
Flu Season Starts With a List Nobody Can Reach
Recall, Reminder or Campaign? Why the Distinction Decides Everything
WhatsApp, SMS or Email for Health Campaigns?
Getting Opt-In and Template Approval Right the First Time
Health Data, Consent and Retention Under the PDPO and PDPA
Running the Campaign Without Overwhelming the Front Desk
What Broadcast Reminders Cannot Do
A family practice in Kwun Tong decides in September to offer seasonal influenza vaccination to its regular patients. The clinical protocol takes an afternoon to agree. The list of eligible patients takes a week to assemble, because it lives partly in the practice-management system and partly in the memory of the two doctors. And then the message itself stalls, because nobody is sure whether it is legal to text patients about a vaccine they did not ask for.
The payoff, by the numbers
Before the detail, here is why this is worth the team’s time:

That last hesitation is the real bottleneck in most Hong Kong and Singapore clinics. Public-health campaigns and clinic recall programmes both depend on reaching people who are not currently in a conversation with you. The channel matters, the consent you hold matters more, and the two questions get tangled together until the season is half over.
There is a workable answer, but it starts with an uncomfortable distinction: a reminder for a second dose the patient already booked and an invitation to come in for a seasonal shot are not the same kind of message. They sit in different template categories, carry different consent expectations, and are priced differently.
This guide separates the two, compares the realistic channels, sets out what the PDPO in Hong Kong and the PDPA in Singapore expect when a message reveals a health context, and gives a campaign sequence that does not collapse the first time a patient replies with a clinical question.
On the WhatsApp Business Platform, any message you send outside an open conversation window must use a template that Meta has reviewed and assigned to a category. For vaccination messaging, three situations look similar to a clinic and are treated very differently by the platform.
A scheduled-dose reminder relates to something the patient has already arranged — a second dose with a date, a booked appointment. It relates to an existing transaction, which is the description of a Utility template.
A clinical recall tells a patient that a course is incomplete or that a due date has passed. This is closer to Utility in intent, but it is proactive, and how Meta's review treats it depends heavily on the wording. Keep it factual and specific to that patient's record.
A seasonal campaign invitation offers a service the patient has not asked for. Assume it will be treated as a Marketing template, and plan the consent, the opt-out line and the cost on that basis. Writing a campaign invitation as though it were a reminder is the most common reason a clinic's templates are rejected, and repeated rejections affect how quickly later submissions clear.
The practical rule: if the patient could reasonably be surprised to receive it, treat it as marketing. That single test resolves most of the ambiguity and keeps the clinic on the safe side of both Meta's policy and local privacy law.
| Message | Likely template category | Consent you should hold | Timing |
|---|---|---|---|
| Second dose already booked | 實用性 | Consent to be contacted about care | 48-72 hours before |
| Course incomplete, dose overdue | Utility, wording dependent | Consent to be contacted about care | Once, then a single follow-up |
| Seasonal shot now available | 市場營銷 | Explicit opt-in to health promotions | Start of season, with an opt-out |
| Clinic-wide public health notice | 市場營銷 | Explicit opt-in | Once per campaign |
Splitting your patient list along that line before you write a single template saves weeks. Most clinics find they can message a large group about booked care immediately, and a much smaller group about campaigns until fresh opt-in is collected.
Clinics in Hong Kong and Singapore realistically choose between three channels, and most end up running two. The differences that matter are whether the message is read, whether the patient can act without leaving the thread, and whether the clinic can show consent afterwards.
SMS reaches every handset and needs no app, which makes it a reasonable fallback for a very short prompt. It is effectively one-way in practice, carries no rich content, and gives the patient no easy path to book.

