Asking a Psychiatric Patient for a Public Review: A Close Look at Built-In Reputation Tools

EMR conversations in psychiatry almost always stop at the same three places: EPCS, note flexibility, and whether the telepsychiatry video is good enough to conduct a mental status examination through. Those are the right questions, and they shape the clinical day more than anything else on a feature list. The feature nobody thinks to ask about during a demo is the one that quietly determines what a prospective patient finds when they search for a psychiatrist in their city at eleven at night, and that is the post-visit feedback and reputation module. Several vendors have started building this loop directly into the chart rather than leaving it to a separate service, and Hero EMR ships one of the more fully specified versions of it, which makes their documentation a useful specimen for thinking through a problem that is genuinely harder in psychiatry than it is anywhere else in medicine. Asking a satisfied patient to leave a public review is unremarkable in dermatology and routine in primary care. In psychiatry, the same request asks a patient to attach their real name, permanently and searchably, to the fact that they are under psychiatric care, and that is not a small thing to ask of someone who came to you in part because the rest of their life does not know.

Why This Feature Lands Differently in Psychiatry

A public review is a disclosure. When a patient posts about a mole check, they have revealed that they have skin. When a patient posts a warm paragraph about their psychiatrist under a photograph and a real first name, they have told every future employer running an idle search, every acquaintance, and every family member with a grievance that they are in psychiatric treatment. Most patients tapping through a thank-you screen are not weighing any of that in the moment, and the practice that placed the link on that screen has, without intending to, arranged for the patient to disclose something the practice itself is legally forbidden from disclosing on their behalf.

That asymmetry is worth holding onto, because it is the thing that should shape how you configure one of these systems, and it is not a reason to refuse to use one. Patients have every right to speak publicly about their own care, plenty of them want to, and there is a real access argument on the other side of the ledger: the psychiatrist with four reviews and the psychiatrist with sixty are not equally findable, and the patient who gives up after two search results does not get treated at all. The point is not that the ask is wrong. The point is that the ask deserves more deliberate configuration in psychiatry than in a specialty where the only thing a review reveals is that someone had a knee.

There is also a second, quieter reason this category matters more to psychiatric practices than the marketing suggests, and it has nothing to do with reviews. A structured post-visit question, sent to every patient, is one of the few mechanisms that will surface a rupture in the therapeutic alliance while it is still repairable. A patient who felt dismissed at a fifteen-minute medication check will very rarely tell you so at the next appointment. They will simply not come to the next appointment. A survey that catches that at week six is worth more to a psychiatric practice than any number of five-star reviews, because it addresses the actual failure mode of the specialty, which is silent attrition rather than public complaint.

How the Loop Actually Works

The architecture is consistent enough across vendors that it is worth describing once in concrete terms, using the Hero EMR implementation as the reference because its documentation is unusually specific about the rules. A visit is marked complete. If the patient is eligible, the system schedules a feedback invitation on the channel, at the delay, and inside the daily send window the practice has configured. The patient receives a short branded email or text carrying a single link, which opens a mobile survey with no login, no app, and no portal account required. They answer, the response is scored, and the score decides which of three screens they land on.

A positive outcome shows a thank-you screen carrying the practice's public review links, one tap from posting to Google or Healthgrades or wherever else the practice keeps a profile. A negative outcome shows a brief acknowledgment that the practice takes the feedback seriously and will follow up, and it never shows the review links; instead it raises an internal alert so someone can call. Anything in between is neutral and gets a plain thank-you and nothing further. The whole design rests on that three-way split, and every configuration decision you make is really a decision about where the two thresholds sit and who lands on which screen.

The workspace itself is organized into four tabs, and the documented setup order is deliberate: build the survey, set the delivery rules, add the review destinations, send yourself a demo, and only then flip the master enable switch. That last sequencing detail is more useful than it sounds. The demo card sends you the genuine invitation message built from your current settings, the link opens a live survey with a preview banner, and nothing you do inside it is recorded, counted in your statistics, or capable of triggering an alert. You can submit a scathing one-star demo at midnight and nobody gets paged. Demo invitations also work while the feature is switched off, which means you can design the entire thing, experience it exactly as a patient would, and share the link with a colleague for a second read before a single real patient is ever contacted.

