Olitor Clinician Dashboard - UW Nursing

"Enable clinicians to effectively and efficiently make dietary recommendations for patients”

Problem Space

Patient data can be overwhelming for clinicians. Olitor is a patient-facing app that tracks patients’ diet and promotes behavioral change. The Olitor dashboard is designed for clinicians to monitor patients’ dietary data in order to make informed recommendations to improve their health.

Project Overview

This project was in collaboration with University of Washington School of Nursing. The goal is to design a user-centered dashboard that helps clinicians improve their current workflow related to promoting dietary behavioral change in patients.

Challenges

Create data visualization and metrics that are accurate, insightful, and actionable.

Role
Product Design, UX/UI

Duration
6 months

Tools
Figma, Miro

Team
Kent Tsai, Sam Yang, Xue Teng, Julia Pavone

1 | Process

2 | Design Question

It is important to orient ourselves throughout this project to think about what types of data clinicians need to see in order to help them make dietary recommendations and how should we surface these data visualization effectively and efficiently.

How might we help clinicians monitor and improve patients’ health effectively and efficiently?

3 | User Research

We conducted 4 types of research methods to understand what types of data and visualization can help clinicians gain insights to patients’ diet and allow clinicians to quickly make informed recommendations.

Literature review
Subject matter expert interview
Survey
Interviews

Key Findings

1. Time is a huge cost. On average, clinicians are only able to spend 10 minutes/week on the dashboard.

2. Clinicians expect this dashboard to integrate with existing EHR systems.

3. Reviewing trends over time help clinicians understand the impact of actions taken.

4. Dashboard visualizations are key for reviewing reports with patients,

4 | Ideate & Wireframe

From our user research insights, we identified and organized every type of information that clinicians want to see. Then we mapped out the user flow to better understand interaction between screens and to lay out a blueprint for our system.

Identify and organize key data types
Create user flow and information architecture
Create low fidelity wireframes to visualize the interactions and layout

Screen Shot 2020-04-17 at 4.06.17 PM.png

5 | Design System and Design Principles

After we have received user and client feedback on the preliminary wireframes, we are ready to add visual elements and brand identity to our design. First, we defined our design principles that will guide our direction throughout the design process. Then, we created a detailed brand guide to establish brand identity and consistency.

8 | Usability Study and Findings

Through our usability studies, we were able to find out if the designs match our users’ mental model and if there are any confusions left to be clarified.

Words matter - Different vocabulary chosen for labels can give off different meaning to the users. For example, there is a difference in how users understood “meeting goal” vs “meeting threshold”.
Individual patient data matters more than aggregate data - Clinicians develop their treatment plan based on individual patient performance over time. Therefore it is important to present individual trend over time.
Clinicians prefers not to use words associated with negative emotions - A bit part of clinician and patient interaction is encouragement and positive reinforcement. Therefore clinicians tend to stir away from words such as “fallen”.

Iteration

Screen Shot 2020-08-28 at 12.07.35 PM.png

Before: The home page metrics show both meeting and below goal. And the goal is displayed as a constant value throughout the patient’s treatment.

Usability Findings: Since clinicians are short on time, they want to know what to focus on right away, which are the group of patients who needs extra help and attention. Clinicians also mentioned that the goal is personalized and can be different depending on the patient’s condition.

Screen Shot 2020-08-28 at 12.07.45 PM.png

After: In the iterated design, I made the summary metrics more focused and concise, focusing on the percentage of patient population who needs extra attention this week and is performing below their goal. In addition, I changed the constant goal value into a variable value that can be adjusted by the clinicians as they see fit.