Email is cheap and handles the material a patient should read properly — eligibility criteria, contraindications, what to bring. Transactional and promotional mail is heavily filtered, so treat it as a supporting channel rather than the one that drives bookings.
WhatsApp carries a high read rate on opt-in messages, around 68 per cent on typical benchmark data, and lets the patient reply, ask a question or book in the same thread. The trade-off is template approval and explicit opt-in, plus per-message cost under Meta's rate card.
| Channel | Strengths | Limitations | Best used for |
|---|---|---|---|
| 短信 | Universal reach, no app, no opt-in friction | One-way, no rich content, no booking path | A short day-of prompt |
| 電子郵件 | Detail, attachments, low cost | Filtered, low engagement on health mail | Eligibility notes and consent forms |
| High read rate on opt-in, reply in-thread, rich formatting | Template approval, explicit opt-in, per-message cost | The invitation and the booking conversation |
The pattern most clinics settle on is WhatsApp for the invitation and the booking loop, email for anything the patient needs to read carefully before attending, and SMS reserved for patients who never opted into messaging.
imBee is an Official Meta Technology Partner and connects the WhatsApp Business Platform to a shared clinical inbox, so a campaign reply reaches a named person rather than an unmonitored number.
Two things stop vaccination campaigns going out on time: consent that was never captured for this purpose, and templates that keep coming back rejected. Handle them in that order, because the second depends on the first.
Collect opt-in at a moment the patient understands. Registration, discharge and the booking confirmation are all natural points. State what you will send and how often. A tick box buried in a terms document is weak evidence and a poor patient experience.

Record the route, not just the fact. You want to be able to say when consent was given and through which form or conversation, because that is the question you will be asked.
Separate care messaging from health promotion. Two consents, recorded separately. Patients who agree to appointment messages have not thereby agreed to campaign invitations.
Write templates that a reviewer can categorise at a glance. One purpose per template, no mixed promotional content in a reminder, and a clear opt-out line on anything promotional.
Avoid clinical claims in the template. Efficacy statements, comparative claims about vaccines and anything that reads as medical advice belong in the consultation, not in a broadcast.
Name the language variants up front. In Hong Kong that usually means Traditional Chinese and English versions of the same template, approved together, so the campaign is not held up waiting on a second submission.
Appointment and recall data is personal data, and in a clinical setting it carries an extra sensitivity: the mere existence of the message indicates that an identified individual is a patient of yours. Handling obligations are correspondingly stricter than for ordinary commercial messaging.
In Hong Kong, the Personal Data (Privacy) Ordinance (Cap. 486) sets out six data protection principles covering collection, accuracy, retention, use, security and access. Using a patient's number to remind them about care they arranged is a use directly related to the purpose of collection. Inviting them to a seasonal campaign they did not ask about is direct marketing, and the Ordinance imposes specific requirements on that, including the individual's ability to opt out.
In Singapore, the Personal Data Protection Act governs consent, purpose limitation and notification, and the Do Not Call provisions add further checks on marketing messages sent to Singapore numbers. A transactional reminder for a booking the patient made is treated differently from a promotional invitation.
| What a regulator would ask | What the clinic should be able to show |
|---|---|
| Did the patient consent to this kind of message? | A timestamped record of consent, with the route and the wording used |
| Was the content within the stated purpose? | Care reminders and campaign invitations held under separate consents |
| Who inside the clinic could see the message? | Named user accounts with role-based permissions, no shared logins |
| How long is the conversation kept? | A documented retention period applied automatically |
| Can the patient opt out or request erasure? | A working opt-out path and a documented erasure process |
| Was any health detail exposed unnecessarily? | Minimal content in the message; clinical detail kept in the record |
A practical discipline that helps in both jurisdictions: keep the broadcast itself thin. A message that says a service is available and invites the patient to reply reveals far less than one that names a condition or a treatment history, and it achieves the same booking outcome.
imBee provides role-based access, retention controls and audit logging so a clinic can support these obligations. The clinic remains the data user and sets its own policies; no platform discharges that duty on your behalf.
The technical send is the easy part. What breaks clinics is the inbound: questions about eligibility, price, timing, and the occasional clinical concern that must reach a clinician the same day.
Send in waves, not one blast. Start with a few hundred patients, watch the reply volume for a day, then scale. A single send to the whole list on a Monday morning guarantees a bad week.

