PT No-Shows and Dropout: Benchmarks, Rules, Fixes
A clinic owner staring at empty slots wants a number to compare against. Across specialties, the average no-show rate in the published literature is 23 percent, from 43.0 percent in African studies to 13.2 percent in Oceania (Dantas et al., 2018). A large US health system found 18.8 percent across ten clinics (Kheirkhah et al., 2016). In PT specifically, a Melbourne hospital trial found 16 percent nonattendance with no reminder, falling to 11 percent with an SMS reminder (Taylor et al., 2012); more recent Australian physiotherapy clinics ran 8 to 10 percent (Mohammed Selim et al., 2025).
Then there is the number every vendor blog repeats with no citation: that 70 percent of patients do not complete their plan of care. That claim does not trace to any study we could find. The one US record review that actually measured plan-of-care completion found a self-discharge rate of 48 to 54 percent at five community physical health clinics, a figure with real limits: retrospective, not PT-exclusive, and the lowest cohort had self-selected into a practice-branded engagement app with attendance rewards (Topp et al., 2023).
This post works from the sourced literature: benchmarks by setting and year, the formulas to calculate your own rates, which interventions have trial evidence, the TCPA and HIPAA rules for texting patients, and what to require from engagement software. Current as of September 2026.
What is a normal no-show rate for physical therapy?
Rates depend on setting, population, and how a study defines “no-show,” so the honest answer is a range, with the definition attached to each figure.
| Setting | Year(s) | Definition | Figure |
|---|---|---|---|
| Global, all specialties, 105 studies | through 2018 | did not attend a scheduled visit | 23% average, 43.0% Africa, 13.2% Oceania (Dantas 2018) |
| VA medical center, Houston, 10 clinics | FY1997-2008 | did not attend | 18.8% mean (Kheirkhah 2016) |
| 2 PT outpatient depts, Melbourne, no reminder | 2012 RCT | nonattendance without cancellation | 16% (Taylor 2012) |
| Same trial, SMS reminder | 2012 RCT | nonattendance without cancellation | 11% (Taylor 2012) |
| 2 tertiary hospital physio clinics, Queensland | 2018-2021 | clinic non-attendance | 8% and 10% (Mohammed Selim 2025) |
| Academic pain center, not PT-specific | 2017 | no-show, no day-before cancel | 24.6% (Odonkor 2017) |
| 5 community physical health clinics, retrospective | 2018-2019 | self-discharged before planned discharge | 48-54% (Topp 2023) |
The most consistent predictors are a long lead time and a prior no-show history (Dantas 2018), younger age, Mondays or Fridays, and initial or evaluation visits (Taylor 2012); Monday was also highest-risk in the VA data (Kheirkhah 2016).
How do you calculate no-show rate, cancellation rate, and plan-of-care completion?
Definitions have to hold constant, because published studies do not agree on them: Taylor counted only “nonattendance without cancellation” (Taylor 2012); Odonkor counted a no-show only if the patient neither showed nor cancelled the day before (Odonkor 2017). Pick one and write it down.
Track, per period: S (scheduled), N (no-shows: didn’t arrive, didn’t cancel before your cutoff, commonly 24 hours), L (late cancellations, inside the cutoff), A (advance cancellations, before it). Arrived, R, is S minus N minus L minus A. A reschedule is not its own category; it’s a cancellation at the old slot plus a new booking, counted both places.
Two no-show rates are both defensible and not the same number. N / S treats every scheduled slot as the denominator. N / (S − A) drops advance cancellations from the denominator first, since those slots were already freed, which always yields a higher rate for the same N. Report which one you use and don’t switch.
Plan-of-care completion needs the same discipline: a case is a planned discharge only if the therapist closed it; everything else needs a self-discharge rule (for example, no visit in 30 days with an open plan of care), so cases don’t sit open forever (for Medicare cases, the plan-of-care certification period is a natural checkpoint for that review). Report visits per case as a median as well as a mean: the one dataset with reported standard deviations showed mean visits per patient of 10.73 to 13.21 with SDs of 9.80 to 12.09, so a handful of long-running cases pull the mean above what a typical patient experiences (Topp 2023).
