Healthcare Cloud Migration: A Practical Guide for Providers
Healthcare Cloud Migration: A Practical Guide for Providers

Healthcare cloud migration means moving clinical, billing and operational systems, including your EHR, from servers you run to infrastructure you rent. HIPAA allows it. The work is in the details: a signed business associate agreement, your own risk analysis, a migration approach matched to each system, and a cutover plan that keeps clinicians working. This guide walks through each of those, in the order you will face them.
TL;DR
- HIPAA permits cloud storage and processing of ePHI if you sign a HIPAA-compliant business associate agreement (BAA) with the cloud provider and still do your own risk analysis.
- Not every system moves the same way. Sort each application into rehost, replatform, repurchase, refactor, retire or retain before you set a budget.
- Most migration damage comes from data conversion, contracts and cutover, not from the cloud itself. Write transition terms and test before go-live.
- Migration does not end at cutover. Monitor continuously, retest after every update and keep IT and clinicians in the same conversation.
Table of Contents
- Why healthcare organizations are moving to the cloud
- What actually moves: EHRs and everything around them
- Step 1: Assess readiness and decide what you are willing to risk
- Step 2: Choose a migration approach for each system
- Step 3: Lock down HIPAA before any data moves
- Step 4: Plan for downtime, legacy systems and data quality
- What migration costs
- Step 5: Validate, optimize and maintain after go-live
- Our perspective
- How we can help
- FAQ
- Sources
Why healthcare organizations are moving to the cloud
The pressure comes from three directions: resilience, cost structure and AI.
On resilience, England’s NHS cloud guidance says cloud adoption can improve availability, agility and resilience and reduce workforce burden through on-demand resources, backups, failover and automation. That is a UK source, but the operational logic applies to any clinic that has ever lost a server closet to a power failure or a ransomware event.
On cost, TechTarget reports that the economics have shifted: SaaS platforms and cloud providers have matured, while demand for AI is raising the need for scalable computing and GPU hardware keeps getting more expensive. Buying that capacity for a single practice or system rarely makes sense. Pay-as-you-go models also let organizations avoid over-investing in unused onsite infrastructure and put more of the budget toward patient care.
The third driver is data. Scattered systems keep information in silos. Moving to shared, cloud-based platforms is how many organizations finally connect scheduling, billing, clinical records and analytics.
What actually moves: EHRs and everything around them
When people search for an EHR in the cloud, they usually mean one of two things.
- A vendor-hosted or SaaS EHR. The vendor runs the software and you subscribe. Your job is contract review, data conversion and workflow change.
- Your existing EHR or custom application, hosted on a cloud platform. You or a partner re-host or rebuild it on infrastructure from a cloud provider. Your job includes architecture, security configuration and operations.
Beyond the EHR, a typical migration inventory includes:
- Practice management and billing systems
- Patient portals and scheduling tools
- Imaging archives and document storage
- Interfaces and integration engines that move HL7 or FHIR messages between systems
- Reporting databases and analytics
- Backups and disaster recovery copies
- Internal tools your staff built over the years in spreadsheets and small databases
That last category is where surprises hide. NIST has noted that organizations often underestimate how much ePHI they hold across applications, mobile devices and medical equipment. Your inventory must find it all before you can protect it or move it.
Step 1: Assess readiness and decide what you are willing to risk
Start with decisions, not servers. NHS England’s cloud security guidance tells organizations to assess risk appetite, identify risks and governance requirements, weigh issues such as lock-in and data residency, and document their decisions. The same discipline works for a U.S. practice.
A workable readiness assessment answers these questions:
- What do we run, and who owns it? One named owner per application, clinical and technical.
- Where is ePHI? Include shadow systems, exports and email attachments.
- What can we not tolerate? Set a maximum acceptable downtime per system. A scheduling tool and a medication administration module do not share the same number.
- What are our dependencies? Lab interfaces, e-prescribing, clearinghouses and devices all connect to something.
- What does our staff already know? A two-person IT team cannot run a complex multi-cloud design.
- What do our contracts say? Existing licenses, hosting agreements and exit terms may constrain your timing.
Write the answers down. That document becomes the backbone of your HIPAA risk analysis and your project plan.
