What are EHR downtime procedures?
EHR downtime procedures are the steps a hospital follows to keep caring for patients when clinicians cannot access all or part of the electronic health record. They cover three phases: preparing before an outage, working through it, and recovering afterward.
A downtime can be planned, such as a system upgrade, or unplanned, such as a hardware or software failure, a power outage, a natural disaster, or a cyberattack. Either way, EHR downtime procedures usually spell out:
✓Who declares a downtime and who leads the clinical and technical response
✓How staff are notified when the usual channels are unavailable
✓Which paper forms replace orders, medication administration records, and results
✓How clinicians view patient information, such as a read-only backup of the EHR
✓How patients are identified and registered during the outage
✓How paper records get back into the EHR once it returns
What is the difference between a downtime plan and downtime readiness?
A downtime plan is a written policy. Downtime readiness is the ability of staff on any shift to carry out that policy without the systems they normally rely on.
Most hospitals already have the plan. The SAFER guide treats written downtime and recovery policies as a required practice (2.3) and adds a second one right beside it: staff are trained and tested on those procedures (2.1). The gap between the two is where a readiness review does its work.
Readiness has four parts. Each one can be checked in a drill.
Part
What it covers
A readiness question
People
Schedules, call lists, and escalation contacts
Can staff find out who is on call without the intranet or email?
Access
Sign-in, devices, connectivity, and power
Can designated staff get into the downtime tools when corporate sign-in is unavailable?
Information
Patient data and when it was last refreshed
Does the clinician know how old the downtime snapshot is?
Workflow
Orders, results, documentation, and handoffs
Does each open request have a named owner through the shift change?
Why do hospitals need to prepare for EHR downtime?
Because every hospital with an EHR will have one. The SAFER guide states that EHR unavailability "will occur in every EHR-enabled healthcare environment," whether from hardware or software failure, a power outage, a natural disaster, or a cyberattack.
The guide lists the patient safety hazards documented during downtime: a higher risk of medication errors, images that cannot be viewed, delayed follow-up of test results, and canceled procedures. It also notes that the impact grows when one system serves several facilities.
93%
increase in large healthcare data breaches over the past five years, per HHS
~94%
of US hospitals reported a direct or indirect impact from the Change Healthcare ransomware attack in 2024
2–3 wks
typical minimum recovery time for a hospital ransomware event, based on publicly reported incidents
Cyberattacks change the planning assumptions in two ways.
Duration
A ransomware event can last far longer than a maintenance window. The SAFER guide recommends paper forms on every unit for at least 8 hours of care and a formal review of any unexpected downtime over 24 hours.
Scope
An attack can take out the tools hospitals plan to fall back on. That is why SAFER practice 2.2 asks for a communication strategy that does not depend on email, a website, or voice-over-IP phones.
How should hospitals prepare for ransomware downtime?
Prepare for a longer outage with fewer working tools than a routine downtime, and practice it. The SAFER guide addresses ransomware in four of its practices.
1
Train people to spot the way in
Malicious email attachments and callers asking for login details are often the first point of entry. SAFER recommends training, simulated phishing tests, and limiting each user's access to what their job requires (practice 2.4).
2
Keep backups an attack cannot reach
Backup storage is kept separate from normal file storage, and a full restore is tested regularly (practice 1.4).
3
Have a way to communicate off-network
Email, the intranet, and voice-over-IP phones may all be down together. A communication plan that doesn't run on the affected network is essential (practice 2.2).
4
Rehearse the response
SAFER recommends simulated ransomware drills that cover both the clinical side and the technical side (practice 2.4), followed by an in-depth review of any real outage over 24 hours (practice 3.3).
How should a team use this guide?
Use it in one working session with clinical, operational, and technical leaders in the same room, then test what you find in a drill.
1
Name the team
SAFER recommends that practicing clinicians sit on the contingency planning team. Add IT, pharmacy, lab, imaging, registration, quality and risk, and emergency preparedness.
2
Score the 13 SAFER practices
Use the self-assessment below. Record a status, an owner, and a follow-up action for each one.
3
Run the six drills
Each of the six scenarios in this guide ends with a short test you can ask to see. A drill shows what a policy review cannot.
