Notes from replacing a practice management platform

What we learned migrating a chiropractic practice off Tebra: what had to move, what did not, and why both systems ran side by side for a week.

The constraint that shapes everything

A practice cannot close for a migration. Patients are booked weeks ahead, the front desk is busy from eight, and the schedule is the business. Every decision follows from that.

It rules out the clean cutover where you switch on Monday and hope. It also rules out a long parallel period, because two systems taking bookings is a double-booking machine.

What actually had to move

Less than expected. Patient records and future appointments had to come across intact. Historical appointments mattered for recall timing but did not need to be live in the new system on day one. Billing history stayed where it was.

Deciding this early is most of the work. The instinct is to move everything, and everything is where migrations go to die.

Checking the move

We exported, imported, then compared the two schedules appointment by appointment for the four weeks ahead. Not a spot check. A full diff, because one missing appointment is a patient standing at a desk.

Mismatches were mostly appointment types that did not map cleanly, and a handful of records with duplicate entries that had existed for years without anyone noticing.

The overlap week

For one week both systems were open, but only one of them took new bookings. The old platform was read-only, kept purely so the desk could check anything that looked wrong.

That distinction is the whole trick. Parallel running fails when both systems can write. It works when one is a reference.

What surprised us

The intake forms were a bigger win than the scheduling. Under the old system intake arrived as a document that someone read and retyped into the chart, on the day, with a patient waiting. Writing intake straight to the record removed a job nobody had ever counted as a job.

The reminder change was second. Not the fact of reminders, which existed before, but being able to set the timing and wording to match how the practice actually talks to its patients. That is the part the practice cited when no-shows fell by sixty per cent.

What we would do differently

Spend longer on appointment types before exporting anything. Almost every mismatch traced back to a type that meant two different things depending on which provider used it, and untangling that after import is far more painful than before.