medcorder: Helping People Talk About Health

When about two years ago my dad was diagnosed with the prostate cancer that ended up killing him, we had no idea what to do. There was a lot of information coming at us, fast. Most of this information was related to my dad in conversations with doctors. It wasn’t written down in any kind of a clear way and contained a lot of medical jargon, possible conditions, measurements, treatment options, and clinical trial specifications.

My dad had a very straightforward and effective solution: he asked doctors for permission to record the conversations on his phone and then sent them (generally as email attachments) to his children. I’d then usually take point on listening to each recording, transcribing the key content, and emailing back the findings and next steps. My siblings would then research different options based on the transcript. It enabled us to be my dad’s “health team”.

Seeing both how effective and rare this approach was, I started imagining an application that would let people do this effortlessly - record & transcribe a conversation with a doctor, share it immediately with loved ones - and allow messages and pictures (e.g. of test results) to be exchanged in this setting as well. I partnered with a dev team to build this as a side project and today it launched on iOS in the US App Store - the Android version is shipping shortly.

I learned a lot in the past year, working with Evrone.com to explore using new cross-platform frameworks like Flutter (including contributing an audio recording module to the community), design language, and flows. We decided to use a Firebase backend with Google’s speech recognition API. The Evrone team introduced me to a broad swath of design tools and process like Trello, Zeplin, Loom, Marvel, InVision, and more.

We hit a turning point in February when we had a first version of the app put together in beta and I realized it was useless - the primary “object” was the recording but there was no real notion that there might be multiple recordings shared with the same set of folks. In reality what we had seen with my dad was a need to share out a bunch of recordings all with the same people and sharing the same context. So I bit the bullet and decided to rescope the app to discussion “rooms” around a health condition where recordings and their transcripts would be instantly shared with the same team of folks.

There’s more work to be done - realtime transcripts featuring speaker diarization and full punctuation for readability, automated summarizations, specialization of vocabulary to properly detect medical terms, support and automatic detection of other languages, push notifications on room updates…but what we ship today is plausibly useful, so I want it out there in the world.

Many thanks to those of you who helped test the app in beta. This app is built to honor my dad. I hope that this incarnation of his good idea can help others navigate their own difficult health conditions.

Much love,

• David