How to Actually Pick an EMR (and Survive the Switch)

The short version

• Picking an EMR is three decisions and one project: what you need (your cosmetic split ranks the modules), what it really costs (the whole stack, not the license), who to trust (references, the good, the bad, and the ugly), and the switch itself.
• Start with the split: cosmetic-to-reconstructive in plastics, cosmetic-to-medical in derm. Rank your seven modules, buy for your top three, and let the physician pick the documentation style.
• Price year one, not the monthly sticker: tiers, add-ons, AI pricing, integration fees, clearinghouse transaction fees, implementation, training, and data migration.
• The switch is a project: eight-plus weeks is sane, clearinghouse enrollments start 30 days ahead, plan on running two systems for two to three months, and yes, the physician trains too.
• Support is not implementation, and it isn’t training either: post-launch, the help desk fixes what’s broken but teaching staff costs hours and often invoices. Name a superuser, protect their time, and ask up front where support ends and paid training begins.

How to choose an EMR: run a method, not a demo schedule

Most practices pick an EMR the same way: book five demos, get dazzled five times, pick whoever dazzled last, then spend a year finding out what the dazzle didn’t cover. There’s a better order of operations, and it fits on an index card.

Picking an EMR is three decisions and one project. What do you actually need? (Your revenue answers that, not the feature list.) What will it really cost? (The year, not the month.) Who do you trust? (People running it, not people selling it.) And then the project: the switch itself, which has ended more honeymoons than any feature gap ever has.

Standing disclaimer, standing true: we use these platforms daily, we like the companies behind them, and this method isn’t about avoiding bad vendors. There aren’t many bad vendors. It’s about avoiding a bad match, which happens to good practices signing with good vendors every single week.

Cosmetic split first, then rank your EMR modules

Before anyone demos anything, write down one number: your cosmetic split. Cosmetic-to-reconstructive revenue in plastics, cosmetic-to-medical in derm. A 90/10 cosmetic practice needs quoting, photos, memberships, and a killer patient app. A 30/70 reconstructive or medical practice needs eligibility, claim scrubbing, denial workflows, and reporting with real depth. Same specialty, opposite shopping lists, and the split decides which one is yours.

Then rank the seven modules: quoting and consult conversion, photo management, documentation style, scheduling, billing depth, POS and inventory, CRM and patient app. Rank them by where your revenue actually comes from, not by what sounded impressive at a conference. Buy for your top three. The platform that nails them beats the platform that’s decent at all seven, and we mapped who nails what in the shortlist guide.

One module deserves its own sentence: documentation style is the physician’s call, nobody else’s. Checkboxes, ambient scribe, or tap-to-code… ask the doc, watch the doc’s face, and weight the answer heavily. The person charting eight hours a day outvotes the committee. That’s not politics, that’s math.

What an EMR really costs: the year-one math nobody quotes

The quote says $500 a month. Year one says otherwise. Build the real number before you compare anything: base tier (and which modules it actually includes), the add-ons you’ll genuinely use, AI pricing (per-note or per-month, in writing), integration fees for every connection you’re planning, clearinghouse transaction fees if you bill insurance, card processing rates if you don’t, implementation and training fees, and data migration from your old system. That’s the year-one total. Compare those.

Then negotiate like you mean it: fight for the core tier you need plus the specific add-ons you want, and don’t jump to the elevated package because the rep made it the path of least resistance. Bundles are how every industry sells; unbundling is how smart buyers purchase.

And keep the whole exercise in perspective, because we watch practices agonize here for months: we ran the spread math in the shortlist guide, and the gap between the cheapest and priciest system is small against a single case. The pricing work isn’t about saving $200 a month. It’s about never meeting a surprise add-on, which is a different and far more valuable thing.

EMR demos and reference calls: your specialty, your scripts, the good, the bad, and the ugly

Now the demos, on your terms. Rule one: your specialty, your visit types. An ophthalmology scribe demo tells you nothing about a rhinoplasty consult, and a generic patient tells you nothing about your Tuesday. Rule two: bring scripts. We’ve published two that work: the AI questions (what’s shipped versus roadmap, dates not adjectives) and the integration questions (does connecting cost money, what’s exposed, can a third party actually get in). Watch how the room handles being asked. That reaction is data.

