Clinic Dashboard

Lyceum Health improves patient care by linking doctors, patients and the pharmaceutical industry

How can we display patient data for doctors and nurses to make informed and faster decisions?

Platform

Desktop

collaboration

UX Lead, Product Manager, Lead Engineer, Data Analyst

duration

8 months

category

Data visualization, Dashboard, Results

The Problem & Users

Assessing a patient’s eligibility for specialty medications is complex and time consuming.

The workload of finding out if a patient is eligible for a specialty medication accounts for ~50% of clinic paper work.

Doctors are extremely busy and don’t see patients often enough.
Doctors do not have a lot of time to look at the software, so clear communication/documentation is important.
Navigating the healthcare system and eligibility rules / exceptions is a nightmare.
Nurses spend too much of their time on coverage-related paperwork and correspondence.

If we could provide the necessary information to healthcare workers and reduce their workload, it would allow them to focus their time on other tasks and get patients on specialty medications faster.

As you can see, this software is not very scannable - there is a lot of information displayed and poor information architecture, therefore it's hard to tell what's important and not.

However, most healthcare workers have been using this software for years, and therefore it was something that they were very familiar with. This was something to keep in mind when designing, to make sure adoption wasn't too hard.

In addition, there was nothing that suggested what specialty medication patients should be prescribed. Healthcare providers had to search for the information they needed and interpret it to determine what to prescribe.

How do we balance improving the user experience but still maintaining familiarity for the users?

Research - designing a dashboard

Data Visualization best practices

Elements (chart, table) are often displayed in a minimized view with the ability to bring up more details in a modal window or go to a page with more detail
Dashboards must always save the user time and help them be more efficient
Information architecture was very important - we kept in mind the F and Z patterns that reflect how a user’s eye scans a page to help determine the placement of the quadrants
We asked ourselves, “what will my users expect from this dashboard?”

Placement of data

Following the common Z pattern helped us determine the best way for users to consume important information. For these reasons:

We placed the high importance alerts at the top of the page to be addressed right away.
Historical data (1) + assessment scores (2) = eligibility engine (3)
Other general health information (4) that may affect a doctor's decision if they want to dive further into what's eligible.

Feature - Showing patient results (Data Visualization)

User Story

As a HCP (Healthcare Professional), I want a portion of the patient profile screen dedicated to Standard Validated Assessments so that I can quickly and easily view historical data, update values, add new assessments, and view trends over time.

Decision 1 - Figuring out how to switch between different assessments

Since we only had one quadrant of the screen to work with to show results, we needed a way to be able to switch the view to different assessments.

Version 1 - Changing assessment by dropdown

Issues with this design:

It required an extra click to open the dropdown and get to the assessment that you want
The dropdown menu would cover the graph when opened

Version 2 - Changing assessment by tabs

Advantages of using tabs:

Easier to navigate to the necessary assessment, and it’s clear which assessments are available
Tabs also allow us to show alerts next to the assessment name (if necessary) and make it evident to the user

Option 1

Show monthly consecutive data points, regardless of whether we have data for it or not, and regardless of whether the patient was supposed to fill in the assessment or not.

Advantages

When a data point is missing it means that the patient has not been filling in their assessments.

Disadvantages

Lots of empty space, especially if we send assessments every six months.
Wasted space if they aren’t going to add data.

Option 2

Show data points according to the frequency of the assessment, even if it wasn't completed (every 6 months).

Advantages

Better perception of time - this is helpful to show how old the data is

Disadvantages

Each assessment has a different frequency
Wasted space if they aren’t going to add data

Option 3

Show only data points when we have the data

Advantages

More focus on the available data

Disadvantages

If the time on the x-axis is inconsistent, and we only show months with a data point, doctors may think that everything is fine, and not realize that data is missing (at a quick glance, they might not be looking at the x-axis)
This does not show explicitly a potential long period of time where data was missing.

After the feedback, we changed the graph from a bar to line graph, and also decided that it should show any data that was entered within the last two years, as we thought this was a good amount of time for users to see some sort of trend. This ensured that no space was wasted, and only relevant data was shown. Having an x axis with the year also helped differentiate how far apart the assessments were being answered.

User Feedback

Alerts/Notifications

Just before we were about to launch, we met with the clinic and asked for their feedback on the dashboard.
‍
"It would be nice to see some kind of alert/notification for the most important tasks. For example, when a patient has a concerning assessment score. This will help us reduce the amount of time that we spend on the page."

This was something that we started designing for, but wasn't included in MVP as there was still a lot to think about (i.e., what alerts should we show, how long should they remain, can they be actioned).

Status accessibility

We received feedback from one of our healthcare providers that they couldn't tell the difference between the status colours because they all looked the same. This is because we only used colour as an indicator, so this was not accessible to colourblind users. We came up with the solution below, so that status’ were not indicated by just colour, but the fill of the circle which would be visible to everyone. We understood that a lot of users may not know what these symbols mean on their own, which is why we also added a tooltip with what the status was on hover.

Results

Shortly after we launched this product, unfortunately the company realized it wasn’t the best market fit, and we transitioned into our newest product RxNexus. Although the outcome wasn’t what we anticipated, working on the dashboard made me realize it’s such a common and important skill to know how to design for in order to help businesses and individuals make informed decisions.

Takeaways

Data visualization is tough

Talk to the experts, get their feedback, and do your research. One thing I wish is that we involved the data analyst a bit sooner. Initially we were ideating a lot on our own, but the data analyst was able to answer a lot of the questions that we had. There are a lot of different ways to show data, but you have to make sure you’re doing it correctly and make it as readable as possible to the user.

As priorities shift often for an early stage startup, communication is key

This project took a while because of competing priorities, and focus had to be shifted to revenue-generating projects multiple times. However clear communication with the team and working together to establish our product development process kept us going and we were able to launch our MVP.