The patient summary is a highly utilized page of a patient’s medical chart. It aggregates current and historical clinical patient data. It provides a central location to add, review, and edit clinical information such as: medications, diagnoses, social history, and allergies. This patient summary redesign project was a 2 year endeavor lead by the design team. As lead designer, my role was to research and design a solution that addresses user feedback and improves usability of the patient summary. I collaborated with user research, design, engineers, and product management. This feature has been released 9/2018.
The patient summary functions as a face sheet of the patient's current and historical data along with interaction with a practice. How can we apply a consistent design system to improve readability and usability of this information rich page? How can we create flexibility with the design that accommodates different types of practices?
Historically, the summary has been treated as an afterthought in the product. No one team owned the summary page thus resulting in inconsistent data detail, interaction pattern, and display. I started off scanning through user forums and CS feedback to gain a general idea of users concerns with the summary. We received feedback around: Lack of customization, Missing sections of patient info: SIA, Immunizations, Print entire patient chart. Using that feedback, we conducted workshops and off-sites with the design team. We had the chance to share knowledge/learnings and explore ideas with no delivery deadline. This gave us an opportunity to play with ideas & explorations without limiting ourselves to the product.
I identified additional design issues with the summary around: dated visual styles (header styling) that contributed to hierarchy issues and pulled too much focus from the data and inconsistent patterns of interaction within the summary. For example using detail panes, modals, new tabs, or going away from the summary which contributes to a loss of context, disruption in workflow, obstructed view of information, and back and forth navigation.
As a result, we formed a hypothesis around the usage of the summary: Users reference past data while inputting new data simultaneously. Before we could think about redesigning the summary, we needed to plan how to execute the implementation of a drastic design update. This lead us to the arduous task of re-evaluating our design systems, which is a completely different project. As a result of revising our design systems, what we created was a drastically refined visual system and component visual improvements. With our platform team, we were able to create a new component for the card based design of the summary. This component was built using the revised design systems. Applying color, spacing, and type guidelines creates a system to building components that establishes consistency.
Government regulation required EHRs to provide new structured fields for collecting patient data which was the right opportunity for the team to fold in the summary redesign effort. Working with product, we had an idea of what additional improvements we wanted to include: customization, card design, screenings/interventions/assessments, and diagnoses groups chronic/acute.
User TestinG & interviews
Collaborating with our user researcher, Victoria, we conducted 3 phases of testing. We ran an initial phase of testing the redesign to confirm I was headed in the right direction. We sought to test my hypothesis that the Summary is used as a point of reference, and inputting and editing tool, and to gauge users’ reactions to a new layout.
We confirmed that the Summary is a highly utilized page, by both back office staff and providers, to reference, input, and edit information. This would be important when deciding where users should be able to enter new data fields.
All participants appreciated the new card design, though some couldn’t necessarily pinpoint what had changed— they just said they liked it. We saw this as positive feedback indicating that the design change would not be shocking or disruptive to users. We also tested designs around customizing the display setting of the summary. We received positive feedback that the design was clear and easy to understand.
Lastly, through a survey to our user research panel with 39 responses, we learned that the 3 top critical pieces of information reviewed on a summary are diagnoses, medications, and encounters. Closely following were past medical history, allergies and messages. Participants would like to be able to rearrange the order in which they view these pieces of information (easier to do with a card layout).
With the new field requirements we had to add to the EHR, we wanted to make sure that the new content was organized in a way that would make sense to users workflows. We conducted phone interviews to learn if, how, and when practices are already tracking the required data points in some other way in Practice Fusion. Through these conversations we found a lot of variation in whether this data gets recorded: for example: exposure to violence is most commonly recorded by psychiatrists, or OBGYNs, but may not be recorded by other specialties. Because there was such variance, we decided that all new fields should be defaulted to off when the new summary would be released, so as not to overwhelm users with all these new fields.
In addition to the contextual interviews, we had participants complete a card sorting activity. Card sorts are used to help decide how information should be categorized, or organized. We listed all existing field from the Summary, as well as all the new categories, and had participants organize them into categories. While some categories were unanimously agreed upon – like Education in Social Health – we had some close splits for fields like Alcohol and Tobacco Use (participants were split on having that in Behavioral Health and Social History).
We landed on these two cards, Behavioral health and Social health. We decided that having referred to an industry definition in which those fields are within Behavioral health and decided to validate. We tested the two mockups shown, we found that even though a participant might have suggested Alcohol use was a behavioral activity, providers reacted negatively to seeing alcohol use and smoking status in “behavioral health.” Even though when asked to group and categorize the fields providers might consider these to be behaviors, providers have been trained to document in very structured ways: in medical school, other EHRs, and paper forms, alcohol + tobacco use are always included in social health – and that’s still being taught in school.
The feedback from this final round of testing was so consistent, we alerted product and engineering our findings and recommended we remove the behavioral health card and keep a longer Social history card. In an effort not to overwhelm users, we recommended that we default the new fields off in the display settings as to reduce friction and allow users to opt into customizing their patient summary to fit their charting needs.
Collaborating closely with our developers, team Artemis, we conducted regular groomings and design reviews to ensure a successful build and release of the redesign. We accomplished a huge undertaking of releasing a redesign, additional fields, and customization to a highly utilized page.
Post release results
Referring to Mixpanel tracking, we saw 33,404 total instances of users customizing their summary display settings since release in September. Our customer success team would share the positive feedback and reception of the new design and customization functionality.