CareLoop — booking that patients actually finished

The portal existed, technically. But it asked for a Medicare number before a preferred time, used words like "service stream", and ended on a confirmation screen with no date on it. Patients quite reasonably concluded it was easier to call.

Every call is a two-minute task for the patient and a six-minute task for reception, multiplied by ~180 appointments a week. The brief from the practice manager was "make the website better". The real brief, which took me a fortnight to understand: give reception their mornings back without abandoning the patients who'll always need the phone.

I shadowed reception for three full days before interviewing a single patient. That was the right call — the receptionists knew exactly where the portal failed because they fielded the fallout: "it said booked but there's nothing written down."

Then eight patient interviews, deliberately skewed toward low digital literacy, plus an audit of two weeks of call logs to categorise why people phoned.

I ran the lo-fi screens past Dee — nine years on that reception desk — before any patient saw them. She crossed out my "appointment type" label in about four seconds: "Nobody knows what that means. Ask them what they're coming in for." That phrase shipped unchanged.

The UI is intentionally quiet: one calm teal, large type, nothing animated. The constraint that shaped it most wasn't aesthetic — it was Medicare claiming rules, which dictated the order I could collect details in. More than any design principle did, honestly.

Five patients (three over 60) and both receptionists, moderated, on the clinic's own iPad and their own phones.

Three months after launch: phone bookings down from 71% to 38%, no-shows from 18% to 11%, and reception estimated about nine staff-hours a week back. Dee's verdict — "the phone's quieter and the ones who call are the ones who should" — is the metric I'm proudest of.

What I'd do differently: I designed for the patient first and treated staff as an edge case. Halfway through I had to invert that — the 'not sure' path, the flagging, the printable slip all came from reception's reality. If I ran it again, staff workflows would be in the brief from day one, not week four.

And a small humility note: the patients who'll always call still call. The goal was never to eliminate the phone — it was to make ringing a choice instead of a workaround.