Dental Disaster Recovery Plan in 2026: RTO, RPO, Scenarios and Tests
The request for "the disaster recovery plan" tends to arrive from outside the office: an underwriter's questionnaire, a lender, a buyer's adviser during a practice sale. The document that goes back is usually a backup report. It shows that data was copied last night. It does not say who declares an outage, how long the practice can be closed before tomorrow's schedule is cancelled, which system comes back first, or when anyone last proved the copy can be restored.
This guide is the plan itself, written for a practice with one to three locations: the numbers to choose for each system, the ten sections of the document, the five scenarios, the roles, what a disaster recovery service adds to a backup, and the tests that keep it honest. If you want to score how ready you are today first, use the 12-point dental backup and restore-test checklist; this page is what you write once you know the score.
A dental disaster recovery plan is a brief written document that sets a recovery time objective (RTO) and a recovery point objective (RPO) for each system, names who declares an outage and who calls whom, gives the step order for five outage scenarios, and schedules restore tests. The backup is one input; the plan is the decisions made in advance.
- Set a recovery time objective and a recovery point objective for each system separately; the practice management database and imaging carry the shortest numbers and decide the backup schedule.
- The plan is ten short sections: inventory, the RTO and RPO table, declaring authority, roles and call tree, a backup facts sheet, five scenario runbooks, the alternate workplace, communication templates, vendors, and the test log.
- Write one page for each scenario: server failure, ransomware, fire or flood, extended power or internet loss, and a cloud software or vendor outage; the what-still-works matrix differs for each.
- A disaster recovery service adds what a backup product leaves out: someone acting on failed jobs, an offline or immutable copy, a restore target inside the RTO, scheduled restore tests with records, and ownership of the plan.
- Prove the plan with a timed full recovery rehearsal once a year, a database restore to an isolated machine each quarter, and a 30-minute tabletop with a call-tree test twice a year.
In this guide
- Backup, checklist, recovery plan or downtime procedures: which one is the plan?
- What RTO and RPO should a dental office set for each system?
- What sections should the plan document contain?
- How should the plan handle each outage scenario?
- Who declares, who restores and who tells the patients?
- What does a disaster recovery service include beyond a backup?
- How often should the plan be tested, and what does each test prove?
- How do you build the plan in 30 days?
- If the practice software is in the cloud, what is left for the plan to cover?
- What to ask your IT provider about disaster recovery
- How Legend Networking can help
- Frequently asked questions
- The bottom line
Backup, checklist, recovery plan or downtime procedures: which one is the plan?
A dental disaster recovery plan is the written procedure for getting the practice's systems and data back after a failure, fire, flood, ransomware or long outage, with a target time and a target data loss for each system. A backup is a copy; the plan says who restores it, where, in what order, and how you know it worked.
Four documents get called "the plan", and they answer different questions. Keep all four, keep them short, and keep them apart.
| Document | The question it answers | Who owns it | Where it lives on this site |
|---|---|---|---|
| Backup configuration | Is the data being copied, where, how often? | your IT provider | what a dental office should back up and how often |
| Readiness checklist | Would the copy restore, and are we ready today? | office manager with the IT provider | the 12-point checklist linked above |
| Disaster recovery plan (this page) | When something fails, who does what, in what order, to meet which targets? | the owner signs it; the office manager keeps it | this guide |
| Downtime procedures | How do we keep seeing patients while systems are down? | clinical lead and front desk | the first four hours of a dental server outage |
The federal Security Rule keeps them apart too. At 45 CFR 164.308(a)(7), the contingency plan standard asks a covered entity to establish "policies and procedures for responding to an emergency or other occurrence (for example, fire, vandalism, system failure, and natural disaster) that damages systems that contain electronic protected health information." Its implementation specifications, quoted from the rule text, are:
| Specification (45 CFR 164.308(a)(7)(ii)) | Rule text | Marked |
|---|---|---|
| (A) Data backup plan | "Establish and implement procedures to create and maintain retrievable exact copies of electronic protected health information." | Required |
| (B) Disaster recovery plan | "Establish (and implement as needed) procedures to restore any loss of data." | Required |
| (C) Emergency mode operation plan | "Establish (and implement as needed) procedures to enable continuation of critical business processes for protection of the security of electronic protected health information while operating in emergency mode." | Required |
| (D) Testing and revision procedures | "Implement procedures for periodic testing and revision of contingency plans." | Addressable |
| (E) Applications and data criticality analysis | "Assess the relative criticality of specific applications and data in support of other contingency plan components." | Addressable |
This page quotes the text and does not read it for your practice; what "addressable" means for your office is a question for your compliance adviser. The useful point is practical: the rule lists a backup, a recovery procedure and an emergency-mode procedure as three things, and this guide is the second one with the third folded in as scenario runbooks.
