How to Transition From Paper to Electronic Health Records

The short version: moving a practice from paper to an EHR takes most independent practices somewhere between four and eight weeks, start to finish, depending mostly on how much historical chart data you convert and how many providers need training. You do not have to scan decades of paper. You do have to decide what “the chart” means going forward, keep your paper records for as long as your state requires, and give your staff more training time than any vendor’s timeline suggests. Everything below is the long version.

Criterions has been moving practices off paper since 1991, and we still do it — it is not a story about the past here. Some of the practices that come to us have never had an EHR. Others have a system they stopped trusting years ago and quietly went back to paper alongside it, which is more common than the industry likes to admit.

Either way the question is the same, and it is almost never “which software.” It is: what actually happens to my practice while we do this, and what happens to the twenty years of charts in the back room.

What actually changes on day one

Less than most practices expect, and in a different place than they expect.

The visit itself changes least. A provider who has taken notes on paper for twenty years is still doing the same clinical thinking; the note is being captured somewhere else. What changes materially is everything that used to depend on someone physically handling a folder — pulling the chart before a visit, walking a lab result to a provider, finding out whether a referral letter ever went out, and answering a records request without going into the back room.

The front desk feels it first and most. So does billing, because the charge now comes off a documented encounter rather than a superbill someone deciphers afterward.

The honest tradeoff: documentation time per visit usually goes up before it goes down. Every practice experiences this and most vendors don’t mention it. It resolves as templates get configured around how your providers actually work — which is why the configuration step matters more than any feature comparison.

The steps, in the order they actually happen

Most published EHR implementation plans are written for hospitals and have eleven phases. For an independent practice it is closer to six.

1. Decide what “the chart” means going forward. Which fields you actually need, what a visit note should contain, and which of your paper forms have a digital equivalent versus which were workarounds for being on paper. This is the step practices skip and the one that determines whether the rest goes well.

2. Configure the system around that. Problem lists, visit templates, and chart layout get built to match how your providers document — not selected from a menu of defaults.

3. Decide what history moves. See the next section. This is a business decision, not a technical one, and it is where most of the cost lives.

4. Connect what the practice depends on. Labs, pharmacy, clearinghouse, and any outside system that currently reaches you by fax.

5. Train, then train again. First round before go-live, second round two to three weeks after — once staff have real questions instead of hypothetical ones. The second round is the one practices cut, and it is the one that matters.

6. Go live, with paper still available. Nobody should be prevented from writing something down in week one.

The first step is the one people get wrong. Practices ask “when do we start scanning” — that is step three, and starting there means you scan things you did not need.

What happens to your existing paper records

This is the question that stalls the decision, and the answer is more forgiving than most practices assume: you almost certainly should not scan all of it.

The practical approach most independent practices land on is a hybrid:

• Abstract forward, not backward. For active patients, key clinical data — problem list, medications, allergies, immunizations, recent results — gets entered into the new chart, usually at or just before that patient’s next visit. The work spreads across months instead of landing as a project.
• Scan selectively. Documents with ongoing clinical or legal relevance get scanned and attached. Twenty years of routine visit notes for a patient you last saw in 2011 generally do not.
• Keep the paper. Your existing records remain your legal record for whatever your state’s retention period requires. Going electronic does not release you from that, and it does not require you to destroy anything.

The alternative — bulk scanning an entire chart room up front — is expensive, slow, and produces a searchable pile of images rather than usable clinical data. Some practices still choose it, usually for space reasons rather than clinical ones. It is a legitimate choice, just rarely the one that pays off.

At Criterions, this part is handled in-house — our implementation team works directly with your staff to gather what’s actually needed rather than defaulting to converting everything, and what ends up moving is entirely your practice’s call, not a template we impose.

How long it takes

Most practices are live in four to eight weeks, start to finish. The variable is almost always how much history you convert and how many providers need to be scheduled around for training — not the software.

What reliably makes it take longer:

• Converting more historical data than the practice actually needs
• Training scheduled around a full patient load instead of blocked out
• Waiting on an outside system — a lab or clearinghouse connection — that was requested late
• Deciding what the chart should look like during implementation rather than before it

What it costs, and where the cost actually sits

Software licensing is usually the part practices estimate correctly and the smallest part of the total. The costs that surprise people:

• Data conversion, if you choose to convert a lot of history. This is the largest discretionary line and the one you control most directly.
• Provider time during training and the first few weeks after go-live, when documentation is slower.
• Temporarily reduced schedules around go-live, which many practices do deliberately and which is cheaper than the alternative.
• Ongoing costs for specific certified functionality — online patient service, direct messaging, and ePrescribing can each carry monthly costs, and that is true across certified vendors, not just here.

