User interfaces for patient-centered communication of health status and care progress

Abstract

The recent trend toward patients participating in their own healthcare has opened up numerous opportunities for computing research. This dissertation focuses on how technology can foster this participation, through user interfaces to effectively communicate personal health status and care progress to hospital patients. I first characterize the design space for electronic information communication to patients through field studies conducted in multiple hospital settings. These studies utilize a combination of survey instruments, and low- and high-fidelity prototypes, including a document-editing prototype through which users can view and manage clinical data to automatically associate it with progress notes. The prototype, activeNotes, includes the first known techniques supporting clinical information requests directly within a document editor. A usage study with ICU physicians at New York-Presbyterian Hospital (NYP) substantiated our design and revealed how electronic information related to patient status and care progress is derived from a typical Electronic Health Record system. Insights gained from this study informed following studies to understand how to design abstracted, plain-language views suitable for patients. We gauged both patient and physician responses to information display prototypes deployed in patient rooms for a formative study exploring their design. Following my reports on this study, I discuss the design, development and pilot evaluations of a prototype Personal Health Record application providing live, abstracted clinical information for patients at NYP. The portal, evaluated by cardiothoracic surgery patients, is the first of its kind to allow patients to capture and monitor live data related to their care. Patient use of the portal influenced the subsequent design of tools to support users in making sense of online medication information. These tools, designed with nurses and pharmacists and evaluated by cardiothoracic surgery patients at NYP, were developed using topic modeling approaches and text analysis techniques. Embodied in a prototype called Remedy, they enable rapid filtering and comparison of medication-related search results, based on a number of website features and content topics. I conclude by discussing how findings from this series of studies can help shape the ongoing design and development of patient-centered technology.