I went to the presentations for the Health 2.0 Code-a-thon at YetiZen in San Francisco, and it’s always inspiring to see a bunch of designers, developers, researchers, and medical professionals in one room working on a number of innovative healthcare apps. This year’s Code-a-Thon was themed “Power to the Patient” and was focused on patient empowerment through healthcare data. With that focus in mind, it’s clear that there are a thousand ways to extract meaning from patient data to improve patient-provider relationships, reduce the need for follow-up care, make any follow-up care more impactful, and more — the question is what’s the best method to achieve such an impact?
I thought this challenge was executed and organized very well, in that developing a finished, working app wasn’t a hard requirement, but rather a bonus. In a two-day event, requiring working code like a traditional hackathon or developer challenge stifles creativity and out-of-the-box thinking. If you jump into code, you skip insightful thinking about user (patient) experiences, and you risk coding for the sake of winning a challenge instead of thinking critically about solving problems — in healthcare, those are problems that actually hurt and kill people. So, kudos to the organizers on focusing on clear deliverables as more conceptual creations.
The Code-a-thon was split into two tracks: (1) “crossing the chasm between consumer devices for wellness and the rest of the healthcare system,” with an objective of creating a “Health Tile” for a variety of clinical scenarios to help improve care; or (2) “redesigning the Explanation of Benefits” with new types of data, highlighting difficult use cases like when patients change plans or have invalid diagnoses.
The judges were Rebecca Mitchell Coelius, MD (Health and Human Services), Lenny Lesser, MD (Palo Alto Medical Foundation), James Gilbert (Optum Cloud Services at UnitedHealth), Zac Jiwa (Health and Human Services), Stacey Chang (IDEO), Ward Detwiler (Henry Ford Innovation Institute), Drew Schiller (Validic), and JeanLuc Neptune (Health 2.0).
Here’s some of the solutions that were presented, in the order presented:
Archimedes showed their app ‘In4Sure’ which helps patients choose an optimal plan within state Health Insurance Exchanges (HIEs) by analyzing their personal health data. A patient provides information and can upload Explanation of Benefit (EOB) documents, and the patient is presented with plans sorted by the one that is “best for you” — based on ranking price, quality, convenience, cost over time, and whether or not their current doctor is in-network with the plan.
Prime is an iOS app that’s a personal social network for health. Using Validic integration, it allows a patient to view and interact with their medical data in an easy to comprehend way. The app was actually built-out, using Django, postgres, Celery, and Chart.js. The team showed a health tile that presented live patient FitBit data trended over a period of weeks, in an easily digestible manner for patients, a care team, or providers to see.
Patient Watch‘s app gathers physiological data from wearable devices to allow healthcare providers to monitor adverse reactions to any procedure or drug. Because common side effects and adverse events are known in advance, the app can use live health data to monitor and alert the provider early. The team cited the example of Prozac, which causes an increase of 12000% suicidal tendencies. A judge asked if patients might be nervous about releasing their data, and they said “we can do a lot of good if [patients] open [up their data.]”
HeartHealthLink uses BlueButton data to gather a baseline for a heart failure patient’s overall health, and uses weight trackers connecting via Validic to trigger alerts for sudden changes in weight securely transmitted via the HIPAA-compliant HealthChatLink. The app runs passively in the background and warns patients when they miss a check-in. The team thinks the model could be expanded to more conditions, providing a simple ‘stoplight’ of green/yellow/red health data to providers.
Light Hearts, recognizing that 25% of patients going into the hospital for congestive health failure treatment end up going right back in within 30 days without having seen a cardiologist or Primary Care Provider (PCP), created an app that creates a health tile that can communicate data back to a PCP relating a tracked metric like weight — a doctor can provide instructions so that if the weight crosses threshold boundaries, an alert is triggered. Asked whether it’s actionable for a provider, the team said: open-ended questions are a black hole for a care provider trying to make clinical decisions in seven minutes. [Doctors] need to have a way to have quantified data not on ten pieces of paper, but in a format they can look at and say ‘hey, this is already actionable.’
Clinical Browser Spade allows medical students and doctors to donate their browser history as it relates to clinical data to an app that analyzes patterns of learning clinical information and concepts. The app searches browser history for medical terms and medical dictionary codes (like ICD-10), selecting only history items relevant to medical learning and not containing personal information. Then, knowing the path of websites that providers best learn medical concepts, a doctor can prescribe URLs to patients to help them also understand those concepts.
Edge Interns presented Mindmentor, which allows patients to screen for mental health issues, providing a clean visual of symptoms and progress over time. A provider can assign activities or questions, and the patient can journal their thoughts, feelings, and moods which are analyzed using NLP to extract highlights. In the future, social support networks could be added to the platform.
(To be continued: My Best Life, WTF! Denied?, MintMD, Meowware HDK, AkibaH Health, Healthy.Be, PBBP)