A therapist referral network is usually the cheapest and highest-converting patient acquisition channel an interventional psychiatry clinic has, and it is the one most clinics run by accident. Therapists sit with treatment-resistant patients every week, they cannot prescribe, and TMS, Spravato and IV ketamine all work better alongside ongoing therapy, so the incentive to send you patients already exists. What is usually missing is a system: a defined list of partners, a reason for them to trust your clinical judgment, a referral path that takes them under two minutes, and a progress update that comes back after the patient starts. This article covers how to build each of those, including the outreach sequence, what belongs in a referral packet, the anti-kickback and HIPAA rules that constrain what you can offer a referring provider, and how to track whether any of it is working.
One caveat before the tactics. Referral development is slow. Most clinics that stick with it see meaningful volume somewhere between six and twelve months in, and the ones that quit almost always quit in month three. If you need patients this quarter, Google Ads for TMS clinics will get you there faster. Referrals are what you build so that in two years you are not renting all your patient flow.
Why therapists are the right first target
Psychiatrists refer in higher volume. Primary care has more patients. So why start with therapists?
Because the relationship is not competitive. A psychiatrist referring a patient for TMS is handing over a case they could arguably manage themselves, and that hesitation is real even when nobody names it. A therapist referring the same patient loses nothing. They keep the therapy relationship, their patient gets access to something they cannot provide, and the combined outcome tends to be better than either alone. You are solving a problem they actually have: the client who has been in the room for a year, is doing the work, and is not getting better.
Therapists are also far easier to reach. There is no gatekeeper, no referral coordinator, no health system procurement process. Many are solo practitioners who answer their own email.
The trade-off is volume. A single therapist might send you two or three patients a year. That is fine. The model is a wide base of low-volume, high-trust referrers rather than a handful of big accounts. Twenty-five active therapist relationships producing two patients each is fifty patients a year from a channel with no media cost.
Start with the referrers you already have
Before you contact a single stranger, work your existing chart.
Most of your current patients have a therapist. Most of those therapists have no idea their client is being treated at your clinic, or how it went. Every one of them is a warm contact who has already, implicitly, trusted you with someone they care about.
Two changes make this systematic. Add three questions to intake (do you currently see a therapist, who is it, and have you worked with anyone previously) and capture the answers as structured fields rather than free text in a note, so they are actually retrievable. Then build the habit of closing the loop with those named providers, which is covered further down.
This is the highest-yield work available to most clinics and it costs nothing. A therapist who watches their own client improve under your care is a more committed referrer than anyone you will ever meet at a networking event.
Building the target list
For new relationships, the useful sources are boring and public:
- Your payers’ provider directories, filtered to behavioral health in your service area. This gives you therapists who already share your patients’ insurance, which removes a common referral blocker.
- Psychology Today’s directory, which most therapists maintain and which lets you filter by specialty. Look for depression, treatment-resistant depression, OCD, PTSD and mood disorders.
- The International OCD Foundation directory if you offer TMS for OCD.
- Group practices, which are more efficient per contact than solo clinicians: one relationship can reach eight therapists.
- IOP and PHP programs, which discharge patients who need ongoing care and often have nowhere good to send them.
Keep the list small. Fifty names you contact six times beats five hundred you contact once, and the second approach is what most clinics do. A named person should own the list, and it should live somewhere with dates and outcomes attached rather than in someone’s head.
Fix the intake path before you generate demand
A referral relationship dies on a single bad handoff. The therapist sends someone, the patient calls, nobody calls back, the patient reports this in their next session. That therapist will never refer again and will likely mention it to colleagues.
So audit three things first. How long does it actually take you to call a new inquiry back, measured from timestamps rather than from what the process document claims? How many steps does a referring provider have to complete to send you a patient? And can your front desk explain what TMS and Spravato are without reaching for a brochure?
The referral path itself should be one page, one fax number, one direct phone line, and a named human whose job it is to answer it. If a therapist has to create a portal account, you have already lost. We covered the downstream version of this problem in why TMS consult leads don’t book, and the same failure modes apply to referred patients, arguably worse, because a referred patient arrives with someone else’s credibility attached.
The compliance rules that shape all of this
This section exists because most referral marketing advice written for clinics ignores it entirely, and the exposure is not theoretical. None of what follows is legal advice, and the analysis genuinely turns on your specific facts — payer mix, entity structure, state law — so this is a conversation for your own healthcare counsel. But you should know the shape of the problem before you design a program.
