Data protection

The most common data breaches in practices

The data breaches GP, dental and private practices most often face, from misdirected letters to phishing, and the practical controls that prevent each.

Ava research teamPublished 7 min read

On this page
  1. Two families of breach
  2. 1. Letters sent to the wrong address
  3. 2. Emails to the wrong recipient
  4. 3. Texts to old numbers
  5. 4. Information added to the wrong record
  6. 5. Staff looking where they should not
  7. 6. Phishing and impersonation
  8. 7. Malware and ransomware
  9. 8. Lost devices, paper and unapproved tools
  10. 9. Overheard conversations and phone disclosures
  11. Controls at a glance
  12. Example: a quarter of near misses
  13. Frequently asked questions
  14. References

Key takeaways

  • Most practice breaches are everyday mistakes: letters, emails and texts sent to the wrong person, and information added to the wrong record.
  • Among cyber incidents, phishing dominates; it was reported by 85% of businesses that identified a breach or attack.
  • Inappropriate access by staff is less frequent but more damaging to trust, and audit logs are the main deterrent.
  • Each common breach has a matching control, and most controls cost little more than a change of habit or a system setting.

Two families of breach

Data breaches in practices fall into two broad families. The first is human error in everyday administration: something sent to the wrong person, something said within earshot of others, something filed against the wrong patient. The second is deliberate attack, usually by email. Both matter because practices hold health information, which UK GDPR treats as special category data[1], so even a small slip can cause real distress.

The public is alert to the risk. In the government's public attitudes tracker, 57% of UK adults identified data not being held securely, or being hacked or stolen, as a risk of data use[2]. What follows is a guide to the breaches practices meet most often, with the controls that prevent each one.

1. Letters sent to the wrong address

Post remains a frequent cause of breaches: letters to an old address, two letters folded into one envelope, a results letter sent to a patient with a similar name. The harm depends on content. An appointment reminder is minor; a sexual health result or a safeguarding letter is serious.

  • Check and update addresses at every contact, not just at registration.
  • Use window envelopes so the address printed on the letter is the one used.
  • Print and fold one letter at a time for sensitive content, and check the name on each page.
  • Avoid sensitive detail in letters where a phone call or secure message would do.

2. Emails to the wrong recipient

Autocomplete is the most common culprit: typing two letters of a name and sending to the wrong 'Sarah'. Group emails with every address visible in the To or CC field are another classic, revealing that each recipient is, for example, a patient of the diabetes clinic.

  • Turn off or clear autocomplete for external addresses where your email system allows.
  • Use a mail merge or a messaging system for group communications, never a visible list of addresses.
  • Add a short delay before sending so mistakes can be recalled.
  • Send sensitive attachments through a secure route, and check the recipient before every send.

3. Texts to old numbers

Mobile numbers change hands. A text to an outdated number reaches a stranger, and if it says 'your smear test result' or 'your mental health review', it reveals health information. Shared family phones create the same problem.

  • Confirm the mobile number whenever the patient is in contact.
  • Keep text content neutral: 'Please contact the surgery' rather than naming the service or result.
  • Ask patients at registration whether the number is personal to them.

4. Information added to the wrong record

Scanning a hospital letter into the wrong patient's notes, merging duplicate records incorrectly or selecting the wrong patient from a search list are integrity breaches. They can lead to wrong treatment as well as a confidentiality problem when the other patient later sees their record.

  • Check at least two identifiers, such as name and date of birth or NHS number, before filing anything.
  • Flag patients with similar names at the same address in the clinical system.
  • Audit a small sample of scanned documents each month.
  • Treat record merges as a two-person task.

5. Staff looking where they should not

Looking at the record of a neighbour, a relative or a well-known local person without a work reason is less common than misdirected post, but it does far more damage to trust. In the Cyber Security Breaches Survey 2025, unauthorised access to files or networks by staff, even if accidental, was reported by 2% of businesses that had identified a breach[3]. In health care, with large numbers of staff able to open records, it deserves more attention than that figure suggests.

  • Tell staff, and remind them yearly, that access is logged and audited.
  • Run periodic audits, including searches for staff looking at records of patients who share their surname or address.
  • Set role-based access so reception see what they need for admin and no more.
  • Treat confirmed snooping as a disciplinary matter and record it as a breach.

6. Phishing and impersonation

For cyber incidents, phishing is in a league of its own. Among businesses that identified a breach or attack in the previous 12 months, 85% reported phishing and 34% reported others impersonating the organisation or its staff[3]. In practices this looks like a fake invoice from a supplier, a message that seems to come from a partner asking for a payment, or a bogus login page for the email system.

Types of breach or attack among businesses that identified one, UK, 2025
Types of breach or attack among businesses that identified one, UK, 2025Bar chart. Phishing: 85%; Impersonation of the organisation or staff: 34%; Malware such as viruses or spyware: 18%; Hacking of online bank accounts: 6%; Ransomware: 6%; Denial of service: 5%; Unauthorised access by staff: 2%.PhishingPhishing: 85%85%Impersonation of the organisation or staffImpersonation of the organisation or staff: 34%34%Malware such as viruses or spywareMalware such as viruses or spyware: 18%18%Hacking of online bank accountsHacking of online bank accounts: 6%6%RansomwareRansomware: 6%6%Denial of serviceDenial of service: 5%5%Unauthorised access by staffUnauthorised access by staff: 2%2%Types of breach or attack among businesses that identified one, UK, 2025Bar chart. Phishing: 85%; Impersonation of the organisation or staff: 34%; Malware such as viruses or spyware: 18%; Hacking of online bank accounts: 6%; Ransomware: 6%; Denial of service: 5%; Unauthorised access by staff: 2%.PhishingPhishing: 85%85%Impersonation of the organisation orstaffImpersonation of the organisation or staff: 34%34%Malware such as viruses or spywareMalware such as viruses or spyware: 18%18%Hacking of online bank accountsHacking of online bank accounts: 6%6%RansomwareRansomware: 6%6%Denial of serviceDenial of service: 5%5%Unauthorised access by staffUnauthorised access by staff: 2%2%

