Key takeaways
- Every reply should read the same whether or not the sender is a patient: no confirmations, no health details, one clear next step.
- Give public information freely in a DM; move anything personal to the phone or portal, where staff can verify identity first[2].
- Saved replies for appointments, clinical questions, billing, emergencies and third parties keep weekend answers consistent.
- Staff the inbox like a phone line: an owner, a backup, an honest away message, same-day escalation and a monthly ten-thread audit.
What actually lands in the inbox
Open a practice's Facebook or Instagram inbox on a Monday morning and the messages sort themselves into a few piles. Someone wants to know if you take their insurance. Someone wants an appointment on Thursday. Someone writes three paragraphs about a rash, asks whether their lab results are back, or is angry about a bill. Now and then a message reads like an emergency, and it was sent at 11 pm on Saturday when nobody was watching.
The volume follows the platforms patients already use. In Pew Research Center's 2025 survey, 71% of US adults said they use Facebook and half said they use Instagram[1]. If your practice has a page, patients will message it, whatever your website says about calling the office.
None of these messages is a problem in itself. Patients may tell you anything they like, in any channel. The risk sits in the reply: what staff confirm, what they ask for and what they promise. This guide covers the replies, the routing behind them and who owns the inbox. What employees post on their own accounts belongs in your staff social media policy, and answering public ratings is covered in responding to online reviews.
Two rules behind every reply
Rule one: never confirm that the sender is a patient. A reply such as “Sorry about the wait at your visit yesterday” tells anyone who can see the thread that this person was seen at your practice. On a shared family tablet or a hijacked account, that is a disclosure to a stranger. You also have no idea who is typing. Before disclosing protected health information, HIPAA requires you to verify the identity and authority of a person you do not already know[2], and a profile name and photo verify nothing.
Rule two: never handle care inside the DM. A practice may let a vendor create, receive or keep patient information for it only with satisfactory assurance, in a business associate agreement, that the vendor will safeguard it[3]. Consumer messaging apps are not set up as your business associates. When HHS named Facebook Messenger in its 2020 enforcement notice, it was a temporary allowance for video visits during the COVID-19 public health emergency, alongside a warning that public-facing tools such as Facebook Live should not be used at all[4]. That emergency is over, and the allowance went with it.
So the job of a reply is narrow: be polite, give public information freely, and move anything personal to a channel where staff can check identity and keep a record. Everything below follows from those two rules.
Public information is fine to give in a DM: hours, address, parking, which insurance plans you accept, whether you are taking new patients and how to sign in to the portal. It is the same answer you would give anyone who phoned.
Message types and the reply that fits
Staff should sort by what the message asks for, not by who seems to be asking. The reply to a stranger and the reply to a patient of twenty years should be word for word the same.
| Message | Do not | Do |
|---|---|---|
| Hours, location, insurance accepted | Ask whether they are a patient | Answer in full, exactly as you would on the phone |
| Appointment request | Ask for a date of birth or the reason for the visit in the thread | Give the booking line or portal link and the hours it is staffed |
| Results, refills, symptoms | Reply with any detail, even “your results look fine” | Send the clinical template with the nurse line and portal link |
| Bill or payment complaint | Quote balances, dates of service or insurer names | Give the billing office number and its hours |
| Emergency language | Wait for business hours or try to assess it | Send the emergency template at once and tell the manager |
| Asking about someone else | Confirm or deny that the other person is a patient | Say you cannot confirm who is a patient and give the office number |
| Public comment with health details | Reply with details or argue in public | Post the general reply and invite them to call or use the portal |
Reply templates staff can paste
Keep these as saved replies in the inbox tool so nobody writes from scratch at 4:55 on a Friday. Each one gives a next step and a way to reach a person, and none of them confirms anything.
- Example, general: “Thanks for reaching out. We can't discuss personal or health information over social media. Please call us at [main number], Monday to Friday 8 to 5, or send a message through the patient portal at [link].”
- Example, appointment: “We'd be glad to help with scheduling. Please call [booking number] or book through [portal link]. For your privacy, please don't send birth dates or health details here.”
- Example, clinical: “For questions about results, medicines or symptoms, please call our nurse line at [number] or use the patient portal, so our clinical team can confirm who they are speaking with and help.”
- Example, billing: “Our billing office can look into this by phone at [number], weekdays 9 to 4. We can't review accounts over social media.”
- Example, emergency: “If this is a medical emergency, call 911 now. If you are in crisis or thinking about suicide, call or text 988, which answers 24 hours a day. This inbox is not watched around the clock.”[5]
- Example, about someone else: “We can't confirm whether anyone is a patient here. If you help care for someone, please call [number] and our staff will explain how we share information with family members.”
Notice what is missing: no “we see you were in last week”, no “let me check your chart”, no request for a birth date to look someone up. If the sender volunteers details, do not repeat them back, even to sound sympathetic. The same discipline applies when staff answer patients by email.
Moving the conversation somewhere you control
A template only works if the next step is staffed. A DM that says “call us” at 7 pm on a Friday sends the patient to a voicemail box, and they come straight back to the inbox, often angrier and more detailed. Decide where moved conversations land and make sure that route picks up.
- Caller
- Your team
- SI Employee
01Caller
Message arrives
A patient asks in a DM whether a refill request went through.
02Your team
Template reply sent
Staff send the clinical template with the nurse line and portal link.
03Caller
Patient calls the office
The call reaches the main number, forwarded after hours.
04SI Employee
Call answered
Office questions answered; the refill question goes to the staff line.
05Your team
Identity verified
Staff confirm who is calling before discussing the refill.