4
Read the full SAFER guide
The practice numbers and page references here point to selected relevant practices. They are a starting point for discussion, and they do not replace your hospital's policies or clinical safety requirements.
SAFER scores each practice on a five-point scale
Substantial progress
61 to 90%
Fully implemented
91 to 100%
What does the SAFER guide recommend for EHR downtime?
The 2025 SAFER Contingency Planning guide recommends 13 practices in three domains: the technology itself, how people use it, and how the organization monitors it. SAFER labels ten of the practices "required," meaning they tie to a federal rule, regulation, or condition of participation, and rates the other three as medium-strength recommendations.
Domain 1: Safe health IT
These five practices cover the infrastructure: recovery sites, power, paper, backups, and patient identification.
1.1 Disaster recovery plans (p. 10)
Required
A disaster recovery plan describes how the servers and networks behind clinical and administrative applications get restored. SAFER's concern is the single point of failure, such as one database server, one internet connection, or backup tapes stored next to the production servers.
›Review the plan at least once a year
›Large, 24-hour organizations keep a remote "warm site" that can be running in under 8 hours
›Test the backup system at least quarterly
›Keep two separate paths to the internet through two different providers
1.2 Backup power (p. 11)
Required
Hospitals need to keep health IT, data, and communication running through a sustained power outage.
›An uninterruptible power supply (UPS) carries the EHR for at least 10 minutes while the generator starts
›The generator can sustain the EHR functions the organization depends on
›Two days of generator fuel are kept on site
›The UPS and generator are tested regularly, ideally monthly, and kept where they are unlikely to flood
1.3 Paper forms (p. 12)
Required
Clinical and administrative work has to continue during a downtime, so paper needs to be ready to stand in for key EHR functions.
›Each patient care area keeps enough forms to care for its patients for at least 8 hours, including forms for orders, medication administration, lab tests, and radiology tests
›A defined process gets paper records into the EHR after recovery; orders are entered as coded data and paper documents are scanned
1.4 Backups of patient data and system settings (p. 13)
Required
Hardware eventually fails. Backups of patient data and of the system's configuration let the organization restore to the point before the failure with minimal loss.
›Patient data gets a complete, encrypted, off-site backup every day; the most important data is backed up as close to real time as possible
›Several backups are kept at different times and stored apart from normal file storage (an air gap); this helps with recovery from ransomware
›A full restore from the off-site backup is tested regularly, ideally monthly
›A read-only backup EHR is updated at least hourly and tested weekly; unit-level copies are plugged into backup power
1.5 Patient identification (p. 14)
Required
Without a downtime process for identifying patients, records can be mismatched or lost, putting confidentiality, data integrity, and patient safety at risk.
›New patients can be registered during downtime with unique temporary record numbers, reconciled once the EHR is back
›Standard identification steps are used for specimen collection, medication administration, imaging, and delivery of results
Domain 2: Using health IT safely
These five practices cover people and procedures: training, communication, written policy, ransomware awareness, and restarting connections between systems.
2.1 Staff training and testing (p. 15)
Required
SAFER points out that at any given time, many organizations have employees who have never worked in a paper-based environment.
›Training requirements make sure each employee knows what to do during a downtime, and clinicians are trained on paper ordering and charting
›Clinical staff have a job aid, such as a pocket reference card or checklist
›The organization runs an unannounced downtime drill at least once a year
›Staff can reliably get login information for the read-only backup EHR, which may differ from their everyday credentials
›A list of every system-to-system connection is kept and reviewed every 6 to 12 months
2.2 Communication independent of EHR infrastructure (p. 16)
Required
The organization needs to reach key people without the computers or network the EHR runs on. SAFER rules out email, a website, and voice-over-IP phones.
›Leaders and clinicians are notified through a method that doesn't share the EHR's infrastructure; a mobile phone call tree is an example
›There is a way to activate the read-only backup EHR and tell clinicians how to get to it
›There is a way to tell clinicians when the EHR is back online and ready to use
2.3 Written downtime and recovery policies (p. 17)
Required
A written policy gives everyone the same understanding of how to care for patients and keep critical business operations going.
