What You Need to Start a Private Psychiatry Practice

To see your first private-practice patient you need six things: an active license, your DEA registration, malpractice insurance, a HIPAA-compliant way to meet (video for telehealth), a way to document and prescribe, and a way to take a card. That is the whole list. An LLC, a lawyer, an accountant, a polished website, and an office all feel mandatory and are not. They are things you add after you have patients, not before, and the practices that launch fastest are the ones that book the patient first and build the systems around the gap.

The instinct to set up everything before going live is the single most common way new owners lose six months. Almost every "must-have" that delays a launch is an avoidance trap dressed up as diligence. Below is what is actually required, what only feels required, the order to do it in, and the one item, e-prescribing, that genuinely takes weeks and has to be sequenced around.

What you actually need before your first patient

Six jobs have to be covered before patient one. Most prescribers already hold the first two.

An active medical license in the state where the patient sits during the visit. For telehealth, that is the patient's location, not yours.
Your DEA registration, with the address set correctly. Register the address where you actually practice. For a home-based telehealth practice, your home address is allowed; you can keep it non-public and list a P.O. box as the visible one. Virtual-office addresses get rejected.
Malpractice insurance, bound before the first visit. A claims-made policy is fine to start; just know that claims-made is cheaper now and bills you a tail premium of roughly one to two times your annual rate when you leave, so the cheap option has a delayed cost built in.
A HIPAA-compliant way to meet. For telehealth that is a compliant video platform. Free tiers like Doxy.me work, and a HIPAA-upgraded Zoom works. You do not need anything fancier to start.
A way to document and prescribe. Notes can be on paper for the first few weeks and scanned in later. Prescribing is the one piece with a real lead time, covered in its own section below.
A way to take a card. Capture a card at intake and charge it the day of the visit. This single habit is what pushes cash-pay collection rates to nearly 100%, against the 70% to 80% that is considered good when you bill insurance or carry balances.

Two boring items belong here too because they are annoying to change later: a dedicated phone number and email, and a fax route. Pharmacies and labs still fax, so you need a way to send and receive one, but a fax-to-email service that signs a BAA covers it. Sort these early and forget them.

That is the real list. Everything past it is optional at launch, and most of it is an excuse to keep preparing instead of practicing.

What feels required but isn't (and when to add it)

Here is the part the standard guides get wrong. They hand you one undifferentiated list of fifteen items and imply you need all of it before you open. You don't. These are the things that feel mandatory, why each can wait, and when to actually do it.

Feels required	Why it can wait	When to actually do it
An LLC or PLLC	You can legally see patients as a sole proprietor. Incorporating does not shield you from malpractice claims, so it is a tax and structure decision, not a prerequisite.	When revenue is steady (a common trigger is around $10k a month, where an S-corp election starts to pay for itself). See do you need a PLLC.
A lawyer	Most solo cash-pay launches need no attorney upfront. The contracts you actually use (consents, policies) come from templates you adapt.	Just-in-time, for a specific question (a tricky non-compete, a state entity rule).
An accountant	You will not have anything to account for until you have revenue.	At your first tax season, or when you are weighing an S-corp election.
A polished website	A three-page site, or even a directory profile, is enough to get patient one. Perfectionism here stalls people for months while an ugly site that exists gets patients.	Once you have a few patients and know who they are.
An office or lease	Telehealth makes a physical office optional entirely. If you want in-person, sublet or borrow a room.	Never sign a multi-year lease at launch. Add space only when demand is proven.
A perfect EHR configuration	Make 80%-good-enough decisions on forms, scheduling, and templates, then fix the leaks when they appear with real patients.	Revisit and optimize around month six to twelve, not before.
An after-hours phone line	A standing answering line for a small cash panel typically rings with nothing urgent for a year. Confirm your coverage expectations with your malpractice carrier, then skip it.	Probably never, for a small solo telehealth practice.
A concierge or membership model	Patients often don't perceive the value, and plain, honest fees tend to out-earn a bundled membership.	Only once you have a full panel and a specific reason.

The pattern underneath the table: until you have seen patients, you do not actually know what your practice needs. One psychiatrist discovered her early panel was almost entirely women in their twenties and thirties with generalized anxiety, so she went deep on that instead of pre-learning all of integrative psychiatry. You cannot plan that from a spreadsheet. You learn it from patient three.

The order to set it up

Sequence matters because one item has a long lead time and a couple of others gate the rest. Do them in this order.

