Key takeaways
- Pick a route per message type: live transfer for emergencies, a secure app for urgent clinical messages, the EHR inbox for anything that needs follow-up.
- Plain texts should carry no patient details, only “new message waiting”; transmission security and audit controls are Security Rule standards[2].
- Every message needs a status trail: delivered, opened, acted on. Unanswered alerts escalate to a backup on a timer.
- Reconcile the service's log against the EHR every morning, and keep office questions out of the clinical message stream altogether.
The last mile is where messages go wrong
An answering service can take a careful message and still lose it in the last mile, the stretch between the operator's screen and a clinician who acts on it. Most services offer several delivery routes, and many practices end up using a different one for each kind of message without ever writing that down. The result is a 10 pm page that went to a physician who swapped call, a morning batch nobody opened, and a patient name sitting in a text thread on a personal phone.
- Plain text or pager message to the clinician's phone.
- Secure messaging app with its own login, where the message lives on the vendor's system and the phone only gets an alert.
- Email, either ordinary or encrypted, to a person or a shared inbox.
- EHR inbox or task, through an interface or a secure portal the service writes into.
- Live transfer or phone patch, where the clinician speaks to the caller directly.
None of these is wrong on its own. The mistake is letting the route be chosen by habit instead of by the message: how urgent it is, how much patient information it carries, and whether it needs to become part of the record.
What the rules ask of each route
Start with the vendor. HIPAA lets a covered entity have an answering service handle protected health information on its behalf only with satisfactory assurance, in a business associate agreement, that the service will safeguard it[3]. Every route the service uses to send you messages sits inside that agreement, so ask for the delivery methods to be named in it.
The Security Rule then sets the technical bar. It requires audit controls that record and examine activity in systems holding electronic patient information, and transmission security to guard against unauthorized access to that information as it moves over a network. Encryption in transit is an addressable specification[2], which means you either use it or document why an equivalent measure is reasonable. For a message carrying a patient's name and symptoms to a phone, it is hard to write that justification.
Encryption also changes what happens when a phone goes missing. Breach notification applies to unsecured protected health information, defined as information not rendered unusable, unreadable or indecipherable to unauthorized people through methods the HHS Secretary specifies[4]. An impermissible disclosure is presumed to be a breach unless a risk assessment of four factors shows a low probability of compromise[4], and notices to patients are due no later than 60 days after discovery[5]. A plain text thread on a lost personal phone puts you on that clock; an alert that says only “new message waiting” usually does not.
Hospitals have a direct federal answer on texting. Since February 2024, CMS has said that texting patient information and orders among the care team is permissible through a HIPAA compliant secure texting platform, with computerized order entry still preferred for orders[6]. That memo applies to hospitals and critical access hospitals, not to independent practices, but its logic is the standard most practices are measured against: secure platform, yes; ordinary text, no.
Comparing the routes
| Route | Protects content in transit | Audit trail | Shows it was read | Lands in the chart | Best use |
|---|---|---|---|---|---|
| Plain text or pager | Not built for it | Carrier logs only | No | No | A content-free alert that a message is waiting |
| Secure messaging app | Yes, when configured | Yes | Usually, read receipts | By export or interface | Urgent clinical messages overnight |
| Encrypted email | Yes | Partly | Rarely | Manual filing | Routine messages for the morning |
| Ordinary email | Not reliably | Mailbox only | No | Manual filing | Messages with no patient details |
| EHR inbox or task | Yes | Yes | Task status | Yes | Anything that needs follow-up in the record |
| Live transfer | Voice call | Recording if enabled | Yes, answered | Note written after | Emergencies and anything that cannot wait |
Most practices settle on three routes: live transfer for anything that cannot wait, a secure app for urgent clinical messages overnight, and the EHR inbox or a secure portal for everything that needs action in the morning. Ordinary email and plain text are kept for content-free alerts and for messages that carry no patient details, such as a vendor confirming a schedule change.
The content-free alert
The pattern that solves most of the texting problem is simple. The plain text carries no patient information at all. It tells the clinician that a message is waiting and how urgent it is, and the message itself lives in the secure app behind a login. If the alert is not opened within an agreed time, the service escalates.
Example alert text: “Brookside Pediatrics: 1 URGENT message waiting in the secure app. Open within 10 min or the service will call you.” No name, no date of birth, no reason for the call.
- Your team
- System
01Your team
Operator files the message
Name, callback number and reason go into the secure portal, marked urgent.
02System
Content-free alert
The on-call phone gets “1 urgent message waiting” with no patient details.
03Your team
Clinician opens the app
A read receipt is logged; the clinician calls the patient from the office line.
04System
No read in 10 minutes
The service phones the clinician, then the named backup on the list.
05System
Filed to the EHR
Message and callback note reach the chart task list before the morning huddle.
Example sequence. The timer, backup order and filing route come from your agreement with the service.
The escalation timer matters as much as the encryption. A secure message that sits unread for an hour is safe from eavesdroppers and dangerous to the patient. Agree the timer by urgency, for example 10 minutes for urgent and the morning for routine, and make sure the backup clinician's number is current every week. The on-call side of this is covered in on-call escalation rules.
Logging and the morning reconciliation
Every message should carry a status trail: sent, delivered, opened, acted on. That trail protects the patient, and it protects the clinician when a family later says nobody called back. Ask the service for a daily delivery report, not just a list of messages.
Morning inboxes are already full. In 2024, 70% of medical groups reported more patient portal messages than the year before, and only 1% reported fewer[7]. One time-and-motion study found inbox management took 85 minutes of a primary care physician's daily EHR time[1]. Overnight messages from the answering service land on top of that, which is why they need an owner: 71% of medical group leaders say someone is designated to handle inbox messages for other providers[8].
