Guide

HIPAA website requirements: what actually applies to a clinic site

Most of what clinic owners are told about HIPAA and websites is either too vague to act on or is being sold to them. This is a plain explanation of when HIPAA attaches to a website, what the Security Rule asks for in practical terms, and what to check before you publish.

This guide is general information about how HIPAA is commonly applied to clinic websites. It is not legal advice, and no article can be. Your obligations depend on your practice, your data and your vendors, so treat this as a starting point for a conversation with your own counsel or compliance officer.

The single most useful thing to understand about HIPAA and websites is that the law does not regulate websites. It regulates protected health information. A website only enters the picture when it becomes a place where that information is created, received, stored or transmitted. Everything else follows from that one distinction, and a great many clinics spend money in the wrong place because they never made it.

What HIPAA covers, and what it does not

HIPAA governs protected health information: individually identifiable health information held or transmitted by a covered entity or its business associates. Two parts of that definition do the work. The information has to relate to a person’s health, care, or payment for care, and it has to be identifiable — tied to a name, a phone number, an email address, an account, a device identifier, or enough surrounding detail to point at one person.

A brochure website usually contains none of that. If your site publishes what your practice does, which clinicians work there, which insurance you accept, where you are, when you are open and what number to call, then it holds information about your business, not about your patients. That site is not a HIPAA problem. It may still have accessibility obligations, state law obligations, advertising rules from your licensing board, and ordinary security responsibilities — but the Privacy Rule and the Security Rule have nothing to attach to.

HIPAA obligations attach the moment the website does any of these:

  • Collects a patient intake form. Name plus date of birth plus reason for visit is protected health information the instant it is submitted.
  • Offers a “describe your symptoms” contact box. The field name is doing the damage. Anything that invites clinical detail will receive it.
  • Hosts a patient portal login. The existence of an account is itself a statement that a treatment relationship exists.
  • Takes an appointment request that names a condition or a service.“Request an appointment” with a service dropdown reveals more than a phone number does.
  • Runs a chat widget that accepts health details, or stores transcripts of conversations with identifiable people.
  • Processes payments tied to care, or displays balances, statements or claim status.

The practical consequence is that the question “is my website HIPAA compliant?” is usually the wrong question. The better one is: what patient information, if any, does my website touch, and where does it go afterwards? If the honest answer is “none, everything goes through the phone and the EMR,” you have removed most of the risk surface rather than trying to secure it. That is a legitimate design decision and often a better one than building a form and then defending it.

Covered entities and business associates

HIPAA reaches two categories of organisation, and knowing which one everybody in your supply chain is determines who owes what.

You are almost certainly a covered entity

Covered entities are health plans, healthcare clearinghouses, and healthcare providers who transmit health information electronically in connection with certain standard transactions — which in practice means any clinic that bills insurance electronically, and many that do not. A dental practice, a therapy practice, a medical group, an infusion clinic: covered entities. The obligations start with you and they do not transfer. You can delegate work; you cannot delegate accountability.

Your vendors become business associates

A business associate is any person or organisation that creates, receives, maintains or transmits protected health information on your behalf. For a clinic website, that potentially includes:

  • The hosting provider, if PHI is stored on or passes through its servers.
  • The website platform or builder, if it processes form submissions containing patient information.
  • The form provider, which is a separate vendor from your host more often than people realise.
  • The chat widget vendor and any AI service behind it, if conversations can contain health details.
  • Analytics, session recording and heatmap tools, if they capture pages or inputs that reveal PHI.
  • Your email provider, once patients email you clinical detail and you keep it.
  • Any subcontractor of any of the above. The chain does not stop at the first vendor.

Crucially, a vendor is not a business associate merely because it works for a clinic. A company that designs your logo, prints your brochures or hosts a site containing no patient data is not handling PHI and does not become one. The test is the data, not the industry. This is the distinction that decides whether you need a contract, and it is the one most often blurred by vendors selling to healthcare.