What RTO and RPO should a dental office set for each system?
Set a recovery time objective and a recovery point objective for each system, because the practice database, imaging, phones and website do not hurt equally when down. A typical starting point is same-day recovery for practice software and imaging, next business day for email and documents, and a data-loss window of one business day or less for the clinical record.
The two terms come from the federal contingency planning guide, NIST Special Publication 800-34. In its words, the RTO "defines the maximum amount of time that a system resource can remain unavailable before there is an unacceptable impact on other system resources, supported mission/business processes, and the MTD", and the RPO "represents the point in time, prior to a disruption or system outage, to which mission/business process data can be recovered (given the most recent backup copy of the data) after an outage." The MTD, the maximum tolerable downtime, is the longer limit the whole practice can stand before the damage stops being recoverable with overtime and apologies. In a dental office the MTD is the point where you start cancelling the next day.
The table below is the heart of the plan. Fill the two blank-style columns for your own office; the "choose between" cells are the realistic options, not a recommendation for every practice.
| System | What stops when it is down | RTO: choose between | RPO: choose between | What drives the choice |
|---|---|---|---|---|
| Practice management database (Dentrix, Eaglesoft, Open Dental, or the local parts of a cloud system) | schedule, charts, ledgers, claims, recall | 4 hours · same day · next morning | 15 minutes · 4 hours · last night | chairs per day; whether paper mode is rehearsed |
| Imaging and CBCT archive (DEXIS, Carestream, Planmeca and others) | x-rays and scans cannot be opened or acquired | same day · next morning · 2 business days | last night · last week (if images are also written to a second location at capture) | specialty and surgical load; whether sensors can capture locally and sync later |
| Phones | inbound calls, confirmations, emergencies | 1 hour (forward to mobiles) · same day | not applicable (voicemail and call logs live with the carrier) | whether the number can be forwarded from a carrier portal without the office |
| Internet and firewall | cloud software, claims, imaging bridges, VoIP | 1 hour (cellular failover) · same day · next day | not applicable | whether a second path exists and who switches it |
| Email (Microsoft 365 or Google Workspace) | patient and vendor correspondence, referrals | next business day | last night · provider retention | how many referrals arrive by email |
| Payments and claims (terminal, clearinghouse, statements) | taking card payments, sending claims | same day · 2 business days | last night | whether a standalone terminal exists |
| Workstations and sensor computers | the chairs those machines serve | one chair same day · all chairs within 3 days | n/a if nothing is stored locally | whether a spare configured machine is on the shelf |
| Shared documents, consent forms, HR files | forms, policies, scanned documents not in the PMS | next business day · 1 week | last night · last week | how much is still stored outside the practice software |
| Door access, cameras, alarm | physical access and evidence | same day · 1 week | n/a | whether doors fail open or locked |
| Website and online booking | new-patient requests | 2 business days · 1 week | last published version | share of new patients booking online |
Two notes on the table. First, the RPO for a cloud practice management system is the vendor's, and you should have it in writing; the local parts (imaging, documents, scanned IDs) still need your own number. Second, a shorter RTO or RPO costs more every month, in the form of a standby server or hosted copy, more frequent backup jobs and more frequent tests. The plan's job is to choose on purpose.
How to set the numbers in one 30-minute meeting
Put the owner, the office manager and one clinical lead in a room with the table above and a copy of next week's schedule. Work the method below with letters rather than dollars; the point is to rank, not to price.
- For each system, ask "what can we not do while it is down?" and write the answer in plain words ("cannot see the schedule", "cannot take a PA").
- Count the chairs that stop. Call a full stop S, a slowdown D, an inconvenience I.
- Ask "at what hour do we cancel tomorrow?" That hour is the system's MTD. Set its RTO comfortably inside it.
- Ask "how much re-entry could the team stand?" If the honest answer is "one day of charting from paper", the RPO is last night. If it is "an hour", the RPO is 15 minutes to an hour, and the backup frequency has to follow.
- Rank systems by S first, then D, then I. That order is your restore order, and it goes into the plan as written.
Ready.gov puts the same idea in one line for any business: "Priorities and recovery time objectives for information technology should be developed during the business impact analysis." For a dental office the business impact analysis is this meeting.