- Reported an increase 70%
Source: Medical Group Management Association (MGMA), 2025[7]
Show the data
| Response | Share of groups |
|---|---|
| Reported an increase | 70% |
Pull the service's overnight log
Every message with its time, route, urgency and delivery status. Do it at the same time each morning, before the first patient.
Match it to the chart
Each message should appear as an EHR task or note. Anything in the log but not in the chart gets filed now.
Check every urgent one
Confirm the clinician opened it and called back, from the read receipt and the callback note. Missing either is escalated to the physician lead.
Clear the routine ones
Refills, appointments and forms go to their owners with a promised time. Patients hear back the same day.
Note the failures
Wrong routes, late alerts and missing callbacks go on a list that the practice manager reviews with the service each month.
What re-keying costs
When messages arrive by ordinary email or fax, somebody types them into the EHR. That is slow, and every retyped message is another chance for a wrong digit in a callback number. The model below prices the nurse time spent re-keying, using the median registered nurse wage.
| Input | Example |
|---|---|
| Overnight messages per weekday morning | 18 |
| Minutes to re-key and file one | 4 |
| Share arriving by email or fax | 80% |
| Nurse hourly wage | $46.90 |
| Calls a month | 391 |
| Staff hours on these calls | 26 |
| Hours handed off | 21 |
| Monthly value of that staff time | $977 |
How it is worked out: Calls a day × 21.7 working days × minutes per call ÷ 60 × share handled = hours; hours × hourly cost.
Assumptions
- Hourly wage of $46.90, the May 2025 median pay for US registered nurses, before benefits.[9]
- Re-keyed messages add to an inbox that already takes about 85 minutes of a primary care physician's EHR day.[1]
- Eighteen messages, four minutes each and an 80% share are made-up inputs; count a week of your own.
Example only: about 21 nurse hours, roughly $980 a month, go into re-keying in this model. An interface or secure portal that files messages directly gives most of it back.
Keep office questions out of the clinical stream
The cheapest secure message is the one never created. A large share of after-hours calls are office questions: hours, holiday closures, directions, the fax number, whether you take a plan. When an operator turns each one into a message, it enters the same delivery path as a post-operative bleed, gets the same alert and needs the same morning filing. Answering those questions on the call keeps them out of the clinical stream entirely.
That is the slice an SI Employee can take today. With your number forwarded after hours, it answers office and new-patient questions from information your practice signs off, gives the approved 911 wording when a caller describes an emergency, and puts anything patient-specific through to the staff or on-call line you choose. It takes no health details, and AssistCall does not sign business associate agreements today, so it is not a message route for clinical content. Keep your secure app and EHR inbox for that. The specialty practice example shows the split, with the nurses' message line kept inside the practice's own phone system. What it does lists the current capabilities.
Questions to put to your answering service
- Which delivery routes do you support, and which are named in our business associate agreement?
- Can plain-text alerts be set to carry no patient details, for every clinician on our list?
- Does your app show us when a message was delivered and opened, and can we export that log?
- What is the escalation timer for an unopened urgent message, and who do you call next?
- Can messages be filed into our EHR or a secure portal directly, or do they arrive by email or fax?
- How do you handle a clinician who swaps call at short notice, and who updates the list?
- What happens to messages on a clinician's phone when they leave the practice?
Read the answers against the wider checklist in HIPAA-compliant answering services, the internal side in secure staff messaging, and the morning routine in handing off after-hours messages. If your practice also texts patients directly, texting patients and HIPAA covers that separate set of rules.
Frequently asked questions
Can our answering service text patient details to the on-call doctor?
It is safer not to. Send a content-free alert by text and keep the message in a secure app. Transmission security is a Security Rule standard and encryption in transit is addressable[2], so plain texts with patient details are hard to justify.
Is email from the answering service acceptable?
For routine morning messages, encrypted email or a secure portal can work if the route is covered by your business associate agreement[3]. Ordinary email suits only messages with no patient details.
What if the on-call clinician never opens the message?
The service should escalate on a timer: a phone call to the clinician, then the named backup. Agree the timer by urgency and check the backup list every week.
What happens if a phone with message alerts is lost?
If the phone held only content-free alerts and the messages sat in an encrypted app, the patient information was likely never exposed. Plain texts with patient details may be unsecured information, which starts a breach assessment[4].
Who should own the morning reconciliation?
One named person, usually a nurse or the practice manager, with a backup. Many groups already designate someone for inbox messages[8]; overnight answering-service messages belong on the same list.
References
Annals of Family Medicine (2017). Tethered to the EHR: Primary Care Physician Workload Assessment Using EHR Event Log Data and Time-Motion Observations (opens in a new tab) ↑
Legal Information Institute, Cornell Law School (current). 45 CFR 164.312 - Technical safeguards (opens in a new tab) ↑
Legal Information Institute, Cornell Law School (current). 45 CFR 164.502 - Uses and disclosures of protected health information: General rules (opens in a new tab) ↑
Legal Information Institute, Cornell Law School (current). 45 CFR 164.402 - Definitions (Breach Notification) (opens in a new tab) ↑
Legal Information Institute, Cornell Law School (current). 45 CFR 164.404 - Notification to individuals (opens in a new tab) ↑
Centers for Medicare & Medicaid Services (2024). Texting of Patient Information and Orders for Hospitals and CAHs (QSO-24-05-Hospital/CAH) (opens in a new tab) ↑
Medical Group Management Association (MGMA) (2025). Ensuring accurate coding and billing for patient portal messages as e-visits (opens in a new tab) ↑
Medical Group Management Association (MGMA) (2026). Who owns the inbox? How medical groups can redesign portal message work (opens in a new tab) ↑
U.S. Bureau of Labor Statistics (2025). Occupational Outlook Handbook: Registered Nurses (opens in a new tab) ↑