Business associate agreements, and the certification myth

A business associate agreement (BAA) is the written contract that has to be in place before a vendor handles protected health information on your behalf. It is not a formality and it is not a marketing asset. It is the mechanism by which your obligations follow your data.

What a BAA has to do

  • Define the permitted and required uses and disclosures of the information.
  • Prohibit uses or disclosures beyond what the contract and the law allow.
  • Require appropriate safeguards to prevent impermissible use or disclosure.
  • Require the business associate to report security incidents and breaches to you.
  • Bind subcontractors to equivalent terms, so the chain does not leak at the second link.
  • Provide for access, amendment and accounting so you can meet patient rights requests that depend on data the vendor holds.
  • Address return or destruction of the information when the relationship ends.

“Our servers are HIPAA compliant” is not a BAA

This sentence appears on a great many hosting websites and it means less than it sounds like. Servers can be encrypted, access-controlled, monitored and physically secure — all genuinely useful — and none of that creates a contractual obligation to you, a duty to notify you of a breach, or any limit on what the vendor may do with your data. Infrastructure is a capability. A BAA is a commitment. You need the commitment, and a vendor that has built the capability should have no difficulty signing for it.

Nothing is HIPAA certified

There is no HIPAA certification. No government body certifies products, no official seal exists, and no audit firm can confer HIPAA compliance on a piece of software. HIPAA is enforced after the fact: regulators examine whether a covered entity and its business associates actually did what the rules require. Third-party attestations like SOC 2 or HITRUST are real and can be useful evidence about a vendor’s controls, but they are not HIPAA certification and no vendor should present them as such.

When a vendor advertises a “HIPAA certified” product, you have learned something useful — just not what they intended. Ask instead: will you sign a BAA, what does it cover, who are your subcontractors, and what independent assessment of your controls can you show me?

Where to go next on the vendor question
The hosting and platform sides of this decision are covered in more depth on HIPAA compliant website hosting and choosing a HIPAA compliant website builder. Both pages start from the same question this guide does: what does the site actually handle?

The Security Rule, translated into website decisions

The Security Rule applies to electronic protected health information and organises its requirements into administrative, physical and technical safeguards. Read straight, it is abstract. Read as a set of questions about your website, it becomes concrete.

Safeguard

Administrative

Who at your practice is responsible for the website? Who can publish changes? What happens to their access when they leave? Is there a written risk analysis that includes the site, and a documented decision about what it does and does not collect?
Safeguard

Physical

Where do the servers live and who can walk up to them? For a hosted site this is inherited from your provider, which is exactly why the provider needs to be under contract rather than merely reassuring. It also covers the laptop your staff edits the site from.
Safeguard

Technical

Access control, audit controls, integrity, authentication and transmission security. These five map almost directly onto configuration choices you can verify yourself.

Access control

Every person who can edit or publish the site should have their own named account, not a shared login. Accounts should carry the least privilege that lets someone do their job, and removal should be part of your offboarding routine rather than something remembered later. If the site has any patient-facing login, the same reasoning applies with more force.

Audit controls

You should be able to answer “who changed this page, and when?” If PHI is involved, you should be able to answer “who accessed this record, and when?” A platform that keeps no history is asking you to take its word for what happened, and that is a poor position from which to investigate an incident.

Integrity

Information must not be improperly altered or destroyed. In website terms this means backups you have actually tested, version history on content, and confidence that a defacement or a bad edit is recoverable rather than final.

Authentication

Verify that people are who they claim to be. Multi-factor authentication on the publishing account is the cheapest meaningful control available to a small practice, and website admin credentials are a routine target precisely because they are so often unprotected.

Transmission security

Protect information in motion. For a website this is TLS, and the standard is every page, not just the page with the form on it. HTTPS everywhere prevents an attacker on the same network from modifying an unencrypted page to alter where a form posts, prevents session identifiers leaking, and removes the browser warnings that erode patient trust. Certificates should renew automatically; expired certificates are one of the most common self-inflicted outages in healthcare web hosting.