What your RPO means for the backup schedule
The RPO you choose decides how the backup must run; the two cannot be chosen separately.
| RPO you chose | What the backup must do | What it adds | What it depends on |
|---|---|---|---|
| Last night (about 24 hours) | one nightly copy of the database and images, with a second copy off site | the lowest cost; one day of re-entry at worst | that the job completes and someone sees the result every morning |
| 4 hours | several database snapshots a day plus the nightly full copy | more storage; a backup method that can copy the database while staff work in it | software that handles an open database correctly |
| 15 minutes to 1 hour | continuous or near-continuous copying of the database to a second machine or hosted copy | a standby that can take over; monitoring of the copy itself | somebody watching the replication, because a silent failure makes the standby useless |
Open Dental's manual is unusually plain about the baseline for its own product: "At minimum, backups should be made daily. If you at least have good daily backups, then the worst-case scenario is having to re-enter one day of data." The same page adds a warning that applies to every practice system: "RAID is not a backup solution and should not be relied on for backups or disaster recovery plans." For Dentrix and Eaglesoft, ask Henry Schein One and Patterson in writing which backup method they support for an open database and whether the vendor's own backup service covers the imaging folders.
What sections should the plan document contain?
A workable dental disaster recovery plan has ten short sections: the system inventory, the RTO and RPO table, declaration authority, roles and the call tree, a backup facts sheet, five scenario runbooks, the alternate workplace, communication templates, the vendor and contract list, and the test log. Ten pages is plenty; one page goes in the front-desk binder.
The template below is the document. Each row is a section heading; the middle column is what to write under it.
| Section | Fill with | Who fills it | Update when |
|---|---|---|---|
| 1. Scope and system inventory | every system from the RTO table, where it runs (server room, hosted, cloud), the version, the login owner | IT provider drafts, office manager checks | any system is added or replaced |
| 2. RTO and RPO table | the completed table from the 30-minute meeting, signed by the owner | owner | yearly, or when chairs or locations change |
| 3. Declaration authority | who may declare an outage a disaster and invoke the plan; the one alternate; the rule for when neither answers | owner | staff change |
| 4. Roles and call tree | the roles matrix and the call order with primary and alternate contact methods | office manager | staff or vendor change |
| 5. Backup facts sheet | what is backed up, how often, where the copies are, how the offline or immutable copy is reached, who holds the recovery credentials and where the sealed copy is kept | IT provider | any backup change; checked at every test |
| 6. Scenario runbooks | the five scenarios in the next section, one page each: first decisions, recovery path, what the front desk does meanwhile | IT provider and clinical lead | after every test or real event |
| 7. Alternate workplace | the colleague's office, the second location or the hosted desktop you would work from; what is needed to make it usable | owner | yearly |
| 8. Communication templates | three pre-written notes: to staff, to patients being rescheduled, to vendors and referrers; who sends each | office manager | yearly |
| 9. Vendors and contracts | the support numbers and account IDs for the practice software, imaging, carrier, phone provider, insurer and IT provider, and what each contract says about response | office manager | contract renewal |
| 10. Test log and review schedule | the date, type, result and gaps of every test; the next review date | IT provider writes, owner signs | every test |
The one-page version for the front desk
The full document lives with the office manager and the IT provider, in print and offline, not only on the server it describes. CISA's ransomware guide makes the same point about incident plans: "Ensure a hard copy of the plan and an offline version is available." The one-page version carries only what the front desk needs in the first hour:
- Who to call first, second and third, with a second number for each.
- The sentence that declares the outage and who may say it.
- Where the paper-mode kit is and which appointments are kept.
- The patient-facing script for rescheduling.
- Where the full plan is.
How should the plan handle each outage scenario?
Write one page for each of five scenarios: a failed server, ransomware, fire or flood that takes the building, an extended power or internet loss, and an outage at a cloud practice-software or clearinghouse vendor. Each page states what still works, the first three decisions, the recovery path that meets the RTO, and what the front desk does meanwhile.