Decide what a non-clinician may answer. Write the boundary down: scheduling, price and location at reception; anything touching eligibility, contraindications or symptoms goes to a clinician with a defined turnaround.
Give the patient a booking path inside the message. If the only way to act is to phone during office hours, the campaign converts poorly and the phones jam.
Set the after-hours expectation explicitly. Patients message at 22:00. Say in the template when a reply will come, and never imply that the channel is monitored for urgent clinical matters. Direct anything urgent to the appropriate service.
Measure bookings, not sends. The number that matters is appointments created per thousand messages, and the opt-out rate alongside it. High delivery with a rising opt-out rate means the targeting is wrong.
After the first wave, review the questions patients actually asked. They are a precise list of what the template failed to say, and fixing the wording between waves usually lifts conversion more than any change of channel.
Reminder and campaign messaging solves a logistics problem: people who intend to attend but have not got around to it. It does very little for the other reasons people do not come, and buying a messaging platform to fix those is a misdiagnosis.
It does not change a considered decision. A patient who has weighed a vaccine and declined is not persuaded by a second message. Repeated sends to that group produce opt-outs and complaints, not bookings. Suppress them.
It does not fix capacity. A campaign that generates more demand than the clinic can absorb converts goodwill into a waiting queue. Model the slots before you model the sends.
It does not substitute for clinical advice. A broadcast cannot assess eligibility, interactions or contraindications, and any template that implies it can is a liability. The message opens a conversation; the consultation makes the decision.
It does not repair a list you cannot lawfully use. If consent for promotional health messaging was never collected, no amount of platform capability makes the send appropriate. The fix sits at registration, one patient at a time.
| Symptom | Likely cause | Where the fix sits |
|---|---|---|
| High delivery, few bookings | No booking path in the message | Template design |
| Rising opt-out rate | Sending campaigns to patients who declined | Suppression list and segmentation |
| Templates repeatedly rejected | Campaign content written as a reminder | Category discipline, one purpose per template |
| Reception overwhelmed | Whole list sent at once | Wave sending and a reply playbook |
| Clinical questions unanswered | No escalation rule to a clinician | Triage boundary with a defined turnaround |
Set that out before procurement and the evaluation stays honest. WhatsApp campaign messaging is a strong fix for reach and convenience in a patient group that already trusts the clinic. It is not a persuasion engine, and it will not create appointment capacity that does not exist.
Can clinics send vaccination reminders on WhatsApp in Hong Kong?
Yes, provided the clinic holds the patient's consent and sends through the WhatsApp Business Platform using an approved template. The Personal Data (Privacy) Ordinance (Cap. 486) governs how contact details may be used, and a reminder about care the patient arranged is treated differently from a promotional campaign invitation.
Is a vaccination campaign message Utility or Marketing?
A reminder for a dose the patient has already booked is normally Utility because it relates to an existing arrangement. An invitation to come in for a seasonal shot the patient never asked about should be planned as Marketing, with explicit opt-in, an opt-out line and marketing pricing.
What consent do we need before sending health campaign messages?
Explicit, purpose-specific opt-in, recorded with a timestamp and the route it came through. Treat consent to receive appointment messages and consent to receive health promotions as two separate permissions, because a patient who agreed to the first has not agreed to the second.
Does the PDPA restrict health campaign messages in Singapore?
The Personal Data Protection Act requires consent, limits use to the purpose the data was collected for, and imposes notification obligations. Marketing messages sent to Singapore numbers attract additional checks under the Do Not Call provisions, while a transactional reminder for a booking the patient made is treated differently.
How much patient detail should a broadcast contain?
As little as possible. A message that says a service is available and invites the patient to reply reveals far less than one naming a condition or treatment history, and it produces the same booking outcome. Keep clinical detail in the consultation and in the clinical record, not in the message.
Why do vaccination templates get rejected?
Most often because campaign content was written as though it were a reminder, mixing a promotional offer into a message submitted as Utility. Other frequent causes are efficacy or comparative clinical claims, missing opt-out wording on promotional templates, and one template trying to serve several unrelated purposes at once.
Should we send the whole patient list at once?
No. Send in waves of a few hundred and watch reply volume for a full day before scaling. A single send to the entire list generates a week of inbound questions, and reception has no way to triage clinical concerns quickly when everything arrives in the same morning.
What should we measure to know the campaign worked?
Appointments created per thousand messages, and the opt-out rate alongside it. Delivery and read counts describe the channel, not the outcome. A rising opt-out rate with high delivery is a targeting problem, and it is the earliest warning that the list needs segmenting more tightly.

Kelly S.
Content Team Lead, imBee
Kelly S. owns content strategy, product positioning, and customer education at imBee. Previously, Kelly led B2B SaaS content programs and supported go-to-market initiatives for customer engagement products. On the imBee blog, Kelly covers conversational commerce, omnichannel messaging, WhatsApp Business, customer experience, and strategies for scaling business communications.
Questions about anything in this article? Talk to our team.


