Are electronic signatures legally valid in the UK?
For the documents a pharmacy uses, generally yes. The position is more settled than the sector's caution suggests.
The Electronic Communications Act 2000 provides for the admissibility of electronic signatures in evidence. The Law Commission's 2019 report concluded that an electronic signature can satisfy a statutory requirement for a signature, where the signatory intends to authenticate the document and any other formalities are met. Consent to treatment, agreement to service terms and permissions of the kind a private pharmacy service uses fall comfortably within that.
Two qualifications belong alongside it. Certain instruments — deeds most commonly — carry additional formalities, including witnessing. Anything of that kind needs specific legal advice, not reliance on a general position. And validity is not the same as evidential weight, which is the distinction the rest of this article is concerned with. A signature that could legally bind somebody is of limited use if the service cannot show who signed, when, and what they were shown.
What a defensible signature record contains
Services buying e-signature capability tend to judge how the signature looks, which is the part that matters least. What determines whether a consent record survives scrutiny is everything attached to it. Who signed, identified to the patient record rather than to an email address alone. When, to a timestamp rather than a date. What version of the document was displayed at the moment of signing — the element most commonly absent, and the one that decides whether you can state what was agreed. The document as presented, retained rather than regenerated: a consent form rebuilt from current templates shows today's wording, not the wording signed. And evidence of integrity, meaning the record demonstrably has not been altered since. Technical context such as device, browser or address adds weight but is not essential. A record holding those five elements answers the question a dispute actually asks. A scanned image of a signature answers almost none of them. That is why digitising paper is not the same as capturing consent electronically.
Why the document version matters most
Consent wording changes, and it changes more often than services expect. A new medicine is added to a plan, a side effect warning is expanded following an MHRA communication, a data sharing arrangement is revised, or the service's own terms are updated.
Where the wording is stored as a single current document, every one of those revisions silently rewrites history. A patient who consented in March appears to have consented to the September wording, and the service cannot show otherwise. Where wording is versioned, with each revision released under a new identifier and an effective date, every signature continues to point at what was actually displayed.
The discipline is identical to the questionnaire versioning described in the consultation flow guide, and for the same reason. When something changes externally, the first question is which cohort was assessed or consented under which version. A service that cannot answer it is reconstructing rather than reporting.
Where consent belongs in the workflow
Two placements. Choosing between them is a clinical judgement, not a convenience one.
At booking for consent concerning the service the patient has chosen: service terms, data handling, cancellation arrangements and permissions. The argument set out in the patient data capture guide applies with particular force here. A patient signing at home can actually read what they are agreeing to. A patient signing at the start of an appointment, with a clinician waiting, is performing a formality. Capturing it at booking improves the quality of the consent, not merely its timing.
At the consultation for consent that depends on the clinical discussion itself. Consent to a specific treatment, having understood its risks, cannot meaningfully be given before the conversation that conveys them. Signing it in advance produces a record that is evidentially neat and clinically hollow. A reviewer will notice that the consent predates the discussion it claims to follow.
Write the distinction into the service design rather than deciding case by case. The default drift is toward putting everything at booking for operational convenience.
What a pharmacy actually needs signed
| Document | When | Separate or combined |
|---|---|---|
| Service terms and cancellation | Booking | Separate, since it is contractual rather than clinical |
| Consent to treatment | Consultation | Separate, and per treatment where treatments differ materially |
| Data sharing with a GP | Booking or consultation | Separate, since it is withdrawn independently |
| Disclosure to a named third party, comprising an insurer, employer or solicitor | When the request arises | Separate and specific to the recipient and the scope requested |
| Marketing permission | Booking, unbundled | Always separate and separately revocable, per the consent rules |
| Photography or imaging | Consultation | Separate, and only where the service genuinely uses it |
The recurring error is the combined agreement: treatment consent, service terms and marketing permission signed as one document. It is quicker at the point of signing. It fails the moment a patient withdraws one of them, because the service cannot honour the withdrawal without appearing to unwind the others. Marketing in particular must be separately given and separately withdrawable — a requirement, not just good practice, as the PECR guide sets out.
Changing wording and withdrawing consent
Three situations arise. Each has a correct handling.
Minor revision. Release a new version, leave existing signatures pointing at their own version, and apply the new wording to signatures taken from the effective date forward.
Material change. Where the substance of what a patient agreed to has changed, re-consent affected patients against the new version rather than assuming the previous signature carries. The materiality judgement is the service's, and it should be recorded. A reviewer will ask why a change was treated as one or the other.
Withdrawal. Record a withdrawal as an event on the timeline, not by deleting the original consent. The original signature remains, because it evidences what was true at the time. The withdrawal sits after it with its own date. Deleting a consent because it has been withdrawn destroys the record of a period during which treatment was properly given.