06Your team
Handled and noted
The answer is given by phone or portal and noted in the chart.
Example sequence. Nothing about the refill is discussed in the DM or by the answering step.
Step four matters after hours. If your line forwards to an SI Employee for lunch, overflow or evenings, it answers the general questions from the information you approve and passes test results, refills, bills and appointments to the staff line you choose, without taking health details. The family practice example shows that split when the phones are busiest, and what it does lists the rest. Whatever covers your phones, the number in your DM templates must reach something that answers.
Some patients will say they would rather keep messaging. If a patient asks to receive communications in a particular way, that is a confidential communications request: you must accommodate reasonable ones and cannot ask them to explain why[6]. Pass those to the privacy officer, who can agree a secure option such as the portal or a phone number the patient chooses, rather than leaving the person on social media duty to improvise. Matching each kind of request to the right route is covered in portal, phone or text.
Public comments need their own playbook
A comment under a post is riskier than a DM because everyone can read it. When a patient writes “Still waiting on my biopsy results!” under your flu-shot announcement, the information is out, but the practice has disclosed nothing. A reply saying “We've sent your biopsy to Dr. Patel” would. Post the general reply, invite them to call, and leave out the test, the doctor and the visit.
Decide in advance whether staff hide, delete or leave comments that contain health details, write it down and apply it the same way every time. Hiding a patient's own details is a reasonable courtesy; deleting criticism looks like suppression and usually invites more. Keep the decision with the person who plans what the practice posts, as set out in social media for medical practices.
Patient stories are the other trap. A patient who writes “You can share my story!” has not given you HIPAA authorization. Using a patient's information in marketing needs a signed authorization, apart from face-to-face conversations and promotional gifts of nominal value[7]. Thank them, send the authorization form, and post nothing until it comes back signed.
Run the inbox like a phone line
Name an owner and a backup
One person checks the inbox at set times, for example 9 am, noon and 3 pm, and a named backup covers holidays and sick days. Unowned inboxes are where emergencies sit unread.
Set an honest away message
Outside those times an automatic reply gives the main number, the 911 line and when the inbox is next checked. Keep its wording identical to the general template.
Say what you cannot staff
If nobody reads messages at weekends, say so in the page's About section and in the away message rather than letting patients guess.
Escalate the same day
Emergency language, complaints that mention harm, and anything from a journalist or lawyer go to the practice manager within the hour.
Log and keep the paperwork
Note the date, the template used and where the conversation was sent. Keep the written policy and training records for six years, as HIPAA requires for compliance documentation[8].
Audit monthly
Pick ten threads at random and check three things: no confirmation that anyone is a patient, no health details repeated back, and a working next step in every reply.
Example: a practice answering about 40 DMs a month runs its ten-thread audit and finds two replies that mention a recent visit. That is a template problem, not a people problem. The fix is to load the saved replies into the inbox tool, retire the free-typed ones and spend ten minutes at the next staff meeting walking through the table above. Repeat the audit the following month to confirm it stuck, and file the result with your HIPAA training records.
Frequently asked questions
Is it a HIPAA violation if a patient sends us health details in a DM?
No. Patients can share their own information however they like. Your obligations start with the reply: do not repeat the details, confirm they are a patient or continue the conversation there. Move it to the phone or portal, where staff can verify identity before disclosing anything[2].
Can we reply “we'll call you” to someone we recognize as a patient?
Recognizing a name is not verification, and promising a call confirms a relationship to anyone who can see the thread. Ask them to call the office or use the portal so the conversation starts where staff can confirm who they are.
Should we just turn off direct messages?
You can, but patients will then comment publicly instead, which is harder to manage. Most practices do better with a staffed inbox, an honest away message and saved replies than with a closed door.
Can we send appointment reminders through Messenger or Instagram?
Not without a vendor agreement that covers that channel. A practice may let a vendor handle patient information only with satisfactory assurance in a business associate agreement[3]. Use your reminder system instead.
A relative is asking about a patient in a DM. What do we say?
Do not confirm or deny that the person is a patient. Give the office number and say staff will explain how the practice shares information with family members and caregivers.
References
Pew Research Center (2025). Americans' Social Media Use 2025 (opens in a new tab) ↑
Legal Information Institute, Cornell Law School (Electronic Code of Federal Regulations text) (current). 45 CFR § 164.514 - Other requirements relating to uses and disclosures of protected health information (opens in a new tab) ↑
Legal Information Institute, Cornell Law School (Electronic Code of Federal Regulations text) (current). 45 CFR § 164.502 - Uses and disclosures of protected health information: General rules (opens in a new tab) ↑
U.S. Department of Health and Human Services, Office for Civil Rights (Federal Register via GovInfo) (2020). Notification of Enforcement Discretion for Telehealth Remote Communications During the COVID-19 Nationwide Public Health Emergency (85 FR 22024) (opens in a new tab) ↑
Substance Abuse and Mental Health Services Administration (SAMHSA) (2026). 988 Suicide & Crisis Lifeline (opens in a new tab) ↑
Legal Information Institute, Cornell Law School (Electronic Code of Federal Regulations text) (current). 45 CFR § 164.522 - Rights to request privacy protection for protected health information (opens in a new tab) ↑
Legal Information Institute, Cornell Law School (Electronic Code of Federal Regulations text) (current). 45 CFR § 164.508 - Uses and disclosures for which an authorization is required (opens in a new tab) ↑
Legal Information Institute, Cornell Law School (Electronic Code of Federal Regulations text) (current). 45 CFR § 164.530 - Administrative requirements (opens in a new tab) ↑