A couple definitions. Notice that absolutely nothing here for your doctor or therapist to do.
Ecological momentary interventions (EMIs) are defined as treatments which are provided to patients between sessions during their everyday lives.
Ecological momentary assessment
(EMA) involves repeated sampling of subjects' current behaviors and
experiences in real time, in subjects' natural environments.
A perspective on the use of ecological momentary assessment and intervention to promote stroke recovery and rehabilitation
Statement of the problem
Despite extensive progress in the field of stroke rehabilitation, many challenges persist that limit long-term recovery, and consequent participation and quality of life of stroke survivors. 1 One important challenge concerns the transfer of skills gained in the therapy context to real-life situations, outside the clinical settings where new or adapted skills are needed. For many stroke survivors, a disparity between motor capacity (i.e. what one can do) and performance (i.e. what one chooses to do in the home and community) is well documented. 2–4 For example, 64% of stroke survivors who exhibit full arm motor recovery as measured by the Fugl-Meyer Assessment report persistent difficulty with or reduced hand use in daily activities. 5 The limited transfer of skills between the rehabilitation context and natural setting is a complex and multifactorial phenomenon, which if not addressed, may lead to long-term functional degradation, higher levels of disability, and limited community reintegration. 3,6 A limitation of traditional rehabilitation practice models is that assessments and treatments are usually performed in highly structured clinic-based settings, either during single visits (assessments) or relatively short encounters (treatments). Behavioral performance data obtained during these sessions may not accurately reflect the behaviors exhibited in the natural environment with all its unpredictable and changing characteristics. 6,7 Therefore, contextually relevant assessment methods and treatment interventions delivered under challenging but natural circumstances may provide valuable information about how to personalize recovery-promoting interventions, and to transfer the skills acquired in therapy to the natural environment. With recent technological advances, ecological momentary assessment (EMA) and ecological momentary interventions (EMI) have emerged to monitor, in real-time, the context in which behaviors are exhibited, and to deliver interventions in those same natural settings. EMA is defined as a range of methods to repeatedly collect real-time data about participants’ experiences and behaviors in the natural environment, such as current emotional state, physical symptoms, behaviors, contexts, beliefs, attitudes, or perceptions. 8 Similarly, EMI refers to interventions that provide treatment to people during their everyday lives – in real time and in natural settings. 9 This technology-enabled opportunity aligns well with the precision medicine initiative in the U.S. 10
Anchored in Thaler’s Nudge Behavioral Theory, 11 we argue that EMI informed by EMA acts as a gentle nudge to encourage stroke survivors to take advantage of their capacity and adopt health-promoting behaviors, including becoming more physically active. This perspective aims to provide an evidence-based rationale to adopt a coupled EMA and EMI strategy – one that captures and enables lasting changes in the choice to engage functional behaviors typically targeted in stroke rehabilitation programs (e.g. upper limb use, mobility activities). This perspective is divided into three major sections. First, we provide an overview of EMA methods, summarize the extensive research about EMA for non-stroke conditions and detail the emergence of EMA in the field of stroke rehabilitation. Similarly, the second section describes the nature of EMI methods, and highlights the emerging evidence supporting EMI for non-stroke conditions where behavior change is critically important for health. Finally, we recommend a synergistic EMA and EMI approach for the promotion of stroke recovery and rehabilitation, we outline the challenges with EMA and EMI and potential solutions, and we offer our thoughts about a few innovative directions for future research.
A brief history of ecological momentary assessment in behavioral medicine
Ecological momentary assessment (EMA), also known as experience sampling method, 12 originated in behavioral medicine. 8 Initially, participants received ‘notifications’ (i.e. auditory signal) from a pager at random times throughout the day to complete a paper diary about their lived experience. Before the advent of smartphones, data were typically collected using mobile pagers, personal digital assistant, or landline phones. In the past few years, major technological advances and ubiquity of mobile devices in the population at large 13 now affords repeated assessments outside the clinical setting using applications on smartphones or tablets (Figure 1(a) illustrates the recent growth of EMA research). Smartphone embedded sensors and wireless wearable sensors (e.g. activity monitors, sleep trackers, and global positioning systems) enable new possibilities to monitor physiological signals and behaviors throughout the day. 14 The integration of EMA into wearable sensor technology has potential, if used wisely, to enrich traditional single-session assessments typically used in stroke rehabilitation programs.
