Toward Intelligent Display of Health Data: A Qualitative Study of Use Patterns

Gains in healthcare efficiency, quality, and safety that are expected through the use of electronic health records (EHRs) will never be achieved if providers are unable to quickly search and interpret information critical to patient care decisions. As the use of EHRs expands, there is a need to better understand the use of patient data to support organization and prioritization decisions and the development of intelligent presentation and search tools. The goal of our research, funded by the National Library of Medicine, is to develop principles for the design of EHRs that will better support clinicians in their work through the application of contextual design methods.

We applied a contextual inquiry approach to data collection by recording providers’ information use activities during actual patient care followed by retrospective verbal protocol interviews. Critical care providers wore an eye tracker during periods of clinical information use (eg, rounding, admissions). Later, while viewing the eye tracking video, they were asked to describe what information they were accessing and why and the use of various information sources. Providers were interviewed about both the situation that was recorded and the translation of specific information use activities to other patients and contexts. As a means of analyzing the interview transcripts, we applied a grounded theory-based content analysis technique. This technique involved open coding of the interview transcriptions in which we grouped similar ideas into themes or categories. Then, through constant comparison, we renamed, reorganized, and redefined emerging themes.

eyetracker 1of2 eyetracker 2of2

Data were collected from 20 participants – nurses, residents, nurse practitioners, fellows, and attending physicians. Codes that were developed to capture recurring ideas included usability issues, the importance of recent data, trends, legal/financial/regulatory documentation, trust in data, task tracking, and more. The timeframe code, for instance, captured thoughts related to the temporal utility of data, including how long different pieces of data are useful (hours, days, throughout a patient’s stay, etc.), the appropriate frequency of different types of data for establishing trends, and how these information needs may vary for different providers, patients, or contexts. A quote (from a nurse) relevant to this code is “Weights [presented] will be yesterday’s weight, today’s weight, and admission weight which is nice because we trend how fluid overloaded or how dry they are.” Emerging themes that may be useful for generating design principles include meaningful presentation methods (eg, “big picture” flow sheets, lab trends), sort and search functions, and display customization. The timeframe code, for example, is expected to provide insight into methods for presenting data in a meaningful way and how they may be tailored for different situations.

Human-centered design methods are critical to the evolution of EHRs into tools that truly assist clinicians in making the right care decisions rapidly. We used contextual inquiry, a human-centered design ethnographic research method, to study information seeking and documentation activities in critical care. Currently, we are engaged in visioning interviews (which follow contextual inquiry in a contextual design approach). These interviews serve two purposes: (1) they provide an opportunity for interviewees to comment on our eye tracking interview findings and (2) they provide the opportunity to discuss innovation ideas in the context of information presentation to support critical care. We have conducted visioning interviews with 4 critical care providers and one biomedical informaticist. Coding of these interviews is focused on identifying agreement or dissent with respect to our findings from the eye tracking-based interviews and comparing and contrasting the feedback on innovation ideas across participants.

These methods are expected to improve our understanding of information use in the context of healthcare delivery. The findings from our research are expected to have implications for the future design of EHRs to support faster information access and to ensure that critical information is not missed.

References

Wright MC, Dunbar S, Moretti EW, Schroeder RA, Taekman JM, Segall N. Eye-tracking and retrospective verbal protocol to support information systems design. Proceedings of the International Symposium of Human Factors and Ergonomics in Healthcare. 2013; 2(1):30-37.

Segall N, Dunbar S, Moretti EW, Schroeder RA, Taekman JM, Wright MC. Contextual inquiry to support the design of electronic health records. Anesthesiology. 2013: A3151.

Upcoming MOCA® Course Dates

April 22, 2017 (FULL)
November 18, 2017
December 9, 2017

Have an Application for Simulation?

Contact Us

Email: simulation@duke.edu
Phone: 919-684-3661
Fax: 919-684-6251

Mailing Address:
Duke Human Simulation and Patient Safety Center
8 Searle Center Dr.
Suite 5010, 5th Floor, Trent Semans Center
DUMC Box 3094
Durham, NC 27710