Implicit feedback: a new way of measuring patients interests in web-based health literature

Abstract

Existing research on evaluating patients’ interests has focused on user surveys, which encounter problems such as high financial and time costs as well as a low user participation rate. To address these problems, this study evaluated a new approach in measuring patients’ interests by applying implicit feedback. Implicit feedback refers to any data that can be collected unobtrusively by observing users’ interactions during web browsing. The behavior-oriented data is then exploited to infer users’ interests. This study built upon the Diabetes Education Page in the MyHealthAtVanderbilt.com patient portal and collected implicit feedback including Page Staying Time and Link Count Measures on 1450 diabetic patients. Our preliminary results indicated strong correlation between implicit feedback and patients’ interests. This thesis describes the implementation and evaluation of the implicit feedback and its potential in predicting patients’ interests in online health literature.