EHRs have a reputation. This year at my annual checkup I was thinking about EHRs and trying to observe how my PCP's office uses theirs. As I sat on the exam table I got the usual questions from the physician's assistant; "Any pain today?" Smoker?" "Allergies?". I could see the screen on her laptop with tiny rows of text that were slowly growing with each question. There was a lot of clicking, opening various windows, extensive typing, etc. It felt like a long process, and she was having to do a LOT of work to get a little information into the system.

Every year when I go in, I try to notice if anything has changed with the EHR. From my vantage, are the clinicians having an easier time with the EHR? I didn't notice any improvement. But it is hard to tell what gains have been made with casual observation. More importantly, it is hard to make improvements without some help.

There are tools used to take snapshots of how a user navigates a software program. A colleague of ours, Jim, recently wrote an article about heatmapping software. It helps web developers and designers improve experiences for people that visit a website. Jim was writing about heatmapping in the context of content-organization on a website, but his article prompted me to think about using heatmapping for EHR optimization.

Heatmapping shows hotspots of user interaction by tracking and visualizing clicks, cursor movements, and scrolling activity. The purpose is to capture and show how people interact with information on a page to accomplish tasks or goals. It is used both to reduce steps and make information easier for the user to find. For example, heatmapping a website could help show whether or not the location of a sign-up button correlates with the frequency it is clicked. This can be translated to EHR use by observing how the location of information and actions affect how easy it is for users to and find and record information. The results can answer questions like: which processes in the workflow are costing the user time and was a change effective in reducing effort; thus, it can help optimize EHR workflows.

We all know that some systems are notoriously hard to use due to poor UX design. These systems were developed to accommodate so many different perspectives, uses, and workflows that user experience was the last to get any attention. One unfortunate result is that clinician who is analyzing and interpreting information for a diagnosis may have to click between different screens to get the information they need. This affects the time and effort of the clinician and ultimately affects patient care. A heatmap shows what the user is doing to get what they need for a task, illuminating where these inefficiencies occur.

Jim argues in his article that heatmaps should be used with other forms of measurement. Using only one tool to paint a picture of usability could lead to incomplete conclusions about usability shortfalls. Navigating is just one action. Information about navigation can tell us a lot about workflow, but a clinician is also searching for and inputting information.

Eye-tracking used along with the heatmaps can give us more information about how the software is used. The results can tell us if the information a user is viewing relates well to the location of their actions. In other words, the eye may track to one part of the screen for information but the next action is to click on a distant part of the screen to perform the next step in the workflow. When we can identify these problems, we can reorganize information and actions to more optimal positions on the page according to the workflow, and we can measure the results to look for improvements.

A heatmap used with a keystroke logger can tell us about the relationship between the placement of information and the process of entering data. As I observed in the visit to my PCP, there are unnecessary steps that can be consolidated to streamline the interview process into fewer clicks. Instead of having to type all of the answers they could transform into menus with pre-determined selectable answers. Especially for the many yes/no questions.

We need to look at the use cases also. Can the system be optimized for different roles? For example, a person working in the lab will want to access different information about a patient. A person working in a clinic will need information for making decisions about diagnosis and care without the system robbing cognitive load. Using heatmapping and other analytical tools, the information can be arranged to provide each person in each place with less friction in their unique workflow.

Improving UX can be effectively accomplished by measuring and analyzing what people actually do to find the hang-ups. Heatmapping is one piece of the puzzle. When paired with other measurements, it can be an effective tool for improving workflows. We will look at more tools and optimization ideas in the future.