Encryption at rest

Data sitting in a database, a backup or a log file should be encrypted, with keys managed separately from the data they protect. This matters most for the systems behind the website — the EMR, the scheduling system, stored credentials — rather than for public marketing pages, which contain nothing worth encrypting.

Website forms: the biggest practical risk

If a clinic website causes a HIPAA problem, a form is usually how. Forms feel harmless because they look like part of the page, but a form is a pipe, and the interesting question is always where the other end of it is.

Plain-text email delivery is the classic failure

The default behaviour of most website form tools is to email you the submission. That single design choice can send a patient’s name, contact details and description of their symptoms across the internet through a mail relay you have no agreement with, into an inbox that may be a personal account, where it sits indefinitely and gets forwarded. It happens because it is the path of least resistance during setup, and it is rarely revisited afterwards.

The vendor chain behind a form is longer than it looks

A form embedded on a clinic site may involve the website platform, a separate form-builder service, a spam-filtering service, an email delivery provider, a notification integration, and a spreadsheet or CRM where responses accumulate. Each is a business associate if PHI passes through it, each needs a BAA, and each has its own subcontractors. Map the chain before you publish the form, not after someone asks.

The honest alternative: do not collect clinical detail on the website

Many practices route intake through their EMR’s patient portal, which is already covered by an agreement and already built for this, and keep the public website to published information plus a phone number, an address and a link into that portal. This is not a compromise. It concentrates PHI in the one system designed to hold it, and it means the marketing site — the part most likely to be redesigned by an agency, migrated between hosts or plugged into new tools — never becomes a place where patient data lives.

If you do decide to collect, collect the minimum. A form that asks for a name and a callback number is a far smaller problem than one that asks what is wrong. Necessary information can be gathered on the phone or in the portal.

Third-party scripts, trackers and chat widgets

Tracking technologies on healthcare websites have drawn sustained regulatory attention, and the underlying concern is straightforward once stated. A tracker does not need to see a medical record to disclose something protected. It needs only to report that an identifiable person visited a page whose subject matter reveals a health concern or a treatment relationship.

Consider the difference between an analytics script on your homepage and the same script on a page titled for a specific condition, on a patient portal login screen, or in a booking flow for a particular service. The homepage says someone looked at a clinic. The others say something about a person. When a script sends a page URL alongside a cookie identifier, an IP address or an advertising ID to a third party who has no BAA with you, that is a disclosure — regardless of whether anyone intended it and regardless of whether the third party wanted the data.

Practical positions that hold up well:

  • Keep advertising pixels and remarketing tags off pages that reveal a condition, a service interest or a patient relationship. Conversion tracking on a clinic site is rarely worth the exposure it creates.
  • Treat session recording and heatmap tools as high risk by default. They are built to capture what somebody typed, which is precisely the problem.
  • Read what each script actually transmits rather than trusting the category. “Analytics” covers tools with very different data flows.
  • Prefer server-side or privacy-preserving measurement that does not transmit identifiers to an ad network, or accept less measurement.
  • Re-check after every redesign. Scripts get added by agencies, by marketing tools and by well-meaning staff, and nobody sends a notification when it happens.

Chat widgets and AI assistants

A chat widget can be one of the safest things on a clinic site or one of the riskiest, and the difference is entirely in what it is permitted to do. A safe assistant answers from published clinic information — hours, location, parking, services offered, insurance accepted, how to reach a person — declines clinical questions instead of attempting them, does not ask for health details, and hands off to a human when it reaches the edge of what it knows. Nothing it handles is protected health information, because nothing about a patient ever enters it.

A risky assistant does the opposite: it invites people to describe their situation, retains transcripts against identifiable contact details, and ships those conversations to a third-party service. That is PHI in a vendor chain, and it needs the same agreements, safeguards and breach handling as any other system holding patient data. We describe how our own assistant is scoped on the AI chatbot page.

Email, breach notification and documentation

Email and patient communication

A mailto: link and a form post carry different risks. A mailto:link hands the message to the patient’s own email client; nothing passes through your website, your host or a form vendor, and the chain of intermediaries where most website breaches occur simply does not exist. What it does not do is make the resulting email disappear. Once a patient sends you clinical detail, that message is protected health information in your mailbox, and your email provider is handling it.