Then references, the audited market data of this industry. Take the vendor’s reference calls… those practices are happy for a reason, and were picked for a reason. Then work your own network: find practices your size, your specialty, on the platform, and ask what drives them nuts. You want the good, the bad, and the ugly, because you’re going to inherit all three. Anyone who only hears the great opinions bought the brochure twice.

The EMR transition plan: timelines, training, and the clearinghouse cut-over

We wrote the first version of this advice back in 2021, and the original has aged annoyingly well: changing systems is one of the most painful things a practice does, and the pain is mostly self-inflicted through scheduling. The 2026 update, with the scars still fresh on our side of the table:

1. Give it eight-plus weeks, and budget the hours, not just the weeks. Eight or more, you’re calm. Four to eight, you’ll make it with bruises. Under four, you’re volunteering for problems. And the calendar is only half of it: implementation eats real hours from real people… vendor calls, template building, workflow testing, data checks. Plan on your project lead giving this five to ten hours a week for the duration, and take something off their plate to pay for it. When that time isn’t budgeted, it gets stolen from corners you’ll rediscover in your denial queue. Expect a quarter of the learning to happen on live patients no matter what you do, so schedule lighter clinics through launch month.
2. Name a superuser, and treat it like the permanent job it is. One person who learns the system deeper than anyone, sits every implementation call, and owns the build. Here’s what nobody tells you: this is not a launch-week role. Your superuser retrains every new hire you make for the next decade, absorbs every feature release, and fields every “how do I” so the vendor’s ticket queue doesn’t. Pick someone who’s staying, protect their hours, and pay attention to what that loyalty is worth. A practice with a real superuser and an average EMR beats a practice with a great EMR and nobody home.
3. Everyone still trains. Everyone. The part-time aesthetician, the scheduler who’s retiring soon, and yes, the physician. A doc who skips implementation calls becomes a full-time cleanup project for the team, in perpetuity. Be the system captain or formally appoint one… those are the only two options that work.
4. Assign workflow owners by seat. Front desk, clinical, scheduling, billing. Have each owner test their workflows front-to-back with test patients before go-live, because one well-meaning scheduling shortcut can quietly break recall workflows, billing QA, and your post-op metrics in a single move. Junk in, junk out, forever.
5. Start clearinghouse enrollments 30 days early, minimum. Enrollment lag is the classic cash-flow ambush: no enrollment, no claims out the door, no money in the account. And plan on running the old and new systems side by side for two to three months while old AR drains… shutting the old system off at go-live is how receivables go to die.
6. Budget for the export. Records out of the old system typically cost real money and arrive less structured than you’d hope. This is exactly why the exit-terms question belongs in every contract before you sign it, a theme by now.
7. Support is not implementation, and support is not training either. The implementation manager who knows your name disappears at go-live, and the support queue that replaces them troubleshoots what’s broken… it does not teach your new front-desk hire the scheduler. That’s training, and vendors draw the line between a support ticket and a training request in different places, often at an invoice. It’s a murky line, and post-launch frustration lives on it. So ask before signing: what training is included after go-live, what does new-hire or refresher training cost, and where exactly does support end? Then budget internal training hours forever… which, conveniently, is your superuser’s job security. Ask month-six customers how this actually plays out, not week-two customers still in the honeymoon.

The EMR selection checklist, start to signature

The complete run, in order. Print it, or make your admin recite it.

1. Write down your cosmetic split.
2. Rank the seven modules; circle your top three.
3. Let the physician pick the documentation style.
4. Shortlist three platforms that nail your circles.
5. Build each one’s year-one total: tier, add-ons, AI, integrations, clearinghouse, implementation, migration.
6. Demo in your specialty with the AI and integration scripts.
7. References: the vendor’s list, then your friends. The good, the bad, and the ugly.
8. Exit terms in the contract: export format, cost, timeline.
9. Calendar the auto-renew notice window the day you sign.
10. Name your superuser and protect their hours… it’s a permanent role, not a launch role.
11. Then run the switch like a project: eight-plus weeks, enrollments 30 days early, everyone trains, two systems until the old AR drains.

There is no best EMR, there’s the best fit… and now you have the method that finds it. Run it once, properly, and you won’t run it again for a decade. That’s the whole pitch for doing this right.

Frequently asked questions

This article is for general informational purposes and is not coding, billing, or legal advice. Verify current rules and your contractor policies before making operational decisions.