Between Visits

The problem

A child in a developmental follow-up program can see six or more specialists in a single clinic visit, then not return for three to twelve months. The families most likely to lose the thread in that gap are the ones facing income, transportation, or language barriers, and they are exactly the families an app-download-and-password product will never reach.

What it is

Between Visits works over text messages, because everyone already has them. After a visit, the family gets a secure link to a plain-language recap: the main takeaways and each specialist’s instructions. Between visits, about seven short check-ins arrive by text. Answers that signal trouble, low confidence, a barrier, a therapy delay, become exception-based alerts for the care team, who also get a pre-visit summary dashboard before the family returns. Everything runs in English and Spanish. No app, no login, no password.

My role

Product and technical lead: the spec, the architecture, and the build, including the unglamorous parts like carrier compliance registration for the SMS layer.

Built for the environment it has to enter

Health systems do not adopt tools that require re-architecture, so the architecture anticipates the destination: designed to slot into Epic after-visit summaries, hospital single sign-on, and HIPAA-eligible hosting without a rebuild. The pilot runs on synthetic data while those integrations are earned, which is the correct order of operations for anything touching patient information.

What I would do differently

Nothing structural yet, which is itself a statement: this is the newest of my projects and the first one where compliance constraints shaped the architecture from the first commit rather than being retrofitted. The honest open question is whether exception-based alerting stays quiet enough for care teams to trust it, and that is what the pilot exists to find out.