Before the runbooks, the matrix that most plans skip. It shows what keeps working in each scenario for a typical office with a local practice management server, a cloud email service and VoIP phones. Mark your own office's version; a cloud practice system changes the first row.
| Scenario | Practice software | Imaging | Phones | Internet | Payments and claims | |
|---|---|---|---|---|---|---|
| Server failure | down | down (archive) · capture may work locally | up | up | up | portal and terminal up; claims from the PMS down |
| Ransomware | down (isolated) | down (isolated) | up unless the phone system shares the network | cut by choice | up, but treated as suspect | terminal up; everything on the network isolated |
| Fire or flood | down and possibly destroyed | down and possibly destroyed | forwardable from the carrier | down at the building | up from anywhere | portal from anywhere; terminal lost |
| Extended power or internet loss | up on UPS for minutes, then down | same | VoIP down without internet; forward from a phone | down | up on mobiles | terminal needs power and a line |
| Cloud software or vendor outage | down (hosted) | local capture up, upload later | up | up | up | depends on which vendor |
Scenario 1: the server fails
What fails: everything hosted on the server; workstations still boot. First decisions: do not keep restarting it; call the IT provider; declare if the provider's first estimate runs past the RTO. Recovery path: repair in place if the fault is a single part and the estimate fits the RTO; otherwise restore the last good copy to a loaner server, a spare machine or a hosted copy, then re-point workstations. Meanwhile: paper mode from the printed schedule. The step-by-step for the day itself is in the first four hours of a dental server outage; the plan only needs to say which path is taken at which hour, and who decides.
Scenario 2: ransomware
What fails: anything on the network is treated as compromised until proven clean, including the backups that were reachable from it. First decisions: isolate (pull the internet, leave machines powered), declare, call the IT provider and the cyber insurer in the order the policy requires, and start the incident log. The plan's job is the backup part, decided in advance: which copy is offline or immutable, how it is reached when the network is untrusted, and which clean machine it is restored to. CISA's guidance is the standard to write into the plan: "Maintain offline, encrypted backups of critical data, and regularly test the availability and integrity of backups in a disaster recovery scenario", because, as the same guide says, "many ransomware variants attempt to find and subsequently delete or encrypt accessible backups to make restoration impossible unless the ransom is paid." The response sequence is in the dental ransomware recovery playbook; who may talk to counsel and the insurer belongs in the roles matrix below. CISA's guide groups prevention by how attackers get in and tells organizations not to expose services such as remote desktop protocol on the web, so the plan's prevention page should point at remote access security, and at network security for dental practices for the segmentation that limits how far an intruder can reach.
Scenario 3: fire, flood or loss of the building
What fails: everything on site, including any backup copy kept in the building. First decisions: safety first; forward the phone number from the carrier portal; declare; tell the IT provider the building is not available so recovery goes straight to the off-site copy and the alternate workplace. Recovery path: restore the hosted or off-site copy to a hosted desktop or a loaner server at the alternate location; reissue workstations; re-establish imaging with whatever hardware survived. The Security Rule's physical safeguards include a specification for this case, "procedures that allow facility access in support of restoration of lost data under the disaster recovery plan and emergency mode operations plan in the event of an emergency" (45 CFR 164.310(a)(2)(i)); in practice that is a list of who may enter a damaged building with the fire marshal's permission and remove the server, if it is salvageable. Long displacement is its own planning problem; the Gulf-coast version is in planning for weeks of displacement after a storm.
Scenario 4: extended power or internet loss
What fails: on power loss, the server and network have the minutes the UPS gives them, then everything; on internet loss alone, cloud software, claims, VoIP and remote support. A server-room battery backup (UPS) buys minutes for a clean shutdown, not hours of clinic time, so the plan records how many minutes yours actually gives. First decisions: shut the server down cleanly before the UPS empties, which is a decision with a time attached in the plan; switch to the cellular failover if there is one; forward the phones. Recovery path: none needed for data if the shutdown was clean; the plan's work is the order of bringing things back and the check that the database opens before patients are seen. The safe shutdown steps are in shutting a dental server down safely on UPS power, and the internet-only case is in what to do when the office internet is down.
Scenario 5: the cloud practice software or a vendor is down
What fails: the hosted system, while everything in the building works. First decisions: confirm it is the vendor and not your own internet; record the vendor's incident reference; switch to the vendor's offline or read-only mode if one exists, or to the printed schedule. Recovery path: the vendor's, which is why the plan records the vendor's published RTO and RPO and the support route in writing. The local parts still need a plan of their own: imaging captured during the outage, documents scanned, payments taken. Our guide to what a cloud practice system still needs locally lists them. A second dependency belongs here too: a clearinghouse outage stops claims, not care, and the plan's answer is a list of top payers with their portal route.
Who declares, who restores and who tells the patients?
The plan names one person who may declare the outage, one recovery coordinator who runs the day, the IT provider's lead as the technical owner, a clinical lead who runs paper mode, and a front-desk lead who runs the schedule and the phones. Every role has a named alternate, and the call tree records two ways to reach each person.
