The largest volume of anything a practice owns is usually its documents, and they do not ride along with the data. Your scanned charts, consents, insurance cards, and correspondence live as files, not as fields, so they sit outside the tiered conversion and move as a job of their own. It is the job clinicians touch every time they open a chart, and the one most often underestimated, because it is slow, exacting work that is easy to leave for last. This is what moving your documents actually takes, and where it goes wrong. It sits inside the full picture in how to prepare for an AdvancedMD data migration. This is the document job, in depth.
Why your documents do not ride along with the data
The conversion that moves your structured records is made for data that fits a template: who your patients are, who covers them, when they are scheduled. A document is not that. It is a file, a scan or an upload, and there is no field for it to land in. So every document you own has to be moved separately, matched to the right patient, and filed where it can be found, and none of that happens inside the conversion. A modest multi-provider practice can hold tens of thousands of these files, which is why the job is heavier than it looks from the outside, and why it needs its own plan rather than a line in someone else’s.
Getting the documents out of the old system
Before you can file anything, you have to get it out, and the old system rarely hands its documents over cleanly. The common outcome is a bulk export: a folder of files with the links to patients stripped away, so you get the documents and lose the filing that told you whose they were and what they contained. Now you have thousands of files named by whatever the old system called them, with no reliable thread back to the chart they belong to. Pulling a test export early, and looking at exactly what comes with each file, is how you find out whether the export carries the patient link and the document type or strips them, while there is still time to solve it.
The matching job is the whole job
Moving the files is easy. Putting each one back where it belongs is the work. Every document has to be tied to the correct patient, given a name that says what it is, and filed into a structure that keeps it findable a year from now when someone needs it for an audit, an appeal, or a records request. The tie to the patient runs on an identifier, and if your patient identifiers change in the move, the match has to be rebuilt against the new ones rather than assumed. Ten thousand documents that import as a string of identical filenames, attached loosely or not at all, are not records you can use. You have moved the files and lost the medicine, and the day a chart request comes in, the answer is somewhere in a thousand look-alike files that no one can search.
Not all documents are the same
Treating the whole pile as one kind of thing is how the important files get buried. Clinical documents are what a provider needs mid-visit and what an audit will ask for. Legal and consent documents are what protect you when a patient or a regulator questions what was agreed. Financial documents, the explanations of benefits and the correspondence with payers, are what you need to win an appeal or defend a balance. Each has a different reader and a different moment it gets pulled, so the filing structure has to make each one findable by the person who will reach for it, rather than merely present somewhere in the system.
Searchable, or just scanned
There is a difference between a document you can find by its contents and one you can only find by its filename, and it lasts for years. A flat image of a scanned page is invisible to search: the text on it cannot be found unless the page was run through character recognition that turns the picture of the words into words. Move your archive as flat images and every future search depends on someone having named and filed each file correctly. Move it as searchable text and the contents themselves become findable. For an archive you will hold for years and reach into under time pressure, that difference decides how usable the whole thing is.
The edges that catch people
A few situations do not fit the clean path and have to be decided on purpose. Documents attached to patients you chose not to bring across are now orphaned, and you have to decide whether those patients come after all or those files go to the archive with the rest of the history you left behind. Duplicate scans, the same insurance card uploaded three times over three years, multiply the volume without adding anything, so someone decides whether removing them is worth the effort or whether you carry them. And records under legal hold, tied to a dispute or a request in progress, cannot be archived away casually and have to be tracked through the move so they stay reachable.
Retention does not reset when you switch
Whatever you do not move still has to be kept. The same retention rules that govern the chart govern its documents, and they do not care that you changed systems. A records request or an appeal that lands two years after go-live has to be answerable from wherever those files ended up, which means the archive has to stay readable and retrievable for the full retention period, rather than merely present on a drive whose password someone forgets. An archive you cannot search, or cannot open, fails the moment someone needs it, and the moment someone needs it is rarely convenient.
Why this bites late, and needs its own owner
The document job has a habit of being noticed at the worst time. It is usually discovered at or after go-live, when a provider opens a chart mid-visit and the history is not there. The volume makes it slow, the matching makes it unforgiving, and because documents belong to everyone and to no one, they are the piece most likely to be left until last and then to run past the date. The fix is to give the documents their own owner and their own timeline, running in parallel with the data and the configuration from the beginning, rather than queued for after the data is done. When ownership is vague, this is the job that falls through, every time.