Access is gated by a single dedicated permission rather than by general administrator status, and the documentation is explicit that being an administrator alone does not grant it, because this workspace controls what the practice publishes about itself and who gets nudged toward a public review. Organization owners hold it automatically, and an administrator can grant it to one specific staff member as a reputation-only permission without handing over anything else. If your practice requires clinical two-step verification, the workspace honors it. For a small psychiatric practice where the office manager is also the biller and sometimes the scheduler, that granularity is the difference between being able to delegate this and having to sit in it yourself.

The Prompt Mode Decision, and the Compliance Problem Underneath It

The single most consequential setting in any of these systems is the one that decides which completing patients are shown your review links. Hero EMR calls it the prompt mode and offers two positions. All patients means everyone who finishes the survey sees the review links regardless of how they scored. Positive only means the links appear exclusively for patients whose response scored above the positive threshold, and everyone else is quietly routed away from your public profiles.

Positive only is the setting every reputation vendor in the market has been selling for a decade, usually under a friendlier name, and it is the setting you should think hardest about before enabling. The Federal Trade Commission's rule on the use of consumer reviews and testimonials, which took effect in October 2024, addresses review suppression directly, and the agency has been consistent in its view that a body of reviews assembled by selectively soliciting only the customers you expect to be happy can mislead the consumers who read it. Whether any particular configuration crosses the line is a fact-specific question your own counsel should answer rather than a vendor, but the direction of regulatory travel here is not ambiguous. It is worth noting that Hero EMR surfaces a compliance warning inside the product when you select positive only, and recommends prompting all completing patients as the safer choice, which is a mildly unusual thing for a vendor to do about its own feature and a reasonable signal about how the risk is understood.

For a psychiatric practice specifically, there is a practical argument on top of the legal one, which is that the gate buys you very little. Your panel is small, your visit volume per patient is high but your patient count is low, and the arithmetic difference between prompting everyone and prompting only your happiest patients is likely to be a handful of reviews a year. You are accepting a genuine regulatory exposure in exchange for a rounding error, and you are doing it in a specialty where the reviews you are trying to accumulate are the ones patients are least willing to write in the first place. Set it to all patients, accept that some neutral responders will look at your Google link and decline, and keep the part of the system that actually earns its place.

Designing a Survey That Does Not Cost You the Alliance

The survey builder in these systems is generally more capable than you need it to be, which is a trap. Hero EMR allows up to eight questions across four types: a rating scale with optional endpoint labels, a single-choice question with between two and ten options, a simple yes or no, and a free-text comment box, with a hard cap of two free-text questions so the survey stays quick. Question text can carry placeholders that fill in the provider's name, the provider's last name, and the practice name, so the wording reads as though it were written for that specific patient and that specific psychiatrist rather than assembled by a marketing department.

Use three questions. A psychiatric patient who has just spent thirty minutes discussing the worst parts of their life does not owe you a six-question instrument on the drive home, and completion rates fall off a cliff past the first screen anyway. One rating scale that asks about the thing you actually care about, one short question about access or the practical experience of reaching the office, and at most one free-text box is a complete survey.

Be deliberate about what you do not ask. Do not ask whether the patient's symptoms have improved. You already measure that with the PHQ-9, the GAD-7, the ASRS, and whatever else you are running on a schedule inside the chart, those instruments are validated and this one is not, and blending a satisfaction survey with an outcome measure produces a number that is useless for both purposes. More importantly, a question about symptoms invites the patient to type clinical detail into a free-text box that then lives in a reputation workspace rather than in the chart, where it is visible to whichever staff member holds the reputation permission and where it is considerably harder to find again when it matters. Ask about the experience of the visit and leave the clinical measurement where it belongs.

