DEVELOPMENT OF DECISION SUPPORT SYSTEM FOR SUBJECTS TESTED POSITIVE FOR CANCER

Monday, October 25, 2010
Sheraton Hall E/F (Sheraton Centre Toronto Hotel)
Michi Sakai, MPH1, Noriaki Aoki, MD2, Satoko Zenitani, MPH1, Sakurako Narita, MPH1, Yoshiko Kakihara, RN1, Fumio Takesue, MD3, Takuro Shimbo, MD4 and Sachiko Ohta, MD5, (1)Center for Health Service, Outcomes Research and Development-Japan (CHORD-J), Tokyo, Japan, (2)University of Texas - Houston, Houston, TX, (3)Nara Prefecture Government, Tokyo, Japan, (4)International Medical Center of Japan, Tokyo, Japan, (5)Health Informatics Management Professionals (HIMAP), Tokyo, Japan
Background: Many Japanese patients experience difficulty in identifying an “optimal” medical facility for their health conditions, which may be a possible drawback of free access under the national health insurance system in Japan.

Purpose: We designed, developed, and implemented a decision support system (DSS) for subjects seeking medical facilities after a positive cancer screen.

Method: Care-seeking behavior after positive cancer screen was simulated from an established health-belief model. The DSS was designed to allow patients to systematically identify optimal medical facilities, based on screening test results for the five major cancers in Japan (i.e., lung, gastric, colorectal, breast, and hepatocellular carcinoma). The DSS was designed as SaaS (Software as a Service) with LAMP (Linux, Apache, MySQL, PHP). Jumula was used as a contents management tool with MySQL database software.

Result: Our DSS algorithm is as follows: (1) Interpretation of patient condition was based on user input of detailed screening results and risk factors. (2) The system then generated details of further work-up, such as accuracy, cost, risks, and duration, from evidence-based guidelines to anticipate patient concerns. (3) Medical facilities were suggested in accordance to patient condition. Users could also perform online searches for medical facilities and specify preferences for location, number of specialists, and performance. Recommended facilities were listed based on distance from patient homes using Google Map. (4) User data and all results were summarized in a printable format for further discussion with family or healthcare professionals. The system was initiated on May 1, with approximately 2000 hits per week between May 1 and May 31, 2010.

Conclusion: We successfully integrated behavioral science, decision science, and current information technology to provide information tailored to individual patient needs.