The Anti-Kickback Statute is the one that matters most
The federal Anti-Kickback Statute (42 U.S.C. § 1320a-7b(b)) makes it a crime to knowingly and willfully offer, pay, solicit or receive anything of value to induce or reward referrals of items or services payable by a federal healthcare program. Two features make it broader than clinics expect. “Remuneration” means anything of value, not just cash. And courts have read it so that if one purpose of the payment is to induce referrals, the statute is implicated even if there are other legitimate purposes.
It applies to anyone, not only physicians, and violations can pull in the False Claims Act, because claims tainted by a kickback are treated as false claims. Civil monetary penalties reach $100,000 per kickback plus three times the remuneration under 42 U.S.C. § 1320a-7a.
There is no dollar threshold under AKS. This is the part people get wrong. There is no gift amount that is automatically fine.
Stark is narrower, but its dollar limits are a useful yardstick
The physician self-referral law, known as Stark, is a strict liability statute, and intent is irrelevant, but it is narrower. It applies to referrals by physicians for designated health services to entities they have a financial relationship with, so whether it reaches a given outpatient psychiatry arrangement depends on the facts.
Its exceptions do, however, tell you what regulators consider a small enough courtesy to be uninteresting. CMS sets these limits annually, and for calendar year 2026 the nonmonetary compensation limit at 42 C.F.R. § 411.357(k) is $535 per physician per year in aggregate, with medical staff incidental benefits under § 411.357(m)(5) capped at less than $46 per occurrence. Even if Stark does not apply to you, those numbers are a reasonable internal ceiling to work to.
Fitting inside a Stark exception does not protect you under AKS. OIG restated this directly in its fraud and abuse FAQs, noting that providing things like sporting event tickets and other entertainment to referral sources would violate the statute where the intent is there, regardless of whether a Stark exception is satisfied. The two laws are separate tests and you have to pass both.
What this means in practice
The line to hold is that you are giving referring providers information and access, not value.
| Generally low risk | Get counsel before doing |
|---|---|
| Clinical education, one-pagers, candidacy criteria | Any per-referral payment, in any form |
| A modest working lunch during a genuine educational meeting | Free or discounted services, staff or space for referrers |
| Progress updates on shared patients | Gifts, tickets, travel, meals beyond nominal value |
| Being easy to reach and fast to respond | Marketing or admin support provided to a referring practice |
| Genuine co-authored or co-presented CME | Paid medical directorships or consulting with a referral source |
Medical directorship and consulting arrangements with referral sources are not automatically improper, but they have to be at fair market value for real services actually performed, documented in writing, and not tied to referral volume. OIG has issued fraud alerts specifically about these arrangements. Do not structure one without counsel.
HIPAA permits more than most clinics think
A lot of clinics avoid sending progress updates to referring therapists because they assume they need a signed release for every disclosure. Under HIPAA, that is generally not the case.
45 C.F.R. § 164.506(c)(2) permits a covered entity to disclose protected health information to another healthcare provider for that provider’s treatment activities without patient authorization. HHS is explicit about this in its permitted uses guidance on exchange for treatment, which also notes that “treatment” is defined broadly at § 164.501 to include consultation between providers about a patient and the referral of a patient from one provider to another. A therapist who continues to treat your shared patient is squarely within that.
Four things narrow it, and they matter in psychiatry more than in most specialties:
- Psychotherapy notes are carved out and require specific authorization under § 164.508(a)(2). Your treatment summaries should not draw on them.
- Substance use disorder records from a Part 2 program are governed by 42 C.F.R. Part 2, which is stricter than HIPAA and does generally require consent.
- State law is frequently tighter than HIPAA on mental health records, and the stricter rule wins.
- Marketing is not treatment. A newsletter promoting your services is a different activity from a clinical progress note, and mixing them into one communication is how a permitted disclosure becomes an unpermitted one.
Most clinics still collect a release at intake, and that remains sensible practice. It manages patient expectations, covers Part 2 and state law variation, and avoids an awkward conversation later. Just know that the treatment permission is the reason you can communicate at all, and asking a patient’s permission to keep their therapist in the loop is an easy conversation, not a barrier.
One correction worth making, because it circulates widely: fax is common in healthcare and is a permitted channel, but it is not “end-to-end encryption” and describing it that way is wrong. Standard fax has real confidentiality weaknesses, mostly misdialing and unattended machines. Use it because your referral partners use it, confirm the number, and cover-sheet everything. For anything electronic, use a secure channel with a business associate agreement in place where one is required. Ordinary consumer email is not that channel. We take the same line on patient-facing communication in asking psychiatry patients for Google reviews without breaking HIPAA.
