HIPAA & Cybersecurity

HIPAA contingency plan: what a practice has to have on paper

The Security Rule contingency plan: required backups, a restore order, and a way to see patients while the EHR is down.

10 min read
October 18, 2026
Updated October 18, 2026
Practice team reviewing a HIPAA contingency plan and backup procedure
Quick answer
HIPAA contingency plan

A HIPAA contingency plan is the Security Rule standard at 45 CFR 164.308(a)(7). A practice must keep a data backup plan, a disaster recovery plan, and an emergency mode operation plan. Testing and a criticality analysis are addressable, so the decision is documented. The plan names the systems, the people, and the restore order.

The EHR is down at 7:40 a.m. and the first patient is in the room. A contingency plan is the document that tells the front desk, the clinician, and the person who calls the vendor what happens in the next hour, and what happens if the hour becomes a day.

The HIPAA Security Rule treats contingency planning as an administrative safeguard, not as an IT project that can wait for a quiet quarter. The standard sits at 45 CFR 164.308(a)(7). It applies to electronic protected health information the practice creates, receives, maintains, or transmits. A paper schedule in a drawer is not the subject of this standard. The EHR, the imaging share, the billing system, the email that carries referrals, and the backups of all of them are.

HHS describes the Security Rule safeguards, including this one, in its security laws and regulations overview. NIST SP 800-66 Revision 2 is the companion document practices and assessors actually use when they turn the regulation into a procedure. It is a guide, not a second law, and it is the right place to check that a procedure covers what the standard names. The PDF is NIST SP 800-66r2.

What the contingency plan standard requires

Five implementation specifications hang off the standard. Three are required. Two are addressable. Required means the practice implements them. Addressable means the practice assesses whether the specification is reasonable and appropriate in its environment, implements it if it is, and documents an equivalent measure if it is not. Addressable is not a synonym for optional. An auditor who finds a blank page where the assessment should be has found a gap, even when the specification itself was not a fit.

The three required specifications are the data backup plan, the disaster recovery plan, and the emergency mode operation plan. The two addressable specifications are testing and revision procedures, and an applications and data criticality analysis. A one-page policy that says the practice “has backups” does not satisfy any of the five. Each specification asks for a procedure someone can follow on the morning the system is actually down.

Write the plan for the people who will use it. The physician covering Saturday does not need the vendor’s storage architecture. That physician needs to know whether yesterday’s notes are on paper, whether e-prescribing is available, and who is allowed to tell patients the clinic is moving to downtime workflow. The office manager needs the vendor phone number, the order in which systems are restored, and the point at which the practice stops trying to catch up inside the EHR and starts a formal downtime log.

Data backup plan: the copy that is not in the same room

The data backup plan has to establish and implement procedures to create and maintain retrievable exact copies of electronic protected health information. “Retrievable” is the word that does the work. A backup that has never been restored is a hope. A backup that lives on a drive in the same closet as the server is not a plan for a fire, a flood, or a ransomware job that encrypts every volume the server can see.

For a small practice the practical questions are few and they should be written down. What is copied. How often. Where the copy rests. How long it is kept. Who can ask for a restore. How long a restore of the EHR database has taken the last time anyone tried. If the last test restore was never done, the plan should say that, and it should name the date the first test will happen. A plan that pretends a test occurred is worse than a plan that schedules one.

Include the systems that are easy to forget because a vendor holds them. The EHR hosted by the vendor, the imaging archive, the patient portal, the clearinghouse reports, and the email archive are still the practice’s electronic protected health information when the practice is the covered entity. The backup plan should say, for each one, whether the vendor’s copy is the practice’s copy and what the contract lets the practice retrieve if the relationship ends or the vendor has its own outage. That conversation belongs with managed backup and recovery, not in a ticket opened during the outage.

Disaster recovery plan: the order the practice comes back

The disaster recovery plan establishes procedures to restore any loss of data. It is the sequence, not the slogan. A useful sequence names the first system, the second, and the one that can wait until Monday. Most practices restore identity and the EHR before they restore the scanner, the phone attendant, and the marketing site. Write that order down so two well-meaning people do not spend the first hour arguing about it.

Name a person and a backup person for each step. “IT will handle it” is not a name. In a three-clinician office the office manager and one physician are often the only people who can reach the vendor and approve a restore that might roll the database back an hour. The plan should say who makes that call, and it should say what the practice is willing to lose. Restoring last night’s backup discards this morning’s work. If that trade is acceptable, write it. If it is not, the downtime log has to be designed so this morning’s work can be re-entered after the restore.

Restoration is also a security event. The person who brings the EHR back has to be sure the copy being restored is the copy from before the incident, not a copy that already contains the encryption or the unauthorized account. The disaster recovery plan and the security incident procedure should point at each other. They are different documents. They fail in the same hour.