How the outcome gets decided is a separate choice. You can designate one key question, typically the recommend-us rating, and let its answer alone drive the result, or you can blend several scorable questions into a weighted average. For psychiatry, one key question is almost always the better structure, because a weighted average will dilute the patient who rated the relationship question at zero and everything else at eight, and that patient is precisely the one you needed to hear from. Then set your two thresholds, remembering that the negative cutoff is not a grading curve but a tripwire: set it generously enough that ambivalence reaches you, because ambivalence in this specialty is what attrition looks like a month before it happens. One useful design detail is that surveys are versioned, so editing and saving creates a new version while any invitation already in flight keeps the exact version it was sent with, which means a patient never sees the instrument change underneath them and your reporting stays tied to what each person actually answered.

Delivery Rules That Matter More in Psychiatry Than Anywhere Else

The delivery tab is where a system like this either fits psychiatric practice or quietly embarrasses you, and several of the settings deserve more thought than the defaults will give them. Start with timing, which offers either a delay measured in minutes after the visit is completed or a fixed time on a later day. The hour-later option is popular because the visit is fresh, and it is the wrong choice for psychiatry. A patient who has just disclosed something difficult, or who left a session in tears, or who was told that the medication is not working and the plan is changing, should not receive a satisfaction ping in the parking lot. Send at a fixed time the following day, or the day after, and accept a slightly lower completion rate in exchange for not being the practice that asked how it went ninety minutes after a hard session.

The cooldown setting is the one most likely to be misconfigured, because it was designed for specialties where the same patient returns twice a year. You see stable patients every four to twelve weeks and unstable ones every week or two, so a short cooldown means you will survey the same twenty people repeatedly and hear from nobody else, which is both irritating to them and statistically worthless to you. Set it long, in the range of six months, so a given patient is asked once or twice a year at most. The system independently enforces that a patient can only ever have one live survey at a time, which helps, and it automatically deactivates a link if the underlying visit is later reopened or cancelled, so feedback is only ever collected against a genuinely completed encounter.

The age rule deserves a flag of its own for anyone doing child and adolescent work. Only patients eighteen and older are surveyed, minors are never invited, and the check fails closed, so a missing or uncertain date of birth is treated as ineligible rather than waved through. That is the correct default for a consent question this thorny, but it also means that in a pure child and adolescent practice the module will sit almost entirely inert, because it never surveys the guardian who actually experienced the scheduling, the waiting room, and the prior authorization. If your panel is mostly minors, this feature is not going to do for you what it does for an adult practice, and you should know that before you spend an afternoon configuring it.

Exclusions are set by visit type and by provider, which covers the most important cases: exclude your intakes, exclude anything you have coded as urgent or crisis, and exclude any visit type where a feedback request would be tonally wrong. What you cannot do is exclude an individual patient, and in psychiatry there is always at least one patient for whom an automated message inviting them to rate their doctor is a genuinely bad idea. The practical workaround is to build that exclusion at the visit-type level, which is imperfect and worth raising with any vendor you are evaluating. The remaining delivery settings are more straightforward: channels can be email, text, or both, with each patient's own communication preferences and opt-outs overriding your setting so an opted-out patient is never texted, and every message carries STOP instructions. Invitations only go out inside a daily window you set in your practice timezone, and that window must sit inside platform-wide safe contact hours of 8:00 in the morning to 8:30 at night, with each patient's quiet hours honored on top. Link expiry is configurable from three to sixty days, links are single-use, and an expired or already-used link shows a friendly page rather than a broken one.

Negative Feedback Alerts Are Not a Clinical Risk Pathway

When a response scores negative, the system raises an in-app alert, optionally emails selected staff, and marks an unacknowledged-alerts badge on the workspace and on the admin menu so somebody notices without having to go looking. Alert emails deep-link to the specific response rather than dropping the recipient at a dashboard, and the alert stays open until a person acknowledges it, which is the mechanism by which the team signals it has been handled. The recipient picker only lists staff who already hold the reputation permission, so alerts cannot be routed to someone who was not already trusted with the workspace. All of that is well built, and for most of medicine it is the end of the analysis.

In psychiatry it is not, because of what patients write in free-text boxes. A comment field attached to a mental health visit will eventually contain a sentence expressing hopelessness, or a statement about not seeing the point of continuing, or something more explicit. That comment arrives in an administrative queue, not in your clinical inbox. It is not routed to on-call, it does not page anyone, it may sit until the next business morning, and depending on how you have delegated the permission it may be read first by an office manager rather than a clinician. None of that is a defect in the product, which is doing exactly what a reputation tool is supposed to do, but it is a real operational hazard if you switch this on without thinking it through.