The matrix gives each decision one owner. "Decides" is the person whose yes is needed; "does" is who carries it out; "is told" is who hears before patients do.
| Decision or task | Decides | Does | Is told |
|---|---|---|---|
| Declare the outage and invoke the plan | owner (alternate: office manager) | office manager | all staff, IT provider |
| Cancel or keep today's and tomorrow's schedule | owner with the clinical lead | front-desk lead | patients, referrers |
| Isolate the network (suspected ransomware) | IT provider's lead, or the owner if unreachable | whoever is at the firewall or server | owner, insurer |
| Restore from backup, and to which machine | IT provider's lead, within the plan's RTO | IT provider | owner, office manager |
| Accept the restored data (the RPO check) | office manager with the clinical lead | office manager compares the last entries | all staff |
| Talk to the cyber insurer and counsel | owner | owner | office manager |
| Patient and referrer communication | owner | front-desk lead, from the templates | clinical lead |
| Enter the paper-mode work afterwards | office manager | assigned staff | IT provider (for any re-sync) |
| Close the event and log the lessons | owner | office manager with the IT provider | all staff |
The call tree
NIST's planning guide describes the tool exactly: "The call tree should account for primary and alternate contact methods and should discuss procedures to be followed if an individual cannot be contacted." Write it as a table with one rule attached: how long to wait before moving to the alternate. Pick the minutes yourself; the plan only has to state them.
| Order | Role | Primary | Alternate | If unreachable after the plan's wait time |
|---|---|---|---|---|
| 1 | Declaring authority | owner, mobile and home | office manager | the office manager declares and records the time |
| 2 | IT provider | the helpdesk number on the contract, with the account ID | the provider's escalation contact | the office manager opens the ticket in writing as well |
| 3 | Clinical lead | associate or lead hygienist | second clinician | front-desk lead starts paper mode anyway |
| 4 | Front-desk lead | lead receptionist | billing coordinator | office manager takes the desk |
| 5 | Vendors as needed | practice software, imaging, carrier, phone provider | account manager | note the reference numbers in the log |
| 6 | Insurer and counsel (ransomware or breach suspicion) | the policy's hotline | broker | owner keeps the policy's notice deadline in view |
NIST also suggests that a plan test can cover this list: an organization "could test call tree lists to determine if calling can be executed within prescribed time limits." That is a ten-minute test and it goes on the schedule below.
What does a disaster recovery service include beyond a backup?
A backup product copies data and reports whether the job ran. A disaster recovery service adds what the plan depends on: someone acting on failed jobs, an offline or immutable copy, a standby machine or hosted environment to restore to within the RTO, scheduled restore tests with written results, and ownership of the runbook and backup facts sheet.
| What you get | Backup product only | Backup plus a disaster recovery service | Inside a managed IT plan |
|---|---|---|---|
| Copies of the database and images | yes | yes | yes |
| Someone sees and acts on a failed job the same morning | you, if you read the email | the provider | the provider |
| Database-aware copy (the practice software's database is copied in a usable state) | depends on the product | yes, and stated in writing | yes |
| Offline or immutable copy that a network intruder cannot reach | sometimes, as an option | yes, named in the backup facts sheet | yes |
| Somewhere to restore to inside the RTO (loaner, standby or hosted) | no | yes, with the RTO it supports | yes |
| Scheduled restore tests with a written result | no | yes | yes |
| The plan document, roles and runbooks kept current | no | usually the recovery parts | yes, with the rest of the office's IT |
| Workstations, network, phones and security looked after too | no | no | yes |
A backup product with nobody watching it is how an office learns, on the morning of a crash, that its only copy stopped updating weeks earlier and the missing days of x-rays and payments are gone. The service is the watching, the restore target and the test, not the copy.
If you are comparing your current arrangement against a service, our guide to data protection services for dental offices sets the recovery layer beside the other eight. The service page for dental data backup describes how Legend approaches the recovery side and what is and is not included; and where the practice wants the plan kept as part of everything else, managed IT services for dental practices is the arrangement that owns the runbook, the tests and the hardware the data is restored to.
The IT side: what the provider is actually doing
Four technical choices sit under the words "backup and recovery", and the plan should record which ones apply to your office.
- File copies vs image copies. A file-level backup copies the database and the image folders. An image-level backup copies the whole server, operating system and settings included, so recovery to a loaner is a restore rather than a rebuild. CISA's guide recommends keeping "golden images" of critical systems for exactly this reason. Image copies shorten the RTO; file copies are smaller and simpler. Many offices keep both.
