People's health information is scattered across fragmented systems, making it hard to get the right info at the right place and time in order to make good health decisions. I am designing and studying a platform that emulates sites like ifttt.com and Microsoft Flow (which currently have very little integration with health systems) to see whether this style of end-user-programmed integration of different systems and personal informatics automation can help people fill gaps in functionality and improve overall access and usability of health IT, and ultimately make better health decisions.
Screenshot of early version of this platform
Graphics Help Patients Distinguish Between Urgent and Non-Urgent Deviations in Laboratory Test Results (JAMIA)
Working with Dr. Sameer Saini at the Ann Arbor VA, I have designed and developed an online decision aid to help clinicians decide whether the benefits of a colonoscopy are worth the risks for a patient given some individual characteristics. This aid can help clinicians make more personalized decisions about screening and highlights some particular examples of patients whose risk and benefit from screening runs counterintuitively from established guidelines.
My dissertation research looked at the ways that the design of intelligent decision aids (decision aids that use artificial intelligence or similarly sophisticated computation to make recommendations to decision makers) can create decision making biases. I found that the customizability of the system, the transparency, and users expectations of its efficacy can cause users to agree with system recommendations regardless of what recommendations are given.
Going forward, I plan to extend this research to the design of tools for data science. How can we design tools that enable data scientists to both uncover important insights and make good decisions from data?
I have worked with Rick Wash to study how the design of crowdfunding platforms influences users' ability to coordinate their collective efforts and successfully fund projects. Our research has mainly involved controlled lab experiments that simulate crowds on a crowdfunding site, where we have found that the style of crowdfunding (All-or-nothing vs. Keep-it-all) and the timing of donations have important consequences for project outcomes.
I have studied the design of Online Communities with my PhD advisor Rick Wash, where have found that early contributions to online communities set an expectation for newcomers about what is required, and that if new users do not see the existing content on a site they are likely to contribute more. We have also found that communities should be designed to attract members rather than content, as communities such as WikiProjects on Wikipedia will have more long-term growth if they attract many new people rather than getting a smaller number of people to contribute more content.
Some other projects I have worked on have looked at the design of instant messaging and self-service kiosks, as well as some theoretical work about privacy concerns from targeted advertising and understanding how users determine the "source" of their interaction with a computer in an online environment.
You can download my cv here.