Does accepting the AWS BAA make our HIPAA on AWS environment compliant?
Accepting the AWS BAA is necessary before PHI goes into AWS, but it does not make your environment compliant. AWS offers a standard Business Associate Addendum that you review and accept in AWS Artifact, and AWS says PHI should be processed, stored and transmitted only in services listed as HIPAA-eligible in its HIPAA Eligible Services Reference. AWS also states there is no HIPAA certification for a cloud service provider. The addendum sets out what AWS commits to as your business associate. It does not decide whether your accounts, workloads, identity settings, logging, backups or encryption choices satisfy the Security Rule.
Your organization owns its controls, its risk analysis and the evidence that shows both work. Under HHS guidance, each business associate must perform its own risk analysis, and OCR's cloud guidance says the cloud configuration in use affects the risk analysis and risk management plans of all parties. That means the risk analysis for your AWS environment has to cover the specific services, regions, account structure and shared responsibilities you chose. Confirm with your AWS account team and your own architects which services hold ePHI, and check each one against the HIPAA Eligible Services Reference before data arrives. Any service outside that list is a decision to document, not a detail to assume away.
No independent auditor or assessor has to be involved for the BAA itself, but any report or attestation about your environment can come only from an independent auditor, assessor or certification body, not from AWS or from an implementation partner. A partner can prepare configurations, access reviews, logs and the risk management evidence an assessor will ask for, and the HHS Guidance on HIPAA and Cloud Computing is the natural next read on how these responsibilities divide. Ask any partner to state in writing which evidence it prepares and which parts remain your decisions.
Which AWS services count as HIPAA eligible, and how to check
Signing the BAA does not make an AWS account HIPAA ready. Eligibility is set service by service, and the ePHI that flows through a workload reaches more components than the application itself. The checks below help an architect confirm coverage before any PHI lands in the environment.
Confirm each service in the reference
Every AWS service that creates, receives, maintains or transmits ePHI should appear in AWS's HIPAA Eligible Services Reference. AWS says PHI should be processed, stored and transmitted only in listed services, so treat the list as the gate for architecture decisions. Record the service name, the date you checked and the reference version in your design file. Services not on the list may still run in the account, but they must stay away from PHI. Confirm the current list with AWS, because it changes over time.
Trace every component touching ePHI
Eligibility of the main service does not cover its supporting parts. Map where ePHI can appear, including application logs, CloudWatch output, backups, snapshots, queues, analytics stores, search indexes, caches and support tooling. Check each one against the reference, and confirm that any feature used with PHI is itself listed. A common gap is a logging or analytics component that receives PHI by default and was never classified. Document the data flow so an assessor can follow it.
Check regional and feature limits
Eligibility can depend on region and on specific features within a service, so a service listed as eligible may not be eligible in every region or for every configuration. Confirm the regions where PHI will be stored and processed, including replicas and disaster recovery sites. Check that the features you intend to enable, such as particular encryption options or managed integrations, are covered for the service and region you chose. Record these limits in the architecture decision record.
Confirm third-party and marketplace components
AWS's BAA covers AWS services, not software you install on EC2 instances or buy from the AWS Marketplace. A SaaS product that runs on AWS and handles PHI for you needs its own BAA with you, and it is a business associate in its own right. Ask each vendor whether it signs a BAA, which AWS services it uses, and whether it has evaluated its own risk analysis. Open-source packages you deploy yourself fall under your own controls, not AWS's.
Re-check when services change
Eligibility is a recurring check, not a one-time sign-off. Re-review the reference whenever an engineer adopts a new service, enables a new feature, opens a new region or changes how a data flow works. Set a change control step that requires a documented eligibility check before a new service receives PHI. Your risk analysis should be updated when the architecture changes, since OCR's cloud guidance says the configuration affects the risk analysis of every party.
Keep your own safeguards in scope
Eligible services remove none of your obligations. Access control, user authentication, audit logging, configuration of encryption and backups, and workforce training remain yours under the Security Rule. Decide which team owns each setting, write it into your policies, and test it. Ask AWS for the shared responsibility details for each service you adopt, and confirm which controls the account owner must configure. An eligible service configured poorly still exposes ePHI.
When a SaaS partner on AWS needs its own BAA
When your organization buys a SaaS product that stores or processes ePHI, the business associate agreement is between you and the SaaS vendor. You do not sign a BAA with AWS just because the vendor runs on it. The vendor is the business associate that handles PHI on your behalf, and AWS is a subcontractor from the vendor's point of view. Your obligations under the Security Rule (45 CFR 164.308(b) and 164.314) are met by having that agreement in place and by assessing the vendor, not by reviewing AWS's terms alone.
