Hospital downtime is any period when clinical or business systems are unavailable or too slow to use, whether from a cyberattack, an IT failure, or a vendor outage. Headlines put its cost near $7,500 per minute, and no hospital study behind that number was found. The earliest per-minute figure located comes from a 2013 outage study of data centers in several industries [11][12]. What the sources do measure is narrower and more useful: ransomware attacks that took electronic systems down in 41.7% of cases, a 2024 survey in which 74% of responding hospitals reported direct patient care impact after the Change Healthcare attack, and self-reported recovery coverage of 78% in Censinet's 2025 benchmarking study [8][9][10]. HIPAA requires a contingency plan, CMS requires an emergency preparedness program, and a federal guide and HHS proposals say more about what the plan should hold [1][3][5][6]. This post traces the per-minute figure, reports what the research measures, sorts rules from guidance, and describes where Censinet's risk capabilities meet downtime planning [13][14].

What each source measures, and what it does not show
SourceKindWhat it measuresWhat it does not show
Emerson and Ponemon, 2013Industry study, trade republicationCost of unplanned data center outages across industries, 67 U.S. data centers [11]Hospital downtime, ransomware, or patient harm [11][12]
JAMA Health Forum, 2022Peer-reviewed cohort study374 ransomware attacks on health care delivery organizations, 2016 to 2021 [8]A dollar cost, or a hospital-only split in the abstract [8]
AHA survey, 2024Trade association surveyEffects on nearly 1,000 hospitals of the Change Healthcare attack [9]Minutes of downtime or a cost per minute [9]
HIPAA, 45 CFR 164.308(a)(7)RuleContingency plan duties for covered entities and business associates [1]A cost, or a recovery time target [1]
CMS, 42 CFR 482.15RuleEmergency preparedness program for participating hospitals [5]A cyber scenario named in the text read [5]
ONC SAFER Contingency PlanningFederal guidanceThirteen self-assessment practices for EHR unavailability [6]A requirement; it is a self-assessment tool [6]
KLAS and Censinet benchmarking, 2025Industry researchSelf-reported framework coverage, 69 organizations [10]Actual recovery times or outage outcomes [10]
HHS proposed rule, 2024ProposalChanges HHS would make to the Security Rule [3][4]A final requirement; the current rule remains in effect [3][4]

‍

Where the Per-Minute Figure Comes From

The Emerson Network Power and Ponemon Institute study released on December 4, 2013 reports that unplanned data center outages cost slightly more than $7,900 per minute, up 41% from about $5,600 in 2010 [11]. The cost study covered 67 U.S. data centers, and the release gives an average incident cost of approximately $690,200 over an average of 86 minutes [11]. It describes the figure as an average across industries and names telecom providers, e-commerce, defense contractors, financial institutions, hospitality, the public sector, transportation, and media. Healthcare is not among the named sectors [11]. The release says the cost covers direct, indirect, and opportunity costs, and it does not explain how the per-minute number was calculated [11].

A Becker's Hospital Review report from the same day says hospitals lose about $7,908 every minute their data center is down, which conflicts with the release's description of an all-industry average [11][12]. The full Ponemon report was not read for this post, so the conflict is unresolved. The closest source found for a per-minute figure is a 2013 study of data centers across industries, and it is not a study of hospitals. The $7,500 figure in the headline was not found in any source read. A data center outage is also a narrower event than hospital downtime, so the study does not measure care delays, diversions, or the effect of a multiday attack on a hospital.

The 2013 per-minute figure, as reported
ItemWhat the sources report [11][12]
PublishedDecember 4, 2013, Emerson Network Power and Ponemon Institute news release, as republished by Healthcare Facilities Today; Becker's Hospital Review report the same day
Per-minute averageSlightly more than $7,900 per minute across industries (the release); about $7,908 per minute for hospitals (Becker's)
Change from 2010Up 41% from about $5,600 per minute
Per incidentApproximately $690,200 on average, for an average incident of 86 minutes
Sample67 U.S. data centers in the cost study
Industries namedTelecom providers, e-commerce, defense contractors, financial institutions, hospitality, public sector, transportation, and media; healthcare is not named
Cost componentsDirect, indirect, and opportunity costs; the calculation of the per-minute number is not explained
ConflictThe release describes an all-industry average, and Becker's attributes the figure to hospitals; the full report was not read
The $7,500 figureNot found in any source read for this post

