) to be the most accurate when compared to other accelerometers in detecting total step counts when validated against video-recordings of a timed walk test (ICC = 0.97) (Fulk et al., Reference Fulk, Combs, Danks, Nirider, Raja and Reisman2014). The SAM is an ankle-worn device with accelerometer thresholds which can be adjusted to suit different walking patterns (Mudge, Stott, & Walt, Reference Mudge, Stott and Walt2007). Measurement information is stored in the internal memory of the device, and can be downloaded at a later time to a computer (Chen & Bassett, Reference Chen and Bassett2005). However, the SAM has remained primarily a research tool and this could be due to the cost of the unit and the expertise required to operate, download and analyse the data, In addition, the SAM is relatively large in size when compared to other hip mounted accelerometers. Thus, when attached to the ankle, the SAM could be potentially cumbersome and intrusive compared to waist mounted systems which can be easily hidden under clothing.
Multi-site Mounted
Information on postural transitions are useful as it captures time spent in lying, sitting and standing which provides information on the person's activity patterns. The Intelligent Device for Energy Expenditure and Activity (IDEEA) is one technology that is able to detect postural changes by using five thumb-nail sized sensors attached to the upper sternum, bilateral mid-thighs and feet. The recording device is attached at the waist. The IDEEA was reported to be 98.7% accurate in detection of different postures in the healthy population compared to timed observation of postural changes (Zhang, Werner, Sun, Pi-Sunyer, & Boozer, Reference Zhang, Werner, Sun, Pi-Sunyer and Boozer2003). In stroke, one study reported the device to be 99% accurate in measurement of time spent in walking-related activity by utilising a more sensitive algorithm, provided by the manufacturer, to detect slow walking speed (>0.4m/s) (M. A Alzahrani, Dean, Ada, Dorsch, & Canning, Reference Alzahrani, Dean, Ada, Dorsch and Canning2011). No ICCs were reported in this study. While this multi-sensor and site system may be accurate, its clinical utility for long-term activity monitoring is very restricted as patient compliance is likely to be poor.
Although accelerometers have gained popularity in physical activity research, they have several limitations. A major limitation is the discrepancy in outcome measures and algorithms used between devices. Apart from number of steps per day, many studies have used ‘activity counts’ (M. A. Alzahrani, Dean, & Ada, Reference Alzahrani, Dean and Ada2009; M. A Alzahrani et al., Reference Alzahrani, Dean, Ada, Dorsch and Canning2011) as an outcome measure from accelerometers. An activity count is a combined, unit-less measure of amount and intensity of activity over a specified time period (Ainsworth, Cahalin, Buman, & Ross, Reference Ainsworth, Cahalin, Buman and Ross2015). Activity counts are determined by proprietary algorithms. This means that different brands of accelerometers may output different activity counts for the same acceleration signal (Vähä-Ypyä, Vasankari, Husu, Suni, & Sievänen, Reference Vähä-Ypyä, Vasankari, Husu, Suni and Sievänen2015). Thus, comparing results across studies which have utilised different accelerometers is difficult (Bonomi, Goris, Yin, & Westerterp, Reference Bonomi, Goris, Yin and Westerterp2009). The majority of commercially available systems do not allow for export and third-party analysis of the raw accelerometer traces. Given that the inherent technology is similar between devices, and that the proprietary algorithms themselves and not the technology are likely to be the cause of their poor validity, the inability to export the accelerometer data is a potentially insurmountable hurdle for their use in people living with stroke.