Decide the escalation rule before you enable delivery, write it down, and make sure whoever clears the badge knows it. The rule can be simple: any free-text comment containing a statement about self-harm, hopelessness, or safety is forwarded to the on-call clinician immediately and the patient is contacted the same day through your normal clinical pathway, not through a service recovery script. In a small practice, the cleanest answer is often to keep the reputation permission with a clinician rather than delegating it to the front desk, accepting the extra few minutes a week that costs. And it is worth stating the obvious counterpart: this queue is not a safety net, it must never substitute for your actual risk pathway, and a patient in crisis is not going to communicate that through a satisfaction survey. The point is only that if it does happen, the message should not sit in a marketing tool over a weekend.

The Response Trap, and Why the Private Half Is Worth More Than the Public Half

There is one more reason the alert loop matters more than the review links in this specialty, and it is the thing most practices discover far too late. When a negative public review does appear, you cannot meaningfully respond to it. The moment you write anything that confirms the reviewer was a patient, including a sentence as mild as thanking them for coming in or noting that your records show something different, you have disclosed protected health information about an identifiable individual to the entire internet. The Office for Civil Rights has pursued enforcement against practices for precisely this, and the settlements are not small. A psychiatric practice has less room here than most, because the underlying disclosure is more damaging and the temptation to correct the record is stronger.

The practical consequence is that your only genuine lever against a bad review is preventing it, and preventing it means reaching the unhappy patient before they reach a keyboard. That is exactly what the negative outcome branch of this loop is built to do, and it reframes how the whole feature should be evaluated. A practice that judges a post-visit feedback system by how many Google reviews it produced in the first quarter has measured the wrong half. The right measure is how many alerts were raised, how quickly someone called, and how many of those patients were still on the schedule three months later.

That reframing is also what makes the feature defensible in a specialty where the public ask is fraught. You are not primarily running a review generation campaign. You are running a structured, scheduled, rate-limited check-in that happens to include a review link for the patients who are pleased and who choose, on their own, to say so publicly. Configured that way, with the gate off, the cooldown long, the intake and crisis visit types excluded, and the escalation rule written down, it is a reasonable thing to switch on in a psychiatric practice. Configured as a lead generation machine, with an hour-after-visit send and a positive-only gate, it is a liability wearing a friendly interface.

What to Ask Before You Switch Any of This On

If you are evaluating platforms rather than configuring one you already own, the questions worth asking are narrower than a feature list suggests. Ask whether the feedback loop lives inside the EMR or is a separate service with its own business associate agreement, because a bolt-on means a second vendor holding survey responses that identify your patients by name. Ask how survey responses are access-controlled, and specifically whether general administrator status is enough to read them or whether it takes a dedicated permission. Ask what the minimum age behavior is and whether it fails closed. Ask whether exclusions can be set by visit type and by provider, and whether there is any way to exclude an individual. Ask whether editing a survey affects invitations already in flight. And ask whether you can send yourself the real invitation on the real channel before the feature is enabled, because a vendor that cannot let you experience the patient side without exposing a patient to it is telling you something about how carefully the rest of it was built.

Hero EMR answers most of those questions well, which is why it is worth reading their documentation on this even if you have no intention of switching platforms; it is a clear picture of what a thorough implementation looks like, particularly the fail-closed age check, the permission separation, the survey versioning, and the fact that the vendor warns you off its own gating mode. It is not, however, a reputation platform, and it does not pretend to be one. It requests feedback and routes the outcome. It does not manage your listings across directories, it does not pull existing reviews back in so you can triage and respond to them from one place, and it does not do the multi-location work that a practice with several sites and a marketing coordinator genuinely needs. If that describes you, a dedicated service like Birdeye, Podium, or Weave remains the better tool and the EMR module will feel thin by comparison. If you are one psychiatrist, or three, with a single Google profile and a panel small enough that every unhappy patient is a meaningful fraction of your week, the loop built into the chart is very likely the whole job, and the money you would spend on a standalone platform is better spent almost anywhere else.