›The policy states when a downtime is called, who is in charge, how everyone is notified, how often updates go out, and how information collected during downtime gets into the EHR
›Reviewed at least every 2 years
›Says when to switch to the warm site, ideally before an unplanned outage reaches 2 hours
›Paper copies are kept on clinical units and in a safe off-site location
2.4 Ransomware prevention training (p. 18)
Required
Malicious email attachments and callers asking for login details are often the first point of entry for a ransomware attack.
›Users learn to check where a link goes and who sent it, and to confirm unexpected requests through a separate channel
›Access follows "least privilege," meaning each user can reach only what their job requires
›The organization runs simulated phishing tests
›Simulated ransomware drills that cover both the clinical response and the technical response
2.5 Stopping and restarting system interfaces (p. 19)
Medium
An interface is a connection that passes data between systems. If one is stopped or restarted the wrong way, data in transit can be lost or corrupted without any warning.
›A written procedure covers how to start, stop, restart, test, and monitor each interface, internal and external
›Staff confirm the interface queues are empty before stopping or restarting a system
›Users are told when a clinical interface is not working
Domain 3: Monitoring safety
These three practices cover how the organization tracks downtime and learns from it.
3.1 Testing, monitoring, and auditing (p. 20)
Required
Testing and monitoring can prevent downtime events and limit their impact, whether the cause is a natural disaster, a technology failure, or a cyberattack.
›Downtime events are monitored and reported, along with system response time for important clinical tasks
›A written policy describes downtime testing for hardware, software, processes, and people, with a log of every test
›Top leadership monitors unplanned downtimes and whether the follow-up is working
3.2 Functional downtime (p. 21)
Medium
A functional downtime is when the system is technically up but too slow to use safely. Slow response can cause "type ahead" errors, where keystrokes land in the wrong field once the screen catches up.
›Response time is measured on a routine basis; one definition SAFER offers for a functional downtime is any hour where average response time is over 5 seconds
›Users have an easy way to report slowness to the IT help desk
3.3 Review of long downtimes (p. 22)
Medium
An unexpected downtime that lasts more than 24 hours is a chance to learn how to manage and prevent the next one.
›A group of clinicians and IT professionals reviews the event, how it was managed, and the likely root causes
›The review covers the effect on care quality, safety, and timeliness
›The organization considers bringing in outside experts in system reliability
Six drill scenarios to try
SAFER asks hospitals to train and test staff on downtime procedures (2.1) and to run simulated ransomware drills (2.4), but it does not script the drills. These six scenarios are examples of what a drill can cover. Each one lists the SAFER practices it relates to, and where it asks for more than SAFER describes, your clinical, operational, and technical teams would need to validate against hospital policy.
Drill 1
Access: get your team working when systems fail
Time: 30–60 min
Who: IT, charge nurse, unit manager
Turn off: EHR, corporate sign-in, email, hospital network
Goal: Give authorized staff a tested way to keep working when normal hospital systems are unavailable.
Provide secure access without the EHR, corporate sign-in, email, or primary hospital network, using approved alternative connections and devices. Keep essential contacts and instructions available, preserve identifiable users and an audit trail, and provide a usable manual fallback when all connectivity is lost.
Ask to see it work: Disable the EHR, corporate sign-in, email, and hospital network during a drill. Have designated staff access the continuity environment and complete an urgent request using the approved alternative.
Related SAFER practices: 2.2 (p. 16), 1.1 (p. 10), 1.3 (p. 12)
Drill 2
Patient identity: know who every patient is
Time: 45–60 min
Who: Registration, charge nurse, bed management
Turn off: EHR registration module
Goal: Keep a trustworthy patient census, including new arrivals and patients whose identity is initially unknown.
Identify existing patients, assign unique downtime identifiers to new or unidentified arrivals, and track location and responsible team. Use two hospital-approved patient identifiers to match records, specimens, medications, and transfers. Location is not an identifier. Review the matching of temporary identifiers to permanent records after recovery.
Ask to see it work: Register two patients with similar names and one unidentified arrival, move them between units, then reconcile their downtime identifiers without combining records or losing their locations.
Related SAFER practices: 1.5 (p. 14)
Drill 3
Patient information: see the essentials behind each care decision
Time: 30–45 min
Who: Charge nurse, attending, pharmacist
Turn off: Live EHR; use only the read-only backup
Goal: Make essential patient information available with its source, age, and uncertainty clearly visible.