Patients are permitted to email you, and to receive replies by unencrypted email if they have been warned of the risk and still prefer it. What is not acceptable is for the practice to treat ordinary email as a default channel for clinical information without that conversation, or to leave years of patient email accumulating in an unprotected personal account.

Breach notification in outline

When unsecured protected health information is acquired, accessed, used or disclosed impermissibly, the Breach Notification Rule requires notification unless a risk assessment shows a low probability that the information was compromised. Notice goes to the affected individuals, to the Secretary of Health and Human Services, and — above a size threshold — to prominent media in the affected area. Business associates must notify the covered entity, which is one reason the reporting clause in a BAA matters so much: without it you may not learn that anything happened.

Specific deadlines and thresholds are set by regulation and are worth confirming with your compliance officer rather than from an article. What every practice should have in advance is a written procedure naming who is called, who assesses, who notifies and in what order — decided before an incident, not during one.

Documentation and risk analysis

The Security Rule expects an accurate and thorough assessment of risks to electronic protected health information, and expects it to be maintained rather than performed once. Website-relevant documentation includes: an inventory of what the site collects, a list of vendors that touch it and the status of each BAA, an inventory of third-party scripts and what they transmit, the reasoning behind decisions you made, and the date each item was last reviewed. Undated documentation is worth very little. A short, current, honest record beats a long, stale one.

A checklist you can act on

Work through this in order. The first three items resolve most of the uncertainty, because they establish whether the rest applies to you at all.

  1. Write down every place your website collects information from a visitor: forms, chat, portal logins, booking flows, payment pages, file uploads. Include anything an agency added.
  2. For each one, decide honestly whether it can receive health information. If a free-text box exists, assume it will.
  3. Remove or relocate anything that collects clinical detail without needing to. Fewer collection points is the cheapest risk reduction available.
  4. Trace where each remaining submission goes — every service, every inbox, every spreadsheet — and write the chain down.
  5. Get a signed BAA from every vendor in that chain, including your host, your form provider and your email provider. Confirm subcontractors are covered.
  6. Turn off plain-text email delivery of any form that can contain patient information, or stop collecting that information through the website.
  7. Serve every page over HTTPS, with automatic certificate renewal, and redirect HTTP to HTTPS site-wide.
  8. Audit third-party scripts page by page. Remove advertising pixels, remarketing tags and session recording from anything that reveals a condition, a service interest or a patient relationship.
  9. Give every staff member who can edit the site their own account, enable multi-factor authentication, and add access removal to your offboarding checklist.
  10. Confirm the site has version history and tested backups, so a bad edit or a defacement is recoverable.
  11. Check that your chat widget or AI assistant answers only from published clinic information and refuses clinical questions. Read a transcript sample if transcripts exist.
  12. Publish a Notice of Privacy Practices and make it findable, and keep your website privacy policy accurate about what the site actually collects.
  13. Review your patient communication policy: what may be sent by email, what may not, and how patient preferences are recorded.
  14. Write a short incident procedure naming who is contacted, who assesses, and who notifies. Keep it where someone can find it at short notice.
  15. Record a website-specific risk analysis, even if the conclusion is that the site handles no PHI. Date it.
  16. Set a calendar reminder to repeat items 1, 8 and 15 after every redesign, every vendor change, and at least annually.

Where ClinicSite fits

This section is about our product rather than about the regulation, and it is short on purpose. What is above stands on its own regardless of what you build with.

  • Generated ClinicSite websites do not collect patient information. Published sites run under a content security policy that blocks form submission outright and blocks outbound network calls. Contact runs through tel: and mailto: links rather than form posts.
  • Every published site is served over HTTPS with certificates issued and renewed automatically.
  • The platform runs on Microsoft Azure.
  • Connected SimplePractice EMR credentials are encrypted at rest with AES-256-GCM, each ciphertext bound to its tenant.
What we do not claim
ClinicSite publishes no HIPAA certification and no SOC 2 report. Neither exists, and as this guide explains, HIPAA certification is not a thing anyone can hold. If you are considering connecting an EMR, ask us about a business associate agreement before you do — get in touch and we will tell you plainly where we stand.

