Reminder Timing: What the Trial Data Says

Every scheduling vendor has an opinion about send time. Twenty four hours out. Forty eight hours out. Morning of. The opinions are stated with a confidence that the underlying research does not have, so before we give you our own recommendation, here is exactly what the trials measured and where they stop.

What the trials actually tested

The strongest evidence on appointment reminders is the Cochrane review CD007458, published 5 December 2013 as an update of a July 2012 review. It pooled 8 randomised controlled trials covering 6,615 participants.

The headline numbers: attendance was 67.8% with no reminder, 78.6% with an SMS reminder, and 80.3% with a phone call reminder. SMS versus no reminder gave a risk ratio of 1.14 (95% CI 1.03 to 1.26) across 7 studies and 5,841 participants. SMS versus a phone call gave a risk ratio of 0.99 (95% CI 0.95 to 1.02) across 3 studies and 2,509 participants. In two of the studies, the cost per text per attendance was 55% and 65% lower than the phone call equivalent. Cochrane rated the evidence low to moderate quality and noted the studies were heterogeneous.

Read that as three findings. Sending a reminder beats sending nothing by about 10.8 percentage points of attendance. A text is statistically indistinguishable from a call, since the confidence interval for that comparison straddles 1.0. And the text costs roughly a third to a half of what the call costs per attended appointment.

Now the part vendors skip: none of those trials randomised the send time. The variable under test was channel. Cochrane can tell you to send something and to send it by text. It cannot tell you that 24 hours beats 48 hours, because nobody in that pool ran that experiment. Anyone quoting a Cochrane number in defence of a specific send time is quoting it for something it does not measure.

The one timing variable that does have data

There is timing evidence in the literature, but it is about a different clock. A 2024 systematic review in Health Science Reports, PMC11231932, examined no-show rates in outpatient clinics using open access scheduling. It covered 16 articles and did not pool them into a meta-analysis.

Its consistent finding is that longer booking-to-visit lead time is associated with a higher no-show rate. One included study reported no-show rates of 8% for lead times of 0 to 3 days, 16% for 4 to 6 days, and 22% for 28 to 30 days. That is 14 percentage points of spread, and the 28 to 30 day bucket is 2.75 times worse than the 0 to 3 day bucket, from the same clinic, on the same book.

On the intervention side, 10 of the 16 papers (62.5%) showed a significant no-show decrease under open access scheduling, 4 (25%) showed no significant reduction, and 2 (12.5%) showed no change. The reported examples span very different baselines: one clinic went from 3.33% to 1.89%, another from 42% to 27%.

So the interval that has real evidence behind it is the gap between when the appointment is booked and when it happens, not the gap between the reminder and the appointment. That reframes the problem. If your no-show rate is high and you are booking three weeks out, changing your reminder from 48 hours to 24 hours is tuning the smaller of the two levers.

The timing table

Here is the artifact, with the evidence status labelled on every row. Where a row says operating recommendation, that is our judgement and not a trial result.

What to do after the miss

The MGMA Stat poll of 12 August 2025, with 265 applicable responses, found that 73% of medical practices said no-show rates stayed the same (60%) or decreased (13%) in 2025 compared with 2024, while 27% reported increases. The poll reports no absolute no-show percentages, so it tells you about direction, not level. Its practical value is the recovery sequence it recommends:

• Within 24 hours of a no-show, send a friendly, blame free rebooking link.
• Within 72 hours, make a live call for clinically important follow ups.
• Use two way texting that lets people confirm, ask a question, or tap to reschedule, and escalate to a live call for high risk appointments.

That sequence is a recommendation from a professional association, not a trial result, and it comes from a medical setting. The logic transfers cleanly to a service bay or a law office intake calendar, but you should treat it as a starting configuration to measure, not as a proven number.

How to test this on your own book

You do not need a randomised trial to settle this for your business. You need two months of clean data. Split your bookings by lead time first, into 0 to 3 days, 4 to 7 days, and 8 or more days, and calculate the no-show rate in each bucket. If your spread looks anything like the 8% to 22% pattern in the review, your problem is the calendar, not the reminder.

Then alternate send times week by week, 48 hours in odd weeks and 24 hours in even weeks, and compare. Your own book is the only sample where the answer is not extrapolated. Before you start, work out what a missed slot actually costs you, because a two point improvement is either meaningful or noise depending on that number. We laid out that calculation in the no-show math for a service business.

The unglamorous conclusion is that the send time debate is the smallest of the three levers. Sending a reminder at all is worth about 10.8 percentage points of attendance in the pooled trials. Moving from a month of lead time to a few days was worth 14 points in one clinic. The choice between 24 and 48 hours is worth an amount nobody in these sources has measured. Set up appointment reminders and scheduling so that all three are under your control, then spend your attention in that order. Most of the work is plumbing between your calendar, your phone system, and your customer records, which is the same business automation problem in a different costume.

If you want a second pair of eyes on your own numbers before you change anything, our free 15 minute audit will tell you which of the three levers is actually costing you money.