‍

What the Research Measures

Ransomware and Care Delivery

A cohort study in JAMA Health Forum by Hannah Neprash and colleagues analyzed 374 ransomware attacks on U.S. health care delivery organizations from 2016 to 2021, using the Tracking Healthcare Ransomware Events and Traits database [8]. The annual number of attacks more than doubled, from 43 in 2016 to 91 in 2021 [8]. Ransomware disrupted care delivery in 166 of 374 attacks studied, or 44.4%, and electronic system downtime was the most common disruption. Electronic system downtime occurred in 156 attacks (41.7%), canceled scheduled care in 38 (10.2%), and ambulance diversion in 16 (4.3%) [8]. The attacks exposed the personal health information of nearly 42 million patients, and the authors found that attacks increasingly affected large organizations with multiple facilities [8]. The study covers hospitals, clinics, and other delivery organizations, and the abstract does not split results by type [8]. It reports counts and care disruption and gives no dollar cost per minute.

One Vendor Outage: The AHA Survey

The American Hospital Association surveyed all U.S. hospitals after the Change Healthcare cyberattack, and nearly 1,000 responded between March 9 and March 12, 2024 [9]. The AHA page says Change Healthcare processes 15 billion health care transactions annually and touches 1 in every 3 patient records [9]. In the survey, 74% of responding hospitals reported direct patient care impact, 94% reported financial impact with more than half calling it significant or serious, and 83% reported impact on cash flow [9]. Also, 67% said switching clearinghouses is difficult or very difficult, 81% said workarounds were only somewhat successful, and another 11% had not found them successful [9]. The page describes most hospitals as implementing workarounds that are labor intensive and costly [9]. The percentages are given in graphics on the AHA page and the survey PDF was not read, so check them against the PDF. This was an outage at a third party, and the survey measures effects on care, cash, and workarounds, not minutes of downtime.

Measured effects of ransomware and of one vendor outage on hospitals
MeasureResult
JAMA Health Forum: attacks analyzed [8]374 ransomware attacks on U.S. health care delivery organizations, 2016 to 2021
JAMA Health Forum: annual attacks [8]More than doubled, from 43 in 2016 to 91 in 2021
JAMA Health Forum: care delivery disrupted [8]166 of 374 attacks (44.4%)
JAMA Health Forum: electronic system downtime [8]156 attacks (41.7%)
JAMA Health Forum: canceled scheduled care [8]38 attacks (10.2%)
JAMA Health Forum: ambulance diversion [8]16 attacks (4.3%)
JAMA Health Forum: patients exposed [8]Personal health information of nearly 42 million patients
AHA survey: respondents [9]Nearly 1,000 hospitals, responses collected March 9 to March 12, 2024
AHA survey: direct patient care impact [9]74% of hospitals
AHA survey: financial impact [9]94% of hospitals, with more than half reporting significant or serious impact
AHA survey: cash flow [9]83% of hospitals
AHA survey: switching clearinghouses [9]67% said it is difficult or very difficult
AHA survey: workarounds [9]81% found them only somewhat successful; an additional 11% had not found them successful

‍

Other Downtime Research

ASPR TRACIE, the Healthcare Emergency Preparedness Information Gateway, keeps a topic collection on electronic health record downtime procedures [7]. Its listing describes a 2014 survey finding that 96% of institutions had unexpected downtime in the previous three years, and a 2018 analysis of patient safety event reports in which 46% of 76 downtime-related reports involved no downtime procedure or one that was not followed [7]. These come from ASPR TRACIE's descriptions of the studies, and the original papers were not read.

What the Rules Require

The HIPAA Contingency Plan

The Security Rule's contingency plan standard at 45 CFR 164.308(a)(7) requires a covered entity or business associate to establish policies and procedures for responding to an emergency or other occurrence, such as fire, vandalism, system failure, or natural disaster, that damages systems containing electronic protected health information [1]. It has five implementation specifications. Three HIPAA contingency plan specifications are required, and the two that cover testing and criticality analysis are addressable. The required three are the data backup plan, the disaster recovery plan, and the emergency mode operation plan [1]. An addressable specification calls for the entity to assess whether it is reasonable and appropriate, implement it if so, and otherwise document why and adopt an equivalent measure, under 45 CFR 164.306(d)(3) [15]. The section's source note shows its last amendment in 2013 [1].

HHS's Office for Civil Rights wrote about the standard in a March 2018 cyber newsletter. It says malicious software such as ransomware can render an organization's data unreadable or unusable, and that restoring from backups can be the only way to recover the data [2]. It tells entities to set a testing schedule for the plan, review it regularly and when technical, operational, environmental, or personnel changes occur, and define response periods such as the first hour, day, or week [2].

