A pill schedule board for my father

My father takes nine medications on four different schedules, two of which changed after a hospital stay. We had a paper chart on the refrigerator, and it worked until the day it did not — he took an evening dose twice because the morning column had been ticked by my mother and he did not see it. I built this over about two months of evenings. It is not a health product and I would be uncomfortable calling it one; it is a chart that happens to run on a phone.

It solves one narrow problem: knowing what has already been taken today, by anyone in the house. The existing apps I tried were built for a person managing their own medication, with reminders and adherence streaks, and none of them handled the actual situation — three people who might administer a dose and one person who should not have to remember. The screen shows today only, in large type, with each dose either taken or not, and who marked it. It is for my father, my mother and me.

Everything, including the parts I would not have chosen to build. The schedule model is more involved than it looks — some doses are with food, one is every other day, and one was tapering on a fixed schedule for six weeks, which is a rule I had to represent rather than hardcode. I built the sync so all three phones agree, the offline behaviour, and the large-type interface, which went through four versions with my father as the only tester. I also maintain it, which means updating the schedule whenever his doctor changes something.

It shows today and nothing else. My first version had a week view and a history, and watching my father use it made clear that any extra screen is a place to get lost — history now exists but only I can see it. I made marking a dose require a deliberate press-and-hold rather than a tap, because an accidental tap is exactly the failure I was trying to prevent, and my father is fine with the extra half second. I also chose local-first storage with sync, so the app works completely with no signal; their house has poor reception and an app that needs the internet to tell you your medication is useless.

The hardest part was that my father did not want it. He experienced the paper chart as fine and the app as an accusation, which I had not anticipated at all and which no amount of interface work fixes. What changed it was making my mother the primary user for the first month — she marked the doses, he saw the screen — and letting him start marking his own once it was clearly just a chart. The technical work was the easy half; I spent longer on that than on the tapering schedule.

There has not been a double dose in nineteen months, against three in the six months before. My mother says the thing she values is not needing to ask him whether he took the evening one. It has been through two schedule changes from his doctor and both were about ten minutes of my time. It runs on three phones and will never run on a fourth.

My mother, who tested every version honestly and told me when a screen was worse. My father, eventually. His pharmacist answered two questions about how the tapering schedule should be represented, which I would have got wrong.