Date:

Category:

DevOps Strategy

HIPAA Compliance on AWS: What You Own and How to Set It Up

Written by

Dino Starcic

·

CEO

AWS will sign a Business Associate Addendum (BAA) and offers a long list of HIPAA-eligible services, but no workload is HIPAA compliant just because it runs there. AWS secures the infrastructure. You configure and operate everything that touches protected health information (PHI). For most startups and product teams, the practical question is who will do that configuration work, keep it correct as the environment changes, and have the evidence ready when someone asks. At Cloudwise, senior AWS engineers handle that kind of work through a subscription, so your team can ship product instead of chasing IAM policies.

Cloudwisecloudwise.orgKeep AWS HIPAA Controls on TrackCloudwise senior AWS engineers support day-to-day operations, including infrastructure management, cost optimization, continuous security monitoring, and support.Visit

TL;DR

  • AWS has no HIPAA certification for cloud providers. It signs a BAA and publishes a list of HIPAA-eligible services, and you accept the BAA in AWS Artifact.

  • Under shared responsibility, you own access controls, logging, encryption settings and every component that processes, stores or transmits PHI.

  • Store PHI only in eligible services, encrypt it in transit and at rest, enforce least-privilege IAM with MFA, and keep audit logs.

  • A proposed HHS update to the Security Rule would make many of these controls mandatory. Building to that standard now avoids rework later.

  • Cloudwise engineers can run the AWS side of this work under one monthly subscription, with a typical 24 to 48 hour response time.

Table of Contents

  • Is AWS HIPAA compliant?

  • Does HIPAA apply to you?

  • Step 1: Accept the BAA before any PHI arrives

  • Step 2: Keep PHI inside HIPAA-eligible services

  • Step 3: Understand what you own under shared responsibility

  • The technical controls that matter most

  • Step 4: Do the risk analysis, including your cloud assets

  • Step 5: Verify continuously, not once

  • What is changing in the Security Rule

  • Why a subscription fits this work

  • How Cloudwise handles the AWS side

  • Quick start checklist

  • FAQ

  • Sources

Is AWS HIPAA compliant?

The short answer is that AWS supports HIPAA compliance, and your configuration decides whether you achieve it. AWS states that there is no HIPAA certification for a cloud service provider. Instead, it aligns its HIPAA risk management program with FedRAMP and NIST 800-53, and NIST SP 800-66 maps NIST 800-53 to the HIPAA Security Rule.

So when you read “AWS is HIPAA compliant,” read it as “AWS will sign a BAA and has demonstrated compliance for the services on its eligible list.” The compliance of your application is a separate matter, and it is yours.

An AWS blog on generative AI and HIPAA puts it plainly: customers do not automatically inherit HIPAA compliance by using eligible services. They must configure and implement controls within those services.

Does HIPAA apply to you?

HIPAA applies to covered entities (health plans, healthcare providers and clearinghouses) and to their business associates. If you build software or run infrastructure that creates, receives, maintains or transmits electronic PHI (ePHI) for a covered entity, you are likely a business associate yourself.

The cloud provider is covered too. HHS guidance on cloud computing states that a cloud service provider handling ePHI on a covered entity’s behalf is a business associate, even if it only stores encrypted data and does not hold the key. That is why a BAA with the provider is required before PHI goes into the cloud.

If you are unsure whether your product makes you a covered entity or business associate, ask your healthcare attorney or compliance officer. That is a legal determination, not an engineering one.

Step 1: Accept the BAA before any PHI arrives

The BAA is the contract HIPAA requires so that AWS safeguards PHI and limits how it may use and disclose it. AWS offers a standard BAA that you accept and manage through AWS Artifact, its self-service portal for compliance documents.

Two details matter:

  • Scope. AWS explains that an organization can accept the BAA for all accounts in its AWS organization through the self-service process. Do this at the organization level so a new account never holds PHI without coverage.

  • Encryption obligation. AWS’s architecture whitepaper notes that the BAA requires customers to encrypt PHI stored in or transmitted using HIPAA-eligible services, following HHS guidance on making PHI unusable, unreadable or indecipherable.

The BAA is one document. It does not replace your own risk analysis, policies, training records or breach procedures, which sit outside AWS entirely.

Step 2: Keep PHI inside HIPAA-eligible services

AWS maintains the HIPAA Eligible Services Reference, a list of services approved to create, receive, process, maintain or transmit ePHI. Generally available features of listed services are also eligible unless the page notes otherwise.