- Sign with the vendor: Your BAA should name the SaaS company as the business associate and cover the PHI it touches, including data it keeps in AWS.
- Check for no separate AWS contract: Ask whether you are expected to accept AWS Artifact terms directly. Usually you are not, and the vendor should explain why.
- Review termination and return of data: The agreement should state how PHI is returned or destroyed when the contract ends.
- Confirm breach notice timing: Ask how the vendor will notify you of a security incident affecting your data, including incidents in its AWS environment.
- Ask who holds the risk analysis: The vendor must perform its own risk analysis, and the cloud configuration it chose affects that analysis.
Before signing, confirm that the vendor's BAA covers the specific services and regions it uses, and that it can show the agreement it has with AWS.
If your company builds a product on AWS and handles ePHI for covered entities, you are a business associate in your own right. You accept AWS's Business Associate Addendum in AWS Artifact to cover your use of AWS. That agreement does not replace the BAA you must sign with each covered entity customer. AWS's position is that its customers' covered entities sign with the customer, not with AWS, so your contracts with customers are the agreements that govern your handling of their PHI.
- Accept the AWS addendum: Review and accept it in AWS Artifact before you place PHI in any AWS account.
- Sign a BAA with each customer: Each covered entity or upstream business associate that sends you PHI needs its own written agreement with you.
- Keep PHI in eligible services: The AWS addendum applies to services listed as HIPAA-eligible, so confirm each service in your architecture appears on that list.
- Run your own risk analysis: Your risk analysis covers your application layer, your access model and your operations, not only the AWS infrastructure.
- Flow terms to your own vendors: Any subcontractor that handles PHI for you needs a BAA with you.
Map each customer contract to the AWS services it touches before go-live so no PHI reaches a service outside the agreed scope.
Under the Security Rule, a business associate that delegates PHI work to a subcontractor must obtain satisfactory assurances from that subcontractor, in writing, that it will protect the information in the same way. This flow-down runs through every level that creates, receives, maintains or transmits ePHI. A cloud provider that stores encrypted ePHI without holding the key is still a business associate, so the chain does not end at the application vendor. Each link must be documented, and the terms in each agreement should not be weaker than the one above it.
- Map every subcontractor: List the hosting, logging, backup, support and analytics providers that can reach PHI.
- Confirm written assurances: Each subcontractor should have an agreement with the business associate that matches its PHI role.
- Check for gaps in incident reporting: Breach notice must travel up the chain quickly enough for the covered entity to meet its own deadlines.
- Confirm data return and deletion: Each subcontractor should be bound to the same return or destruction terms.
- Track changes: Ask to be notified when the vendor adds or replaces a subcontractor that handles PHI.
Check that the subcontractor list is current at contract signing and at each renewal, because the chain changes as products grow.
When you evaluate a SaaS vendor on AWS, you can ask for evidence that its agreements and controls are in place. A vendor's word that it is HIPAA compliant is not enough, and HHS does not endorse private certifications, so the useful documents are the contracts, the vendor's own risk analysis summary and the policies that govern its handling of PHI. An independent assessor's report can help, but it does not replace the vendor's obligation to perform its own analysis.
- Request the BAA: Ask for the signed agreement with you, and confirm it names the vendor as business associate.
- Request the AWS arrangement: Ask the vendor to confirm it accepted the AWS addendum and to name the HIPAA-eligible services in use.
- Request the subcontractor list: Ask for the subprocessors that can access PHI and the status of their agreements.
- Request a risk analysis summary: Ask how the vendor identified risks to ePHI and how it tracks the remediation of those risks.
- Request the incident process: Ask for the vendor's breach notice timeline and its ransomware response plan.
Confirm each point in writing before you sign, and retain the answers in your vendor file for the next review.
Which safeguards stay with your organization on AWS
AWS runs the physical data centers and underlying infrastructure, but the Security Rule still applies to you as the covered entity or business associate that creates, receives, maintains or transmits ePHI in AWS. The safeguards that stay with your organization include identity and access management, unique user IDs and multi-factor authentication for every person and workload that can reach ePHI, encryption configuration and key management, audit logging and log review, backup and recovery testing, workforce training, written policies, incident response and your own risk analysis. AWS documents which controls it operates, but it does not configure your accounts, decide who can see patient data, or review your logs for you.
