In 2019, a friend tuned me in to the limitations of the mental health industry - too many needs and too few practitioners. With the effects of the pandemic impacting everyone's lives, this overwhelmed support system needs new tools to better serve the community and lessen the burden on practitioners.

What it does

Healthy Mind allows patients to self-monitor and triage their conditions through audio journaling. These journals are processed through various machine learning algorithms ascertain aspects like positive or negative sentiment, personality traits, and key elements. The app generates a report from this info and stores in in the patient's account for review by their doctors or clinicians. Trends can be identified to alert practitioners that further efforts may be required, or that a patient may be doing well without the need for repeat visits.

The app also provides a mindful experience with pleasing visuals, sounds, and messaging to help put a user's mind in a calm and healthy state. The app can suggest and initiate these sessions if the user's journaling assessments are consistently, or suddenly change to negative.

How I built it

I created a native iOS app in Swift. The app interfaces with the DrChrono service using the supplied REST API. I used various iOS frameworks to record, transcribe, and process the user's audio journal. It uses CoreML to run machine learning models on the transcribed audio to make its various assessments. The app builds a PDF report from these assessments and writes it to the user's DrChrono account using the REST API.

The app also uses native Swift code to generate a mindful experience. It uses core graphics to create pleasing visuals, it plays calming audio, and animates positive messages.

Challenges I ran into

The DrChrono API was not well documented as far as expected parameters, and I did a lot of trial and error to figure out correct properties and formats. The OAuth interface doesn't seem to support iOS deep link callbacks, so I had to find workarounds. Integrating various ML models and processes took a lot of research and time.

Accomplishments that I'm proud of

I'm proud that I was able to work through the various challenges I ran into. The app works well, and I think offers a good user experience.

What I learned

I learned a great deal about the DrChrono REST API. I also learned a lot about processing audio on iOS and utilizing various ML models to process the results.

What's next for Healthy Mind

Healthy Mind needs a more complete method to log in a user. Right now I'm just using one test user to demonstrate the app operation. I would like to get feedback from clinicians on what they would like to see in the generated reports.

Share this project: