To verify a patient's identity over the phone, ask the caller for two identifiers, usually full name and date of birth, check them against the chart before you say anything about the patient, and ask one more question when two records match. That is the whole protocol for most calls. The rest of this post is about the calls where it is not enough: the caller who is not the patient, the two patients with the same name, and the requests that should never be settled by phone at all.
Patient identity verification by phone is the check a practice runs before it discloses or changes anything in a patient's record during a call: the caller supplies identifiers, staff match them to exactly one chart, and the caller's authority to act is confirmed if they are not the patient. HIPAA requires verification when the caller's identity or authority is not already known, and leaves the method to the practice.
This is general information about federal rules, not legal advice. State law on minors, guardianship and sensitive services varies, and your policy should reflect your state.
What does HIPAA require to verify identity over the phone?
Less detail than most people assume. Under 45 CFR 164.514(h), before a permitted disclosure a covered entity must verify the identity of a person requesting protected health information, and their authority to have it, if the identity or authority is not already known. The regulation does not say how. HHS guidance on the right of access says the rule does not mandate any particular form of verification, such as a copy of a driver's license, and leaves the method to the practice's discretion and professional judgment, as long as it does not create unreasonable barriers for patients.
That last point is getting firmer. HHS published a final rule on November 19, 2026 that, among other changes, prohibits verification requirements that unreasonably impede patients from exercising their access rights where less burdensome measures are practicable, with compliance required in July 2027. Making a patient come in to prove who they are, when two identifiers on the phone would do, is the kind of barrier it is aimed at.
Two other standards shape the protocol. The Joint Commission's National Patient Safety Goal on identification requires at least two patient identifiers when providing care, and lists the individual's name, an assigned identification number, telephone number or another person specific identifier as acceptable. It is written for care and treatment rather than phone calls, but its rationale transfers: wrong patient errors occur in virtually all stages of diagnosis and treatment. And for callers who are not the patient, 45 CFR 164.510(b) limits what you share to what is directly relevant to their involvement in the patient's care or payment.
What is a good protocol to verify patient identity by phone?
- Ask, do not tell. Ask for full name and date of birth. Never read details out for the caller to confirm, such as "is this Maria Lopez, born March 3rd?", because that hands the identifiers to whoever is on the line
- Match to exactly one chart. If two records match, ask a distinguishing question: the address on file, the phone number on file, or roughly when they were last seen
- Confirm the caller's role. Are they the patient? If not, who are they and what is their relationship to the patient? That decides what you may discuss
- Act only within the request. Rescheduling an appointment needs less than discussing a referral or a bill
- Escalate when it does not add up. A caller who gets the date of birth wrong twice, or hesitates over their own address, goes to a senior staff member or gets a callback at the number in the chart
Caller ID is not an identifier. The FCC notes that caller ID can be altered to display a different number, and that this has legitimate uses too, such as a doctor calling from a personal mobile phone and showing the office number. A familiar number on the screen tells you which phone is calling, not who is holding it.
The single most useful habit: when something feels off, offer to call back at the number in the chart. A real patient says yes. Someone fishing for information usually does not.
Which identifiers should you use, and which should you avoid?
| Identifier | Use it? | Why |
|---|---|---|
| Full name | Yes, always | One of the two standard identifiers |
| Date of birth | Yes, always | The second standard identifier, and patients know it without looking |
| Address or phone number on file | Yes, to break a tie | Separates two patients with the same name and date of birth |
| Date of last visit or provider name | Sometimes | Useful as a third check, less reliable for patients seen rarely |
| Full Social Security number | No | Not needed to schedule care, and Medicare cards now carry a separate identifier; asking for it by phone adds risk for nothing |
| Caller ID | No | Can be altered, and is often a shared phone |
| Anything you read out first | No | A caller who agrees with what you said has proved nothing |
Medical identity theft is the reason this matters beyond privacy. The FTC describes it as someone using another person's information to get care, prescriptions or insurance claims, and warns that if the thief's health information is mixed into a real patient's record, it could affect the care that patient receives. A careful identity check on the phone is one of the cheaper places to stop it.
How do you verify a caller who is not the patient?