The cloud configuration you choose changes the risk analysis of every party involved. HHS cloud guidance says the configuration affects the risk analysis and risk management plans of all parties, so a risk analysis that describes only the physical facilities misses most of the exposure. Confirm which AWS accounts, regions, storage services, keys and administrator roles hold ePHI, record them in your risk analysis, and check that each is a HIPAA-eligible service under your BAA. Ask your implementation partner, if you use one, to document which controls they configure and which remain with your staff.
OCR's cloud guidance lists the service level agreement topics relevant to HIPAA: availability and reliability, backup and data recovery, how data is returned or deleted after termination, security responsibility, and limits on use, retention and disclosure. Review the AWS terms and any SaaS partner terms against each topic, and confirm with your legal and security leads what recovery objectives and data return steps your organization needs in writing. Where a term is unclear or missing, raise it before you store production ePHI, not after an incident.
Why there is no HIPAA certification for AWS or for your environment
No Security Rule standard requires an entity to certify compliance, and HHS does not endorse or recognize private organizations' certifications. HHS has also stated that such a certification would not stop it from later finding a violation. For AWS, this means the BAA you accept in AWS Artifact is a contract, not a badge. AWS says there is no HIPAA certification for a cloud service provider, and the same logic applies to your own environment: no document from AWS or anyone else replaces the risk analysis you owe under 45 CFR 164.308(a)(1), which must cover the workloads, storage and identities you configure.
AWS states that its HIPAA risk management program aligns with FedRAMP and NIST SP 800-53. That alignment describes how AWS manages its own platform. It is not a HIPAA certification, and it does not tell you that your buckets, IAM roles, logging or encryption settings meet the Security Rule. Ask your team to map each safeguard to the AWS services that implement it and to the configurations you own, so the shared-responsibility boundary is written down rather than assumed.
A third-party assessor can test whether your controls operate as described and produce findings, but the assessor's report describes testing, not HIPAA compliance. Only the covered entity or business associate can be found compliant or not by OCR. Before engaging an assessor, confirm in writing what scope, samples and evidence they will use, and be wary of any vendor or consultant who offers a HIPAA certificate, since HHS does not recognize one.
Who does what in a HIPAA on AWS program
As the covered entity or business associate, your organization keeps ownership of the Security Rule. AWS can run the infrastructure, but it does not perform your risk analysis, adopt your policies, train your workforce or decide who may access ePHI. Each business associate must perform its own risk analysis, and the cloud configuration you choose changes that analysis for everyone in the chain.
- Risk analysis and risk management: You document where ePHI lives in AWS, the threats and vulnerabilities you have identified, and the plan for reducing them.
- Policies and procedures: You write and maintain the policies that govern access, incident response, backups and workforce conduct.
- Workforce and access: You assign a security official, grant and remove user access, and train staff on handling PHI.
- Evidence and records: You retain the logs, reviews, tickets and documentation that show the safeguards operate over time.
- Vendor oversight: You confirm the BAAs and SLA terms for every SaaS partner that touches your ePHI.
Before you rely on any arrangement, confirm with your counsel exactly which AWS account holder signs the BAA and which of your own controls cover the rest.
AWS is responsible for the contract, the services it offers under that contract and the controls inside its own infrastructure. Its HIPAA position rests on a standard Business Associate Addendum that customers review and accept in AWS Artifact. Its obligations do not extend to how your accounts, identities or applications are configured.
- Business Associate Addendum: You accept the standard BAA in AWS Artifact, and the terms apply to the accounts covered by that acceptance.
- HIPAA-eligible services: AWS says PHI should be processed, stored and transmitted only in services listed in its HIPAA Eligible Services Reference, so each service in your design needs checking against that list.
- Infrastructure controls: AWS operates physical data centers, hardware and the underlying platform under its own risk management program.
- No certification: AWS states there is no HIPAA certification for a cloud service provider, so any claim of a "HIPAA-certified" AWS environment should be treated with caution.
Confirm the current eligible services list and the BAA terms directly with AWS before you finalize your architecture.
An implementation partner such as DeployOpen prepares the environment and evidence that your team will then own and operate. The work is typically defined in a fixed-scope written agreement that names the accounts, services, configurations and deliverables in advance. The partner does not hold the BAA, decide your risk tolerance or certify the result.
- Account and configuration preparation: Setting up accounts, identities, network boundaries, encryption and logging to match the agreed design.
