How to Build a HIPAA Compliant App in 2026: An Engineering Guide

In this article
- HIPAA applies to your app when you are a covered entity (a provider, health plan or clearinghouse) or a business associate handling PHI for one. A consumer health app outside that chain usually answers to the FTC Health Breach Notification Rule and state privacy laws instead.
- Every vendor that stores, processes or can see PHI needs a signed business associate agreement: cloud, email, SMS, video, support desk, analytics and AI APIs. No BAA, no PHI.
- The Security Rule turns into concrete engineering: a risk analysis, unique logins with MFA, audit logs, encryption, backups you have restored, and automatic logoff. Most leaks happen in push notifications, logs and analytics SDKs.
- There is no official HIPAA certification. In our estimates, HIPAA adds about 15–30% to engineering and $30k–80k a year in compliance upkeep for a startup selling to providers.
Jump to
- First: does HIPAA apply to your app?
- How to build a HIPAA compliant app, step by step
- The Security Rule, translated into engineering
- What you rent under a BAA and what you build
- Where PHI leaks in real apps
- Building a HIPAA-compliant AI healthcare app
- Breach notification: plan it before you need it
- What goes into the first version
- Timeline and budget at a glance
- Mistakes that cost the most later
- HIPAA readiness checklist
- Where Gilzor fits
First: does HIPAA apply to your app?
Health data alone doesn't trigger HIPAA. Who you are and who you work for does. Three questions decide it:
- Are you a covered entity? Healthcare providers that bill electronically, health plans and clearinghouses are.
- Do you handle PHI on behalf of one? Then you are a business associate, with direct obligations under HIPAA.
- Is the data identifiable? Properly de-identified data is not PHI.
The answers put you in one of four positions:
- Examples: a clinic's own patient app, a virtual care practice, a health plan member app.
- What applies: the Privacy, Security and Breach Notification Rules, in full.
- You need: BAAs with your development vendor if it can access PHI, and with every service provider.
- Watch out: patients have a right to access their records. Your app may be part of how you meet it.
- Examples: SaaS for clinics, remote monitoring platforms sold to providers, billing and scheduling tools.
- What applies: the Security Rule, the parts of the Privacy Rule in your BAA, breach reporting to your customers.
- You need: a BAA with each customer, and BAAs down the chain with your own subcontractors.
- Watch out: hospital buyers send long security questionnaires. Your answers have to be true in the code.
- Examples: period trackers, symptom checkers, fitness apps that sync health data, direct-to-consumer apps not run by a provider.
- What applies: usually not HIPAA. The FTC Health Breach Notification Rule, the FTC Act and state laws such as Washington's My Health My Data Act.
- You need: honest privacy disclosures and user consent before sharing health data.
- Watch out: the FTC treats sharing health data with advertisers without permission as a breach.
- Examples: general wellness content, research tools on de-identified datasets.
- What applies: HIPAA doesn't cover properly de-identified data.
- You need: a documented de-identification method: removing the identifiers HIPAA lists, or an expert determination.
- Watch out: a device ID, an exact date or a small ZIP code can make data identifiable again.
Whether HIPAA, the FTC rule or a state law applies depends on your business relationships and data flows. Confirm your position with healthcare privacy counsel before launch, and again when you add a new customer type.
One more thing to clear up early: there is no HIPAA certification. HHS doesn't certify apps or vendors. Compliance is a set of safeguards, documents and agreements you keep up over time, and anyone selling a "HIPAA certified" badge is selling a third-party opinion.
How to build a HIPAA compliant app, step by step
Nine steps. The first three are paperwork that shapes the code, so they start in week one.
- Map every PHI flowDraw where health data enters, where it is stored, who sees it and where it leaves: the EHR, a lab, a pharmacy, an email. Mark every vendor on the diagram. This map is the input for everything else.
- Run a security risk analysisThe Security Rule requires one. List the threats to each flow, how likely they are and how bad they would be, and decide how you'll reduce each risk. Repeat it when the system changes, and at least once a year.
- Sign BAAs before any PHI movesCloud, database, email, SMS, video, support desk, error tracking, AI APIs. If a vendor won't sign, it doesn't get PHI. Plan for higher plan tiers: many vendors only sign on business or enterprise plans.
- Design access controlUnique accounts for every person, MFA for staff, roles with the minimum access each job needs, an emergency access procedure, and automatic logoff after inactivity.
