A phishing email disguised as a message from a prescriber is harmless if no one responds to it. Even so, responding with any patient information generally constitutes an unauthorized disclosure and likely a breach. Clicking to log in may allow access to all messages received by that user, and an inappropriate refill or other order change could adversely affect a patient.

Of 170 breaches involving email in 2025, 17% resulted from phished user login credentials, according to Paubox’s 2025 report. Emails purporting to be from physicians are well suited to this type of phishing attack, given the expectation that clinic and other personnel will respond promptly to a prescriber's inquiry.

It is useful to consider what measures will prevent such messages from arriving, and how to respond if they do.

 

Why a prescriber's request gets acted on quickly

It is common for various staff to receive and respond to inquiries and other communications from prescribers. An email purporting to request information on current medications would not stand out.

The authors of one study conducted live phishing against a Norwegian hospital described phishing as a dominant attack vector in health care, “Exploiting psychological factors such as urgency and authority.” Fifty-three percent of 753 individuals not included in one of the intervention arms clicked on the phishing link. These findings should be interpreted with caution, being limited to one hospital in Norway.

The only notable predictor of clicking was high workload, based on survey answers from 397 employees in two hospital networks, which were linked to their recorded responses to simulated phishing attempts. However, only 12.8% of those approached completed the Journal of Medical Internet Research survey.

 

How attackers make an email look like a prescriber's

An email could use the prescriber's name but an unaffiliated sending address; an address that differs only slightly from the prescriber's actual domain; or the prescriber's own address if their login details had been stolen.

An attacker needs little information. Public sources provided email addresses for 468 staff working at one children's hospital in London, for use in phishing attempts, although none led to stolen user log-ins. When sending customized phishing emails to more than 6,000 staff at one Italian hospital, the authors found that the staff were "customization of phishing emails makes them much more likely to be acted on" to clicking. Some deviations from the intended protocol occurred.

Phishing emails may include information the target would be aware of. For example, an email with the display name "Lena Ortiz, MD" and specialty "Cardiology" may mention an appointment the recipient knows is booked for one of their patients, and request a list of current medications. Only the display name may be apparent when viewing an email on mobile.

Read also: Display name spoofing: A root cause of many cyberattacks

 

When a reply with patient information becomes a breach

A provider may disclose protected health information (PHI) to another provider for treatment purposes under 45 CFR 164.506, including responding to the patient's cardiologist with relevant information. Where the identity and authority of the person requesting information are unknown, however, reasonable steps are required to confirm both, as noted in 45 CFR 164.514(h).

Any information disclosed in response to such a message would be impermissible. 45 CFR 164.402 states that it would be considered a breach, absent a determination following risk assessment that there is a low probability of compromise. The recipient would be relevant, and it is difficult to conclude that there would be a deliberate request from a criminal makes a low-probability finding hard to support

The value of medical records to malicious actors was recognized by the authors, who identified staff emails from open sources. It would depend on what information the recipient chose to include; confirming the time of an appointment would be less serious than providing details of the patient's medications and conditions.

 

When a click turns one mailbox into a larger incident

Alternatively, the email may seek to get the target to click on a link to enter their login details on a fraudulent website. In one JAMA Network Open study, employees of six US healthcare organizations received approximately 3 million simulated phishing emails over 7 years. Almost 1 in 7 were clicked, although a click doesn’t always mean a password was entered.

Several breaches identified by Paubox involved access to a user's received emails following phishing. All prior emails would be available, potentially including many test results and lists of medications, and could allow further phishing of other staff.

Organizations have to have processes to detect and appropriately address suspected incidents, reduce the impact as much as possible, and record details and actions taken. This would typically include changing passwords, reviewing auto-forwards, and determining what emails may have been accessed.

 

What changes when the email asks for a prescription action

Email alone will not meet the criteria to issue a prescription, and a prescription sent electronically must meet various criteria set out at 21 CFR 1311.100. Any e-prescription generated using non-compliant software would not constitute a valid e-prescription. Non-controlled substances would be subject to applicable state requirements.

Several measures were recommended to address illicit use of prescription medications in one region, including increased vigilance by prescribers and pharmacists to identify fraudulent requests. Any request sent through email to alter a controlled substance prescription should be viewed with suspicion.

 

How sender checks and inbound filtering catch impostors

Domain owners can implement various technical measures (SPF, DKIM, DMARC) to authenticate outgoing messages, and DMARC lets them set a policy for how incoming mail servers should handle failed authentication.

However, such a policy may not be in place. Of all domains involved in a breach, Paubox’s 2026 email security report found that 74% of breached domains had ineffective DMARC protection in 2025, up from 65% in 2024. Thus, when the practice sending the referral has not implemented a sufficiently strict policy, emails are more likely to be delivered.

How to verify a prescriber request outside the email thread

164.514(h) sets no specific method for confirming another provider’s identity, so it is sensible to establish internal policy, for example:

  • Compare the full sender address with the prescriber's address on file rather than the display name.
  • Call the prescriber's office using a number from your records, never one from the email.
  • Send PHI only to the verified address through encrypted email.
  • Forward the suspicious message to IT so it can be removed from other inboxes.

Learn more: HIPAA Compliant Email: The Definitive Guide (2026 Update)

 

FAQs

Is a prescriber impersonation email a breach if nobody responds?

An email nobody acts on involves no disclosure of or access to PHI, so it is not a breach, though it should be logged under your security incident procedures.

 

Can staff email records to a prescriber who requests them?

Staff can share PHI with a verified prescriber for treatment, and sending it through an encrypted email service covered by a business associate agreement protects it in transit.

 

Does encrypting the reply help if the requester is an impostor?

No, encryption protects the message in transit but delivers readable content to whoever controls the recipient address.