HIPAA contingency plan specifications and HHS's proposed changes
SpecificationStatus todayWhat it covers [1]Related HHS proposal (aligned by this post) [3]
(A) Data backup planRequiredProcedures to create and maintain retrievable exact copies of electronic protected health informationSeparate technical controls for backup and recovery
(B) Disaster recovery planRequiredProcedures to restore any loss of dataWritten procedures to restore the loss of certain relevant electronic information systems and data within 72 hours
(C) Emergency mode operation planRequiredProcedures to enable continuation of critical business processes while operating in emergency modeBusiness associates notify covered entities no later than 24 hours after activating a contingency plan
(D) Testing and revision proceduresAddressableProcedures for periodic testing and revision of contingency plansWritten incident response plans with procedures for testing and revising them; no required or addressable distinction
(E) Applications and data criticality analysisAddressableAssess the relative criticality of specific applications and data in support of other contingency plan componentsAnalysis of the relative criticality of relevant electronic information systems, used to set restoration priority

‍

HHS's Proposed Changes

HHS issued a notice of proposed rulemaking on December 27, 2024 [3][4]. It would remove the distinction between required and addressable specifications and make all of them required, with specific, limited exceptions [3]. It would require written procedures to restore the loss of certain relevant electronic information systems and data within 72 hours, an analysis of the relative criticality of those systems to set restoration priority, separate technical controls for backup and recovery, and written incident response plans with procedures for testing and revising them [3]. Business associates would have to notify covered entities no later than 24 hours after activating a contingency plan [3]. These are proposals, and HHS says the current Security Rule remains in effect during the rulemaking [3][4]. Confirm the status of the rulemaking on the day this post is published.

CMS Emergency Preparedness

A hospital that participates in Medicare and Medicaid must maintain an emergency preparedness program under 42 CFR 482.15, built on an all-hazards approach [5]. The emergency plan rests on a documented facility-based and community-based risk assessment, and it covers continuity of operations, including delegations of authority and succession plans [5]. The policies include a system of medical documentation that preserves patient information, and the communication plan covers primary and alternate means of reaching staff and emergency agencies [5]. The program must be reviewed and updated at least every 2 years, training is required at least every 2 years, and the hospital must conduct exercises to test the plan at least twice per year [5]. The text read for this post does not mention cyberattacks or information systems. In this post's reading, a hospital reaches a cyber event through that all-hazards risk assessment, and the rule does not say so itself.

CMS emergency preparedness elements that bear on downtime, 42 CFR 482.15
ProvisionWhat it requires [5]
ProgramA comprehensive emergency preparedness program using an all-hazards approach
(a) Emergency planBased on a documented facility-based and community-based risk assessment; covers continuity of operations, including delegations of authority and succession plans; reviewed and updated at least every 2 years
(b) Medical documentationA system of medical documentation that preserves patient information, protects confidentiality, and keeps records available
(c) Communication planPrimary and alternate means of communicating with staff and emergency agencies; sharing patient information with other providers for continuity of care
(d)(1) TrainingInitial training for staff, contracted individuals, and volunteers, and training at least every 2 years
(d)(2) TestingExercises to test the plan at least twice per year: a community-based full-scale exercise, or a facility-based functional exercise when one is not accessible, plus a second exercise that can be a tabletop exercise; analysis and documentation of each
Cyberattacks and information systemsNot mentioned in the text read for this post

‍

What Federal Guidance Recommends

The SAFER Guides from the Office of the National Coordinator for Health IT include a Contingency Planning self-assessment on planned or unplanned EHR unavailability. The edition read for this post carries an August 2024 footer, and it lists 13 recommended practices in three domains: Safe Health IT, Using Health IT Safely, and Monitoring Safety [6]. It is a self-assessment tool and not a rule, and it cites no study statistics [6]. Its practices include disaster recovery plans reviewed at least annually, paper forms to replace key EHR functions during downtimes, regular backups that are tested, including a read-only backup EHR, staff training and testing on downtime and recovery procedures, and downtime communication that is independent of the computing infrastructure supporting the EHR [6]. Its thresholds include paper forms sufficient for at least 8 hours of patient care per unit, unannounced downtime drills at least once a year, and a read-only backup updated, optimally, at least hourly [6].

