5:30 - 5:50 pmSaturday, September 17
LK 120
Leveraging econometric theory to detect adherence patterns in digital health social networks
LK 120
Leveraging econometric theory to detect adherence patterns in digital health social networks
CEO, Evolution Health Systems
BackgroundDigital behavior change interventions have shown much promise. However, many programs that are successful in experimental conditions suffer from high attrition in the general population. This... Read more

Description

Background
Digital behavior change interventions have shown much promise. However, many programs that are successful in experimental conditions suffer from high attrition in the general population. This presentation leverages two economic models to help illustrate how digital patients engage with online support groups.

Methods
Two periods of activity (July 25, 2008-August 7, 2012; Nov 15, 2005 – June 6, 2014) of actor and network-level data were extracted from AlcoholHelpCenter.net, a free-to-consumer, self-guided, patient-centric intervention designed to assist individuals cut down or quit drinking. In study period one, demographic characteristics and social network usage patters of 2584 registrants were analysed. Study period two assessed social network engagement patterns of 830 actors.

Results
In period one, when comparing actors (n=449) to non-actors (n=2135), there were no observable differences in gender, age range, occupation, or level of education. However, there were statistically significant associations between social network use and engagement with tailored exercises (P<0.01). In period one, social network contributory patterns resembled a power curve, with a high R2 values (.962) and a statistically significant spearman correlation (.987, P<.001). In period two, social network patterns of engagement were analysed over 34 quarters. Number of actors (Mean=43, SD=89) and posts (Mean=507, SD=368) varied, however the Gini coefficient, an economic measure of statistical dispersion used to measure income distribution, remained surprisingly consistent (Mean=.29, SD=.02).

Conclusions
In this digital health program, it was not possible to detect usage patterns based on demographic characteristics. However, engagement in the social network strongly resembled a power law, and the Gini coefficient was consistent. These results suggest that it is possible to predict and stratify individual engagement patterns. Based on these findings, it may also be possible to develop algorithms that can intervene with specific users who are in the process of disengaging with treatment.

While working for an eHealth startup during the late 90’s dot-com boom, Trevor realized that the Internet would transform the delivery of healthcare services. Recognizing this potential, in 2000 he founded Evolution Health and focused on bridging traditional, evidence-based healthcare with Internet and mobile-based tools. Through designing algorithmic platforms that transform behavior-based data into tailored treatment protocols, Trevor has become a respected pioneer in online and mobile health behavior change. As CEO he leads the strategic growth of Evolution Health and plays key roles in business development, technology innovation, and data analysis.

Trevor has an Honors Bachelor’s Degree in History and English from the University of Western Ontario, and a Masters of Science in Community Health (MScCH) with a specialization in Addictions and Mental Health from the Dalla Lana School of Public Health, University of Toronto. In addition, he has an MBA from the Rotman School of Management, University of Toronto and a Global Executive Masters of Business Administration (GEMBA) from the University of St. Gallen, Switzerland. He is currently a Research Associate at Henley Business School, University of Reading, and his doctorate focuses on cost modeling and financial analysis of eHealth and mHealth interventions.

 

Contact Us

We're not around right now. But you can send us an email and we'll get back to you, asap.

Not readable? Change text. captcha txt

Start typing and press Enter to search