For the infrastructure and platform detail behind these points, see HIPAA compliant website hosting and HIPAA compliant website builder.

Frequently asked questions

Does my clinic website have to be HIPAA compliant?
It depends entirely on what the site does. A website that publishes your services, your hours, your address, your clinician bios and your phone number handles no protected health information, and HIPAA imposes no requirements on that content. The moment the site collects, transmits or stores individually identifiable health information — an intake form, a symptom description box, a patient portal login, an appointment request that names a condition — the Security Rule and the Privacy Rule apply to that data and to every vendor that touches it.
Is there such a thing as a HIPAA certified website or host?
No. HIPAA has no certification programme, no accrediting body and no official seal. The Department of Health and Human Services does not certify products, servers, hosts or website builders. Any vendor advertising itself as HIPAA certified is describing something that does not exist. What is real is enforcement: regulators assess whether a covered entity and its business associates actually met their obligations after something goes wrong. A vendor can meaningfully offer a business associate agreement, describe its safeguards and let you audit them. It cannot hand you a certificate.
What is a business associate agreement and when do I need one?
A business associate agreement is a written contract between a covered entity and a vendor that creates, receives, maintains or transmits protected health information on the covered entity's behalf. You need one before that vendor handles any PHI. It has to set out the permitted uses of the information, require appropriate safeguards, require the vendor to report security incidents and breaches to you, extend the same obligations to any subcontractors, and cover the return or destruction of the data when the relationship ends.
A vendor told me their servers are HIPAA compliant. Is that enough?
No. Infrastructure claims and contractual obligations are different things. A host can run encrypted, access-controlled, well-audited servers and still refuse to sign a business associate agreement, and without that agreement you are not permitted to route protected health information to them. Ask for the agreement first. If a vendor will not sign one, treat that as a decision about what you can send them, not as a detail to work around.
Are Google Analytics and advertising pixels a problem on a clinic website?
They can be, and this is an area regulators have paid real attention to. The risk is not analytics in the abstract. It is that a tracker on a page which itself reveals something about a person — a page about a specific condition, a patient portal login screen, a booking flow for a particular service — can transmit an identifier plus the context of that visit to a third party who has no business associate agreement with you. The safe pattern is to keep third-party scripts off any page that could indicate a patient relationship or a clinical interest, and to read what each script actually sends rather than assuming.
Is a mailto: link safer than a contact form?
It has a different risk profile rather than a strictly safer one. A mailto: link opens the patient's own email client, so no form data passes through your website, your host or a form vendor. The message still arrives in your inbox, and if the patient volunteers clinical detail then your email provider is now handling protected health information and needs to be covered accordingly. What the mailto: link removes is the chain of intermediaries between the browser and you, which is where most website breaches actually occur.
Can an AI chatbot on a clinic website be used safely?
A chatbot is safe in proportion to what it is allowed to do. One that answers only from published clinic information — hours, location, services, insurance accepted, how to reach a human — and declines clinical questions rather than answering them collects nothing that would be protected health information. A chatbot that invites people to describe their symptoms, retains transcripts, or passes conversations to a third-party service without a business associate agreement is handling PHI, and everything that touches those transcripts falls within scope.
Do I need a documented risk analysis if my website collects nothing?
The Security Rule's risk analysis expectation applies to your practice's handling of electronic protected health information as a whole, not to your website alone. If your website genuinely collects nothing, the risk analysis for it is short, but it should still exist in writing so you can show the reasoning and the date. Practices are routinely asked to produce documentation of what they assessed and when, and a written record that the public site handles no PHI is far better than an unrecorded assumption.

See what your clinic site could look like

Paste your current website address. You get a full generated replacement to look at — free, no account, no card.

Free preview · no card · previews are deleted after 7 days