One illustration month, made-up but internally consistent numbers, not sourced statistics:
| Metric | Formula | Illustration (S = 1,000) |
|---|---|---|
| No-show rate | N / S | 60 / 1,000 = 6.0% |
| No-show rate, alternate | N / (S − A) | 60 / 880 = 6.8% |
| Late-cancellation rate | L / S | 40 / 1,000 = 4.0% |
| Advance-cancellation rate | A / S | 120 / 1,000 = 12.0% |
| Arrival rate | R / S | 780 / 1,000 = 78.0% |
| Plan-of-care completion | planned discharges / cases closed | 48 / 80 = 60.0% |
| Visits per case, mean | visits by closed cases / cases closed | 800 / 80 = 10.0 |
At a made-up $95 collected per visit, this month’s 60 no-shows and 40 late cancels represent roughly $9,500 of slots that produced no revenue, illustration only. The one sourced cost figure, $196 per no-show, is a single VA system in 2008 dollars and does not transfer to a PT practice today (Kheirkhah 2016).
Does 70 percent of patients really not complete their plan of care?
No, at least not as a sourced claim. Tracing “70 percent” back lands on home-exercise non-adherence, not plan-of-care completion. A 2017 review of 30 quantitative articles states non-adherence to home-based physical therapy “can reach 70%” (Essery et al., 2017). A 2010 review of physiotherapy outpatient adherence cites a 1993 Dutch study for the same figure, that “non-adherence with treatment and exercise performance could be as high as 70%” (Jack et al., 2010, citing Sluijs et al., 1993). The 1993 study itself gives no percentage in its own abstract; the “70 percent” reaches the modern literature secondhand, through later reviews citing it. None of this is a study of what share of US PT patients complete a plan of care.
The 2010 review does report, from an earlier study it cites, that 14 percent of physiotherapy outpatients never returned for follow-up at all (Jack 2010). And the one figure that actually measures plan-of-care completion in a US outpatient setting is the 48 to 54 percent self-discharge rate above, with its retrospective and self-selection limits (Topp 2023).
What actually reduces no-shows and dropout in PT?
Some interventions have randomized trial evidence; others have only correlational support, or a review’s conclusion that the evidence isn’t strong enough yet.
| Intervention | Evidence | Limit |
|---|---|---|
| SMS text reminders | Cut nonattendance 16% to 11% in a PT-specific RCT, NNT 19 (Taylor 2012); a Cochrane review (8 RCTs, 6,615 participants) found pooled RR 1.14 versus no reminder, on par with phone calls and cheaper (Gurol-Urganci 2013); multiple notifications beat one (Robotham 2016); content matters, stating the missed-appointment cost cut one clinic’s rate 11.1% to 8.4% (Hallsworth 2015) | The FCC’s healthcare exemption caps exempt messages at one per day, three per week per patient |
| Short lead times, booking the full plan at evaluation | High lead time and prior no-show history are the most reported determinants (Dantas 2018); initial visits, Mondays, Fridays carry higher risk in PT (Taylor 2012) | Association, not a tested bundled intervention; a practice recommendation, not an RCT result |
| HEP apps with tracking, remote support | Beat a paper handout on adherence by 1.3 of 11 points (Lambert 2017); tracking, goals, and behavior-change messages more than doubled complete adherence at 26 weeks, 36% vs. 16% (Hinman 2026) | In Hinman’s trial the adherence gain did not translate into a function difference; in Lambert’s, function improved by 0.9 of 11 points, which the authors called of unclear clinical importance |
| Feedback, goal-setting, booster sessions | Lack of positive feedback was a main correlate of noncompliance in the foundational Dutch study (Sluijs 1993); feedback, diaries, and collaborative goal-setting are recommended in a review of 20 cohort studies (Jack 2010); booster sessions showed a small-to-medium pooled effect for hip or knee OA, SMD 0.39 (Nicolson 2017) | A review of 12 RCTs found “insufficient data to endorse their use in clinical practice” (Peek 2016) |
| Between-visit check-ins via RTM | Objective remote monitoring is the missing measurement tool this literature describes (Argent 2018); see our RTM billing guide | A monitoring and billing mechanism, not itself a trialed attendance intervention |
| Screening for psychological or social barriers | Depression, anxiety, low self-efficacy, and helplessness predict poor adherence (Jack 2010; Essery 2017); the low-to-medium-attendance group in a 2026 multimorbidity trial was younger with lower self-efficacy and more depressive symptoms (Kaalund 2026) | Correlational; no trial here tested whether screening itself changes attendance |
Can you text patients appointment reminders? TCPA and HIPAA rules
TCPA: consent, the healthcare exemption, and revocation
The FCC’s rule treats a text message as a “call”: the healthcare exemption says so expressly (“the term ‘call’ includes a text message, including a short message service (SMS) call”) and the revocation rule speaks throughout of “calls or text messages” (47 CFR 64.1200(a)(9) and (a)(10), October 2025 CFR edition). It’s unlawful to make a call, or send a text, using an automatic dialing system or an artificial or prerecorded voice to a cell number without prior express consent, absent an emergency (47 U.S.C. 227(b)(1)(A)(iii); 47 CFR 64.1200(a)(1)).
