All articles
Recruitment StrategyOctober 28, 2026·3 min read

Building an SMS Reminder Cadence That Holds Screening Slots

A one-text-the-day-before reminder catches forgetters but misses the bigger no-show groups. Here's a cadence, message structure, and measurement approach that holds screening slots.

By Trialflow Team

Start with the reminder schedule, not the wording

Most sites that add SMS reminders send one text the day before and call it done. That catches the people who simply forgot, but it misses the two larger groups: participants who never fully committed at booking, and participants who hit a logistics problem they could have solved with a day's notice.

A cadence that handles all three usually looks like this:

  1. Confirmation within minutes of booking. Sent while the call is still fresh. It locks in the date in writing and gives them something to screenshot.
  2. A reminder three to four days out. This is the reschedule window. Far enough ahead that a conflict can be moved rather than canceled.
  3. A reminder the afternoon before. Include the fasting, medication, and ID requirements here, not earlier, because this is when they'll act on them.
  4. A short morning-of text with parking and arrival details.

Sites that extend reminders beyond a single touch commonly see the biggest improvement not in no-shows but in advance cancellations — which is a win, because an empty slot you know about 72 hours out can be refilled.

Write texts that assume no context

Your participant booked this visit nine days ago, possibly after speaking with three other clinics. Reminders that open with "Reminder: your appointment is tomorrow at 9:15" make them work to figure out what appointment and whose.

A usable reminder includes:

  • Site name, spelled out, not an abbreviation
  • Day of week plus date plus time ("Thursday, the 14th, at 9:15 AM")
  • Street address, with a cross street or building name
  • One sentence on what to bring or avoid
  • An expected duration
  • A direct reply path

Keep each message under about 300 characters. Longer texts get split across segments, arrive out of order on some carriers, and read as spam.

Make replies actually go somewhere

The fastest way to waste a reminder program is to send from a number nobody monitors. Participants reply "can I come at 11 instead" and get silence, then don't show. If your messages come from an unmonitored line, say so explicitly and give the phone number that is staffed, with the hours it's staffed.

Better: route replies into whatever queue your coordinators already work, and treat a reschedule request as a save, not a loss. Train the team to respond with two specific alternative slots rather than "what works for you?"

Confirm, don't just notify

Asking for a reply — "Reply C to confirm or R to reschedule" — changes the dynamic. Non-responders become a work list. Call the people who didn't confirm by the afternoon before; in most site datasets, that group has a materially higher no-show rate than confirmed participants, so a short call is well spent.

Measure it properly

Track no-show rate separately from cancellation rate, and segment by referral source. Self-referred web leads, database recalls, and physician referrals behave differently, and a cadence that works for one may be excessive for another.

Also watch opt-out rate. If more than a small fraction of participants are texting STOP, you are sending too often or too early in the relationship. Trim the morning-of message first; it adds the least.

Practical guardrails

Get documented consent to text at the point of booking and log it. Keep reminder content purely logistical — appointment details, directions, prep instructions. The moment a reminder starts describing the study, its benefits, or compensation, it becomes recruitment material and needs IRB review. Keep those two message types in separate templates so nobody blends them by accident.

Enrolling studies shouldn't be this hard

Trialflow gives research sites one platform for lead management, AI trial matching, and recruitment analytics.

Request a Demo