Figure 1. Growing number of publications related to ecological momentary assessment and intervention. Number of yearly publications in PubMed related to: (A) ecological momentary assessments (EMA, left graph) and (B) ecological momentary interventions (EMI, right graph). See supplementary material for details
What are the benefits of an ecological momentary assessment approach?
EMA offers many advantages to assess the dynamic nature of behavior. First, the higher resolution time series of assessments that EMA provides reveals a more accurate picture of one’s daily functioning and real-life experiences rarely captured during a single-structured assessment. Second, EMA avoids the recall and attendant bias often associated with retrospective questionnaires by collecting self-reports in that exact moment in time or shortly thereafter. 15 This is especially important for individuals with cognitive impairments, such as stroke survivors, and since the current behavior or mental state may fluctuate highly within and over days. 16 Third, multiple measurements can increase reliability of data, afford examination of time-sensitive phenomena, capture rapidly changing states, allow the development of temporal relationships between variables and generate more stable estimates of constructs for outcomes research. 9,17,18 EMA also allows one to better monitor treatment responses in real-time and in real-world settings. Different sampling methods can be used to collect information throughout the day. Interval and random sampling refer to information collected at definite or random time intervals during the day, respectively. Both sampling methods are used to acquire representative characteristics and patterns of behaviors and experiences across time. Event-based sampling is used for hypothesis-testing to examine antecedents and consequences of specific behaviors by collecting information after the occurrence of a specific event. 19 EMA offers researchers and clinicians the unique opportunity to collect rich, ecologically valid momentary information of how a person functions in daily life.
An ecological momentary assessment approach for non-stroke health conditions
EMA and EMI have been used across a broad range of non-stroke health conditions. There is an extensive body of work pertaining to the use of EMA for a variety of health conditions and the general population across the lifespan (youth 20,21 to older adults 22 ; Figure 2(a) displays the range of conditions for which EMA has been used). EMA is an established methodology used to better understand the psycho-social components of chronic pain, 23 the contextual factors and affects surrounding eating and dietary intake behaviors, 24–27 the dynamics of mood disorders, 28–30 and the physical activity levels in non-disabled individuals. 17,18,31 However, EMA is strikingly underutilized in the field of rehabilitation. 32
Figure 2. Target health conditions for ecological momentary assessment and intervention. (a) Types of health conditions or populations in which ecological momentary assessment (EMA) has been used in the literature. (b) Types of health conditions/population in which ecological momentary intervention (EMI) was informed by EMA (black bar) or not (gray bar) in the literature. See Supplementary material for details about the conditions included in each category and the details of the search used to gather these data. Abbreviations: ADHD: Attention-disorders/hyperactivity disorder, HIV/AIDS: Human immunodeficiency virus infection and acquired immune deficiency syndrome
![]({"type":"image","src":"/na101/home/literatum/publisher/tandf/journals/content/ytsr20/0/ytsr20.ahead-of-print/10749357.2020.1856557/20201203/images/medium/ytsr_a_1856557_f0002_oc.jpg"})
Ecological momentary assessment has been used for stroke survivors
To date, seven studies used EMA across acute to chronic stroke survivors to study post-stroke depressive symptoms, 16,33–38 two other studies report about fatigue, 39,40 and one focuses on the interaction between social interactions, pain, fatigue, and mood. 41 EMA provides insights on the fluctuation of mood in daily life, 34,42,43 the behavioral risk factors for depression, 16,35,36 and the impact of early social contact and support on long-term depression and daily life functioning. 33 EMA was also used in combination with magnetic resonance imaging (MRI) to characterize the association between depressive symptoms in daily life and cerebellar volume. 37 The EMA-MRI study underscores the value of linking brain imaging with information concerning daily-life experiences over time and across contexts through the use of mobile technologies. Research using EMA to monitor functional behavior of stroke survivors in real-world settings is only emerging. One example is this study from our group 44 in which thirty stroke survivors wore wrist activity monitors for 5 days and received six daily auditory prompts via a smartphone to capture arm and hand use, self-efficacy, mood, and social interactions. This preliminary work established that perceived positive social environments and higher self-efficacy were associated with higher paretic arm and hand use in the daily context. While this study did not offer an intervention, the simple act of responding to daily EMA prompts induced a small but significant increase in paretic arm movements (i.e. measurement reactivity). This promising finding provides an initial impetus for the use of context-relevant EMI to encourage functional movement behavior in the natural environment. 44 This study along with the long history of EMA work in behavioral medicine provide the opportunity for a better understanding of the complex relationship between stroke survivors’ impairments, psychosocial and environmental factors that impact the choices to engage in home and community activities.
