How much do appointment reminder calls reduce no-shows? What the research says

Missed appointments (no-shows) are a silent cost line for clinics, dental practices and any office that runs on appointments. A systematic review of 105 studies found an average no-show rate of roughly 23% across all specialties, with regional figures ranging from 13.2% in Oceania to 43.0% in Africa [1]. In Türkiye the problem is visible at national scale as well: according to an April 2024 announcement by the Ministry of Health, 23 million people booked an appointment through the MHRS system in 2023 and did not attend, and a further 6 million did so in the first three months of 2024 [2].

The cost side has also been measured: across the 10 main clinics of a large US veterans' hospital the mean no-show rate was 18.8%, and the average cost per no-show was calculated at $196 in 2008 dollars [3]. This article brings together the peer-reviewed studies that measure the effect of reminders: how much they help, which channel and timing work better, why asking for confirmation makes a difference, and where reminders fall short. Booking appointments over inbound calls is covered in a separate article (avaritcall.com/en/blog/7-24-randevu-alan-asistan); this article deals with outbound reminder calls.

Why do patients miss appointments?

A survey of patients who had missed general practice appointments in the UK found that forgetting was the most common reason: more than 40% of respondents said they forgot the appointment; a quarter said they had tried hard to cancel or that the time was inconvenient; a fifth cited family commitments or being too ill [4]. In the same study, more than 90% of those who missed an appointment consulted again within three months; the need had not gone away, they simply had not come that day [4].

In the systematic review, the most commonly reported determinants were a long lead time between booking and the appointment and a prior no-show history; younger adults, lower socioeconomic status, living far from the clinic and lacking private insurance were also associated with risk [1]. In short, a large share of no-shows stems from logistical causes such as forgetting and being unable to cancel; part of it comes from structural barriers that no reminder will solve.

How much do appointment reminder calls reduce no-shows?

A systematic review of 29 studies (33 estimates) covering phone, SMS and automated call reminders found a weighted mean relative reduction in non-attendance of 34% of the baseline rate; all but one study reported a benefit [5]. In the same review, manual phone calls made by a person were more effective than automated reminders: a reduction of 39% versus 29% of baseline [5].

Single-centre randomized studies point to effects of similar size. In a US study covering 9,835 patients, the no-show rate was 23.1% with no reminder, 17.3% with an automated reminder and 13.6% when clinic staff called, and all pairwise comparisons were statistically significant [6]. In a Geneva primary care clinic serving a vulnerable patient population, a sequential reminder 48 hours before the appointment (phone first, SMS if unreachable, and a letter if there was no mobile number) cut the missed appointment rate from 11.4% to 7.8% and produced a net financial benefit of about €1,850 over three months [7].

Not every system delivers the same result, though: at the veterans' hospital mentioned above, a centralized phone reminder moved the rate only from 16.3% to 15.8% [3]. Having "a reminder" is not enough; the design is what matters.

SMS or phone call: which appointment reminder is more effective?

A Cochrane review (8 randomized trials, 6,615 participants) showed that SMS reminders increase attendance compared with no reminder (seven trials, risk ratio 1.14); in three trials comparing SMS with phone calls there was no meaningful difference (risk ratio 0.99). Attendance was 67.8% with no reminder, 78.6% with SMS and 80.3% with phone calls; in two studies the cost of SMS per attendance was 55% and 65% lower than that of a phone call [8].

On the other hand, in studies where clinic staff personally called patients, the call outperformed automated systems [6][5]. Putting the two findings together, the practical reading is this: what matters is less the channel itself than whether the patient is actually reached and can respond. SMS is cheap and scalable; a live call is more effective but consumes staff time.

Timing and confirmation: which set-up works better?

On timing the evidence is surprising: in the 29-study review there was no difference in non-attendance whether the reminder was sent the day before or the week before the appointment [5]. The number of reminders, by contrast, matters. In a randomized trial of 54,066 patients across 25 primary care clinics, a single automated reminder produced a no-show rate of 5.8% when sent 3 days before and 5.3% when sent 1 day before, while sending both reduced it to 4.4%; among high-risk patients the gap was wider (25.0% and 24.2% versus 20.5%) [9].