- Hardening and documentation: Applying configuration baselines and writing the runbooks and system descriptions your team needs.
- Backup and testing support: Helping you configure backup and recovery, and supporting restore tests you schedule and review.
- Evidence preparation: Organizing configuration records, logs and screenshots so your team can show how safeguards were implemented.
- Ongoing support: Where you need it, a retainer for upgrades, monitoring support and periodic reviews.
Ask any partner to put its scope, deliverables and what it will not do in writing before work begins.
An independent auditor or assessor tests the program and is the only party that issues a report on it. Its work begins from the evidence you provide and the controls you say you operate, then it tests whether those statements hold. HHS does not endorse private certifications, and a third-party assessment does not prevent OCR from later finding a violation.
- Testing: The assessor samples configurations, access records, risk analysis documents and incident records to check operation over time.
- Reporting: Only the independent assessor issues a report or attestation about your program.
- Scope limits: Its conclusions cover the systems and period it examined, so confirm the boundary before relying on the result.
- Independence: The assessor should not design the controls it tests, which is why implementation partners and assessors are usually different firms.
DeployOpen prepares systems, controls and evidence for assessment. It does not issue audit reports, attestations or certifications, and it does not represent any assessment outcome.
What to have in place before PHI goes into AWS
Before any ePHI reaches an AWS account, the organization needs a signed BAA, a clear map of where PHI will flow, and a documented risk analysis that reflects the configuration actually built. The cards below cover the decisions and checks that tie those pieces together, so the BAA, eligible services and your own safeguards line up before go-live.
Executed BAA through AWS Artifact
Accepting AWS's standard Business Associate Addendum in AWS Artifact is the step that puts the AWS relationship on a HIPAA footing, and it is done by the account holder, not by a consultant. Confirm the accepting account is the one that will hold PHI, and record who accepted it, when and under which account ID. Keep a copy in your compliance file, since your auditor will ask for it. Accepting the BAA does not make a workload compliant on its own.
Inventory of ePHI data flows
Before provisioning, list every place ePHI will be created, received, stored, processed or transmitted: application tiers, databases, object storage, backups, logs, analytics exports and support tickets. Each flow should name the source system, the AWS service, the region and any third-party SaaS partner that touches it. Gaps in this inventory become gaps in your risk analysis, so review it with application owners rather than from architecture diagrams alone.
Risk analysis covering the AWS configuration
Each business associate must perform its own risk analysis, and HHS cloud guidance notes that the cloud configuration you choose changes your risk analysis and risk management plan. Document the shared responsibility split, the account and network design, encryption choices and who administers each layer. Your analysis must describe the configuration as built, so re-check it after major architecture changes. Ask your AWS partner to show how the design was reviewed against your analysis.
Architecture limited to eligible services
AWS says PHI should be processed, stored and transmitted only in services listed as HIPAA-eligible in its reference, so the approved architecture should name each service and its eligibility status. Check the list on the day the design is approved, because the list changes. Set guardrails, such as organization policies that block unapproved services and regions, so a developer cannot place PHI in an ineligible service by accident. Any service not on the list needs a documented decision before use.
Identity, access and logging design
The Security Rule expects unique user identification, access control, audit controls and the ability to review activity in systems holding ePHI. Design named identities with multi-factor authentication for administrators, least-privilege roles and centralized logs that are retained and protected from change. Confirm who reviews logs and how often, and test that a single user's access to PHI can be traced end to end. These controls are yours to configure and evidence, even though AWS operates the underlying platform.
SLA and contract terms OCR names
OCR's cloud guidance lists service level topics relevant to HIPAA: availability and reliability, backup and data recovery, how data is returned or deleted at termination, security responsibility, and limits on use, retention and disclosure. Read the AWS terms against each topic and identify what you must cover yourself, such as backup testing and recovery objectives. Where a SaaS partner sits between you and AWS, confirm that its own BAA and SLA address the same topics.
Common questions about HIPAA on AWS
Is accepting the AWS BAA enough to make my ePHI workloads HIPAA compliant?
No. Accepting the Business Associate Addendum in AWS Artifact creates the contractual relationship, but it does not make your workloads compliant. AWS says PHI should be processed, stored and transmitted only in services listed as HIPAA-eligible, and the Security Rule still applies to your configurations, access controls, logging, risk analysis and workforce practices. The BAA is one required piece of the arrangement, and the rest depends on how you build and operate the environment. Confirm the current BAA terms directly in AWS Artifact before you rely on them.
