Building New Computational Models of Momentary Health-Related Behavior
|
If you are the presenter of this abstract (or if you cite this abstract in a talk or on a poster), please show the QR code in your slide or poster (QR code contains this URL). |
Abstract
In the quest to understand, change, and maintain health-related behaviors, behavioral theories that are predictive, prescriptive, and parsimonious are key to developing successful interventions. Mobile technologies (e.g. smartphones, computers, embedded, wearable/wireless sensors) provide increasingly rich data and opportunities for personalized, ‘Just-In-Time’ Adaptive Interventions (JITAI). These interventions can be fine-tuned to ‘adapt’ to individual’s behaviors as they change over time, and feedback can be immediate through sensors or mobile phones based on specific times, sensed/reported events, or locations. The ability to provide personalized and adaptive support to individuals, exactly when it is needed (e.g., just prior to the moment when a person engages in a less favorable behavior like eating fried foods), or as feedback directly after a behavior (e.g. praising a bout of exercise), affords huge potential for fostering healthier behaviors. However, current behavioral theories were not developed to drive such JITAIs.
Individuals now leave ‘digital footprints’ that reveal their behavior on a momentary, dynamic, contextualized and longitudinal basis. Behavior, feelings, thoughts, environment & location, time, social contexts, and temporal interrelationships between behaviors can be captured digitally and often transmitted in real time. These data provide the opportunity for new, dynamic, multi-method, conceptually driven, and data-rich approaches to theory development to support adaptive interventions. Full utilization of these data to support JITAI requires behavioral theories that provide insights about appropriate decisions rules for adaptation that guide which intervention strategies to use at what time. Likely requirements for these momentary theories of behavior include a) accounting for behavioral feedback loops (i.e. how my breakfast this morning influences my snacking behavior at lunch, or how stress today influences sleep tonight), and b) taking into account the lagged and contextualized nature of emotions & cognitions that potentially drive behavior (e.g. a put-down in a negative work environment might influence a person’s self-efficacy to exercise later in the day). This talk will outline a transdisciplinary approach to modeling momentary, contextualized behavioral theory as a dynamical system; an approach that enables improved model specificity, evaluation, forecast, and iteration. A strategy for seeding new models with existing theories will be outlined using the Social Cognitive Theory as an example. Strategies for integrating data and new theoretical constructs that emerge from a variety of sources (e.g., smartphones, wearable/ deployable sensors, digital trails that people leave through social networking and internet use) into this preliminary model to support more dynamic theories will be discussed. Incorporation of “dynamic” insights drawn from a dynamical model will be contrasted with the more static predictions made via traditional conceptualizations of behavioral theory. The final discussion will include the arc of development and cross-validation that can occur via the use of a mixed-methods approach for theory development, including a) model simulation, b) the use of “informative” and “optimized” idiographic system identification experiments to support empirical validation and refinement of dynamical models, c) strategies for identifying and integrating missing features from the preliminary model (i.e. linkages between variables, d) decision-rules for interventions, or contextual moderation) and e) the creation and evaluation of a JITAI.
Individuals now leave ‘digital footprints’ that reveal their behavior on a momentary, dynamic, contextualized and longitudinal basis. Behavior, feelings, thoughts, environment & location, time, social contexts, and temporal interrelationships between behaviors can be captured digitally and often transmitted in real time. These data provide the opportunity for new, dynamic, multi-method, conceptually driven, and data-rich approaches to theory development to support adaptive interventions. Full utilization of these data to support JITAI requires behavioral theories that provide insights about appropriate decisions rules for adaptation that guide which intervention strategies to use at what time. Likely requirements for these momentary theories of behavior include a) accounting for behavioral feedback loops (i.e. how my breakfast this morning influences my snacking behavior at lunch, or how stress today influences sleep tonight), and b) taking into account the lagged and contextualized nature of emotions & cognitions that potentially drive behavior (e.g. a put-down in a negative work environment might influence a person’s self-efficacy to exercise later in the day). This talk will outline a transdisciplinary approach to modeling momentary, contextualized behavioral theory as a dynamical system; an approach that enables improved model specificity, evaluation, forecast, and iteration. A strategy for seeding new models with existing theories will be outlined using the Social Cognitive Theory as an example. Strategies for integrating data and new theoretical constructs that emerge from a variety of sources (e.g., smartphones, wearable/ deployable sensors, digital trails that people leave through social networking and internet use) into this preliminary model to support more dynamic theories will be discussed. Incorporation of “dynamic” insights drawn from a dynamical model will be contrasted with the more static predictions made via traditional conceptualizations of behavioral theory. The final discussion will include the arc of development and cross-validation that can occur via the use of a mixed-methods approach for theory development, including a) model simulation, b) the use of “informative” and “optimized” idiographic system identification experiments to support empirical validation and refinement of dynamical models, c) strategies for identifying and integrating missing features from the preliminary model (i.e. linkages between variables, d) decision-rules for interventions, or contextual moderation) and e) the creation and evaluation of a JITAI.
Medicine 2.0® is happy to support and promote other conferences and workshops in this area. Contact us to produce, disseminate and promote your conference or workshop under this label and in this event series. In addition, we are always looking for hosts of future World Congresses. Medicine 2.0® is a registered trademark of JMIR Publications Inc., the leading academic ehealth publisher.
This work is licensed under a Creative Commons Attribution 3.0 License.