MedMind is a personalized population health management platform driven by artificial intelligence. The platform provides general wellness and chronic disease management, personalized health recommendations, and appointment scheduling for patients and family members, while allowing care coordinators to identify at-risk patients. MedMind was created in collaboration with Martin Shapiro.
The cost of Medicare patients is rising with chronic disease management. Congestive heart failure alone accounts for 34% of all Medicare costs—over three times the average cost for Medicare patients.
The Centers for Medicare and Medicaid Services (CMS) created a 5-star rating system that provides federal bonus payments to qualifying health plans. But between 2015 and 2017, Humana's rating has fallen from 4.2 to 3.6; this decline has lost them roughly $240 million in Medicare bonuses.
For Humana's Health Innovation Challenge, Martin and I conceptualized MedMind, an artificial intelligence platform that increases Medicare profits through personalized patient engagement and predictive population health management. We projected that MedMind would yield a 20% admission reduction at $312m in savings per year.
On the patient side, the platform keeps users up to date on preventative healthcare like breast cancer screenings and flu shots that can be scheduled directly. MedMind also partners with Circulation to manage transportation logistics for appointments.
MedMind includes a chronic disease program to manage congestive heart failure. The platform checks for red flag symptoms and drastic changes in vital signs, weight, and exercise; these trends are automatically measured and synced via connected devices.
In the provider web portal, care coordinators can identify patients who need the most intervention. The messaging feature suggests educational resources and specialist referrals that care coordinators can attach directly in their messages.
The messaging feature also includes the option to automate educational resources and auto-populated responses to patients based on their messages. The platform performs A/B tests to optimize methods of engagement— text to email to phone call—as well as types of content.
We aimed for familiarity—even for people who weren't familiar with mobile applications. And we tested the old-fashioned way: Martin sat in the waiting room of a hospital, asking strangers to move through our prototype and provide feedback.
As a medical application, MedMind needed to be usable by all people with a diverse range of abilities. We quickly disregarded common design patterns like like clickable tiles or hamburger menus. Instead, we opted for obvious action indicators like text labels and button outlines, making them as simple and obvious as possible.
We carefully selected our color scheme without placing too much emphasis on the concept of color altogether. The limited color palette highlights different parts of the platform as a visual component, but no parts of the platform are dependent on distinctions between colors.
Our goal was to create an engaging application that people could navigate with ease. The primary concern was making sure that we kept our content as readable and concise as possible while still providing sufficient information and presenting enough warmth and friendliness in the platform's messaging.