Clinician using a downtime workflow while the EHR is unavailable

Emergency mode operation plan: seeing patients while the EHR is down

The emergency mode operation plan is the one clinicians feel. It establishes procedures to enable continuation of critical business processes for protection of the security of electronic protected health information while operating in emergency mode. In a clinic that means: how do we see the patient, how do we keep the record private, and how do we get the information back into the EHR when it returns.

Decide in advance which visits can proceed and which should be moved. A medication refill for a known patient is a different problem from a new patient whose history lives only in the EHR. The plan should say where the downtime packet is, what it contains, and that the packet does not walk home in someone’s car. Paper created during the outage is still protected health information. It needs a place, a person, and a path back into the chart.

Critical processes are not the same list for every specialty, which is why the addressable criticality analysis exists. An ophthalmology office that cannot reach diagnostic images is in a different emergency from a primary care office that cannot reach the medication list. The analysis is the practice saying, in writing, which applications and which data have to be available for the clinic to function at a minimum, and which can wait. Do that analysis once, review it when a new system is added, and do not leave it as a sentence that says “the EHR is critical.”

Testing, revision, and the criticality analysis

Testing and revision procedures are addressable. They ask the practice to implement procedures for periodic testing and revision of contingency plans. A tabletop once a year is enough for many small practices if the tabletop is real: the EHR is treated as unavailable, the downtime packet is actually opened, and someone writes down what was missing. A test that nobody writes down did not happen for the purpose of the file an auditor will ask to see.

Revise the plan when the test finds a hole, when a vendor changes the restore process, when the practice adds a location, and when the person named in the plan leaves. A plan that still lists a manager who left in March is evidence the document is not in use. Put a review date on the first page and keep the last two versions. The revision history is part of showing the plan is maintained.

Tie the test to the systems the criticality analysis named. If the analysis says e-prescribing and the schedule are critical, the test should touch both. If the analysis never mentioned the patient portal, do not spend the test on the portal and call the critical systems covered. This is the same discipline as the systems that have to come back first: the list is short on purpose.

What to hand an auditor, and what not to invent

A surveyor or a consultant doing a Security Rule review will ask for the contingency plan, evidence of backups, and evidence that someone has thought about emergency mode. They will not be impressed by a template downloaded the week before the visit with no practice name, no system names, and no dates. They will also not be impressed by a binder that contradicts the way the office actually behaves. If the front desk has never seen the downtime packet, the plan is fiction.

Keep the plan next to the risk analysis. The risk analysis is a different required specification, at 164.308(a)(1), and it is where threats such as ransomware, a failed server, and a vendor outage should already be listed. The contingency plan is how the practice responds to the threats it has already decided are plausible. A risk analysis that never mentions availability, and a contingency plan that never mentions the risks, read as two documents written by two people who did not talk.

Qventive’s work on HIPAA security rule safeguards treats the contingency plan as an operating procedure for Northern New Jersey practices, not as a policy paragraph. The address is 200 Passaic Street in Hackensack, and the question we ask in an assessment is the same question the regulation asks: if the EHR stopped this morning, could this office still protect the record and still see the patients it is safe to see. If the answer is a shrug, the plan is not done.

Do not pad the document with promises the practice cannot keep. Do not write that backups are tested daily if they are tested quarterly. Do not write that emergency mode has been drilled if the only drill was a conversation. Accurate and short will survive a review. Optimistic and long will not. When the plan is true, the next step is to put the review on the calendar and to make sure the person named on page one still works here.

HIPAA contingency plan: common questions

Is a HIPAA contingency plan required for a small practice?
Yes. The contingency plan standard at 45 CFR 164.308(a)(7) applies to covered entities and business associates that hold electronic protected health information. Practice size does not remove the standard. The procedures can be short. They still have to exist, name the systems, and be something the office can follow.
Which parts of the contingency plan are required and which are addressable?
Data backup, disaster recovery, and emergency mode operation are required implementation specifications. Testing and revision, and the applications and data criticality analysis, are addressable. Addressable still requires an assessment and a documented decision. It does not mean the practice may ignore the specification.
Does a cloud EHR vendor’s backup satisfy the data backup plan?
Only if the practice knows what the vendor copies, how far back the copies go, and how the practice requests a restore. The vendor’s backup can be the practice’s retrievable copy. The plan still has to say so, and someone at the practice has to have asked for a restore at least once so the procedure is real.
How often should the contingency plan be tested?
The regulation says periodic testing and revision. It does not set a number of days. Pick an interval the office will keep, write it in the plan, and keep a note of what the test found. A yearly tabletop that opens the real downtime packet is more useful than a quarterly checkbox nobody performs.
Where should downtime notes go when the EHR returns?
Back into the record of the patient they belong to, entered by someone the plan names, with the date of the visit and a mark that the note was captured on paper during downtime. Loose pages in a folder are still protected health information, and they are invisible to the next clinician until they are filed.
Qventive Healthcare

Let’s meet

A free practice technology assessment looks at the contingency plan you have, the backups you can actually restore, and the downtime path the front desk would use tomorrow morning.