Another limitation of accelerometers is that they typically do not detect non-walking related activities such as gardening, swimming or cycling, and thus researchers are unable to gain a complete profile of physical activity behaviours (Butte, Ekelund, & Westerterp, Reference Butte, Ekelund and Westerterp2012). New technologies have been developed which combine measures of acceleration with physiological parameters such as heart rate and body temperature to calculate a measure of energy expenditure in everyday living. Measurement of energy expenditure, which is the basic unit of physical activity, might be a better indication than steps per day in evaluating if people, after stroke, were meeting the recommended guidelines for physical activity. The SenseWear® Armband (SWA) is a multi-sensor device that is worn on an armband around the upper limb. The device can measure movement (through a biaxial accelerometer) in number of steps per day and energy expenditure through physiological parameters such as heat flux, skin temperature and galvanic skin response (Moore et al., Reference Moore, Hallsworth, Bluck, Ford, Rochester and Trenell2012). Proprietary algorithms process the raw data into energy expenditure level; however the contribution of each of the physiological measures to the prediction equation for calculation of energy expenditure is unclear (Chen & Bassett, Reference Chen and Bassett2005). In healthy older adults, the SWA version 5.1 and 6.1 have demonstrated a strong correlation (r = 0.901 and r = 0.893 respectively) against doubly labelled water (Mackey et al., Reference Mackey, Manini, Schoeller, Koster, Glynn, Goodpaster and Cummings2011). In contrast, validation studies of the SWA in people after stroke have shown that the device (worn on the unaffected upper limb) is not valid for step counts (ICC = 0.352) and has poor to fair correlations for energy expenditure during treadmill walking (ICC > 0.409) (Kuys, Clark, & Morris, Reference Kuys, Clark and Morris2014; Manns & Haennel, Reference Manns and Haennel2012; Vanroy et al., Reference Vanroy, Vissers, Cras, Beyne, Feys, Vanlandewijck and Truijen2013). Although multi-sensor devices are potentially useful, more research is required to determine their suitability for people with stroke.
Portable global positioning systems
Global positioning systems (GPS) technology uses space-based satellite systems to calculate time and location data. It is currently used in various fields such as sports and environmental epidemiology (Rodríguez, Brown, & Troped, Reference Rodríguez, Brown and Troped2005) and is common in the communication devices used in everyday life. However, the use of GPS is still novel in physical activity research, particularly in clinical populations (Clark, Weragoda, Paterson, Telianidis, & Williams, Reference Clark, Weragoda, Paterson, Telianidis and Williams2014). Global positioning systems are able to provide information on the frequency and duration of outdoor physical activity and have been validated against direct observation in people after stroke, with excellent agreement at 94% for number of outdoor trips (McCluskey, Ada, Dean, & Vargas, Reference McCluskey, Ada, Dean and Vargas2012). Through synchronisation with Google Earth, or for more advanced analysis Geographical Information Systems, the data can also be used to explore relationships between outdoor physical activity and the neighbourhood attributes. Self-report diaries and physical activity questionnaires can be used to complement the data obtained from GPS. Combining accelerometer and GPS data will allow measurement of activity patterns and location of physical activity (Rodríguez et al., Reference Rodríguez, Brown and Troped2005). However, GPS is not without its limitations. These include its inability to be used accurately indoors, typically poor signal in built-up environments and on cloudy days, and the power consumed inhibiting long-term monitoring without regular recharging. Further research is required to determine the best design that has adequate battery life to last for the monitoring period.
mHealth Technologies
Smartphone and Mobile Applications
The smartphone is an example of a combined tri-axial accelerometer and GPS system that is widely used. Smartphones are mobile phones built on a mobile operating system that allow easy access to the internet and have advance connectivity features such as mobile applications (apps) that expands their function (ACMA, 2013). Some examples of smartphones are Apple iPhones, Android phones, Windows mobile phones and Blackberry. It is now estimated that 96% of the population worldwide has a mobile phone subscription with mobile phone penetration of 121% in developed countries and 90% in developing countries (“The World in 2014: ICT Facts and Figures,” 2014). In 2013, approximately 56% of all phones worldwide are smartphones and the average user would have spent 80% of their time on the smartphone using apps (“Infographic: 2013 mobile growth statistics,” 2013).
Fitness apps have gained popularity amongst the general population (Bort-Roig et al., Reference Bort-Roig, Gilson, Puig-Ribera, Contreras and Trost2014). Apps vary in function and may contain features such as social support, demonstration and instructions on how to perform the activity and feedback to promote physical activity behaviours (Yang et al., Reference Yang, Maher and Conroy2015). Intervention trials in weight loss research have utilised apps that combine external (data transferred by wireless connection via Bluetooth) and in-built measurement technologies. External measurement technologies utilise self-report on activity context, non-walking related activities that are not detected by accelerometers and other health-related behaviours (Bort-Roig et al., Reference Bort-Roig, Gilson, Puig-Ribera, Contreras and Trost2014). By combining information on self-report and objective in-built technologies, the person is able to obtain a complete report of their physical activity levels, diet and sleep.
