INSIGHTS · HEALTHCARE TECHNOLOGY
What a Healthcare Website Should Collect—and What It Should Leave to the Clinic
Inquact Labs ·
Design website enquiries around the immediate administrative purpose, with clear boundaries for clinical information, professional judgement and the systems that handle them.
On this page
Introduction
An appointment-request form can begin with a name, a contact number and a preferred time. As different teams suggest improvements, it may acquire fields for symptoms, diagnosis, medication, previous treatment, medical history and uploaded reports.
Each addition may seem useful in isolation. Together, they change the nature of the system. A website that once routed an administrative enquiry may now hold clinical information, with a wider set of people, vendors and processes involved in handling it.
That is not automatically the wrong design. Some legitimate healthcare services need to collect clinical information online. But that collection needs a specific purpose, appropriate safeguards and a governed workflow. It should be a deliberate system decision, rather than an unnoticed extension of a general contact form.
The practical starting point is simple: collect the information required for the immediate administrative task. Leave clinical information in appropriate clinical systems and professional workflows unless there is a justified, secured and properly governed reason to collect it through the website.
This article concerns a healthcare organisation’s own website. Inquact Labs’ business enquiry channel is for discussing digital-system requirements, not for patient appointments or submitting medical information.
1. Start with the purpose
Before deciding which fields to display, ask: What does the organisation actually need to accomplish at this stage? An appointment request is different from a clinical assessment, even when both eventually concern the same visit.
An administrative purpose might be to request an appointment, record a preferred department or location, provide a callback number, answer a general enquiry, capture a preferred time, or route the request to the relevant administrative team. Write that purpose in plain language before building the form.
Then connect each field to a decision or action. If the team only needs a location and callback number to arrange the next step, a detailed treatment history may add responsibility without improving that immediate outcome. If a field is genuinely necessary for a different, more specialised service, define that service and workflow separately.
Related guidance: [1]
2. What an administrative website form may reasonably need
The following are illustrative options, not a universal required-field list. A clinic should choose only the fields its administrative process can justify:
- The patient or enquirer’s name, where needed to handle the request.
- A contact number for the agreed callback or confirmation process.
- An email address only where that communication channel is genuinely needed.
- A preferred clinic or hospital location.
- A requested department or speciality, without treating that preference as a diagnosis.
- A preferred appointment date or time, clearly presented as a request rather than confirmed availability.
- Whether the person is enquiring for themselves or someone else, if this affects administrative handling.
- A short administrative note where necessary, with clear instructions about what not to include.
3. What should trigger a second question
Administrative information is not automatically low-risk: a selected speciality, department or service can reveal something about a person’s health or circumstances. It still needs privacy and security protections. W3C’s form guidance supports clear instructions and identification of required and optional fields, not the choice of healthcare fields above. Free-text instructions should make clear what visitors should not include.
Before adding any field, work through a short decision test:
- Why do we need it for this immediate purpose?
- Who will use it, and for which action?
- Where will it be stored or copied?
- How long will it be retained?
- Who can access it?
- What happens if it is inaccurate?
- Can the same administrative outcome be achieved without collecting it here?
If the team cannot answer these questions, pause the addition. Clarify who will use the field and what action it enables, or complete the task without collecting it.
4. Information that deserves stronger justification
Symptoms, diagnoses, medical history, medication information, laboratory results, diagnostic reports, medical images and treatment history deserve particular attention. Insurance or identity documents also need justification when they are not necessary for the immediate step.
These can be legitimate inputs to an online healthcare service. Before collecting them, establish who receives the information, who reviews it, where it is stored and who can access it. Consider the consequences of inaccurate routing or unintended copying.
Where clinical information is needed, use a patient or clinical workflow with the organisation’s approved identity checks, access controls and review arrangements. A general contact form and shared email inbox should not become a clinical repository by default.
5. Separate administrative routing from clinical judgement
A website can route a request by the visitor’s selected department, location, appointment preference, callback request, useful language preference or general service category. That is administrative routing: helping a request reach a team that can review it.
The administrative workflow described here does not determine a diagnosis, clinical urgency, treatment, medication or suitability for a procedure. A routing label is not a clinical assessment.
An organisation operating clinical triage or decision-support needs a different level of system design: validation for its intended use, professional oversight, governance, monitoring and escalation. Adding an AI text classifier to an appointment form does not establish those controls.
