Empowering diabetes clinics to make a shift from reactive to proactive healthcare
Company: Tidepool | Role: Product Design Lead | Current Status: In Pilot Phase with 3 Clients
Problem
Unmet need for diabetes clinics to provide timely and targeted care to at-risk patients
Solution
Enable clinics to monitor real-time patient data and identify at-risk patients via a dashboard
Outcome
Forecasted revenue in the first commercial year is $1.7M
Objective
Business-driven
With this dashboard, the business goal was to target higher-value diabetes clinics with enormous patient volumes and increase revenue generation for the company.
Team
1 Product Manager, 1 Product Designer,
7 Engineers
Our mighty team of 9 worked closely to build this dashboard within 4 months. We still continue to iterate, build new features and improve the user experience for this tool.
My Role
Lead Product Designer
I was knee-deep in the entire development process from problem discovery, ideation, product definition, prototyping, and user testing to launch.
Product Evolution over four release versions
Understanding the Problem Space
We started off by meeting subject matter experts to understand the complexity of devices and technologies used today to manage diabetes. After that, we conducted 10 user interviews with prospective customers which revealed one key insight:
There is a heavy shortage of clinical staff needed to treat the increasing number of diabetic patients.
This puts pressure on clinics to prioritize patient care based on clinical needs so that they can reduce costs, achieve better patient outcomes and improve overall patient satisfaction.
When asked about their workload, all endocrinologists echoed the same issue:
"I don’t have enough time! I have a 3-month long patient waitlist, and I’m just trying to get through each appointment and keep my head above water.”
Identifying Problems worth solving
After doing a thorough competitor analysis and analyzing the user insights, we narrowed down the product scope to focus on the following problems:
1
Clinicians don't have a way to monitor the real-time health data of all their patients in one place
2
The current process of prioritizing patient care is manual, time-consuming, and resource intensive
3
Clinicians lack effective tools that help them identify at-risk patients
"I will get messaged 3 times a week by the type A high-stress patient who is doing pretty well, but I don’t have resources to proactively reach out to patients who actually need help.”
— A Clinician from Yale New Haven Hospital, CT during a user interview
Product Goal
Currently, the platform collects data from patients’ diabetes devices (like CGM or Insulin Pumps). It generates easy-to-read graphs of aggregated blood glucose statistics so clinicians can make informed treatment decisions.
The new dashboard aims to expand product capabilities to monitor patient data in real time and enable prioritization of the entire patient list based on standard clinical guidelines.
User Stories
Requirements
As a clinician, I want to be able to monitor patients' health data in one place so that I can quickly see how my patients are doing between their clinic visits.
Provide a quick, glanceable summary of clinically critical data points required to make a determination about patients' diabetes management. These data points include Last Upload Date, % CGM Use, GMI, and % Time in Range
As a clinician, I want to be able to filter patients based on standard clinical guidelines so that I can identify at-risk patients and prioritize proactive outreach.
Provide the ability to filter based on the clinically recommended blood glucose targets i.e the amount of time the patient's blood glucose levels were high, low, or in the normal range
As a clinician, I want to be able to see when my patients have synced their device data so that I can avoid making decisions based on outdated data.
Provide the time stamp for when the patient synced data and the ability to filter based on that
As a clinician, I want to be able to group patients with similar characteristics (e.g. pediatric, older, higher risk, pregnancy) so that I can keep track of patients that need more attention.
Provide functionality to assign tags to patients with similar characteristics and filter based on those tags
Design Principles
Assumption
The clinician user has prior experience using any diabetes management software to analyze their patient's diabetes data.
Design Rationale
The user experience is optimized to minimize time spent on prioritization of patients in a clinic and improve workflow efficiency. This aligns with the product goal of encouraging active engagement and sustained usage of the product.
Design principles crafted for this experience are:
Discoverability
Making necessary patient data and frequently taken actions easy to find and navigate
Learnability
Minimizing the learning curve and time spent on onboarding and learning how to use new features
Efficiency
Ensuring critical tasks can be performed efficiently with reduced errors
Solution
A Population Health Dashboard that provides a quick overview of real-time diabetes data of all patients in one place.
With this new dashboard, diabetes clinics can view and analyze diabetes data for their entire patient population under one roof. They can filter and identify at-risk patients and prioritize outreach resources for those who need them the most.
Time in Range Visualization
A graphical representation of the patient's CGM* data enables quick and easy risk assessment of all patients
More Green = Under Control
More Reds or Purples = High Risk, needs investigation
How much time a patient with diabetes spends in different blood glucose ranges is the most critical indicator of risk. Clinicians can glance at their Time In Range statistics from the bar chart and hover over for accurate % values.
*CGM is a Continuous Glucose Monitoring device that measures your glucose levels 24 hours a day when worn.
Filter for Last Upload
Ability to check the validity of patient data to inform clinical decision making
Clinicians can filter patient data through different time periods and device sources to understand if the patient data is recent and valid for decision-making.
Filter for Patient Risk
Ability to filter patients by risk based on standard clinical guidelines
Clinicians look at specific Blood Glucose ranges to determine the patient's severity. They can filter patients based on these thresholds and prioritize care for those at higher risk.
This was one of the most interesting and challenging designs to come up with because of the sheer complexity of defining patient risk. Coming up with a solution that would resonate with most of the use cases was tricky and involved a lot of design iterations.
I created 10+ alternate designs for this feature and tested them with users over 8 weeks before we landed on this design
Patient Tags
Ability to group patients together for closer monitoring by adding Patient Tags
Lorem ipsum dolor sit amet, consectetur adipiscing elit, sed do eiusmod tempor incididunt ut labore et dolore magna aliqua. Ut enim ad minim veniam, quis nostrud exercitation ullamco laboris nisi ut aliquip ex ea commodo consequat.
Filter for Patient Tags
Ability to find patient groups based on tags assigned to them
Lorem ipsum dolor sit amet, consectetur adipiscing elit, sed do eiusmod tempor incididunt ut labore et dolore magna aliqua. Ut enim ad minim veniam, quis nostrud exercitation ullamco laboris nisi ut aliquip ex ea commodo consequat.
Impact
Opportunities for Improvement