However, a recent systematic review found that the majority of intervention trials in promoting physical activity after stroke utilised questionnaires to assess physical activity. The type of interventions explored in these trials were catergorised into face to face tailored counselling, telephone counselling, supervised exercise programmes and unsupervised home exercise programmes (Morris et al., Reference Morris, MacGillivray and Mcfarlane2014). Though both assessment and behaviour change techniques can be incorporated into apps to promote physical activity, this is yet to be explored in people after stroke. Furthermore, before these technologies can be used in stroke intervention trials, the accuracy of objective measurement of physical activity via in-built accelerometers in smartphones needs to be examined. In addition, there are battery-life limitations to smartphones which inhibit their use as a daily activity monitor. Continuously logging sensor data from GPS and accelerometers to memory cards causes a major drain on the battery life of smartphones. These sensors need to be constantly operating in the background without interrupting the usual function of a phone (Morillo, Gonzalez-Abril, Ramirez, & de la Concepcion, Reference Morillo, Gonzalez-Abril, Ramirez and de la Concepcion2015). If GPS and accelerometer logging is added to its normal use most phones are highly unlikely to last the required amount of time each day. However, as technology becomes increasingly energy efficient these issues may be overcome.
Smartwatch
Smartwatch is a novel wearable mHealth technology that has recently entered the mainstream commercial market. Apart from functioning as a wrist-watch, the technology is also characterised by it's ability to connect to the internet (Rawassizadeh, Price, & Petre, Reference Rawassizadeh, Price and Petre2014). The smartwatch can operate as an independent unit or it can be integrated with the smartphone through Bluetooth connectivity forming a body sensor network (Årsand, Muzny, Bradway, Muzik, & Hartvigsen, Reference Årsand, Muzny, Bradway, Muzik and Hartvigsen2015). The smartwatch, similar to other wearable sensors, has the potential to revolutionise mHealth by collecting data beyond activity tracking such as walking speed, quality and movement disorders (Samson, Reference Samson2015; Sharma et al., Reference Sharma, Mankodiya, De La Torre, Zhang, Ryan, Ton and Jain2014). Additionally, the smartwatch meets the commercial market demand for wearables with multi-functions (e.g. access to health apps such as GoogleFit or Apple Healthbook, able to answer calls and read messages). However, this technology is currently expensive and only in the developmental design stages for health monitoring in clinical populations. The smartwatch also suffers from limitations as manufacturers attempt to balance wearability with computing capabilities. These constraints include a small screen size, short battery life and a restricted ability to transfer data between devices (Rawassizadeh et al., Reference Rawassizadeh, Price and Petre2014). Thus, it is premature at present to expand on the applicability and usability of smartwatch technology in stroke or other neurological populations.
User-centred design for long-term sustained engagement
The ‘digital divide’
Although mHealth is a promising avenue for assessing and delivering physical activity interventions, much consideration into design needs to be given before it can be accessible or suitable in the stroke population. People after stroke could be of varying age groups and have different degrees of physical or cognitive impairments. Despite a rapidly growing mobile technology market, in Australia only 33% of people who are 55 years and above own a smartphone compared to 74% in the younger age groups. However, an industry survey estimates that this ‘digital divide’ between the generations is rapidly narrowing and is predicted to equalise by 2020 (Deloitte, 2014). The age-related differences in technology use could be due to changes in income level with aging, perceived need or importance of technology, and/or physical or cognitive impairments (Rogers & Fisk, Reference Rogers and Fisk2010). Some smartphones have assistive features such as voice control (e.g. Siri on Apple iPhone) and dictation features for people with physical impairments, however these are not uniform amongst the different brand of smartphones. There has been no study to date investigating the use of smartphones and their limitations in the stroke population.
