Balancer: AI-powered Bipolar Medication Decision Support App

Balancer is nearing MVP (Minimum Viable Product) stage! In other words, we are close to releasing version 1.0 to the world. We are currently working on the ability to click on a medication result and view the source journal article(s) that were used by the system in its decision to rank the medication at that level in the list of possible choices. Knowing the rationale behind the rankings will provide prescribers with clarity, confidence, and materials for further research if they desire. This feature suggestion was brought to our team by our community partner, Dr. David Steinman, professor of psychiatry at UPenn.

Other tasks on our MVP roadmap include making the full list of journal article sources available to the user on a Data Sources page, updating the look and feel of our chatbot, and a few minor clean-up/enhancement efforts.

We are well on our way to achieving our goal of shortening the journey to medication stability for patients with bipolar disorder. However, we could always use some help.

Help wanted:

• Prescribers (psychiatrists, residents, nurse practitioners, physician's assistants) who treat patients with bipolar disorder and who are willing to consult with us or test our product
• People who might be able to refer us to the prescribers mentioned above
• Front end, back end, and full stack developers
• UX specialists, specifically UX designers and UX researchers
• Anyone interested in supporting us in research efforts, like locating journal article sources and uploading them to our system
• Anyone interested in our mission but not yet sure how to help

If you're interested in getting involved, please see the project description at https://codeforphilly.org/projects/balancer for how to get started. Thank you in advance for your support!

Exciting news: We have built and hosted a functional website (for medical personnel only, not open to the public) that allows psychiatrists and other medication prescribers to select patient characteristics on a form and submit the form to get a list of recommended medications. We are now focused on fine-tuning the form results by training our own AI model on medical journal articles centered around treatment for bipolar disorder.

Help wanted:

• Prescribers who treat patients with bipolar who are willing to consult with us / test our product
• Front end / back end / full stack developers
• UX specialists, specifically designers and researchers

If you're interested in getting involved, please join the #balancer channel on Slack and introduce yourself or email kjost@codeforphilly.org.

We have done several UX interviews with bipolar medication prescribers at UPenn, built the application front end, and are getting closer to an alpha testing phase. We are now working on perfecting the list of medications that get suggested when a prescriber selects various combinations of patient characteristics.

Help Needed

• Early-career prescribers, such as psychiatry residents, to test the application and validate our system's medication recommendations
• Non-technical volunteers to help with things like non-UX research and newsletter creation
• Donations for things such as help desk software, UX research interview transcription software, and marketing expenses: https://opencollective.com/code-for-philly/projects/balancer

Please email kjost@codeforphilly.org or join the #balancer channel in the Code for Philly Slack if you know of any way to help!

We have found a Technical Lead for the project and are actively recruiting other volunteers, such as:

Technical folks

• Data scientists
• AI/ML, full-stack, and backend developers
• UX researchers and designers

Non-technical folks

• Researchers (no specific skills required) to gather bipolar-related content from the internet, such as clinical trial reports, to feed the AI
• Psychiatrists/Medication Prescribers to be Subject Matter Experts (SMEs) and help us get this thing right!
• Anyone interested in mental health who thinks they could help

Please email kjost@codeforphilly.org or join the #medassist channel in the Code for Philly Slack to get involved.