This is where most phone mistakes happen, because what you may say depends on who is calling.
| Caller | What the rules allow | What to do |
|---|---|---|
| Parent of a minor | Usually acts for the child, but state law on minors' consent can carve out some visits | Verify the child's identifiers and the parent's name against the chart, and check for restricted visit types |
| Spouse, adult child or friend | Information directly relevant to their involvement in care or payment, under 45 CFR 164.510(b) | Check the chart for people the patient has named; when unsure, take a message for the patient |
| Guardian or health care power of attorney | Treated as the patient within the scope of their authority | Confirm the document is on file before relying on it |
| Pharmacy or another provider | Disclosures for treatment are permitted | If you do not recognize the caller, call back on a published number |
| Insurer, attorney or police | Different rules, often requiring documentation | Transfer to the practice manager; never settle at the front desk |
HHS has an FAQ for the family caller specifically. If a caller says they are a family member or friend involved in the patient's care or payment, HIPAA does not require proof of their identity, and a provider may set its own rules for verifying who is on the phone. What it shares is limited to what that person needs to know about the patient's care or payment. For someone who is not family or a friend, HHS says the provider must be reasonably sure the patient asked that person to be involved.
What should never be settled over the phone?
- Changing the phone number, address or email on file without a second check, since that redirects every future contact
- Reading results, diagnoses or medication lists to anyone whose identity you have not confirmed
- Requests from anyone claiming legal authority who has no documentation on file
- Adding a new person to the patient's list of approved contacts on the say so of the caller rather than the patient
None of these is forbidden by phone in every case. They are the requests where a mistake does lasting damage, so they deserve a slower path: a callback to the number in the chart, a portal message, or a conversation at the next visit.
What if a caller will not give their date of birth?
Explain why you are asking, once: "I ask everyone for a date of birth so I do not open the wrong person's chart." Most callers accept that. If they still decline, you can give general information that is not about any patient, such as hours, location and accepted plans, and offer a callback at the number on file for anything else. Do not drop the check because the caller is impatient. That is the exact moment it is doing its job.
How does an AI phone agent verify identity?
The same way a careful receptionist does, with less tolerance for a near match. On athenahealth, MedPhone confirms the caller's identity against the patient record before it books, reschedules or cancels anything, and it transfers to staff rather than proceeding on a probable match. Booking against the wrong record is a worse outcome than handing the call to a person.
The limitation is judgment about third parties. Whether a particular spouse or adult child may hear about a particular patient is a professional judgment the rule assigns to the practice, and we recommend that practices set MedPhone's transfer rules so those calls go to staff. MedPhone also gives no clinical advice, so a verified caller with a clinical question is transferred, not answered.
Sources
- 45 CFR 164.514(h), verification requirements, Legal Information Institute
- 45 CFR 164.510(b), disclosures to family members and others involved in care, Legal Information Institute
- HHS, Individuals' right under HIPAA to access their health information, including verification
- HHS, HIPAA FAQ: proof of identity when a family member calls
- Covington, Inside Privacy: HHS finalizes changes to the HIPAA Privacy Rule, November 2026
- The Joint Commission, National Patient Safety Goals, Ambulatory Health Care, January 2026
- FTC, What to know about medical identity theft, 2024
- FCC, Caller ID spoofing consumer guide
Questions people ask
How do you verify a patient's identity over the phone?
Ask the caller for their full name and date of birth, match both to exactly one chart, and ask a distinguishing question such as the address on file if two records match. Confirm whether the caller is the patient before discussing anything, and offer a callback to the number in the chart when something does not add up.
Is name and date of birth enough to verify identity by phone?
For most routine calls, yes, provided the caller supplies them rather than confirming details you read out. Add a third check, such as the address on file, when two charts match or when the request is higher risk, such as changing contact details.
Does HIPAA require two patient identifiers?
No. HIPAA requires a practice to verify identity and authority when they are not already known, and leaves the method to the practice. The two identifier standard comes from the Joint Commission's patient safety goals for clinical care, and it is a sensible default for phone calls too.
Can a front desk talk to a patient's spouse on the phone?
Within limits. Under 45 CFR 164.510(b), a provider may share information directly relevant to a family member's involvement in the patient's care or payment, and HHS says proof of identity is not required when a caller says they are family involved in care. Check the chart for the patient's wishes first, and take a message when unsure.
What if a caller refuses to give a date of birth?
Explain that you ask everyone so you do not open the wrong chart. If they still decline, give only general information such as hours and location, and offer a callback at the number on file for anything about the patient.
Should a medical office ask for a Social Security number to verify identity?
Generally no. Name, date of birth and a detail from the chart are enough for phone verification, and Medicare has replaced Social Security based numbers on its cards with a separate Medicare Beneficiary Identifier. Collecting a Social Security number by phone adds exposure without adding much certainty.
Written by
Agni Patel is the founder and CEO of MedPhone, an AI phone agent for medical practices. Its rule is to transfer to staff rather than act on a probable match, and he would like your front desk to have one as clear.