A practice with fifteen years of paper files decides to go digital, and assumes that means fifteen years of files being typed in. The quote comes back, the timeline comes back, and the project stalls before it starts.
Almost none of that data needs to be in the new system on day one, and some of it never does.
Split the archive by use, not by age
The question is not how old a record is. It is how likely somebody is to need it at the desk, quickly, in the next year.
Migrate: active patients, meaning anyone seen in roughly the last one to two years, plus anyone with an ongoing condition regardless of last visit. Also current balances, current HMO details, and any allergy or adverse reaction record, which must be available instantly and is dangerous to leave in a cupboard.
Retain accessibly: everything else, in an organised physical archive with a documented retrieval process. If a patient returns after five years, you retrieve their file once and enter what is still clinically relevant at that visit.
Consider scanning without transcribing: a middle path that is often the right answer. Documents are scanned to PDF and attached to the patient record, searchable by patient rather than by content. Far cheaper than structured data entry, and it means the file is on screen when the patient is in front of you.
What must be structured
A scan is a picture. You cannot report on it, filter it, or have the system warn you about it. A small set of fields genuinely needs to be typed as data rather than captured as an image:
- Patient identifiers and contact details
- Allergies and adverse drug reactions
- Current active problems and current medications
- Outstanding balances and HMO enrolment
- Anything you will need to run a report on
Everything else can start as an attachment and be structured over time, as patients present.
Enter the rest as patients arrive
This is the approach that works in practice. On the day a returning patient attends, whoever registers them enters their core details properly. Within a year the active population is fully in the system, spread across ordinary working days, with no separate data entry project and no overtime bill.
Check the data before you move it
Migration is the one moment when duplicates, dead phone numbers and misspelled names are easy to catch. Moving a mess into a new system produces a new system containing a mess, and it will be blamed on the software.
Keep the paper until you are sure
Do not destroy original records on the strength of a migration until it has been verified against a sample, and never before your retention obligations allow. Retention requirements exist independently of what system you use, and the archive has to satisfy them whether or not the data is also digital.