Records patients need for insurance and other third parties
Private pharmacy services tend to design their records for the regulator and forget the audience that asks more often: the patient, needing to prove something to a third party. Private medical insurers require evidence before reimbursing a claim — the treatment, its clinical justification and the patient's authorisation of it. Travel and life insurers ask patients to declare treatments, then verify the declaration. Employers and occupational health services request confirmation of ongoing treatment. Solicitors request records in personal injury and clinical negligence matters. Immigration and visa processes sometimes require evidence of prescribed treatment. In every one of those, a patient returns to the pharmacy asking for their file, frequently years afterwards. What they receive determines whether their claim succeeds. A service holding a complete signed record produces it in an afternoon. A service holding a tickbox, a regenerated consent form showing today's wording and consultation notes spread across an inbox produces something the insurer will query. The patient experiences that as the pharmacy failing them, not as a records design decision made three years earlier.
Two distinct obligations sit inside that, and services regularly conflate them.
Producing records to the patient is straightforward, and it is their right. A patient may request their own data. A service that can assemble the consultation record, the clinical decisions, the consents and the supply history as a coherent file has discharged the duty properly. The practical test: does the file read as a chronology to somebody who was not there? It is the same test the patient data capture guide applies to the record generally.
Disclosing records to a third party is a different act, and it needs its own authority. Never treat that authority as implied by the patient having asked for something. When an insurer, employer or solicitor requests records directly, ask for the patient's specific written authority. Then disclose the minimum the request genuinely requires, not the whole file. That authorisation is itself a consent worth capturing as a signature, with the recipient and the scope named. A later question about why information was released then has a documented answer.
Design directly for what insurers most commonly query. They ask whether the treatment was clinically indicated, which the consultation record and the prescriber's reasoning answer. They ask whether the patient consented, which the signature answers only if it carries the version of what was explained. And they ask when it was authorised relative to when it was supplied, which the timeline answers. A record built to satisfy those three questions serves the patient's claim, the regulator's inspection and any later dispute with the same material. That is the practical argument for building it properly once.
How long to keep signed consent
For as long as the record it supports. In practice that means the same retention schedule as the patient record, not a shorter one.
Consent evidences a clinical or contractual event, and its usefulness arrives late — frequently years after the treatment concluded, when a question is raised about whether something was agreed. The insurance and third-party requests described above are the most common form that question takes. They arrive on no schedule the service controls. A service that retains treatment records for the applicable period but purges consent earlier has kept the half describing what was done and discarded the half explaining why it was permitted.
One migration point deserves emphasis. Signed consent must survive a change of system. It belongs in any export taken before a migration, and it should be verified as readable afterwards, per the data export guide. A signature that exists only inside a system the pharmacy has left is not evidence it can produce.
Capacity, carers and third parties
Electronic capture makes one thing easier to get wrong: knowing who actually signed.
Where a form is completed remotely, the service is relying on the person at the other end being the patient. For most private services that risk is acceptable, and the identity verification the flow already performs mitigates it. Where it is not — services where a proxy has an incentive — the identity step should come before the signature, not after it.
Three situations require the record to show more than a signature. A carer or representative signing should be identified as such, with their relationship recorded, rather than appearing as the patient. Any question about capacity belongs in the clinical record, addressed before consent is taken rather than documented afterwards. And a patient who cannot use the electronic route needs an alternative — an accessibility obligation as well as a practical one, on the reasoning in the accessibility article.
What to check before you rely on it
Seven questions, applicable to any e-signature arrangement including one built into a pharmacy system.
Does the record identify the signer against the patient record rather than only an email address? Is the timestamp stored, not merely the date? Is the version of the document recorded, and is the document retained as presented rather than regenerated? Can the record demonstrate it has not been altered since signing? Can signed consent be exported, and does it remain readable outside the system? Are separate consents captured separately so one can be withdrawn without the others? Could you assemble a complete, chronological file for a patient who needs it for an insurance claim, and do you capture specific authority before releasing anything to a third party? And is there a documented alternative for patients who cannot sign electronically?
A service that can answer all seven has consent it can rely on. A service that can answer only the first two has a signature, which is not the same thing.
Key takeaways
- Electronic signatures are generally valid for the documents a pharmacy uses. The Electronic Communications Act 2000 provides for admissibility, and the Law Commission confirmed they can satisfy statutory signature requirements, subject to additional formalities for deeds.
- The signature is the least important part of the record. Who signed, when, which version, the document as presented and evidence of integrity determine whether it survives scrutiny.
- Version the consent wording, because storing a single current document silently rewrites what every previous patient appears to have agreed to.
- Service terms and permissions belong at booking, where a patient can read them. Consent to treatment belongs in the consultation, since it cannot precede the discussion that informs it.
- Capture consents separately rather than as one combined agreement, because they are withdrawn independently and marketing permission must be separately revocable.
- Handle revision, material change and withdrawal distinctly, and record a withdrawal as an event rather than by deleting the original signature.
- Patients need these records for insurance claims, employers and solicitors, frequently years later. Insurers ask whether treatment was indicated, whether it was consented to and when it was authorised.
- Disclosing to a third party is a separate act. It needs the patient's specific written authority, disclosed to the minimum scope, and that authorisation is itself worth signing.
- Retain signed consent for as long as the record it supports, include it in any migration export, and provide a documented alternative for patients who cannot sign electronically.
Sources and further reading
FAQs
Consent you can actually produce.
Bring your current consent wording and we will show what a versioned signature record looks like against it. See it working, then bring us your own case and we will walk through capture at booking, the audit trail and what you would send an insurer.
See Dataforge PMR