Step 2: Choose a migration approach for each system
There is no single “lift everything” button. The Department of Veterans Affairs’ migration framework lists six options: rehosting (lift-and-shift), replatforming (lift-and-reshape), repurchasing (drop-and-shop), refactoring (rewriting or decoupling), retiring and retaining.
| Approach | What it means | Healthcare example | Watch for |
|---|---|---|---|
| Rehost | Move as-is to cloud servers | An on-premises practice management server moved to a virtual machine | Costs can run higher than expected if the system is never tuned |
| Replatform | Small changes to use managed services | Moving a database to a managed database service | Testing of interfaces |
| Repurchase | Replace with SaaS | Swapping a legacy scheduling tool for a hosted one | Data conversion and retraining |
| Refactor | Rewrite or decouple | Breaking a monolithic custom app into services with APIs | Highest effort, highest long-term payoff |
| Retire | Switch off | An unused reporting database | Records retention rules |
| Retain | Leave in place | A device-bound system that cannot move | Secure connectivity to the cloud |
A common pattern for a small or midsize organization: repurchase what is commodity (email, scheduling, document storage), rehost what is stable and hard to change, and refactor only the systems that give you a competitive or clinical edge.
Step 3: Lock down HIPAA before any data moves
HIPAA does not forbid the cloud. HHS guidance says a covered entity or business associate may use a cloud service to store or process ePHI if it enters into a HIPAA-compliant BAA with the cloud provider and otherwise complies with the HIPAA Rules. Both parties must comply with the requirements that apply to them, and the covered entity must conduct its own risk analysis and set up risk management policies.
That last point catches people. A BAA is a contract, not a security program. The provider secures its platform; you still configure access, encryption, logging and backups correctly, and you still train your staff.
What your BAA needs to settle
If you want a plain-language primer, this business associate agreement guide covers the basics. For the compliance officer’s view of where agreements go wrong with OCR, see our piece on business associate agreement risk. Beyond the basics, check these points:
- Access is never cut off. HHS states that a business associate may not block a covered entity’s access to ePHI, such as with a software kill switch, even during a payment dispute.
- Exit terms are explicit. The same HHS guidance says a business associate must keep ePHI accessible and usable on demand and, when an agreement ends, return or dispose of PHI as the BAA requires. Spell out formats and timelines.
- Data location is documented. HHS guidance does not bar a provider from storing ePHI on servers outside the U.S., as legal commentary on the guidance notes, but the risks of doing so belong in your risk analysis. Many organizations simply require U.S. regions.
- Subcontractors are covered. If your cloud provider uses other vendors that touch ePHI, those relationships need agreements too.
Keep visibility as data moves
NIST’s cloud security guidance warns that organizations must keep visibility into cloud data protection to stay consistently compliant, because cloud technologies can dynamically move data across geographic locations. In practice: tag workloads that hold ePHI, restrict regions, turn on audit logging from day one and review who can reach what.

For a broader checklist of what an audit looks for, read our HIPAA compliance essentials, and our perspective on building HIPAA-ready systems.
Step 4: Plan for downtime, legacy systems and data quality
The three problems that most often hurt a healthcare migration are downtime, legacy data and contracts.
Downtime. Plan cutovers for your quietest window, keep a read-only copy of the old system available, and write a paper downtime procedure that staff have actually rehearsed. Decide in advance what triggers a rollback and who makes the call.
Legacy systems. Old applications often depend on outdated operating systems, hard-coded addresses or vendor-specific interfaces. Test each interface in a staging environment before you trust it. If a system cannot be moved safely, retain it and connect it through a secured, monitored link while you plan its replacement.
Data conversion and contracts. The Office of the National Coordinator for Health IT (ONC) notes that switching EHRs can be costly and disruptive, with risks of lost access to records or incomplete or corrupted data. Its advice is to negotiate transition provisions for data transfer, conversion and support, and to specify a standardized structure and format for transferred data along with a deadline for conversion. Those clauses cost nothing to ask for before signing and a great deal to lack afterward.
A few habits make conversion safer:
- Run a trial conversion on a copy of production data and have clinicians review sample charts.
- Reconcile record counts and key fields between old and new systems.
- Decide what historical data moves in full, what is archived and what is retired under your retention policy.
- Keep the legacy system readable until you have verified the new one.
What migration costs
Budget for four layers, not one:
- Assessment and planning: inventory, risk analysis, architecture.
- Migration work: engineering, data conversion, interface rebuilds, testing, training.