Source: GOV.UK (Department for Science, Innovation and Technology), 2025[3]

Show the data
ItemValue
Phishing85%
Impersonation of the organisation or staff34%
Malware such as viruses or spyware18%
Hacking of online bank accounts6%
Ransomware6%
Denial of service5%
Unauthorised access by staff2%
  • Require multi-factor authentication for email and remote access. Only 40% of businesses required two-factor authentication[3], so many small organisations are exposed.
  • Train everyone to report suspicious emails and reward reporting rather than blaming clicks. Just 19% of businesses had carried out staff training or awareness raising on cyber security[3].
  • Verify any change of bank details or urgent payment request by phone, using a known number.

7. Malware and ransomware

Malware was reported by 18% and ransomware by 6% of businesses that identified a breach or attack[3]. Ransomware is rarer but can stop a practice working for days. Unsupported computers, such as old machines running imaging or specialist equipment, and missing backups are the usual weak points.

  • Keep operating systems and software updated, and isolate devices that cannot be updated.
  • Keep at least one backup that ransomware on the network could not reach, and test restoring it.
  • Have a paper fallback for appointment lists and urgent contacts.

8. Lost devices, paper and unapproved tools

Laptops, phones, memory sticks and printed clinic lists are still lost or stolen. Encryption turns most of these from reportable breaches into non-events. A newer risk is unapproved software: 11% of GPs said they used AI tools they had obtained independently[4], and staff may paste patient information into free online services without realising it leaves the practice.

  • Encrypt every laptop and phone used for work, and enable remote wipe.
  • Limit printed lists and shred them at the end of the session.
  • Publish a list of approved tools and explain why others are not allowed.

9. Overheard conversations and phone disclosures

Not every breach involves a system. Reception desks close to waiting areas, clinicians discussing a patient in a corridor and callers who claim to be a relative all create disclosure risks. Phone calls are the hardest: a confident caller who knows a date of birth can often persuade a busy receptionist to read out test results or appointment details.

  • Agree a standard identity check for phone callers and apply it every time, including to people who sound familiar.
  • Never confirm whether someone is a patient to an unverified caller.
  • Offer a private space for conversations at the desk, and keep screens angled away from the queue.
  • Call back on the number in the record when in doubt, rather than giving information to an incoming caller.

Controls at a glance

Matching common breaches to low-cost controls
ControlBreaches it prevents
Confirm contact details at every contactMisdirected letters and texts
Two-identifier check before filingWrong-record entries
Autocomplete off and send delayMisdirected emails
Neutral text and letter wordingDisclosure through misdirected messages
Access audits and role-based accessStaff snooping
Multi-factor authenticationAccount takeover after phishing
Tested offline backupsRansomware loss of access
Device encryptionLost or stolen devices
Approved tools listData pasted into unapproved services

Whatever happens, write it down. UK GDPR requires every personal data breach to be documented with its facts, effects and remedial action[5]. The log is also your best tool for spotting which of these breaches recur in your practice.

Example: a quarter of near misses

Example (fictional): a dental practice asked staff to log near misses as well as breaches for one quarter. The log showed nine entries: five letters caught before posting with the wrong enclosure, two emails nearly sent to the wrong lab, one text to a number the patient had given up and one phishing email reported by a nurse. The practice switched to printing letters singly, set a send delay on email and added a number check to its reminder process.

For what to do when a breach does happen, see our guide to reporting a breach within 72 hours. For safer email in particular, see emailing patients safely, and for new systems, DPIAs for new practice technology.

Frequently asked questions

What is the most common data breach in a GP practice?

Everyday mistakes, such as letters, emails or texts sent to the wrong person, and documents filed in the wrong record. Among cyber incidents, phishing is by far the most common type[3].

Is a text to the wrong number a data breach?

Yes, if it discloses personal data. Neutral wording, such as asking the patient to contact the surgery, reduces the harm if it happens.

Do near misses need to be recorded?

They are not breaches, but recording them is good practice. They show where processes are fragile before patients are affected.

Which single control prevents the most cyber breaches?

Multi-factor authentication on email and remote access, combined with staff who feel safe reporting suspicious messages.

References

  1. legislation.gov.uk (The National Archives) (2026). UK GDPR Article 9: Processing of special categories of personal data (opens in a new tab) ↑

  2. GOV.UK (Department for Science, Innovation and Technology) (2024). Public attitudes to data and AI: Tracker survey (Wave 4) report (opens in a new tab) ↑

  3. GOV.UK (Department for Science, Innovation and Technology) (2025). Cyber security breaches survey 2025 (opens in a new tab) ↑

  4. Nuffield Trust (2025). How are GPs using AI? Insights from the front line (opens in a new tab) ↑

  5. legislation.gov.uk (The National Archives) (2026). UK GDPR Article 33: Notification of a personal data breach (opens in a new tab) ↑

Book a demo

See Ava handle your calls.

Book a short demo. We’ll walk through your calls and workflow, and agree what Ava should take care of.

Book a demo