Should we text, call or mail our annual recall reminders, and how do the three compare on cost?

Text is the cheapest per attempt and the fastest to answer, running well under a dime once you count the message fee and the seconds of staff time. A printed letter costs roughly a dollar all in, mostly postage and paper. A live phone call is the most expensive by a wide margin, because six to eight minutes of a scheduler's time at a loaded wage is real money, but it is also the only channel that converts a hesitant patient in one contact.

The honest answer is that you do not choose. You sequence them, cheapest first, and reserve the expensive channel for the patients the cheap channels did not move. That ordering is what makes the arithmetic work.

Below is the cost structure for each, the consent rules that actually apply to screening outreach, and a method for building a defensible cost per booked exam.

Cost structure of each channel: staff minutes, postage, per message fees

Build each channel from its parts rather than accepting a headline price. These figures are assumptions for illustration. Substitute your own vendor pricing and your own wage before you take any of this to your owner.

Assume a scheduler at $22 an hour, and assume loaded cost with payroll taxes and benefits at roughly 1.3 times that, so about $28.60 an hour, or 48 cents a minute.

Channel	Direct cost	Staff minutes	Cost per attempt
SMS, automated	About 2 cents per segment, two segments	0.1 minute of review	About 9 cents
Letter, first class	Stamp, envelope, paper, toner, roughly 90 cents	0.5 minute to stuff and log	About $1.14
Postcard	Card stock plus postcard rate, roughly 55 cents	0.3 minute	About 69 cents
Live call, reached	Negligible	7 minutes including scheduling	About $3.36
Live call, no answer	Negligible	2 minutes with voicemail and logging	About 96 cents

Notice the line that surprises people: a call that does not connect still costs about what a postcard costs. Dial, wait, leave a message, note the attempt. A scheduler making sixty of those a week and reaching a third of them is spending roughly two and a half hours a week on the twenty conversations and another eighty minutes on the forty that went nowhere.

Also notice that the letter is not the expensive channel people assume. Postage dominates it, and postage is predictable. The volatile cost is always staff minutes.

Keep reading: Why do so many of our screening no shows never come back, and what can we do about it?

Consent and TCPA considerations for texting patients about screening

The Telephone Consumer Protection Act governs automated calls and texts to mobile numbers. The relevant question for a screening center is what kind of message you are sending.

The FCC has recognized an exemption for certain healthcare messages made under HIPAA rules for treatment and care coordination purposes. Appointment reminders, exam confirmations and preventive care notices fall in that territory. Advertising, marketing of services, and anything about billing or debt collection do not. A text that says her annual screening is due and offers to book it reads very differently from a text promoting a new service line, and the wording you use decides which bucket you are in.

Practical guardrails that keep you comfortably inside the line:

• Capture the mobile number from the patient herself, at registration, with a checkbox acknowledging she may receive appointment and screening reminders at that number.
• Identify your facility by name in the first message. Anonymous texts about breast imaging are alarming.
• Include opt out instructions and honor STOP immediately and permanently, across every list.
• Keep frequency low. A recall sequence should be a handful of messages over weeks, never a daily drip.
• Keep protected health information out of the message body. "Your annual screening is due" is fine. Prior results are not.
• Log consent with a date and a source. If someone asks in two years how you obtained the number, you need an answer.

Revocation is the piece that gets centers into trouble. A patient can withdraw consent by any reasonable means, including replying to the text or telling the front desk in person. If a verbal opt out at the desk does not reach the messaging system, you will keep texting a woman who asked you to stop. Give the front desk a single field that suppresses her everywhere.

Which channel reaches which age band in a screening population

A screening population skews older than a general patient population by construction, since routine annual screening typically begins in the forties and continues for decades.

Rather than assume, measure. Segment your recall cohort by decade, then look at what you already hold for each: how many have a mobile number on file, how many have a verified mailing address, how many have neither. That single table tells you more about channel mix than any general claim about demographics.