- Build audit logging into the first featureEvery read, create, update and delete of PHI records who, what, when and from where. Write logs to an append-only store that app admins can't edit.
- Encrypt and back upTLS for every connection, encryption at rest for databases, files and backups, keys in a managed key service. Restore a backup to a clean environment before launch and write down how long it took.
- Seal the leaksCheck push notifications, logs, analytics events, crash reports, emails and support tickets for PHI. This is where most real-world exposures happen. More below.
- Write the policies and train the teamAccess reviews, incident response, device rules, workforce training and sanctions. Keep them short enough that people read them, and keep the records for six years, as HIPAA's documentation rule requires.
- Test, then prove itRun a penetration test on the apps and the API, fix the findings, and keep the report. Hospital customers will ask for it, along with your risk analysis summary.
The Security Rule, translated into engineering
The Security Rule groups its safeguards into administrative, physical and technical. Some are "required" and some "addressable", which means you implement them or document why an equivalent measure is reasonable. In practice, teams building new apps implement nearly all of them.
HHS proposed an update to the Security Rule in January 2025 that would make most safeguards mandatory, including encryption and MFA. Whatever its final form, a new app should be built as if it applies.
| Safeguard | What it means in the code and the cloud |
|---|---|
| Unique user identification | No shared logins anywhere: app, admin console, database, cloud console |
| Access control | Role-based permissions checked on the server for every request, scoped per organization |
| Person authentication | MFA for staff and admins; strong session handling and biometric re-entry for patients |
| Automatic logoff | Session timeouts on web and mobile, shorter for shared clinical devices |
| Audit controls | Append-only access logs for PHI, reviewed regularly, with alerts on unusual access |
| Integrity | Database constraints, checksums on files, no direct production edits without a trail |
| Transmission security | TLS 1.2 or newer everywhere, including internal service calls and webhooks |
| Encryption at rest | Encrypted databases, file storage, backups and device caches, with managed keys |
| Contingency plan | Automated backups, a tested restore, and a written plan for running during an outage |
What you rent under a BAA and what you build
You can rent most of the infrastructure from vendors that sign BAAs. Signing the BAA doesn't make you compliant: you still configure and use the service correctly.
| Layer | Rent, with a BAA | You still build or configure |
|---|---|---|
| Cloud | HIPAA-eligible services from a major cloud provider | Use only the services listed in the BAA; network rules, encryption, logging settings |
| Communication | Email, SMS, chat and video providers on BAA plans | Message templates with no PHI, consent records, opt-outs |
| Identity | An auth provider that signs a BAA | Roles, organization scoping, session rules, emergency access |
| Support | A help desk on a BAA plan | Rules for what agents may ask for and paste into tickets |
| Monitoring | Error tracking and logging services that sign BAAs | Scrubbing PHI from events before they leave your servers |
| AI | AI APIs under a BAA, with the covered configuration | Prompt logging, access control on outputs, human review of clinical content |
Where PHI leaks in real apps
Breaches in health apps rarely start with a hacker breaking encryption. They start with data going somewhere nobody thought about. Check each of these:
- Push notifications. Apple's and Google's push services aren't covered by your BAAs, and the text shows on locked screens. Send "You have a new message", never a diagnosis, drug or test name.
- Application logs. A debug line that prints a request body puts PHI into your logging service. Log IDs, scrub payloads, and make sure the logging vendor signs a BAA anyway.
- Analytics and ad SDKs. HHS guidance on tracking technologies and FTC enforcement against health apps both point the same way: pixels and SDKs that send health details or screen names to third parties are a disclosure. Event names like "viewed_hiv_results" count.
- Crash reports. They capture screen state and sometimes memory. Configure scrubbing, or use a reporting service under a BAA.
- Email and SMS. Links to the app are fine; results in the body are not, unless the patient asked for it and you documented the request.
- The device itself. PHI cached unencrypted, sensitive screens visible in the app switcher, tokens in plain preferences. Use the Keychain or Keystore and blur sensitive screens.
- Developer and test environments. Copies of production data on laptops or in staging. Use synthetic data outside production.
Building a HIPAA-compliant AI healthcare app
AI features add new places for PHI to land: prompts, outputs, embeddings, fine-tuning sets and evaluation logs. The rules don't change, the surface does.
- BAA for the exact service. Several major AI and cloud providers sign BAAs for selected API products and configurations, often with zero data retention. Consumer chat apps are not covered.
