This presentation is part of: Poster Session - Public Health; Methodological Advances
MEASURING HOW PATIENTS USE INTERNET-BASED DECISION AIDS
Griffin Weber1, Jen Fortin, MPH1, Carol Landau, PhD2, Michele Cyr, MD3, and Nananda F Col, MD, MPH, MPP1. (1) Harvard Medical School, Medicine, Boston, MA, (2) Rhode Island Hospital, Psychology, Providence, RI, (3) Brown Medical School, Medicine, Providence, RI
Purpose: The Internet is an important vehicle for disseminating decision aids, yet little is known about how web-based decision aids are used. We sought to assess how women use a comprehensive web-based, interactive decision aid. Methods: We developed the web-based Women's Interactive System for Decisions on Menopause (WISDOM), which generates personalized risk reports (cardiovascular disease, breast cancer, hip fracture) and shows the impact of prevention options on absolute risks. WISDOM includes an embedded Markov model and a database of menopausal treatments that are accessed via an interactive symptom/treatment chart that can be sorted by treatment type, symptom concerns, and efficacy. Tiered text explanations about treatments are included. Summary reports can be sent to clinicians. Healthy women between 45 and 65 years were recruited as part of a randomized trial evaluating the impact of WISDOM versus standard care. The navigation patterns of WISDOM users were captured in log files. Results: To date, 69 women have enrolled, 35 were randomized to WISDOM, and 25 visited the website. The mean age was 51 years, 97% were white, and 86% were college graduates. Most (68%) completed all aspects of WISDOM; of these, 88% chose to send summary reports to their clinicians. Patients spent an average of 31 minutes (m) using WISDOM, with 11m allocated to inputting data, 4m interacting with the risk report, 3m with the symptom/treatment chart, and 1.5m with prevention options. Users were most interested initially in learning about diet and vitamins (43%), followed by lifestyle changes (33%), herbals (14%), and prescription meds (10%). Six (24%) clicked the chart an average of 12 times to read about a treatment-symptom interaction; 2 viewed more detailed information an average of 3 times. The 19 participants who viewed the prevention options page averaged 38s each on breast cancer and hip fracture, and 24s on cardiovascular disease. Conclusions: Participants interacted with a wide range of features and spent over 30 m on the site. Most, but not all, women sent summary reports to their clinicians, yet few were primarily interested in information about prescription meds. Entering the detailed personal information needed to provide tailored recommendations consumed a substantial portion of users' time, but participants used the feedback it generated.