Start the slow credentials first. Bind malpractice and confirm your DEA address now, because these have real lead time and everything else waits on them being right.
Begin e-prescribing enrollment immediately. This is the longest pole in the tent (see the next section). Starting it on day one, in parallel with everything else, is the difference between launching on schedule and waiting on a token.
Stand up the patient-facing core. Your video platform, intake forms with card authorization, a way to take payment, and a place to write notes. This is the part you can have working in a day or two.
Wire up the boring infrastructure. Phone, email, and a fax route.
Test the whole thing end to end with a fake patient. Cover this in the launch section below; it is the step most people skip and most regret.
Book and see patient one. You can bridge non-controlled prescriptions while e-prescribing finishes provisioning.

The reason to decompose it this way is psychological as much as logistical. Every "simple" task on this list (register a DEA, set up forms) hides about ten sub-steps, and trying to swallow them whole is how the setup phase becomes death by a thousand paper cuts. Do one sub-step a day. Expect the full build to take a couple of months alongside a job. It is one-time, front-loaded work, and it ends.

The e-prescribing lag, and how to bridge it

E-prescribing is the one item that will not be ready in two weeks, so plan around it instead of waiting for it. Setting up e-prescribing with EPCS (the credential for electronically prescribing controlled substances) involves identity proofing, a token, and PDMP registration, and the identity-proofing step alone commonly takes a few weeks. New owners are routinely surprised by this because every other piece of the stack comes up in a day.

You can bridge part of the gap, but not all of it. For non-controlled medications, you can phone or fax prescriptions to the pharmacy, and you can hand a paper prescription to an in-person patient. That covers a real share of a general psychiatry panel and lets you start seeing patients while your e-prescribing provisions in the background.

What you cannot safely bridge is controlled substances. EPCS is required to e-prescribe controlled medications, and the workarounds that exist for routine prescriptions do not extend to stimulants or benzodiazepines in most settings. (Whether you can prescribe those controlled substances by telehealth at all is a separate, moving question, covered in the current DEA telehealth rules.) So sequence honestly: if your intended panel leans heavily on controlled prescribing, get EPCS live before you book those patients, or start with non-controlled cases first and add the controlled-substance patients once the credential clears. Do not promise a stimulant refill you are not yet set up to send.

This is also where an all-in-one platform earns its keep. The course community spent years stitching prescribing onto their EHR with a separate tool because their practice software had no e-prescribing at all. Eureka has built-in e-prescribing and EPCS, so the prescription writes from the same chart that holds the note and the intake, and there is no second prescribing app to bolt on and reconcile. You still complete identity proofing, so the lead time is real either way; budget for it and start it first.

When do you actually launch?

You launch when your systems are tested and working, not when your schedule is full. Those are two different milestones, and waiting for the second one is how people delay for a year. A new practice does not need a pipeline of patients to open. It needs a few that can get through the door cleanly.

The test that tells you the door works is simple: run your own booking-to-intake flow as a fake patient before you accept a real one. Book a dummy appointment, fill out your own intake forms, walk the calendar and the reminders and the payment step end to end. People who do this consistently catch several broken automations they would otherwise have discovered on a real patient: a reminder that never fires, an intake packet that doesn't gate the appointment, a card field that doesn't save. Fixing those in a test run costs nothing. Fixing them on patient one costs trust.

After that, go. One PMHNP moved her launch up by a full month the moment her systems passed that test, on the simple logic of "just get the patient in the door." That is the right instinct. The emotional arc of the first year is real (the first few months feel like terror, then impostor syndrome, then somewhere past month six, quiet confidence), and the way through it is patients, not more setup. Plan for a realistic ramp of one to two new patients a week and roughly six months to a livable paycheck, and do not let a slow start convince you the practice is broken.

The throughline of every launch story worth learning from is the same. One clinician saw her very first paying patient with no merchant account and no office furniture, then set up billing afterward. Another committed to a consult for a condition she had not treated in years, booked it a few days out, and used the gap to prepare. None of them had everything ready. They had enough to see the patient in front of them, and they built the rest from real demand. The fastest way to a finished practice is to start an unfinished one.

If you want the dollars-and-cents version of this list, the companion piece breaks down what it costs to set up a practice line by line. For the patient side, getting your first patients and what to say on the consult call cover where they come from and how to convert them, and what to charge sets the fee behind your "way to take a card." Two of those, your NPI setup and your entity question, are the only paperwork that is genuinely yours to get right early.