What is the nature of an ecological momentary intervention?
Recent advances in technology allow improved possibilities to accurately collect contextual and behavioral data outside the clinical and laboratory-setting, and to deliver treatment directly in the natural environment when needed. EMI extends the repeated within-environment prompting methods of EMA by providing interventions in the real-life context. 45 EMI is broadly defined and ranges from unstructured clinical recommendations (e.g. reminders to exercise) to structured interventions (e.g. personalized education and tailored recommendations about medication management). 9 It can be implemented as a stand-alone intervention or incorporated with usual care to supplement existing interventions, using smartphones with embedded apps. Two approaches are commonly used to individually tailor EMI: 1) EMI content based on pre-intervention assessments or EMA; or 2) EMI delivered at specific moments when individuals are especially in need of additional support (i.e. intervention initiated by a specific event or circumstance). 9 For the latter, decision rules are critical to define when a specific intervention will be deployed. 46 Unlike EMA, EMI is a relatively new approach that emerged due to advances in technology offering clinicians and researchers tools to deliver interventions at critical times (Figure 1(b) highlights the growth of EMI research in the past 10 years). The lack of consistent terminology to describe interventions delivered in real-time and in the natural environment may have biased our count toward an underrepresentation of the actual number of published studies classified as EMI. While often not labeled as such, EMI encompasses mobile health technologies (mHealth) defined as mobile and wireless technologies to support the achievement of health objectives. 47 For example, accelerometers or smartwatches can capture walking behavior and provide real-time feedback about step counts. Another subset of EMI is ‘just-in-time adaptive interventions, 48 in which certain elements are adapted over time to an individual’s changing status and context, derived from mobile and sensing technologies. 49 Along with the growing field of digiceuticals, 50 EMI is a rapidly evolving enterprise stimulated by innovative advances in information technology. EMI will soon allow the provision of high-quality, relevant, and personalized interventions that can be used in the home and community to “nudge” individuals toward positive and lasting behavior change in their own environment. 11
What are the benefits of an ecological momentary intervention approach?
EMI extends interventions beyond the standard rehabilitation context by providing timely treatment in the context of an individual’s everyday life, as compared to treatment that can only be delivered at discrete times and places (e.g. in-clinic therapy session, 2x/week), shifting the focus of rehabilitation to the home and community. 9 Since EMI delivers interventions using solely a cellphone, it can help to reach individuals with limited access to rehabilitation care or those with limited mobility. EMI also provides its clients an opportunity to apply new skills and behaviors to their lived experience. 9 Until recently, EMA and EMI have typically been used separately. 9 However, the benefits of combining EMA with EMI would offer research-minded clinicians an opportunity to customize an intervention to the biological clock of an individual’s fluctuating state and context acquired through EMA on any given day. For example, Granholm et al. 51 prompted individuals with schizophrenia to complete multiple daily assessment items with response-driven feedback designed to promote medication adherence, socialization, or coping with auditory hallucinations. This combined approach shows promise for stroke rehabilitation which like schizophrenia is a complex and heterogeneous condition.