- Database-aware copies. Practice management systems run on a database engine (MySQL under Open Dental, for example), and a plain copy of an open database can be unusable. Open Dental's manual notes that "it is recommended to run an Open Dental Backup only when other users are not working in Open Dental", and that copying while in use "is possible, but only with certain software." Your backup facts sheet should say how your database is captured, and the restore test should open the restored database, not just count the files.
- Replication and standby copies. For an RPO measured in minutes, a second database kept in sync is the tool. Open Dental's manual says plainly that replication "requires that proper safeguards be put in place to monitor and repair replication problems" and that "having a setup and response plan is important." Replication is a plan component that needs its own monitoring, not a substitute for the nightly copy.
- Where the restore happens. A loaner server from the provider, a spare machine on the shelf, a virtual machine on another box in the building, or a hosted environment. Each has a different RTO and the plan should name which one applies to which scenario; the building-loss scenario can only use the last one.
What to verify before you sign a disaster recovery service
- The RTO and RPO the service supports for your practice software and imaging, written per system, and what they cost to shorten.
- The restore target: which machine or environment, how it is kept ready, and whether it has been used for a test on your account.
- The test schedule and the form of evidence you will receive (a dated record with the time taken and the data checked).
- Which copy is offline or immutable, how it is reached, and who holds those credentials.
- The backup method for your specific practice software and imaging, confirmed in writing with the software vendor if the vendor publishes requirements.
- A signed business associate agreement, and the retention period for your copies after you leave.
- Who writes and updates the plan document, and whether the practice receives its own copy.
- What happens on the morning a job fails: who sees it, by when, and what they do.
If the comparison widens to the whole IT relationship, how to choose a dental IT company gives the scorecard and the best dental IT companies list sets out who serves practices nationally. Recovery is one row on that scorecard, but it is the one that decides whether you open on Tuesday. If ransomware is the scenario that worries you most, how to choose a dental cybersecurity provider covers the security half of the same decision.
How often should the plan be tested, and what does each test prove?
Test the plan with five different tests, because each proves a different thing: the daily job check proves the copy ran, the monthly small restore proves files come back readable, the quarterly database restore proves the practice software opens on restored data, the yearly full rehearsal proves the RTO, and a twice-yearly tabletop proves people know their roles.
| Test | What it proves | How often | Who runs it | Pass evidence kept in the test log |
|---|---|---|---|---|
| Backup job check | the copy completed and covered what the facts sheet says | every business day | IT provider (you see the monthly summary) | the job report; failures and the fix noted |
| Single-file restore | a named document or x-ray comes back readable | monthly | IT provider, witnessed by the office manager | file name, date restored, opened by whom; see restoring a deleted patient file or x-ray |
| Database restore to an isolated machine | the practice software opens on the restored database and the last entries match the RPO | quarterly | IT provider | the time taken, the last chart entry visible, the scorecard in the 12-point checklist |
| Full recovery rehearsal | the RTO can actually be met on the restore target | yearly | IT provider with the office manager timing it | the timed run sheet below |
| Tabletop walk-through | people know who declares, who calls, what the front desk does | twice a year, 30 minutes | office manager | attendance, the gaps found, the plan edits made |
| Call-tree test | everyone can be reached inside the plan's wait time | twice a year, with the tabletop | office manager | who answered, how long it took |
NIST's guide defines the two exercise types in one line each: a tabletop "is discussion-based only and does not involve deploying equipment or other resources", while functional exercises let staff "validate their operational readiness for emergencies by performing their duties in a simulated operational environment." The full rehearsal is the functional one. Open Dental's manual describes the right place to do the database test: "verify the quality by restoring backups to a secure, off-site computer not connected to the office network", so that the live database is never overwritten by mistake.
How to run a full recovery rehearsal against your RTO
A rehearsal is a restore of the practice software and imaging to the restore target named in the plan, timed from a pretend declaration, on a day with no patients. The point is one number: minutes from "declared" to "a front-desk person opens tomorrow's schedule on the restored system and finds yesterday's last entry."
