Smoking is still the number one preventable cause of cancer death. New approaches are needed to engage smokers in the 21st century in smoking cessation. I propose to develop S4S (Smokers for Smoker), a next- generation patient-centered computer tailored health communication (CTHC) system. Unlike current rule- based CTHCs, S4S will replace rules with complex machine learning algorithms, and use the collective experiences of thousands of smokers engaged in a web-assisted tobacco intervention to enhance personally- relevant tailoring for new smokers entering the system. This NCI K07 will provide a mentored research experience, giving me the opportunity to acquire new competencies in cancer health communication for behavior change, research design and statistical methods underlying clinical trial implementation and evaluation. I will adapt collectiv intelligence algorithms that have been used outside healthcare by companies like Amazon and Google to enhance CTHC. Using knowledge from scientific experts, current CTHC collect baseline patient "profiles" and then use expert-written, rule-based systems to tailor messages to patient subsets. Such theory-based "market segmentation has been effective in helping patients reach lifestyle goals. However, there is a natural limit in the ability of a rule-based system to truly personalize content, and adapt personalization over time. Current CTHC have reached this limit, and I propose to go beyond.
My first aim i s to develop the Web 2.0 "S4S" recommender system.
My second aim i s to evaluate S4S within the context of a NCI funded web-assisted tobacco intervention (Decide2Quit.org). In my efforts, I will guided by my primary mentor (Dr. Houston, MD MPH) and two other mentor teams: The Cancer Health Behavior and Communication Team (Stephenie Lemon, PhD and Kathleen Mazor, EdD), and the Clinical Trial Design and Analysis Team (Jeroan Allison, MD, MS and Arlene Ash, PhD). My comprehensive training also includes coursework, seminars, and conferences.
A high priority research area identified by the NIH is to understand the role of different media in increasing consumer demand for and use of effective, individually oriented tobacco cessation treatments for diverse populations. Specifically, a research focus is on understanding how to best tailor interventions. Although effective, there is a natural limit in the ability of current rle-based tailoring systems to truly personalize content, and adapt personalization over time. This NCI K07 will advance computer tailoring by adapting machine learning collective intelligence algorithms that have been used outside healthcare by companies like Amazon and Google to enhance the personal relevance of the health communication.
|Sadasivam, Rajani S; Luger, Tana M; Coley, Heather L et al. (2014) Robot-assisted home hazard assessment for fall prevention: a feasibility study. J Telemed Telecare 20:3-10|