An ecological momentary intervention approach for non-stroke health conditions
Early EMI research provided proof-of-concept that EMIs could be beneficial for behavior change in various health conditions. 9 More recent research enabled by rapid advances in technology demonstrates increased potential of EMI. 46 EMI has primarily been used in the area of behavioral medicine and psychology to target smoking cessation, excessive drinking, weight loss, and promote healthy behavior change (see Figure 2(b) for the variety of non-stroke health conditions in which EMI has been used). EMI most frequently incorporates different approaches, such as reminders, encouragements, or various forms of education. Meta-analyses reveal moderate to strong evidence that mobile devices can positively influence multiple health-promoting behaviors, such as physical activity, 52 smoking cessation, 53 and diabetes management. 54 Findings from a large randomized control trial that tested four digital interventions to encourage physical activity in the general population, revealed that all four interventions increased mean daily step count. 55 Surprisingly, a simple daily prompt to read the guidelines from the American Heart Association website was more effective in encouraging physical activity than e-coaching based upon an individual’s personal activity patterns or reminders to exercise. This suggests that the exact mechanism through which an intervention might alter user behavior remains unclear. 55
EMI is well suited to implement elements of cognitive-behavioral therapy in the natural environment, 9,56 such as encouragements and prompts to motivate participants to practice skills and engage in exposure activities. 56 EMI was shown to have small to medium effects on mental health and positive psychological well-being. 46,56–58 The support of a mental health professional to guide individuals to use EMI-based apps was identified as essential for successful EMI. 46,57 Mixed results concerning the effectiveness of EMI have emerged for the treatment of psychotic disorders and behavior change. 9,45,59 These mixed results can be attributable, in part, to the significant variability in content, timing, and level of personalization of EMI across studies. One group suggested that many behavioral EMI programs were developed using an atheoretical approach, with minimal empirical evidence, or limited to absent treatment guidelines. 48 While the technology enables new treatment possibilities, an effective EMI is more likely to emerge from robust evidence, a personalized approach and a strong theoretical foundation, rather than primarily on the promise enabled by the technology.
Ecological momentary intervention for behavior change in stroke survivors
EMIs have yet to be implemented in the field of stroke rehabilitation (Figure 2(b)). While not labeled as EMI, a few studies have used mHealth technologies to deliver feedback about movement behavior or physical activity levels outside clinical settings. 60–64 The provision of quantitative feedback from activity monitors can include objective measures of activity (e.g. movement or step count, time spent in moderate intensity activity), graphs of daily activity, or encouragement of activity goals (e.g. keep going, you can do it, your half way toward your exercise goal for today) to motivate engagement in physical activity. 61 Increased levels of everyday arm and hand use and physical activity among stroke survivors rely heavily on effective behavior change approaches. Consistent with successful behavior change motivated by Nudge Behavioral Theory, 11 we propose that EMI could provide a viable nudge to encourage stroke survivors to adopt healthier behavior.