- Overlap: for a while you pay for the old and the new environment at once, so schedule the shutdown of legacy hosting as part of the project.
- Ongoing operations: cloud usage, monitoring, security tooling, support and periodic reassessment.
Cloud bills grow with usage, so the savings usually come from tuning and from retiring old infrastructure rather than from the move itself. Ask any partner for a total-cost view over several years, not a migration quote alone.
For reference, here is how we price our own work. A custom MVP or simple app runs $10,000 to $50,000, and a mid-complexity platform runs $50,000 to $150,000, with enterprise work scoped individually. Our custom software cost guide explains what moves a project up or down in that range.
Step 5: Validate, optimize and maintain after go-live
Go-live is the midpoint. The ONC’s SAFER guide on system management says organizations should use continuous technical monitoring, multidisciplinary testing before go-live and after updates, and regular EHR updates, with IT staff and clinicians collaborating on changes to protect workflows and patient safety. It describes system management as configuring, validating and maintaining hardware, software and APIs.
Turn that into a routine:
- Week one: daily huddles between IT and clinical leads, a visible issue log, extra support at the point of care.
- First quarter: review performance, access logs and cloud spending. Right-size what you over-provisioned.
- Every update: standardized regression tests, including interfaces, order sets and reports, before the change reaches clinicians.
- Every year: repeat the risk analysis, review the BAA list, and run a restore test from backup.
Our perspective
We build healthcare software for a living, and the migrations that go well share one trait: someone did the unglamorous discovery first. Before we write a line of code or touch an environment, we want to know where the data lives, who depends on it and what a bad Tuesday looks like for the clinic.
We also think the cloud is not the goal. A cloud-hosted system that nobody can use is no better than the server closet it replaced. The right question is which workflows should get faster, safer or cheaper, and the platform follows from that. For many small providers the best first move is a smaller one: replace one fragile system, prove the process, then continue.
Last, compliance is a design input, not a coat of paint. Encryption, access control, audit logging and a clear BAA chain are far cheaper to build in at the start than to retrofit after an audit finding. Engineers who have worked at scale know that, which is why we staff projects with people who have done it before.
— Cameron
How we can help
If you are weighing a move, our Discovery Audit is a $999 one-off engagement that maps your systems, data flows and risks before you commit to a build. For ongoing work, we deliver HIPAA-conscious custom software, system integrations and cloud architecture. Details on the audit and our related engagements are on our Discovery Audit page, and you can see what compliant delivery looks like in our HIPAA-compliant patient portal case study.

For healthcare teams that want a ready-made foundation, see AiRN, our HIPAA-compliant healthcare software and AI offering.
FAQ
Is it HIPAA compliant to put an EHR in the cloud?
Yes, if you meet the conditions. HHS says a covered entity may use a cloud service for ePHI when it has a HIPAA-compliant BAA with the provider and otherwise complies with the HIPAA Rules. You must also complete your own risk analysis and keep risk management policies current.
Which cloud migration strategy is best for healthcare?
There is no single best one. Replace commodity tools with SaaS, rehost stable systems that are hard to change, and refactor only where a rewrite gives clear value. Retire what you no longer need, and retain what cannot safely move.
How long will our clinic be down during migration?
It depends on the system and the approach. Well-planned cutovers for individual systems can be short, and a read-only copy of the old system plus a rehearsed paper downtime procedure limit the impact. Set a maximum acceptable downtime for each application before you begin.
What should we do first?
Inventory your systems and find where ePHI lives, then document your risk decisions. That inventory drives the HIPAA risk analysis, the choice of approach for each system and the budget. A short discovery engagement can do this for you if your team lacks the time.
Sources
- Why cloud adoption is important to the NHS, NHS England
- Why health systems are moving IT infrastructure to the cloud, TechTarget
- Healthcare in the Cloud, NetSuite
- What’s New with HIPAA, NIST
- Step 2 Assess, NHS England Digital
- Cloud Computing Migration Framework, Department of Veterans Affairs
- Guidance on HIPAA & Cloud Computing, HHS
- May a business associate block or terminate access, HHS
- HHS Releases Guidance on HIPAA and Cloud Computing, Hunton
- Trusted Cloud Security Practice Guide, NIST NCCoE
- Transition Issues: Switching EHRs, HealthIT.gov
- SAFER: System Management, HealthIT.gov