The outreach sequence
You will see “it takes seven touches” repeated constantly in referral marketing. Treat it as a rough prompt to be persistent rather than a finding. It comes from general sales folklore and there is no healthcare-specific evidence base behind the number. What is reliably true is that one contact produces almost nothing and clinics give up far too early.
A workable sequence over a first quarter:
| Stage | What you send | Purpose |
|---|---|---|
| Week 1 | Introductory letter from the medical director, faxed and emailed | Establishes clinical credibility and who you are |
| Week 2 | Referral packet with candidacy criteria and the referral path | Makes referring possible before they are ready to talk |
| Week 4 | Request for a short call or virtual meeting | Clinician-to-clinician contact, which carries the real weight |
| Week 6–8 | Something genuinely useful: new evidence, a capacity update | Stays visible without asking for anything |
| Week 10–12 | Invitation to a lunch-and-learn or case discussion | Converts awareness into an actual conversation |
| Ongoing | Monthly-ish contact, plus progress updates on shared patients | Maintenance |
The introductory letter should fit on one page and lead with clinical substance, not marketing. Who your medical director is and where they trained. What you treat and with what. Which insurers you are in network with. Your current wait time to first appointment. This is the single most persuasive line in the letter, because a therapist’s real problem is that everywhere they call is booked out three months. Then a specific ask: a fifteen-minute call.
What goes in the packet: a one-page candidacy sheet for each treatment, written for a clinical reader rather than a patient; a short note on what to tell a client about what the treatment involves; the referral form and direct line; insurance and typical authorization criteria; and a named contact with a real phone number.
Lunch-and-learns, done properly
These work about as well as the follow-up around them. Without it you have bought someone a sandwich.
Keep the meal modest and incidental to a genuine educational session — that is both the compliance posture and, frankly, what therapists prefer. The session earns the time; the food is not the draw. Virtual works well and costs less; deliver lunch to the practice and meet by video over their break.
Structure forty-five minutes roughly like this: ten minutes on who you are and how you work; twenty on genuine clinical content, ideally the referral decision itself: when has a client failed enough medication trials to be a candidate, what does the evidence actually support, what are the realistic response rates; ten on logistics, walking through exactly how a referral happens; and five asking directly whether anyone has a client in mind right now.
That last question is the one most clinics skip, and it is the one that produces referrals. People agree in principle and then forget. Ask them to think of a specific person before they leave the room.
Closing the loop is the whole game
If you do one thing from this article, do this one.
When a therapist refers a client, the referral usually vanishes. They do not hear whether the person was seen, whether they qualified, whether anything helped. From their side it feels like sending someone into a void, and it makes the next referral less likely.
Send three communications on every referred patient. An acknowledgement within a day or two confirming you received the referral and have made contact. A short note after the initial evaluation with the plan and expected timeline. And an end-of-treatment summary with outcomes: symptom scale movement, response, and what you recommend going forward.
Keep them short and clinical. A therapist reading that a client they sent you moved from a PHQ-9 of 21 to 8 over a course of TMS has just watched their clinical judgment pay off, in writing. That is more persuasive than any brochure, and it costs you a few minutes of a coordinator’s time.
Make it a workflow with an owner and a trigger, not a good intention. It is the first thing to disappear when a clinic gets busy, and it is the last thing that should.
The referral conversation differs by treatment
Clinics tend to pitch “interventional psychiatry” as one thing. Therapists are deciding about one patient at a time, and the three main modalities present very differently to them.
For TMS, the therapist’s questions are about time and coverage. Daily sessions over roughly six weeks is a serious ask of a working client, and insurers typically require more failed medication trials than the evidence strictly supports. Tell them your usual authorization criteria and your current wait time and you have answered most of it.
For Spravato, the constraint is the REMS program. Patients are monitored on site for two hours after dosing and cannot drive home, which is a logistics conversation the therapist will have with their client long before you meet them. Explain the visit structure plainly. Our comparison of IV ketamine versus Spravato marketing covers why these two get confused constantly, including by referrers.
For IV ketamine, the honest issue is cost. It is generally not covered, and a therapist who refers a client into an unexpected several-thousand-dollar conversation will not do it twice. Be direct about pricing in your materials. Therapists are protective of their clients’ finances and will respect the candor.