What is worth more than a bare reminder is the chance to confirm or reschedule. In the groups that received reminders, cancellation rates were also significantly higher than in the control group [6]; in the Geneva study, 28% of cancelled appointments could be reallocated to other patients [7]. Türkiye's Ministry of Health has applied the same logic at national scale: under the plan announced in April 2024, patients with an MHRS appointment were to confirm "I will attend" by 23:59 the day before, and unconfirmed slots were to be given to someone else [2]. In the practice that started on 13 May 2024, the patient confirms by 20:00 the day before via mhrs.gov.tr, the MHRS mobile app or the ALO 182 call centre; an unconfirmed appointment is cancelled by the system, and patients over 65 and cancer patients are exempt [10]. A practical plan for a small clinic or office:

• Verify the phone number when booking; in the Geneva study only 61.7% of patients had a mobile number on record [7].
• Make two contacts: one reminder 3 days before the appointment and a second contact the day before [9].
• Offer three options every time: I will come, I cannot come, let's move it to another day.
• Release the slot of a patient who says "I cannot come" to a waiting-list patient the same day [7].
• Keep patients with a previous no-show or a long booking lead time on a separate list and call them more often [1].

A reminder brings the appointment back to the patient's mind; a confirmation clears the clinic's calendar. The slot of a patient who says today that they cannot come can be given to someone else tomorrow.

Can AI make appointment reminder calls?

The gap the research reveals is clear: automated calls work, but they lag behind live human calls [6][5]. The plausible explanation is interaction: a staff member can answer "Could we do Thursday instead of Tuesday?" on the spot, while a one-way automated system cannot. Conversational AI agents target this middle ground; but there is no evidence for that yet. The "automated call" class in the studies covers recorded-message and IVR systems; whether a conversational AI agent ends up closer to that class or to live staff in its results has not yet been measured. To be honest, while preparing this article we did not find a peer-reviewed randomized trial comparing reminder calls made by conversational AI agents with live staff; the evidence available today is the measured gap between automated and live calls, and the value of confirmation and rescheduling. Two practical notes: for reminder calls that are informational in nature and relate to an existing service, the Regulation on Commercial Communication and Commercial Electronic Messages does not require separate consent; if marketing content is added to the call, İYS consent is required (details: avaritcall.com/en/blog/yapay-zeka-ile-giden-arama-kampanyasi-yasal-cerceve). In a clinic call the agent should only take the date, time, confirmation and rescheduling information; no health information should appear in the transcript, because health data is a special category of personal data.

AvaritCall brings this flow into a single call: the AI agent calls the patient, restates the appointment details, takes the "I'll come / I can't come / let's reschedule" answer in natural conversation and can pass the outcome on through calendar and CRM integrations; a cancellation or rescheduling request appears instantly in the panel and, via push notification, in the iOS app. Turn-taking, hesitation and interruption behaviour is specifically optimized for Turkish phone conversation. In Türkiye the average response latency is ~0.8 s (at Opus voice quality and without truncating context; the time from when the caller stops speaking until the agent starts its reply). No voice recording is made; only a text transcript is stored, and the platform operates in compliance with KVKK and GDPR. Pay as you go: $0.08 per minute of conversation (STT, LLM and TTS included, a single line item), $5 credit at sign-up, no commitment. Details at avaritcall.com/en/pricing.

No-show causes a reminder does not solve

A reminder wins back the patient who forgot; it does not win back the patient who stays away out of fear, or who faces a transport or financial barrier. In the UK survey, a quarter of those who missed an appointment had already tried to cancel [4]; for them the problem is not the reminder but how reachable the cancellation channel is. A patient with a wrong or missing number receives no reminder through any channel [7]. In some systems the gain stays below one percentage point [3]; do not expect much before you have measured with your own data.

The limits on the product side are equally clear: recognition accuracy drops in very noisy environments; concurrent call capacity depends on the plan, and very high concurrency requires an enterprise package; the system is not designed to replace emergency lines and does not produce medical decisions, so a patient asking a medical question must be handed over to a person. A reminder call is an appointment-operations tool, not a clinical decision tool.

In summary, the peer-reviewed evidence shows that reminders reduce no-shows by roughly a third of the baseline rate [5], that live calls are more effective than automated ones [6], that SMS and phone give similar results once the patient is reached [8], and that a two-contact-plus-confirmation set-up works better than a single reminder [9][6]. This article is not legal advice; MHRS rules and legislation may change, and sources were accessed on 30 August 2026.