| Clock | Step | Who | Record |
|---|---|---|---|
| T+0 | Declare the pretend outage; start the log | office manager | time, scenario chosen |
| T+0 to T+15 | Run the call tree as written | office manager | who was reached, how |
| T+15 | IT provider confirms which copy is being used and its timestamp | IT provider | copy date and time (this is your real RPO) |
| T+15 onward | Restore to the named target, not to the production server | IT provider | start and finish time |
| at restore complete | Open the practice software; open an image from the last clinic day | office manager and clinical lead | last visible chart entry; image opens |
| at restore complete | One workstation points at the restored system and books a test appointment | front-desk lead | time, any error |
| T+end | Compare the finish time with the RTO; list every gap | all | met or missed by how much; gaps and owners |
| within a week | Edit the plan and the backup facts sheet | office manager with the IT provider | the new version date |
Two results matter. If the time beat the RTO, the plan is honest. If it did not, you have a choice to make on purpose: shorten the recovery (a readier restore target, image copies) or lengthen the RTO and rehearse paper mode for longer. Either is a plan; a missed number nobody wrote down is not.
How do you build the plan in 30 days?
Build the plan over four weeks: inventory and the RTO table in week one, roles, call tree and the backup facts sheet in week two, scenario runbooks and communication templates in week three, and a tabletop plus the first database restore test in week four, ending with the owner's signature and a review date.
- Week 1: inventory and numbers. The IT provider lists every system with where it runs and who owns the login. The owner, office manager and clinical lead hold the 30-minute meeting and fill the RTO and RPO table. Nine of the ten systems take minutes; the practice management database takes the discussion.
- Week 2: people and facts. Fill the roles matrix and the call tree with real names and two contact methods each. The IT provider writes the backup facts sheet, including the offline or immutable copy and where the recovery credentials are sealed. Decide the alternate workplace and write down what makes it usable.
- Week 3: scenarios and words. Write the five one-page runbooks from the matrix above, naming the recovery path for each at each hour. Draft the three communication templates. Collect the vendor and contract list with account IDs and the response terms each contract actually promises.
- Week 4: prove and sign. Run the 30-minute tabletop with the whole team and the call-tree test. Have the IT provider run a database restore to an isolated machine and record it. Fix what the two tests found. The owner signs the plan; the office manager puts the one-page version in the binder and sets the yearly rehearsal date.
Practices that are also replacing a server or moving software should fold this into the project; a migration is the one time everything is already inventoried. Opening a second location changes the alternate-workplace section and the restore order, and a practice being bought or sold will be asked for the signed plan and the test log, so keep both where they can be produced.
If the practice software is in the cloud, what is left for the plan to cover?
Most of it. A cloud practice system (Dentrix Ascend, Curve or Denticon) moves the database and its backups to the vendor, so the vendor's RTO and RPO replace yours for that one system. Everything else in the building still needs the plan: internet and its failover, sensor and scanner computers, locally kept images, scanned documents, phones, payments and the people.
The plan changes in four places. The practice-software row of the RTO table records the vendor's published targets and support route rather than your own. Scenario 5 becomes the most likely event and deserves the most rehearsal, including the vendor's offline or read-only mode if one exists. Scenario 4 matters more, because without internet nothing hosted is reachable, so the second internet path and the phone forwarding become the recovery plan. And the imaging row usually stays local: a sensor captures to a computer in the operatory, and that capture needs a copy and a sync path when the connection returns. Put plainly, a hosted practice system is only as available as the office's internet connection.
What to ask your IT provider about disaster recovery
About the targets
- Which RTO and RPO have you written down for our practice software and for imaging, and when were they last met in a test?
- Which of our systems would you restore first, and is that order written where our office manager can read it?
About the copies
- Which copy of our data could not be reached by someone who had control of our network, and how do you reach it?
- How is our practice software's database captured so the restored copy opens, and has the vendor confirmed the method?
About the restore
- Where would you restore our server to tomorrow morning, and when was that target last used for our account?
- If the building were not available, what would the first two hours look like?
About the evidence
- Show us the dated record of our last restore test, with what was restored and the minutes it took.
- What does the morning after a failed backup job look like on your side, and when do we hear about it?
About the document
- Who keeps our plan current, and do we hold a printed copy and the sealed recovery credentials ourselves?
Send the list in writing and keep the answers with the plan. A provider who answers every line with a date and a number is running a recovery service. Answers about the product rather than about your office are the sign that the plan has not been written yet, which is a reasonable moment to read about dental IT support as a whole and about the nine components of dental practice IT infrastructure the plan has to cover.
How Legend Networking can help
From its Dallas headquarters, Legend Networking has served over 3,500 practices since 2004, which makes 22+ years in dental IT. Legend supports the IT environment around four practice systems (Dentrix, Eaglesoft, Open Dental, Archy) and five imaging systems (DEXIS, Vatech, Carestream, Planmeca/Romexis, Dolphin Imaging). Monitoring runs 24/7, the engineers are in-house with no outsourced first-level triage, onsite dispatch is by appointment across served dental markets, and support is remote-first nationwide; the locations page lists the cities Legend serves.