The ordinary case: a patient hands your front desk their cell number. That constitutes prior express consent for HIPAA healthcare calls under the FCC’s 2015 ruling, “if the covered entities and business associates are making calls within the scope of the consent given, and absent instructions to the contrary” (FCC 15-72, para. 141). Scope isn’t unlimited: the ruling’s own example is that “if a patient provided his phone number upon admission to a hospital for scheduled surgery, then calls pertaining to that surgery or follow-up procedures for that surgery would be closely related to the purpose for which the telephone number was originally provided” (footnote 474). A reminder for the booked visit is inside scope; an unrelated marketing text is not.
Marketing or advertising texts, cash-pay pitches, wellness-membership upsells, package promotions, need prior express written consent: a signed (including electronic) agreement disclosing the automated messages and the number, stating consent isn’t a purchase condition (47 CFR 64.1200(f)(9)). A “health care” message exception from written consent exists at (a)(2), but it’s written for treatment-related, exigent messages, not promotions; don’t rely on it for a cash-pay pitch.
A separate exemption removes consent entirely for certain healthcare messages, but only if free to the patient and not counted against their plan’s minutes or texts (47 CFR 64.1200(a)(9)). Most clinics can’t verify a patient’s carrier plan, so this is hard to guarantee. Its eight conditions are still worth a safe-harbor template, while relying on documented consent, not the exemption, to actually protect the practice:
| Condition | Requirement |
|---|---|
| (A) | Sent only to the wireless number the patient provided |
| (B) | States the provider’s name and contact information |
| (C) | Limited to: appointment/exam confirmations and reminders, wellness checkups, hospital pre-registration, pre-operative instructions, lab results, post-discharge follow-up, prescription notifications, home healthcare instructions |
| (D) | No telemarketing, solicitation, advertising, billing, or debt-collection content; complies with HIPAA |
| (E) | Concise: about one minute for voice, 160 characters or less for text |
| (F) | One message per day per patient, max three per week combined |
| (G) | Easy opt-out in every message; for texts, “STOP” is the exclusive opt-out method |
| (H) | Opt-out requests honored immediately |
Whether or not a message qualifies for that exemption, patients can revoke consent any time, by any reasonable method. Specific methods count automatically: an automated opt-out mechanism, or “stop,” “quit,” “end,” “revoke,” “opt out,” “cancel,” or “unsubscribe” in reply to a text (47 CFR 64.1200(a)(10)). Honor it within a reasonable time, no later than ten business days, and never designate an exclusive means of revoking. One confirmation text is allowed; sent within five minutes of the opt-out, it’s presumed within the patient’s existing consent, and nothing further should follow (FCC 24-24).
The stakes are per message: a patient can recover actual loss or $500 per violation, whichever is greater, and a court may treble that to $1,500 for a willful or knowing violation (47 U.S.C. 227(b)(3)).