Keep the boundary understandable to both visitors and staff. A request received is not an appointment confirmed; a requested speciality is not a diagnosis; and a routing suggestion is not clinical advice.
6. Where automation helps
Administrative automation can reduce repetitive handling without attempting to make clinical decisions. One illustrative sequence is:
Website request → acknowledgement → administrative classification → appropriate team or CRM queue → human review → appointment or callback action → reminder or follow-up → outcome recorded.
Useful tasks include routing, acknowledging receipt, maintaining workflow status, scheduling agreed reminders, prompting staff follow-up and producing operational reports. An acknowledgement should describe what actually happened: receipt of a request, not a confirmed appointment or a medical conclusion.
Define which actions are automatic, which need human approval and what happens when routing is uncertain or a system is unavailable. Prevent repeated processing from creating duplicate messages or actions. Important decisions should retain suitable human review, and AI should not independently send diagnoses, medical instructions or clinical recommendations as part of this administrative flow.
7. Design the data journey, not just the form
A short form is only one part of a responsible design. Trace where the submission goes after the visitor selects Send: the application, database, email, CRM, notifications, staff devices, vendor systems, backups and logs. Note any movement between systems or jurisdictions and have the relevant teams review what that means for the organisation.
For each destination, identify the purpose, owner, permitted users and retention/deletion process. Decide how deletion requests or retention decisions affect copies and backups, and who handles incidents. Choose vendors and integration settings with those responsibilities in mind.
Use TLS to protect information in transit between systems; OWASP’s TLS guidance addresses this transport protection. It does not protect stored information or determine who can read a submission after it reaches an inbox, CRM, application or database.
For stored information, a practical starting point is to review encryption at rest and key access for each destination. For staff and integration accounts, limit access to the tasks each account needs, use MFA where supported and appropriate, and review privileged access and credentials. These are design recommendations, separate from the TLS guidance; they do not guarantee security.
Logs should support diagnosis of system failures without routinely copying entire submissions or medical details into a second repository. OWASP’s logging guidance specifically treats health information as data that should usually be excluded, masked or otherwise protected in logs.
8. Consent is not a substitute for minimisation
A consent checkbox does not explain why a field is needed. Keep collection tied to the stated task and tell visitors how their information will be used, rather than inviting anything that might be relevant.
Whether consent is required, which legal basis applies and what notice is sufficient depend on the service and applicable jurisdictions. Obtain legal and professional input for those decisions.
NIST’s Privacy Framework provides privacy-risk-management background, not certification or proof of compliance.
Related guidance: [1]
9. A practical design pattern
For an administrative enquiry, the following pattern makes responsibility at each stage explicit. Services collecting clinical information online will need a process designed for that purpose.
Public website
Collect the minimum administrative information justified for the immediate task.
Administrative workflow
Route, acknowledge and classify the request without turning classification into a clinical assessment.
Human review
Staff confirm the appropriate administrative next action and escalate matters requiring professional judgement.
Appropriate healthcare system
Collect or access clinical information through the organisation’s governed clinical process when it is required.
10. Questions to ask before launching the form
Before launch, check that the design decisions work together:
- Can the team explain the purpose of the form and justify the fields it has kept?
- Are instructions clear about what visitors should and should not submit?
- Have owners confirmed access, vendor handling, retention, deletion and backup arrangements across the full data journey?
- Have appropriate transmission, storage, authentication and access safeguards been reviewed?
- Which decisions or communications require human review?
- What is the fallback when automation, routing or an integration fails?
- Does the privacy notice explain the actual handling process, with consent addressed where applicable?
- How will staff restrict and appropriately route accidentally submitted clinical information, rather than spreading it across ordinary support tools?
- Have the workflow and its failure paths been tested without using real patient information?
Conclusion
Before adding another field, identify the immediate action it enables and whether the information belongs in the public administrative workflow at all. If the team cannot explain the action or who handles the information, resolve that before adding the field.
For an administrative enquiry, keep the handoffs clear: public website → administrative workflow → human review → a governed clinical process when clinical information is needed. Assign responsibility for each handoff before launch.
Sources and further reading
Technical and design background, not a complete legal or clinical framework. These sources do not establish compliance for a particular organisation.
- NISTPrivacy Framework
- W3C Web Accessibility InitiativeForm Instructions
- OWASPLogging Cheat Sheet — Data to exclude
- OWASPTransport Layer Security Cheat Sheet