Physician perceptions of clutter in electronic medical records

Abstract

Electronic medical records (EMRs) are now used by more than 95% of US hospitals (American Hospital Association (AHA) Annual Survey Information Technology Supplement, 2013). EMR systems typically provide a wide range of functionalities, including computerized physician order entry and the storage and presentation of patient medical data. The expectation has always been that these EMR functions would contribute to increased efficiency and safety of operations in hospital environments (Blumenthal & Glaser, 2007). However, display clutter in EMRs can lead to negative performance effects that can compromise the efficiency and safety of medical environments (e.g., Moacdieh & Sarter, 2015; Murphy, Reis, Sittig, & Singh 2012). However, it is not clear to what extent physicians view clutter as an impediment to their work, and, if so, whether it is solely the amount of visual data that leads to their perception of “clutter”. To this end, the aims of this study were to determine 1) whether physicians believe the nature and amount of EMR visual data affect their use of EMRs, 2) whether physicians think improvements are needed, and 3) to what extent it is the amount of data that leads to clutter versus some other qualitative aspect of the data. An anonymous survey was conducted among emergency medicine residents at the University of Michigan Department of Emergency Medicine. The response rate was around 60%, with 31 residents responding (age range 21-40 years). Residents had to respond to 18 questions. The first five questions asked for demographic information and participants responded using a dropdown menu. The next eight questions asked participants for their opinions about their satisfaction with their current EMR and the effects of visual data load on their work; participants responded using a 5-point Likert scale (strongly disagree or not at all (1) to strongly agree or extremely important (5)). The next three questions were free text and allowed residents to suggest design improvements to their current EMRs. Finally, the last two questions asked residents to rate, on a 100-point scale, the amount of clutter and the amount of information on sample screenshots from their current EMRs. This data was then correlated with each of the clutter image processing algorithms of Rosenholtz, Li, & Nakano (2007): feature congestion, subband entropy, and edge density. In general, results showed that physicians place a lot of importance on the design of visual information. Of the residents who responded, 52% indicated that visual data representation was “extremely important” for safety and the same percentage also said it was “extremely important” for efficiency. Also, 41% of residents agreed or strongly agreed that problems with visual data presentation have led to medical errors in their experience. In the free text space, physicians described many improvements that could be made to their EMR displays, particularly the reduction of excess irrelevant data. In addition, the correlation coefficients between the algorithm values and the ratings of amount of information were lower than the coefficients for ratings of clutter. This suggests that it is not just the quantity of information that factors into physicians’ perception of clutter; other factors, such as color variation and organization, play a role as well. In conclusion, this study showed that there is more to EMR clutter than merely excess data, and physicians appear to be aware of the dangers of clutter in their EMR displays.