HIPAA: what a reminder text can and can’t say
A reminder for a booked visit falls within HIPAA’s definition of treatment, “the provision, coordination, or management of health care” (45 CFR 164.501, October 2024 CFR edition), so it needs no authorization. But minimum necessary still governs content: a covered entity must make reasonable efforts to limit PHI to what’s necessary for the purpose (45 CFR 164.502(b)). Standard carrier SMS isn’t encrypted end to end, and HIPAA’s transmission security standard treats encryption as “addressable,” not required (45 CFR 164.312(e)): the practice, not a texting vendor, must assess whether encryption is reasonable in its environment and, if it declines to encrypt SMS, document why and adopt an equivalent safeguard (45 CFR 164.306(d)(3)).
A compliant reminder carries only logistics: provider name, date, time, how to confirm or cancel, and STOP. It should not carry a diagnosis, body region or condition, a procedure name, or a balance due. Patients also have a standing right to request an alternative channel or location; a covered provider must accommodate a reasonable request, may require it in writing, and may ask how payment will be handled (45 CFR 164.522(b)).
Marketing texts need their own authorization
HIPAA defines marketing narrowly, as “a communication about a product or service that encourages recipients of the communication to purchase or use the product or service” (45 CFR 164.501). Using PHI, including a phone number, to send one requires the patient’s signed authorization, with two narrow exceptions: a face-to-face communication, and a promotional gift of nominal value (45 CFR 164.508(a)(3)). A cash-pay pitch or wellness-membership upsell texted to your patient roster is exactly what this rule targets; it needs its own authorization, separate from appointment consent.
What to require from patient engagement software
The software question is the consent, content, and reporting questions above, built into the system instead of left to staff judgment:
| Feature | Why it matters | EHR-native vs. bolt-on |
|---|---|---|
| Portal with PIN or passwordless access | Turns a would-be no-show into an advance cancellation, since patients can confirm, cancel, or reschedule without a call | EHR-native ties changes to the chart and schedule directly; bolt-on needs a sync job |
| Two-way secure messaging | Keeps clinical content off standard SMS, sidestepping the 164.312(e) encryption question | Native messaging lives with the visit record; bolt-on is a second system to search in a records request |
| Online self-scheduling and rescheduling | Targets the highest-risk visits, initial evaluations, Mondays, Fridays, with an easy alternative to no-showing | Works either way if it writes back to the front desk’s schedule |
| Reminder engine with consent capture, channel preference, STOP handling | Documented consent, not the exemption, actually protects the practice; STOP must be honored and logged immediately | Native ties consent to the patient record; bolt-on risks an opt-out list that drifts out of sync |
| Templates limited to the (a)(9)(iv)(C) purposes | Keeps content inside the safe-harbor template even when relying on consent, not the exemption | Template control matters more than platform |
| HEP with video, completion logging, clinician-visible adherence | Only 15 percent of patients in one study could recall and demonstrate their program (Peek 2019) | Native HEP data feeds the chart the clinician documents in; bolt-on is a second login and record |
| RTM data capture flowing to the chart and to billing | Between-visit adherence data is the missing measurement tool (Argent 2018) | Native avoids re-entering monitoring minutes for billing |
| Reports: no-show, cancellation, arrival, plan-of-care completion, visits per case (mean and median), by clinician and location, with a definable self-discharge window | You can’t manage what isn’t defined | Native reporting pulls from the same schedule and chart data as everything above |
| Intake and consent forms capturing the phone number and texting consent with a timestamp | Documents exactly the consent the TCPA analysis depends on | See our HIPAA-compliant form builder guide |
EHR-native tools keep the schedule, chart, HEP, RTM data, consent, and reports in one place. Bolt-on tools mean separate consent records and opt-out lists someone reconciles by hand, exactly the gap a records request or a TCPA complaint will find.
None of this works without one consistent record of who was scheduled, who showed, and who kept up with exercises between visits. PrismEHR’s patient portal gives patients PIN-based access to view and complete prescribed exercises, message their provider securely, and book appointments online, and its exercise completion tracking and home exercise program compliance monitoring sit in the same record as the reporting built for physical therapy practices. For the between-visit adherence measurement this literature says is missing, see our RTM billing guide. And because the consent analysis above starts with a timestamped record of how a patient’s number was collected, CuratoForm captures that consent at intake alongside the signature, with responses that can populate the PrismEHR chart for clinician review.