What should not be a surprise is the license itself. Ask any vendor what you are paying for that you will not use — with a bundled platform the honest answer is often “quite a lot,” and with modular licensing it should be “nothing.”

Training staff who have never used an EHR

The training that fails is the training delivered all at once, before anyone has a real question.

What works better is two passes. The first before go-live, covering the paths each role actually uses — a front desk person does not need the clinical documentation training a provider needs, and giving it to them wastes the attention you need later. The second two to three weeks in, once everyone has hit something specific and confusing. That second session is where the real questions surface, and it is the one that gets cancelled because everyone is busy.

Plan for one person to become the internal expert, whether or not you name them. In every practice somebody becomes the person others ask. Better to know who that is and train them deeper than to discover it three months later. Criterions’ implementation team runs this training in-house and stays involved through go-live, rather than handing you off to a separate training vendor partway through.

What ONC certification means, and why it matters when you’re leaving paper

ONC Health IT certification means an EHR has been tested by an accredited body against federal criteria — for data exchange, patient access, security, and clinical functionality. It is not a quality rating and it is not a review. It is a floor.

For a practice coming off paper it matters for a practical reason more than a regulatory one: certified systems are required to support standardized data exchange. That is what makes it possible for lab results to arrive in the chart, for records to move to another provider, and for your data to leave with you if you ever change vendors. A system that cannot exchange data in a standard format is a system your records are stuck inside — which is a worse version of the problem paper already gives you.

Criterions EHR is ONC Health IT certified. Full certification details, including criteria, Real World Testing results, and API documentation, are published on our certifications page.

Moving off paper when AI is already in the room

A practice going electronic in 2026 is making a different decision than one that went electronic in 2012, and it is worth being clear about why.

Ambient AI documentation — software that listens to the visit and drafts the note — changes the calculation for exactly the providers who resisted EHRs hardest. The historical objection to leaving paper was never really about paper. It was about typing during a visit, and about a screen between the provider and the patient. Ambient documentation removes that specific objection, because nobody types and nobody narrates: the conversation becomes the draft, and the provider reviews and signs it.

For a practice coming straight from paper, that means the workflow you are adopting may look less like the EHR your colleagues complained about a decade ago than you expect. It also means the sequencing question is real — some practices go electronic first and add ambient documentation once staff are comfortable, others start with both. Either is defensible; doing it accidentally is not.

Criterions supports DAX Copilot for ambient documentation, so a practice can add it once staff are comfortable rather than needing to decide on day one.

What goes wrong, and what it looks like when it does

The failures are predictable and mostly not technical.

• The chart was never defined. Configuration happened during implementation instead of before it, so the templates encode whatever seemed right that week.
• One provider never really adopted. They keep a paper workflow alongside the system, and now the practice runs two records — which is worse than either alone.
• Too much history was converted, slowly, by staff who also had a day job.
• Training happened once, before anybody had a real question.
• An interface was requested late, so results arrived by fax for the first two months and everyone concluded the system did not work.

None of those are software problems, which is why comparing feature lists is a poor way to choose. The question worth asking a vendor is not what the system does. It is what they do when the third item on that list happens to you.

A checklist you can actually use

Before you choose a system

• Write down what a visit note needs to contain, in your own words
• List every paper form you use, and mark which exist because you’re on paper
• Decide roughly how much history you want in the new chart
• List every outside system that reaches you — labs, pharmacy, clearinghouse, referral sources
• Confirm the system is ONC certified, and ask what the exchange standards actually mean for getting your data back out
• Ask what you’re paying for that you won’t use

Before go-live

• Chart layout and visit templates configured to your providers, not to a default
• Interfaces requested and confirmed working, not just ordered
• Role-based training completed — different for front desk, billing, and providers
• A second training session already on the calendar for two to three weeks out
• A reduced schedule for the first few days, deliberately
• Someone identified as the internal go-to person
• Paper still available, and nobody discouraged from using it in week one

After go-live

• Second training session actually held, not cancelled
• Templates revised based on what providers are working around
• Historical abstraction happening at visits rather than as a separate project
• Paper records retained per your state’s requirement

If you want a checklist-style walkthrough of switching between EHR systems specifically — not coming from paper — see our companion post, the EHR Transition Checklist.