Show the latest available allergies, medications, important conditions, code status, recent results, and care team through approved records or downtime snapshots. Display source and update time, flag unknown or outdated information, and make verified downtime changes visible to the next caregiver.
Ask to see it work: Start with an older patient snapshot, document a newly verified allergy during downtime, and confirm that the next clinician sees both the change and the original information's age.
Related SAFER practices: 1.4 (p. 13)
Drill 4
Clinical coordination: put an owner on every urgent request
Time: 45–60 min
Who: Charge nurse, covering clinician, incoming shift
Turn off: Paging system and EHR messaging
Goal: Help staff reach the responsible clinician and see who has accepted responsibility.
Route requests to the current responsible clinician or service, distinguish delivery from acceptance, and escalate unanswered requests according to hospital policy. Transfer open work between shifts with explicit ownership, while preserving established emergency activation channels.
Ask to see it work: Send an urgent request to an unavailable covering clinician, confirm escalation reaches an alternate, and hand the unfinished request to the next shift with an accepting owner.
Related SAFER practices: 2.2 (p. 16)
Drill 5
Diagnostics: follow every order through to clinical action
Time: 60–90 min
Who: Nurse, ordering clinician, lab, radiology
Turn off: LIS/RIS interfaces; use paper requisitions
Goal: Keep diagnostic requests, specimen handling, results, and required follow-up connected.
Support approved clinical orders from authorization through acceptance, collection or scheduling, performance, reporting, and review. Preserve patient and specimen identity, distinguish coordination messages from clinical orders, and record acknowledgment and required action for critical results.
Ask to see it work: Run an urgent laboratory order through downtime collection and reporting, then verify that a critical result reaches an accountable clinician and its required follow-up is documented.
Related SAFER practices: 1.3 (p. 12), 1.5 (p. 14)
Drill 6
Medication safety: keep medication decisions and doses accounted for
Time: 60–90 min
Who: Nurse, pharmacist, ordering clinician
Turn off: eMAR and pharmacy system; use paper MAR
Goal: Give clinicians and nurses a reliable view of medication orders, administrations, and outstanding doses.
Connect validated pharmacy and medication administration record capabilities, including an approved manual MAR when needed. Preserve authorized orders, pharmacy review, relevant allergies, last administration times, omissions, and changes in one agreed authoritative record, with required high-risk medication checks and reviewed reconciliation after recovery.
Ask to see it work: Process a medication change across a shift handoff, verify the last administered dose and pharmacy review, then restore systems without creating a duplicate order or administration.
Related SAFER practices: 1.3 (p. 12), 1.5 (p. 14)
Downtime readiness self-assessment checklist
The 2025 SAFER Contingency Planning guide has 13 recommended practices in three domains. Pick a status for each, name an owner, and note the follow-up. The "what to look for" column summarizes selected implementation guidance from the guide, so read the full worksheet page before you score.
Domain 1: Safe health IT
Practice
What to look for
Strength
1.1 Disaster recovery plans are in place and reviewed at least annually (p. 10)
Backup site tested at least quarterly. Redundant internet path through two providers.
Required
1.2 A generator and enough fuel support the EHR in an extended power outage (p. 11)
Battery backup and generator tested monthly. Two days of fuel on site. Both kept away from flood risk.
Required
1.3 Paper forms are available to replace key EHR functions (p. 12)
At least 8 hours of forms on each unit for orders, medication administration, lab, and radiology. A process to enter and reconcile paper records after recovery.
Required
1.4 Patient data and configuration settings are backed up and tested (p. 13)
Daily off-site encrypted backup. Backups kept separate from normal file storage. Read-only backup EHR updated at least hourly and tested weekly.
Required
1.5 Policies ensure accurate patient identification before, during, and after downtime (p. 14)
Temporary record numbers for new patients, with reconciliation after recovery. Standard identification steps for specimens, medications, imaging, and results.
Required
Domain 2: Using health IT safely
Practice
What to look for
Strength
2.1 Staff are trained and tested on downtime and recovery procedures (p. 15)
Unannounced downtime drill at least once a year. A reference card or checklist for clinical staff. Staff know how to reach and sign in to the read-only backup.