What to track
Referral programs get cut because nobody can show they worked. Four numbers prevent that.
Active referring providers, meaning those who sent at least one patient in the last twelve months. Referrals per active provider. The conversion rate from referral to started treatment, which should run well above your paid channels. If it does not, the problem is your intake, not your outreach. And time from referral to first appointment, because it is the number your referrers experience directly.
Capture referral source as a required field at intake, and ask the patient rather than relying on the paperwork. Patients name the person who told them about you, which is often not whoever signed the form.
A CRM helps but is not the point. A shared spreadsheet with contact dates, referral counts and next actions, reviewed monthly by someone whose job it is, beats an expensive system nobody updates. The discipline is what matters.
What kills referral relationships
In rough order of how often it happens:
- A referred patient who never gets called back. One instance ends the relationship permanently.
- Silence after the referral. No acknowledgement, no outcome, nothing.
- Taking over the therapy. If a therapist suspects you are absorbing their client rather than co-treating, referrals stop immediately. Say explicitly, in writing, that you expect the client to continue therapy with them.
- Overpromising outcomes. Interventional treatments have meaningful non-response rates. A therapist who sends three clients on the strength of an inflated claim and sees no benefit will conclude you are a salesperson.
- Wait times you did not disclose. If you are eight weeks out, say so.
- Contacting them only when volume dips. It is transparent, and it is the thing most clinics do.
A realistic first year
Months one and two are internal: fix intake, build the packet, add the intake questions, work your existing patients’ therapists. Months three to five are first contact with a list of forty to sixty. Months five to eight bring the first meetings and the first referrals, which will feel disappointingly few. Months nine to twelve are where compounding starts, as early referrers who received progress updates begin sending second and third patients and mentioning you to colleagues.
Anyone promising a referral pipeline in ninety days is selling you something. The channel is slow, and that slowness is exactly why it holds — a competitor cannot outbid you for a relationship the way they can outbid you on a keyword. It works best alongside the faster channels rather than instead of them, which is the argument in patient acquisition for TMS clinics, and it should sit inside a plan rather than run as a side project, which we cover in how SEO fits into a healthcare marketing plan.
If you want help building the outreach system, the referral materials and the tracking behind it, that is the work CuraReach does for interventional psychiatry clinics. Book a strategy call and we will start with the referrers already sitting in your patient records.
Frequently asked questions
Can I pay a therapist for referring patients to my clinic?
No. Paying for referrals of patients covered by any federal healthcare program implicates the Anti-Kickback Statute, which prohibits offering anything of value to induce referrals and carries criminal exposure plus civil penalties. There is no safe per-referral amount, and many state laws extend similar prohibitions to commercially insured and cash-pay patients. Build the relationship on clinical value and access instead, and take any proposed financial arrangement to healthcare counsel first.
Do I need a signed release to send a progress update to a referring therapist?
Under HIPAA, generally not. 45 C.F.R. § 164.506(c)(2) permits disclosure to another provider for that provider’s treatment activities without authorization, and HHS defines treatment to include referral between providers. Psychotherapy notes, substance use disorder records under 42 C.F.R. Part 2, and stricter state mental health laws are the exceptions. Most clinics still collect a release at intake as good practice.
How long does it take to build a therapist referral network?
Expect first referrals around months five to eight and meaningful, compounding volume between nine and twelve months. Clinics that work their existing patients’ therapists first tend to see something sooner, because those relationships are already warm.
How many referral partners should an interventional psychiatry clinic have?
Aim for a working list of forty to sixty contacts and twenty to thirty active referrers rather than a large directory you cannot maintain. Therapists refer in low volume individually, so the model is breadth of relationships with consistent contact, not a few large accounts.
Can I buy lunch for a referring provider?
A modest meal incidental to a genuine educational meeting is common practice, but there is no fixed safe harbor amount under the Anti-Kickback Statute, which turns on intent. The Stark limits CMS publishes each year are a useful internal ceiling: $535 per physician annually in nonmonetary compensation for 2026, and under $46 per occurrence for medical staff incidental benefits. Keep it modest, keep the education real, document it, and get counsel on your specific program.
What is the difference between therapist referrals and psychiatrist referrals?
Psychiatrists send higher volume but hesitate more, because referring a treatment-resistant patient means handing over a case they could arguably manage. Therapists send fewer patients each but refer more readily, since interventional treatment complements rather than replaces what they do. Most clinics get traction faster with therapists and should build that base first.