The working rule from AWS’s HIPAA page: you may use any AWS service in an account designated as a HIPAA account, but you should only process, store and transmit PHI in the eligible services defined in the BAA. A service not on the list can still sit in your architecture if it never touches ePHI.

In practice, this means drawing a boundary:

  1. List every service in the workload.

  2. Mark which ones ever see PHI, including logs, queues, caches, analytics and backups.

  3. Confirm each marked service is on the current eligible list.

  4. Re-check when you adopt a new service or feature.

The list changes over time, so check it directly rather than relying on a snapshot in an article. Teams get caught when a developer wires in a convenient service for debugging or analytics and PHI quietly flows through it.

Step 3: Understand what you own under shared responsibility

AWS runs the physical facilities, the host operating system and the virtualization layer. You run what is above that. AWS’s shared responsibility documentation for HIPAA says customers are responsible for the guest operating system, application software, security group configuration and, for HIPAA, all system components that process, store or transmit PHI. That includes access controls, log settings and encryption settings on services such as S3 and DynamoDB.

AWS and customer HIPAA responsibility layers

The AWS Security Blog’s July 2026 guidance on the Security Rule’s technical safeguards organizes this around 45 CFR §164.312: access control, audit controls, integrity, authentication and transmission security. It includes a responsibility matrix, guidance on defining an ePHI boundary, and a foundation checklist. It is an implementation reference, not a legal interpretation, but it is a good map of the work.

The technical controls that matter most

The Security Rule describes outcomes, not AWS settings. Here is how those outcomes usually translate into configuration.

Safeguard

What it means on AWS

Access control

Least-privilege IAM roles, MFA, no hardcoded access keys, short-lived credentials

Encryption at rest

KMS encryption on S3, RDS, EBS volumes, snapshots and backups

Encryption in transit

TLS on every connection, including between internal services

Audit controls

CloudTrail enabled in every region, logs stored in a protected location, alerts on unauthorized API calls

Public exposure

S3 Block Public Access at the account level, private subnets, VPC endpoints

Integrity and recovery

Versioning, tested backups, documented restore procedures

Encryption

AWS recommends encrypting PHI both in transit and at rest and notes that AWS KMS gives customers flexibility in how they do it, per the shared responsibility paper. Under the current rule, encryption is an “addressable” specification, meaning you must implement it or document an equivalent alternative. In practice, treat it as required. The proposed rule discussed below would remove the addressable label.

The gaps that tend to appear are unencrypted snapshots, backups and volumes, and internal traffic that skips TLS. Turn on default encryption for new resources so the safe option is automatic.

Access control and identity

Overly permissive roles, missing MFA and long-lived access keys are the usual IAM failures. Define roles per function, require MFA for humans, use roles rather than static keys for workloads, and review access on a schedule so you have evidence that you did. Our AWS security checklist for startups walks through the baseline IAM and account-level settings that this builds on.

Logging, without leaking PHI into logs

Audit controls require that you can record and examine activity. Enable CloudTrail everywhere PHI lives, protect the logs from tampering, and alert on suspicious API calls. Decide how long you retain records, because HIPAA documentation requirements run six years. Our guide to cloud log management on AWS covers what to collect and where to keep it.

One trap: application logs. Request bodies, exception dumps and patient identifiers in query strings can end up in CloudWatch or third-party log tools. Scrub identifiers in the application framework and keep logs free of PHI wherever you can.

Network and storage exposure

Public S3 buckets are the classic healthcare breach pattern. Enable S3 Block Public Access at the account level rather than bucket by bucket, put databases in private subnets, and use VPC endpoints so traffic to AWS services stays off the public internet. Where users need file access, use short-lived signed access rather than a public policy.

Step 4: Do the risk analysis, including your cloud assets

The Security Rule requires a documented risk analysis of every system that holds ePHI. That includes your EC2 instances, S3 buckets, RDS databases and cloud backups, assessed with the same rigor as anything on premises. The analysis is not an AWS feature, and AWS cannot write it for you. If you also run on-site systems, such as healthcare data center infrastructure, those belong in the same analysis.

A useful risk analysis produces an inventory of assets, the threats and vulnerabilities for each, the likelihood and impact, and a remediation plan with owners and dates. Keep it current as the environment changes. When a customer’s security review or an auditor asks for it, this is the document they want first.

Step 5: Verify continuously, not once

A configuration that was correct in March can be wrong by June after a few deployments. Verification should be automated where possible.