ONC SAFER Contingency Planning practices that bear on downtime
PracticeWhat the guide says [6]
1.1 Disaster recovery plansPlans in place and reviewed at least annually, covering hardware duplication, network redundancy, and data replication
1.2 Electric generatorA generator and fuel to support the EHR during an extended power outage
1.3 Paper formsPaper forms are available to replace key EHR functions during downtimes
1.4 Data backupPatient data and configuration settings regularly backed up and tested, including a read-only backup EHR
1.5 Patient identificationPolicies that ensure accurate patient identification when preparing for, during, and after downtimes
2.1 TrainingStaff are trained and tested on downtime and recovery procedures
2.2 CommunicationDowntime communication is independent of the computing infrastructure that supports the EHR
2.3 Written policiesDowntime and recovery processes, including reactivation and how downtime data is entered into the EHR
Thresholds namedPaper forms sufficient for at least 8 hours of patient care per unit; unannounced downtime drills at least once a year; a read-only backup updated, optimally, at least hourly

‍

What Censinet's 2025 Benchmark Shows

Censinet's 2025 Healthcare Cybersecurity Benchmarking Study, produced with KLAS Research and the American Hospital Association, was published in April 2025 and draws on 69 healthcare and payer organizations surveyed from September to December 2024 [10]. Respondents reported average coverage of 85% for the NIST Cybersecurity Framework 2.0 Respond function and 78% for Recover, against an average of 72% across all six functions [10]. Within Recover, incident recovery communication scored 83% and incident recovery plan execution scored 76%, against 73% aggregate coverage for all other categories [10]. The report says many organizations have more mature processes for immediate response than for long-term recovery and business continuity, and it calls recovery plan execution an area for improvement [10]. The 2025 benchmarking study found higher self-reported coverage for responding to an incident than for restoring operations afterward. Supply chain risk management (52%) and asset management (53%) had the lowest coverage, and the report links asset management to an inventory of third-party hardware, software, data, and storage [10]. These are self-reported averages from 69 organizations, so they describe coverage of framework practices and not recovery times or outages.

Self-reported NIST CSF 2.0 coverage, 2025 benchmarking study (n=69)
MeasureAverage coverage [10]
Respond function85%
Recover function78%
Detect function74%
Protect function69%
Govern function64%
Identify function64%
Average across the six functions72%
Incident recovery communication (Recover)83%
Incident recovery plan execution (Recover)76%
All other NIST CSF 2.0 categories, aggregate73%
Asset management (Identify)53%
Supply chain risk management (Govern)52%

‍

What Hospital Teams Can Do With This

Start from three anchors: a documented contingency plan under HIPAA, an emergency program that tests the plan twice a year under CMS, and downtime procedures that clinicians can run on paper [1][5][6]. Beyond those, each step is a hospital choice with no rule behind it. List the systems each unit cannot work without, set a recovery target for each, record the minutes of downtime the hospital logs and the cost lines it can measure, run an unannounced drill with vendor systems switched off, and add each critical vendor product to an inventory with a named owner. In this post's reading, a hospital's own record of downtime minutes and costs will serve a budget request better than any published average, because the published averages measure other settings.

Where Censinet Fits

Censinet's Third-Party Risk page describes routing assessment findings into a Risk Register, where ownership, timelines, and prioritized remediation are tracked through resolution [13]. It also describes breach and ransomware notifications focused on the customer's inventory, a workflow to review, triage, and track those alerts to resolution, and a centralized vendor and product inventory with live profiles that hold assessment history and current risk status [13]. Censinet's Enterprise Risk page describes a Risk Register as a system of record with defined ownership, status tracking, and longitudinal history, and benchmarking against peer organizations [14]. Neither page uses the word downtime, so applying them to downtime is this post's reading. In that reading, the inventory is where a team records which vendor products each critical clinical system depends on, which matters because 67% of hospitals in the AHA survey found switching clearinghouses difficult or very difficult [9][13].

Censinet capabilities and where downtime planning fits
CapabilityWhat the page describesWhere downtime planning fits (this post's reading)
Risk Register routing [13]Assessment findings go to the Risk Register, where ownership, timelines, and prioritized remediation are tracked through resolutionA finding about a vendor product behind a critical system gets an owner and a due date
Vendor and product inventory [13]Centralized inventory with live profiles that hold assessment history and current risk statusRecord which vendor products each critical clinical system depends on
Breach and ransomware notifications [13]Notifications focused on the customer's inventory, not the broader marketSpot a vendor event that touches a product the hospital runs
Incident review workflow [13]Review, triage, and track alerts through to resolutionKeep a record of each vendor event and what was done
System of record [14]Defined ownership, status tracking, and longitudinal history for findings and corrective actionsHold downtime lessons and fixes in one place across departments
Peer benchmarking [14]Compare program maturity, assessment coverage, and risk findings against peer organizationsCompare recovery planning coverage with peers, as in the 2025 benchmarking study

‍

Conclusion

Key points

No hospital study behind a $7,500 per-minute downtime cost was found, and the closest source is a 2013 study of data centers across industries. What the research measures is care disruption: 44.4% of the ransomware attacks in a JAMA Health Forum study disrupted care delivery, and 74% of hospitals responding to the AHA survey reported direct patient care impact after one vendor attack. HIPAA requires backup, disaster recovery, and emergency mode operation plans, and HHS has proposed a 72-hour restoration requirement that is not yet in force. CMS requires exercises twice a year, and Censinet's 2025 benchmark shows lower self-reported coverage for recovery than for response.