How to run the document workstream
Name an owner for the documents on day one, and treat the move as its own track. Pull a test export early and confirm what comes with each file, the patient link and the document type, so you know what you are working with. Decide the filing structure up front, by document type and by the reader who will need each kind, so files land somewhere findable rather than in one undifferentiated pile. Decide whether the archive will be searchable or flat, because it is far cheaper to choose before the move than to re-process afterward. And start early and run it alongside everything else, since the volume means it takes longer than any single other piece and cannot be compressed at the end.
Before you move the documents
On the document side, these should be true.
- The documents have a named owner and their own timeline, running in parallel from the start.
- A test export has been pulled and checked, so you know whether the patient link and document type come with each file.
- The filing structure is decided, by document type and by who will need each kind.
- You have decided whether the archive will be searchable text or flat images.
- Orphaned files, duplicate scans, and any records under legal hold are accounted for.
- Whatever you archive rather than move is confirmed readable and retrievable for the full retention period.
If the documents do not have an owner and a plan of their own, they are the part of the migration most likely to be found missing at go-live, by the provider who reaches for a chart and finds it empty.
Common questions
Do scanned documents transfer in a practice management data conversion?
No. The conversion moves structured records that fit a template. Documents are files, so they sit outside it and move as a separate job. Each file has to be tied to the right patient, named, and filed so it stays findable, and that work runs in parallel and needs its own owner.
How are documents matched to patients in a migration?
Each file is tied to a patient by an identifier. If the export strips the patient link, or your patient identifiers change in the move, the match has to be rebuilt against the new ones. Files that import without a reliable link arrive as a loose pile no one can search by patient.
What happens if documents import without being matched and named?
You get the files and lose the filing. Thousands of look-alike filenames attached loosely or not at all are not usable records. When a chart request or an audit arrives, the answer is buried in files no one can search, so the migration preserved the document and lost its usefulness.
Should my document archive be searchable?
For an archive you will hold for years and reach into under pressure, yes. A flat scanned image can only be found by its filename and folder. Text run through character recognition becomes findable by its contents. That difference decides how usable the archive stays over time.
What document types matter most in a migration?
All of them, but they serve different readers. Clinical documents are needed mid-visit and in audits, legal and consent documents when an agreement is questioned, and financial documents like explanations of benefits when you appeal or defend a balance. The filing has to make each findable by the person who needs it.
What happens to documents for patients I do not bring across?
They are orphaned. You decide whether those patients come after all or their files go to the archive with the rest of the history you left behind. Either way it is a decision to make on purpose, because orphaned files that are neither moved nor archived are simply lost.
Do I have to keep documents I do not migrate?
Yes. Retention rules govern the documents the same as the chart, and they do not care that you changed systems. Whatever you archive has to stay readable and retrievable for the full retention period, so a records request two years out can still be answered from wherever the files ended up.
How long do practices have to keep medical records?
It varies by state and record type, commonly several years for adults and longer for a minor’s chart. The rule follows the record, not the system, so changing software does not reset the clock. Confirm your own requirement before deciding what to archive rather than move.
How many documents does a typical practice have to move?
More than most expect. A modest multi-provider practice can hold tens of thousands of scanned files across charts, consents, insurance cards, and correspondence. The volume is why the document move takes longer than any single other piece and cannot be compressed at the end of the project.
When should the document migration start?
At the beginning, running in parallel with the data and the configuration. The volume makes it the slowest single piece, so queuing it for after the data is done is how it runs past the go-live date. Give it its own owner and its own timeline from day one.
Why do document migrations get underestimated?
Because documents belong to everyone and to no one, so no single person owns them until something goes wrong. The job is slow and exacting, and it is usually noticed only at go-live, when a provider opens a chart and the history is missing. A named owner and an early start prevent that.
Where this leaves you
Your documents are the largest, slowest part of a migration and the one your clinicians feel first. Give them an owner, decide the filing before you move a file, and keep the archive searchable, and the day a provider opens a chart the history is there instead of scattered across a thousand look-alike files.