How to cut no-shows in a therapy practice: what the evidence says
Evidence-backed protocols to reduce no shows in a therapy practice: reminder cadence, same-day recovery texts, waitlist fills, and what to skip.
7 min read
If you want to reduce no shows in a therapy practice, the intervention with the best evidence behind it is also the least interesting one: a text message sent before the appointment. A Cochrane review of mobile phone reminders for healthcare appointments put the risk ratio for non-attendance at roughly 0.77 against no reminder at all. It is cheap and it is dull. Most of what follows is either a refinement of that finding or an attempt to catch the people it misses.
Two honest caveats first. That Cochrane number comes from healthcare broadly, and behavioral health has a failure mode general medicine mostly does not: people miss sessions partly because of the thing they are being treated for. A reminder fixes forgetting. It does nothing about avoidance, a panic spike at 9am, or a copay landing the same week as rent. And no-show rates swing so hard by population and payer mix that other practices' benchmark numbers are close to useless for setting your target. Measure your own baseline for eight weeks before you judge any change you make.
How many reminders is too many?
Start with two. There is at least one large study supporting the second one specifically in mental health: research published in The Permanente Journal, run across a big health system, found that adding a second text reminder cut mental-health no-shows by about 11% relative to a single reminder. A modest gain on top of a modest gain, which is what real operational wins usually look like.
A cadence that works for a weekly outpatient caseload:
| Timing | Channel | Content |
|---|---|---|
| 72 hours before | Email, or text if the client has no working email | Date, time, clinician first name, link to confirm or reschedule |
| 24 hours before | Text | One line, date and time, same reschedule link |
| 2 hours before, first appointments only | Text | Join link for telehealth, or parking and suite number for in-person |
The third touch is scoped deliberately narrow. Intake and first-session slots are where most of your lost revenue lives, and a same-day nudge there is worth the message. Applying it to every established weekly client is how you train people to reply STOP, and a STOP takes away the two reminders that were actually working. Cap it: no more than three automated messages to one person in a rolling seven days, counting every kind of automated message.
What should the reminder actually say?
Short and clinically empty. HHS guidance on texting patients allows it with reasonable safeguards, and the safe pattern everyone converges on is a neutral message plus a portal link for anything sensitive. No diagnosis, no measure scores, no medication names, no group name that gives away what the group is for.
Something like: "Reminder: your appointment with Dana is Thu Jul 30 at 2:00pm. Reply C to confirm or tap the link to reschedule." That is enough for the client and useless to anyone reading over their shoulder on a bus.
If you run a substance-use program under 42 CFR Part 2, tighten this further. The 2024 final rule has been fully in force since February 2026, and Part 2 treats affiliation with the program as sensitive on its own. Your sender name should be something neutral, and you want documented consent on file before automated outreach starts. Do not put the program name in the from-line.
What happens in the ten minutes after someone misses?
This is the part most small practices leave undone, and it is the highest-yield thing on the list after reminders exist at all. The same-day "we missed you" text with a self-serve rebooking link is the pattern patient-communication platforms built their reputations on, and the mechanism is obvious once you say it out loud: the client already feels bad, and every hour that passes makes calling the office harder, so you hand them a way to fix it that requires no conversation.
Concretely, a three-touch recovery sequence:
- Fifteen minutes past the scheduled start, automated text: "Sorry we missed you today. Grab a new time here when you're ready." Plus the booking link. Leave the fee out of it entirely; the statement can raise that later.
- Day 3, if they have not rebooked, a second text at a different time of day than the first. If the first went at 2:15pm on a Tuesday, send this one mid-morning. People have consistent dead zones in their week.
- Day 7, if still nothing, a task lands on a human. Actual phone call from the front desk or the clinician, and a voicemail if nobody picks up. Two misses with no contact is a clinical event, not an admin one, and the clinician should know about it before day 10.
The day-7 human step is the one people quietly drop when they get busy. Put it in the system as a task with a due date rather than trusting anyone to remember it.
Do cancellation fees do anything?
Probably less than you hope. I have not seen behavioral-health evidence that fee policies move attendance, and the theory is shaky: if someone missed because they were dissociating or their car died, a $75 charge makes their month harder without making next Tuesday more likely. Fees do have a real job, protecting income against chronic late cancels from clients who can comfortably pay. Treat it as that.
A workable policy: 24-hour notice required, first occurrence always waived with a note, second occurrence waived at clinician discretion, third charged. Write the rule down, put it in the intake packet, and apply it the same way for everyone so it never becomes a conversation about whether this particular client deserves a break.
Filling the hole you already have
Recovery is slow. Filling a gap this week is fast. Keep a short-notice list of eight to twelve clients who have explicitly opted in to same-week openings, and when a slot opens more than four hours out, text the whole list at once with first-come booking. Two rules keep this from backfiring: cap it at one blast per person per week, and never send it to anyone in acute crisis, where an unfilled offer reads as rejection.
The other structural lever is time-to-first-appointment. Whatever your reminder setup, a client booked eighteen days out is a different risk profile than one booked four days out, and shortening that gap tends to do more for intake attendance than any message you can send during it. If you have a waitlist, this is where the waitlist should be spent.
What to skip
Automated voice calls as the primary reminder. They get treated as spam, they leave voicemail that discloses more than a text does, and nobody under 40 answers an unknown number.
Overbooking. It works in high-volume primary care with 15-minute slots. In a practice running 50-minute sessions, the day both clients show up costs you more in clinician trust than the empty slot did in revenue.
Long policy texts. If your reminder is four sentences explaining the cancellation fee, the useful information (time and date) is below the fold on a lock screen.
Personalized clinical content in reminders. "Time for your PTSD group" is a HIPAA problem and, under Part 2, potentially worse than that.
How will you know any of it worked?
Track no-shows weekly, split at minimum into new clients and established clients, because those two numbers move for different reasons and averaging them hides both. Give any change eight weeks before you judge it. Small-practice weekly counts are noisy enough that a good week and a bad week look like signal when they are not.
Worth watching alongside attendance: your outcome measures. The AIMS Center treat-to-target model suggests PHQ-9 and GAD-7 at baseline, every two weeks during active treatment, and monthly in maintenance, with a flag when someone shows under 50% improvement by week 10 to 12. My working assumption, and it is an assumption rather than a finding, is that stalled scores and rising cancellations usually show up in the same clients, and the measure often moves first. If you are already collecting the scores, the flag is free.
None of this is exotic. It is two reminders, a same-day recovery text, a day-3 follow-up, a day-7 human task, and a short-notice list, all running without anyone remembering to run them. The reason small practices do not have it is almost never disagreement about whether it works. It is that setting it up by hand takes an afternoon nobody has.
That automation is what Valence's care pathways are for: two-touch reminders and the no-show sequence above as a template, with quiet hours enforced at 8am to 9pm in the patient's local time, per-patient weekly send caps, automatic opt-out handling, and a per-patient log of every message that went out. Texts are $0.03, email is free. You can see how the pathway automation is put together, or read our HIPAA checklist for solo practitioners before you turn any messaging on. If you are still deciding on a platform, we also keep an honest rundown of SimplePractice alternatives.