Written by Ian Lynch (22 years in dental IT), who oversees Legend's NOC, onsite dispatch and client technology roadmaps. Google reviewers rate Legend 4.8 from 173 reviews. To see the recovery layer inside a wider scope, read a multi-location practice's onboarding that included backup and disaster recovery.
Frequently asked questions
In plain words, what is a disaster recovery plan for a dental office?
A disaster recovery plan for a dental office is a brief written document that says how the practice gets its systems and data back after a server failure, ransomware, fire or flood, a long power or internet loss, or a vendor outage. It sets a recovery time and a data-loss window for each system, names who declares and who does what, and schedules the tests that prove the backup restores.
What does the HIPAA Security Rule text list under the contingency plan standard?
The rule text at 45 CFR 164.308(a)(7) lists five implementation specifications: a data backup plan, a disaster recovery plan and an emergency mode operation plan, each marked "Required", and testing and revision procedures and an applications and data criticality analysis, each marked "Addressable". The page quotes the text; how it applies to a particular practice is a question for the practice's compliance adviser.
How do RTO and RPO change which backup a dental office needs?
The recovery time objective decides where the data must be restorable to: a nightly copy alone cannot meet a four-hour RTO if there is nothing to restore it onto. The recovery point objective decides how often the copy runs: a last-night RPO needs one nightly copy, a four-hour RPO needs several snapshots a day, and an RPO in minutes needs a continuously synchronised second database with its own monitoring.
What RTO do most dental practices choose for their practice management software?
For a single-location practice the realistic choice is between same-day and next-morning recovery for the practice management database, because the deciding hour is when the next day's schedule has to be cancelled. A surgical or busy multi-chair office may set four hours and pay for a readier restore target. The right number is the one set in the 30-minute meeting and then met in a timed rehearsal.
Who should be in a dental office call tree?
The call tree names the declaring authority (usually the owner, with the office manager as alternate), the IT provider's contract number and escalation contact, the clinical lead who runs paper mode, the front-desk lead who runs the schedule and phones, the practice-software, imaging, carrier and phone vendors, and the cyber insurer's hotline. Every entry has two contact methods and a wait time before the alternate is called.
How is a ransomware recovery plan different from a hardware-failure plan?
After a hardware failure the backups are trusted and the network is clean. After ransomware, every copy reachable from the network is suspect, so the plan names an offline or immutable copy, says how it is reached while the network is isolated, and sends the restore to a clean machine. It also adds roles: who contacts the insurer and counsel, and who keeps the incident log.
Do we still need a disaster recovery plan if our practice software is in the cloud?
Yes. With a cloud practice system the vendor's recovery targets cover the database, and you should hold them in writing. Everything in the building still needs your plan: internet and its failover, phones, the sensor and scanner computers, locally stored images and scanned documents, payments, and the people who run paper mode. The vendor-outage and internet-loss scenarios become the ones to rehearse most.
What is a tabletop exercise for a dental practice?
A tabletop exercise is a 30-minute discussion in which the office manager reads out a scenario, such as "the server will not start on a Monday morning", and each person says what they would do, in order, from the plan. No equipment is touched. It shows whether people know their roles, whether the call tree works, and which sections of the plan are missing or out of date.
How often should the recovery plan itself be rehearsed, not just the backup?
Rehearse the plan twice a year as a tabletop with a call-tree test, and once a year as a timed full recovery to the restore target named in the plan. Backup tests run on their own schedule: a daily job check, a monthly single-file restore and a quarterly database restore to an isolated machine. Rewrite the affected sections after every test and after every real event.
How many pages does the recovery plan need, and what goes on the one-page version?
Ten short sections, usually ten pages or fewer, plus a one-page front-desk version: the inventory, the RTO and RPO table, the declaring authority, the roles and call tree, the backup facts sheet, five one-page scenario runbooks, the alternate workplace, three communication templates, the vendor list and the test log. Long plans do not get read on the morning they are needed.
The bottom line
If the only document you could hand an underwriter today is a backup report, ask for a Free Dental IT Assessment: a review of your computers, software, network security, backups and performance, with written findings and prioritised recommendations, including whether the backup and recovery arrangement you have can meet the RTO and RPO you would choose. It is the fastest way to turn this page into your own plan, with dental IT support that owns the restore target and the tests.