Required
2.2 Downtime communication is independent of the infrastructure that supports the EHR (p. 16)
A way to notify leaders and clinicians that does not rely on email, a website, or voice-over-IP. A way to announce the read-only system and the return to normal.
Required
2.3 Written downtime and recovery policies ensure continuity of care and critical business operations (p. 17)
Who calls a downtime, who leads on the clinical and technical side, and how often updates go out. Reviewed at least every 2 years. Paper copies on units and off site.
Required
2.4 Users are trained on ransomware prevention (p. 18)
Training to spot malicious emails and fraudulent callers. Simulated phishing. Simulated ransomware drills from the clinical and technical side.
Required
2.5 Policies describe how to stop and restart data exchange across interfaces (p. 19)
A procedure to stop, restart, test, and monitor interfaces. A way to tell users when a clinical interface is not working.
Medium
Domain 3: Monitoring safety
Practice
What to look for
Strength
3.1 A testing, monitoring, and auditing strategy is in place for downtime events (p. 20)
Downtime events tracked and reported. A log of downtime testing. Leadership reviews unplanned downtimes and the follow-up.
Required
3.2 Functional downtimes (unacceptably slow response) are identified and addressed (p. 21)
Response time measured for key clinical tasks. An easy way for users to report slowness.
Medium
3.3 Unexpected downtimes over 24 hours get an in-depth review (p. 22)
A multidisciplinary review using root-cause analysis or a similar method. Effect on care quality, safety, and timeliness reviewed.
Medium
Drill checklist
Tick each one off once you have watched it work with the EHR, corporate sign-in, email, and hospital network unavailable.
☐Access: designated staff reach the continuity environment and complete an urgent request
☐Patient identity: two similar-name patients and one unidentified arrival are registered, moved, and reconciled
☐Patient information: a newly verified allergy and the age of the original snapshot are both visible to the next clinician
☐Clinical coordination: an unanswered urgent request escalates to an alternate and is handed to the next shift with an owner
☐Diagnostics: an urgent lab order runs through collection and reporting, and a critical result reaches an accountable clinician
☐Medication safety: a medication change crosses a shift handoff and recovery without a duplicate order or administration
What needs to happen when systems come back?
Recovery is its own phase, and it carries its own risks. Three steps keep downtime work from being lost or duplicated.
1
Announce the return
SAFER practice 2.2 asks for a way to tell clinicians the EHR is back online and ready for use. Staff should not find out by trying to log in.
2
Reconcile the record
Enter orders as coded data, scan paper documents that are not otherwise entered, and match temporary patient identifiers to permanent records (practices 1.3 and 1.5). Restart interfaces in an orderly way so data in transit is not lost (practice 2.5).
3
Review what happened
For any unexpected downtime over 24 hours, SAFER practice 3.3 recommends a multidisciplinary review of the event, its effect on care quality, safety, and timeliness, and what would prevent or soften the next one. Log shorter events and drill results too (practice 3.1), so leadership can see the trend.
Frequently asked questions
What counts as an EHR downtime?
Any period when clinicians or other users cannot access all or part of the EHR, planned or unplanned. SAFER also describes a functional downtime, when the system is running but too slow to use safely.
How often should hospitals run EHR downtime drills?
At least once a year, unannounced, according to the SAFER guide. SAFER also recommends simulated ransomware drills and regular tests of the supporting systems: the backup site quarterly, and backup power and data restores monthly.
How many hours of paper forms should each unit keep?
Enough to care for its patients for at least 8 hours, including forms for orders and for documenting medication administration, lab tests, and radiology tests.
How often should a downtime policy be reviewed?
SAFER recommends reviewing the EHR downtime and recovery policy at least every 2 years and disaster recovery plans at least once a year.
Does a HIPAA contingency plan cover the SAFER practices?
Not fully. SAFER notes that a contingency plan created for the HIPAA Security Rule addresses many of its recommended practices, but not all of them, and that completing the SAFER checklist does not equal HIPAA compliance. It encourages hospitals to coordinate the two.
Who should be on the downtime planning team?
Practicing clinicians alongside IT. SAFER says clinicians should be on the contingency planning team because the substitute workflows used during downtime depend on their input.