What you will usually find in a breast center is a mobile coverage rate that is high in the forties and fifties and drops in the seventies and eighties, alongside address quality that runs the other direction. Older patients move less and their addresses stay accurate for longer. Younger patients change addresses often and change phone numbers rarely.

That asymmetry is the actual argument for keeping mail alive. It is not nostalgia. It is that mail is your only reliable channel for a meaningful slice of the women who most need to keep coming.

Keep reading: How many screening slots per day can one mammography unit realistically hold in our center?

Sequencing all three instead of choosing one

Here is a sequence you can run as written, anchored on the date the patient becomes due.

1. Day 0. Text: facility name, her screening is due, a link or number to book. Cost about 9 cents.
2. Day 7. Letter or postcard to everyone who has not booked. Cost about 69 cents to $1.14.
3. Day 21. Second text, different wording, to non responders with a valid mobile. Cost about 9 cents.
4. Day 35. Live call to everyone still unbooked. Cost about 96 cents to $3.36.
5. Day 60. Final letter, then rest the patient until next cycle.

The point of this order is that each cheap step removes patients from the expensive step. If the first text books fifteen percent of the cohort and the letter books another ten, your day thirty five call list is already a quarter smaller than the cohort you started with, and it is composed of exactly the people who need a human voice.

What a live call still does that no automated message can

Three things, specifically.

It handles the objection. A woman who is putting off her mammogram is usually not confused about whether she is due. She is anxious, or she remembers the compression being painful, or she is worried about the cost, or she is caring for someone else and cannot see a free morning. A message cannot hear any of that. A scheduler can, and can respond with the specific fact that unblocks her: that screening is covered without cost sharing under most plans, that a 7:15 slot exists, that the exam takes twenty minutes.

It books in the moment. Every handoff between contact and booking loses people. A call that ends with a confirmed appointment has zero handoffs.

It updates the record. In the same conversation you can confirm her address, her mobile, her current provider and her insurance. That data hygiene has value beyond this one exam, and it is nearly impossible to collect any other way.

See how MammoLoop handles this for breast imaging and women's screening centers

Handling wrong numbers, returned mail and address hygiene

Every channel degrades if you do not clean it, and undeliverable contacts silently inflate your cost per booking because you keep paying to reach nobody.

For mail, use the address correction service on your outgoing pieces so returns come back with a reason and, where available, a forwarding address. Work returns weekly. A returned letter should immediately suppress that address and open a task, not sit in a tray.

For text, watch for hard failures from your messaging vendor. A number that reports as invalid or unreachable should be flagged after the first failure and removed after the second, not retried monthly forever.

For phone, a disconnected number is a data event worth recording with a date. Three channels failing on the same patient is a strong signal she has moved out of your area, and she should drop to a low priority tier rather than consuming a call slot every cycle.

Building a per booked exam cost you can defend to the owner

Run the sequence above against a cohort of one thousand due patients, using the assumptions already stated. Replace every number with your own.

Step	Patients touched	Unit cost	Step cost	Bookings
Text 1	800 with mobile	$0.09	$72	120
Letter	880 unbooked	$1.14	$1,003	88
Text 2	700 unbooked with mobile	$0.09	$63	49
Live call	743 unbooked	$1.75 blended	$1,300	186
Total			$2,438	443

That is about $5.50 in outreach cost per booked exam. Set it against your average screening reimbursement, and against the downstream diagnostic and biopsy work that a screening program legitimately generates, and the case makes itself.

Two adjustments make the number honest. Subtract no shows before you claim the bookings, because a booked exam that does not arrive earns nothing. And be clear that the call step, which produced the most bookings here, also consumed roughly twelve hours of a scheduler's week for that cohort. The dollars are modest. The hours are not.

Where the hours actually go

The cost per booking is rarely what stops a recall program. What stops it is that the person running it has a waiting room to manage, and the outreach sequence lives entirely in her head and her calendar.

MammoLoop runs the sequence for you: staged texts and letters on the timing you set, callback tracking so nothing gets worked twice, and a no show recovery queue that pulls missed appointments back into the loop. The live calls that still need a human get made to a much shorter list, by someone who can see exactly what has already been tried.