- Treat AI data as PHI. Vector databases, prompt logs and model outputs get the same access control, encryption and audit logging as your main database.
- Minimum necessary. Send the model only the fields the task needs. A summary of today's visit doesn't need the patient's address.
- Human review for clinical content. Generated notes, letters or triage suggestions need a clinician's sign-off. If the AI diagnoses or recommends treatment, FDA device rules may apply. Our guide on how to build a medical app covers that question.
Breach notification: plan it before you need it
If unsecured PHI is exposed, the HIPAA Breach Notification Rule sets the clock. Covered entities notify affected individuals without unreasonable delay and no later than 60 days after discovery. Breaches affecting 500 or more people also go to HHS and, when 500 or more residents of a state are affected, to the media. Smaller breaches are reported to HHS yearly. Business associates notify the covered entity.
Encryption that meets HHS guidance is your best protection here: data rendered unreadable to unauthorized people is generally not "unsecured", so a lost encrypted laptop is usually not a reportable breach. Consumer apps under the FTC rule have their own notification duties to users, the FTC and, for larger breaches, the media.
For the app team, that means an incident runbook, logs good enough to tell who was affected, and a named person who decides.
What goes into the first version
A compliant MVP doesn't need every enterprise control. It needs the ones that are hard to add later.
| At launch | Can wait |
|---|---|
| Risk analysis, PHI flow map, signed BAAs | A third-party assessment or SOC 2 report, until buyers ask |
| Unique accounts, MFA for staff, role-based access | SSO with each hospital's identity provider |
| Audit logs for every PHI access | Automated anomaly detection on access patterns |
| Encryption in transit and at rest, a tested restore | Multi-region failover |
| PHI-free notifications, logs and analytics | Your own self-hosted analytics stack |
| Policies, training records, an incident runbook | A full compliance management platform |
| Penetration test with findings fixed | A bug bounty program |
Built by Gilzor
Results we’ve shipped




Talk to the people who build it. Tell us about your project and get a free estimate of scope, timeline and cost.
Timeline and budget at a glance
The yearly figure covers BAA-tier hosting and vendors, an annual risk assessment and a penetration test, on top of the usual 15–20% of the build for maintenance. See the healthcare app development cost guide for the full breakdown, and the telemedicine app development cost guide if your product includes virtual visits.
Mistakes that cost the most later
- Picking vendors before checking BAAs. Replacing an analytics, email or auth provider after launch means a migration. Ask about the BAA in the first sales call.
- Using services the BAA doesn't list. Cloud BAAs cover specific services. A new managed service turned on by a developer may not be one of them.
- Audit logs left for the week before the first hospital deal. Retrofitting logging touches every endpoint, and you can't reconstruct who saw what last year.
- Health details in analytics events. Screen names and event names are data too. Rename them before they leave the app.
- Backups never restored. A backup you haven't restored is a guess. Test it, time it, write it down.
- "HIPAA certified" on the website. There is no such certification, and hospital security teams know it. Describe your safeguards instead.
HIPAA readiness checklist
Tick what is already true for your app. It shows how far you are from handling real patient data.
HIPAA readiness
FAQ
How do I build a HIPAA compliant app?
How long does it take to build a HIPAA-compliant app?
How much does it cost to make an app HIPAA compliant?
Is there a HIPAA certification for apps or developers?
How do I build a HIPAA compliant mobile app?
Can I use AI APIs in a HIPAA-compliant healthcare app?
Where Gilzor fits
We build healthcare web and mobile apps, backends and integrations, and run QA and app audits. For a US blood testing laboratory, our team built native apps for patients, doctors and phlebotomists and worked with US attorneys on HIPAA authorization and the EULA. We don't claim a HIPAA certification, because none exists; we design the safeguards in from the first sprint. We work from Poland and Cyprus, with a few shared hours a day with the US East Coast.
Send us your PHI flow, even a rough sketch, and your current vendor list. We'll mark where the data goes, which agreements are missing and which safeguards belong in the first version.
No sales pitch
Get a straight answer for your project
Tell us what you’re building. We’ll reply with options, a rough cost and timeline. If we’re not the right fit, we’ll say so.

CTO of Gilzor. Responsible for architecture and the engineering standards our teams work by.
LinkedIn →Gilzor · Mobile Development partner
Need a team for your mobile app?
Services
Mobile DevelopmentNative and cross-platform iOS and Android apps, from MVP to scale.→By company type
Selected projects






The team behind them