Patients’ perspective on usability
Apart from the technical considerations of enhancing the feasibility of smartphone use to assess and promote physical activity, further exploration is required in establishing design features and support for use in the stroke population with physical and cognitive impairments. There is a paucity of research investigating the needs, wants and limitations in utilising wearable technologies for long term use in the stroke population. Guidance for design could be gleaned from studies in the general and elderly populations. A recent randomised controlled trial (SMART MOVE) investigated the effectiveness of a physical activity app (Accupedo-pro Pedometer) in increasing activity amongst adults in the primary care population and found that there was a significant increase in steps/day in the intervention group (Glynn et al., Reference Glynn, Hayes, Casey, Glynn, Alvarez-Iglesias, Newell and Murphy2014). A follow-up qualitative study exploring patient perspectives of this cohort found that patients reported usability barriers such as requiring help to download and setup the app, resetting the app in the morning, the need to carry the phone around and increased battery consumption resulting in charging phone more frequently (Casey et al., Reference Casey, Hayes, Glynn, Olaighin, Heaney, Murphy and Glynn2014). However, the ability of the app to run automatically in the background, a large and simple display, and having the app icon on the home screen, were highlighted as features that promoted usability (Casey et al., Reference Casey, Hayes, Glynn, Olaighin, Heaney, Murphy and Glynn2014).
Similarly, qualitative research in ‘gerontechnology’ application of mHealth to older people has resulted in identification of three key aspects for design which are: 1) identifying usability issues with current technologies, 2) evaluating suitability of training and instruction materials, and 3) understanding the unmet needs of the elderly not provided for by current technology (Rogers & Fisk, Reference Rogers and Fisk2010). Recommendations for wearable devices design in the older adults include attention to aesthetics, need for the device to be lightweight, comfortable, waterproof, easy to operate, inexpensive and have a long battery life (Murphy, Reference Murphy2009). It should also accurately assess physical activity and provide immediate feedback (Fini, Holland, Keating, Simek, & Bernhardt, Reference Fini, Holland, Keating, Simek and Bernhardt2014). Ideally, raw data from the devices should be able to be transmitted wirelessly or downloaded with ease onto a user-friendly platform. These recommendations could also be applicable for a stroke population.
Existing literature on user perspective of technology use in people after stroke is mainly around assistive devices, virtual reality and robotics for rehabilitation. However, these technologies are different in nature and purpose to wearable and smartphone technologies. Thus, there needs to be research specific to usability problems of wearable sensors and smartphones to aid in the in development of these technologies for people after stroke.
Clinicians’ perspective on supporting technology in stroke
The translation of new technologies from research to clinical settings, and repurposing of existing technologies for clinical use could be influenced by both the therapists’ and patients’ perspective. Current research reports that some of the barriers cited by physiotherapists for using technology in clinical practice were lack of time, insufficient knowledge with use of and interpreting information provided by technological measures, and a lack of technical support and processes to facilitate use of technology (Liu et al., Reference Liu, Miguel Cruz, Rios Rincon, Buttar, Ranson and Goertzen2014; Pak et al., Reference Pak, Jawed, Tirone, Lamb, Cott, Brunton and Inness2015). Integrating the data from the wearable sensors and smartphone to existing electronic patient records could also enable the results to be more accessible in the clinic for therapists (Veerbeek, Voshaar, & Pot, Reference Veerbeek, Voshaar and Pot2012). One study found that despite the initial difficulties that new technologies could present to clinicians, if the clinician believed that using the new technology will improve patient outcomes then they are more likely to use the technology in clinical practice (Liu et al., Reference Liu, Miguel Cruz, Rios Rincon, Buttar, Ranson and Goertzen2014). Thus, time and cost needs to be factored in to train both clinicians and patients to ensure uptake and sustainability of a new technology. Future research on new technologies to assess and promote physical activity should routinely report therapists’ and participants’ feedback which will improve our understanding of the feasibility of the devices for clinical use.
Summary
Direct measurements of physical activity, such as accelerometers, have increased in popularity amongst the general population in the form of wearable activity trackers. There is also a growth in literature validating different types of wearable sensors in stroke research. However, the use of wearable sensors and smartphone technology for assessment and promotion of physical activity in people after stroke is still in its infancy. These devices have yet to be translated to everyday clinical use due to the cost and technical expertise required in interpreting and analysing the data. Furthermore, there is a need for the use of a combination of technologies to measure different aspects of physical activity so that an accurate and complete profile of the person's activity can be generated. Wearable sensors and smartphone technology hold great promise, but need to be validated for measuring activity in people after stroke and other neurological conditions. Incorporation of behaviour change techniques, such as goal-setting, physical activity tracking, visualisation of progress and motivation, into customised apps could complement tailored counselling sessions for physical activity interventions.
Financial Support
This research received no specific grant from any funding agency, commercial or not-for-profit sectors.
Conflict Of Interest
None