Is AWS HIPAA certified?
No. There is no HIPAA certification for a cloud service provider. HHS does not endorse or recognize private certifications for the Security Rule, and such certifications do not relieve covered entities or business associates of their legal obligations. AWS states that it aligns its HIPAA risk management program with FedRAMP and NIST SP 800-53. Ask for AWS's own HIPAA documentation and treat any claim of HIPAA certification, from AWS or from a vendor, with skepticism.
What happens if we store PHI in a service that is not HIPAA-eligible?
The service falls outside the BAA's coverage for that data, so you lose the contractual protection AWS provides for it, and your own compliance position weakens. The practical control is to restrict PHI to services on the HIPAA Eligible Services Reference and to enforce that restriction in your design, such as through approved service lists, account structure and policy guardrails. Check the reference before every new service is adopted, because eligibility can change over time. Confirm the current status with AWS before you move any ePHI.
If our SaaS vendor runs on AWS, does its BAA cover AWS?
Not directly. AWS states that a SaaS partner's covered entity customers sign a BAA with the partner, not with AWS. Your contract is with the SaaS vendor, and that vendor is a business associate that must in turn manage its own relationship with AWS. Ask the vendor to confirm that it has accepted the AWS BAA where needed, which services hold ePHI, how its subcontractors are bound, and how it handles breach notification. Review the vendor's own risk analysis, because the cloud configuration affects the risk analysis of every party.
What must our organization still do itself on AWS?
Your organization keeps the safeguards that depend on how you use the platform. These typically include access management and unique user IDs, multi-factor authentication, encryption and key management decisions, audit logging and review, backup and recovery testing, workforce training, incident response, your written policies and your risk analysis and risk management plan. AWS provides the underlying infrastructure and tooling, but the Security Rule obligations attach to the covered entity or business associate that holds the ePHI. Map each requirement to an owner before go-live.
How much effort does a HIPAA-aligned AWS environment take, and what drives it?
Effort depends mainly on the number of accounts and services that touch ePHI, how mature your existing risk analysis is, how many integrations and third-party tools handle PHI, and how much logging and evidence you need to retain. Each new service or data flow adds review, and each change to identity or network design adds testing. The work is partly one-time (architecture, risk analysis, policies, initial configuration) and partly recurring (access reviews, log review, patching, training and periodic risk updates). Ask any quote to itemize these components and state what is excluded.
Can an implementation partner run our HIPAA program on AWS for us?
A partner can prepare and operate parts of the program, but it cannot take over the legal obligations. DeployOpen, for example, plans, implements, hardens, documents and supports self-hosted open-source platforms in your own cloud accounts, and it runs security testing. Under a fixed-scope engagement, scope, deliverables and timelines are agreed in writing before work starts. Your organization still owns its controls and the Security Rule duties, and a partner does not issue HIPAA certifications or attestations. Agree in writing which safeguards the partner operates and which remain yours.
How DeployOpen prepares a HIPAA on AWS environment
DeployOpen plans, builds and documents self-hosted open-source platforms inside your own AWS accounts, so the work follows the structure HIPAA on AWS requires: your BAA with AWS, PHI kept in HIPAA-eligible services and a clear line between AWS responsibilities and yours. Our role is to prepare systems and evidence; your organization owns its controls and any assessment is performed by an independent party.
How we scope
- We confirm which workloads will hold ePHI and which AWS services they depend on before any build starts.
- We check each planned service against AWS's HIPAA Eligible Services Reference and flag anything that is not listed.
- Engagements are fixed-scope, with scope, deliverables and timelines agreed in writing before work begins.
How we build
- We configure accounts, encryption, access controls and logging inside your AWS environment, with the BAA status confirmed in AWS Artifact.
- We document which safeguards AWS provides under the shared model and which ones your team must operate.
- We test configurations against your written risk analysis and record the results as evidence for your review.
How we hand over
- We deliver architecture records, configuration baselines, runbooks and a responsibility matrix your staff can maintain.
- We walk your security official and engineers through the controls, owners and review cadence before handover.
- We list the items you must confirm with AWS, your SaaS partners and your own risk analysis, since those sit outside our deliverables.
How we support
- Ongoing support, upgrades and continuous testing are available as retainers after handover.
- We can re-check configurations when AWS services change, your workloads expand or your risk analysis is updated.
- Any independent HIPAA assessment is commissioned and performed by a third party of your choosing, not by DeployOpen.