  • AWS publishes an Operational Best Practices for HIPAA Security conformance pack for AWS Config, which checks resources against a set of HIPAA-related rules.

  • Use Config rules and Amazon Macie to catch public buckets and PHI turning up in unexpected places.

  • Enforce guardrails at the organization level, such as Service Control Policies that prevent anyone from disabling Block Public Access or CloudTrail.

  • Keep evidence: screenshots and exports of settings, access reviews, restore test results and incident records.

This ongoing maintenance, not the first setup, is where small teams struggle. Findings from a scan need an engineer who can fix them, and the fix needs to be written down.

What is changing in the Security Rule

On December 27, 2024, HHS’s Office for Civil Rights issued a Notice of Proposed Rulemaking to update the Security Rule. It proposes removing the distinction between required and addressable specifications and would require, among other things:

  • encryption of ePHI at rest and in transit

  • multi-factor authentication, with limited exceptions

  • vulnerability scanning at least every six months

  • penetration testing at least every 12 months

  • network segmentation

  • procedures to restore data within 72 hours

  • annual written verification from business associates

While the rulemaking proceeds, the current Security Rule remains in effect. AWS’s July 2026 guidance notes that the final rule had not been published as of June 2026, and it recommends treating all specifications as required for new workloads. Building to the stricter standard now is cheaper than retrofitting later, and it matches what a careful risk analysis would already point you toward.

Why a subscription fits this work

HIPAA on AWS is not a project with an end date. It is a set of controls that must stay correct while your product changes weekly. That favors a steady engineering resource over a one-time engagement or a hire you may not be able to afford.

A full-time senior DevOps or security engineer is a large fixed cost for an early-stage team, and one person is a single point of failure. Hourly consulting makes every small fix a negotiation. A subscription with a shared backlog fits the shape of the work: encryption gaps, IAM clean-up, logging, monitoring and evidence collection arrive as a stream of tasks, not one deliverable.

How Cloudwise handles the AWS side

Cloudwise is a subscription service that gives you senior AWS and DevOps engineers for day-to-day cloud operations. You submit tasks to a shared backlog and prioritize them. Typical response time is 24 to 48 hours, with no hourly billing and no long-term contracts, and you can cancel anytime.

How Cloudwise handles the AWS side

The work that supports a HIPAA program on AWS falls within what we offer: security and reliability work including IAM, monitoring and observability, AWS infrastructure management, migrations and modernization and compliance and audit readiness. A backlog for a HIPAA-bound team might include enabling default encryption across accounts, tightening IAM roles, setting up CloudTrail and alerting, or moving a workload into eligible services.

The Simplest Subscription is $7,499 per month and covers unlimited AWS and DevOps task requests. Browse the full list of services to see what fits your backlog.

Cloudwise

The legal side stays with the right professionals: your attorney or compliance officer decides how HIPAA applies to you, and you own the policies, training and breach procedures that live outside AWS.

Quick start checklist

  1. Confirm with your compliance officer or attorney whether you are a covered entity or business associate.

  2. Accept the AWS BAA in AWS Artifact for the whole AWS organization.

  3. Inventory services and confirm PHI only touches HIPAA-eligible ones.

  4. Turn on account-level S3 Block Public Access and default encryption.

  5. Enforce TLS everywhere, including internal traffic.

  6. Apply least-privilege IAM and MFA, and remove hardcoded keys.

  7. Enable CloudTrail in every region and protect the logs.

  8. Keep PHI out of application logs.

  9. Test backups and document restore steps.

  10. Complete and update your written risk analysis.

  11. Run the HIPAA conformance pack in AWS Config and fix what it flags.

Sources

Recommended

Questions Answered

Frequently asked questions

Clear answers to common questions about our AWS & DevOps subscription and how it works.

What AWS services are HIPAA eligible?

Add Icon

Is the cloud HIPAA compliant?

Add Icon

What AWS services are HIPAA eligible?

Add Icon

Is the cloud HIPAA compliant?

Add Icon

Do I need a BAA to use AWS with PHI?

Add Icon

Is the proposed HIPAA Security Rule update in effect?

Add Icon

Do I need a BAA to use AWS with PHI?

Add Icon

Is the proposed HIPAA Security Rule update in effect?

Add Icon

Do I need a BAA to use AWS with PHI?

Add Icon

Is the proposed HIPAA Security Rule update in effect?

Add Icon

Read More Articles