Evidence is currently lacking to support the use of feedback from activity monitors to motivate physical activity behavior in community-dwelling stroke survivors. 61,63,64 For paretic arm and hand movements, Da-Silva et al. 60 demonstrated the feasibility of delivering daily feedback (i.e. vibration prompt when activity levels fell below a predetermined level) from a wristband accelerometer to prompt greater use and independent practice of the paretic arm. Another study investigated the impact of delivering feedback collected through wrist-worn activity monitors on perceived and actual paretic arm use. 62 During a 3-week period, feedback was provided during seven follow-up sessions and consisted of discussions and graphs about the amount of use of the paretic arm. Findings from this study suggested that a high dose of feedback improved participants’ perception of how much and how well they used their paretic arm, but there was no effect on overall recovery measured with the Wolf Motor Function Test. There are at least three possible explanations for why quantitative feedback from activity monitors may have shown limited effectiveness on recovery outcomes. First, activity monitors used in the natural environment are largely limited to the provision of simple quantitative metrics (e.g. step or arm movement counts). 65 However, a large body of evidence suggests that enduring motor skill acquisition (i.e. habit formation) is best achieved with performance-relevant quantitative and qualitative information that is meaningful to the user. 66–69 Second, activity monitors are typically insensitive to behavioral and environmental context within which the behavior occurs. Yet, emerging evidence suggests that factors such as self-efficacy, attentional focus, and environmental contexts are important mediators of successful learning. 70–72 Decision-making also plays a crucial role in the adoption of a lasting behavior change. Nudge theory emphasizes the impact of choice architectures (i.e. carefully considering how choices are presented). 11 Third, durable behavior change requires learning and the development of health-promoting habits of behavior. As such, a critical limitation of several recent studies aiming to modify behavior of stroke survivors in the natural environment is that they target nonspecific generic goals (e.g. increase the number of steps taken per day), rather than the decision-making process toward a more personalized goal to achieve a beneficial movement strategy while walking (e.g. take a longer step with the non-paretic limb). With these kinds of generic goals, the focus is on the end game rather than the deliberate decision-making process (i.e. the choice) and behavior that ultimately leads to a relatively permanent and qualitative change in the target behavior (i.e. learning). This generic goal approach reflects ignorance pertaining to the well-known learning-performance distinction. 73,74 When we ignore the learning-performance distinction, we may well achieve temporary changes in behavior and even perception, but rarely do those changes persist beyond the cue that prompted them (i.e. feedback). One promising ongoing trial to encourage real-world arm use involves the delivery of multimodal feedback and reminders from data captured with activity monitors along with personalized goals. 72 Participants can visualize activity data, set themselves reminders, and engage with a game-like app that encourages recovery-promoting behaviors (i.e. growth of a virtual tree with arm activity). To provide for more effective EMI, researchers will need to provide support for health-promoting choices, through more precise and personalized strategies, and careful consideration of how choices are presented to encourage these choices as default options. Other behavior change strategies such as self-monitoring, biofeedback, and information about health consequences should also be considered. We suggest that a coupled EMA and EMI strategy could better empower, motivate, and educate the user about health-promoting behaviors. This combined and synergistic approach aligns well with recent evidence and theoretical perspectives about how attentional focus, intrinsic motivation, and mindset can be shaped to drive goal-action coupling for learning, acknowledge and encourage competence, build efficacy for self-management, and provide autonomy support (e.g. choice) in the recovery process. 6,71
Challenges for stroke rehabilitation and potential solutions
Beneficial strategies and key challenges should be considered by clinicians and researchers prior to adoption of EMA and EMI for stroke rehabilitation. Privacy, confidentiality and security are often highlighted as one of the most important challenges in EMA and EMI. 9,19,45,59 Cost-effective solutions commonly used in Internet/eHealth, telemedicine, and cybersecurity research should be integrated as the field progresses. 75 Other methodological and practical considerations about EMA and EMI include participants’ compliance, measurement reactivity, risk of device loss or damage, usability, and the complexity of analysis. 9,19,45,59 User engagement is crucial to minimize low compliance or drop-out. Strategies such as gamification features and integration of behavioral science theory and research in the design of EMA and EMI should be considered to encourage user engagement. 76,77 Practical considerations, such as the selection of EMA and EMI applications that are compatible with different operating software to be installed on participants’ smartphones, should also be integrated (see Burke et al. 78 for concrete suggestions to address technological and human participant challenges). In their current forms, EMA and EMI rely heavily on language to capture daily experiences and provide interventions, which may limit usability for stroke survivors with language or cognitive impairments. Alternative ways of providing EMA and EMI could improve their suitability for a broader stroke population, such as for stroke survivors with aphasia or memory limitations. Modifications could include supportive images or pictograms to supplement the verbal prompts or feedback delivered. The rapid development of the technology can be seen as a challenge, but it could also enable new possibilities that effectively address current limitations.
Conclusions and future directions
Preliminary evidence and technological advances support the implementation of EMA and EMI for a variety of health conditions. Such a combined and synergistic approach has yet to be fully adopted in the broad area of stroke recovery and rehabilitation. Evidence from work done with other populations offer promise to encourage behavior change of stroke survivors in the natural environment. However, the effectiveness of EMI on functional behavior change in stroke survivors remains to be established. EMA provides promise to gain important insights about the dynamic relationship between physical and cognitive impairments, and the psychosocial and environmental factors prevalent in the natural setting. There is ample evidence that these factors present as major impediments to sensorimotor recovery and positive behavior change. Therefore, we propose that EMA methods could be used to address clinically relevant, hypothesis-driven questions that are limited by standard methods and single-session assessments. We argue that EMI in the context of stroke rehabilitation should be informed by EMA to allow more precision and to drive a more desirable and durable behavior change, especially in light of the availability of more sophisticated body worn sensors. In the near future, it will be possible to use these sensors to capture activity-specific movement behavior along with physiological data and then provide a brief personalized, context, and state-relevant prompt delivered through one’s smartphone. EMA and EMI could potentially be used to identify and target underlying factors influencing arm and hand use and mobility activities in the natural environment, such as self-efficacy. Elements of a personalized intervention might include response-driven qualitative and quantitative feedback about performance, self-management tips, and/or strategies to solve identified barriers to functional movement behavior offered at strategic times.
To guide research and clinical practice, we present a case example of how a coupled EMA and EMI strategy might be used to encourage stroke survivors to overcome “learned non-use” to begin to use their paretic arm and hand for daily activities. The findings from our previous work using solely EMA 44 suggest that simply asking people if they have used their arm and hand leads to a statistical increase in that use over a 5-day period. We hypothesize that EMA questions can help draw stroke survivors’ attention to their arm and hand use. To capitalize on this simple but reasonable assertion, we propose an EMA and EMI coupled strategy that begins by asking stroke survivors if they recently used their paretic arm and hand for any activity and then follows up by offering encouragement to reach their self-selected goals by using qualitative and quantitative feedback about their arm and hand use behavior. In addition, EMI could periodically prompt the user to evaluate their progress toward goal achievement and empower them to set-up new goals, as they progress. Psycho-social data collected through EMA could be incorporated into EMI, such as offering tailored self-management strategies to overcome barriers to goal achievement. This coupled EMA and EMI strategy could be used in conjunction with usual care to encourage stroke survivors to apply the skills learned in therapy to real-life situations and act as a real-time nudge to encourage healthier behavior.
There are numerous opportunities in the field of stroke rehabilitation and recovery to advance a combined and synergistic EMA and EMI methodology. One obvious research opportunity includes the identification of user-specific critical feedback characteristics that are most likely to drive a targeted health-promoting behavior change. Moving forward, it will be imperative to identify ways in which behavior change can be maintained over time and to explore the provision of “booster doses” of EMI with the intent to prevent drift and forgetting, similar to that used in fall prevention programs. 79,80 The growth of eHealth and tele-rehabilitation in response to the global coronavirus pandemic provides further motivation for the development of effective coupled EMA and EMI strategies in the field of stroke rehabilitation.
The timing is right for stroke researchers and clinicians to take advantage of these methodologies to better understand functional behaviors outside restricted clinical settings and deliver evidence-based, personalized and simple interventions that drive health-promoting behaviors. Recent clinical practice guidelines 81 provide a solid foundation upon which to innovate a combined EMA and EMI approach – one that in all probability, if implemented thoughtfully could be highly impactful by enabling more precise, and personalized data-driven interventions, that reduce the disparity between motor capacity and performance and shift the focus of rehabilitation and recovery and empower the person in the home and community where they live.
No comments:
Post a Comment