Next steps

Record the hospital's own downtime minutes and costs, document the HIPAA decisions on testing and criticality analysis, add downtime scenarios to the CMS exercises, inventory critical vendor products with owners, and check the status of HHS's proposed Security Rule changes.

References

  1. Cornell Legal Information Institute. 45 CFR 164.308, Administrative safeguards, paragraph (a)(7) (no currency date shown; source note shows 68 FR 8376, February 20, 2003, as amended at 78 FR 5694, January 25, 2013). https://www.law.cornell.edu/cfr/text/45/164.308‍
  2. HHS Office for Civil Rights. Cyber Newsletter, "Plan A… B… Contingency Plan!" (March 2018). https://www.hhs.gov/sites/default/files/march-2018-ocr-cyber-newsletter-contingency-planning.pdf‍
  3. HHS Office for Civil Rights. HIPAA Security Rule Notice of Proposed Rulemaking to Strengthen Cybersecurity for Electronic Protected Health Information, Fact Sheet (NPRM issued December 27, 2024; content last reviewed December 27, 2024). https://www.hhs.gov/hipaa/for-professionals/security/hipaa-security-rule-nprm/factsheet‍
  4. HHS Office for Civil Rights. HIPAA Security Rule NPRM (content last reviewed December 27, 2024). https://www.hhs.gov/hipaa/for-professionals/security/hipaa-security-rule-nprm/index.html‍
  5. Cornell Legal Information Institute. 42 CFR 482.15, Condition of participation: Emergency preparedness (no currency date shown; amendment history through 84 FR 51817, September 30, 2019). https://www.law.cornell.edu/cfr/text/42/482.15‍
  6. Office of the National Coordinator for Health IT. SAFER Guides, Self-Assessment: Contingency Planning (August 2024 footer). https://www.healthit.gov/sites/default/files/playbook/pdf/2-contingency-planning-final.pdf‍
  7. ASPR TRACIE. Electronic Health Records and Downtime Procedures, topic collection (cover dated December 2, 2025; the text says it was refreshed in 2023). https://asprtracie.hhs.gov/technical-resources/66/electronic-health-records-and-downtime-procedures/‍
  8. Neprash HT, McGlave CC, Cross DA, et al. Trends in Ransomware Attacks on US Hospitals, Clinics, and Other Health Care Delivery Organizations, 2016-2021. JAMA Health Forum 2022;3(12):e224873 (published December 2, 2022; read from the abstract record and the journal's news release). https://pesquisa.bvsalud.org/bvsms/resource/pt/mdl-36580326 and https://www.eurekalert.org/news-releases/975364‍
  9. American Hospital Association. AHA Survey: Change Healthcare Cyberattack Significantly Disrupts Patient Care, Hospitals' Finances (March 15, 2024). https://www.aha.org/2024-03-15-aha-survey-change-healthcare-cyberattack-significantly-disrupts-patient-care-hospitals-finances‍
  10. KLAS Research and Censinet. Healthcare Cybersecurity Benchmarking Study 2025, Executive Summary (April 2025; figures read from the study report). https://censinet.com/2025-healthcare-cybersecurity-benchmarking-study-executive-summary‍
  11. Emerson Network Power and Ponemon Institute. Study finds significant increase in cost of unplanned data center outages since 2010, news release (December 4, 2013), as republished by Healthcare Facilities Today. https://www.healthcarefacilitiestoday.com/posts/Emerson-Network-Power-Ponemon-Institute-Study-finds-significant-increase-in-cost-of-unplanned-data-center-outages-since-2010--3402‍
  12. Becker's Hospital Review. Report: Data Center Outages Cost $8k Per Minute (December 4, 2013). https://www.beckershospitalreview.com/healthcare-information-technology/report-data-center-outages-cost-8k-per-minute‍
  13. Censinet. Third-Party Risk. https://www.censinet.com/capabilities/third-party-risk‍
  14. Censinet. Enterprise Risk. https://www.censinet.com/capabilities/enterprise-risk‍
  15. Cornell Legal Information Institute. 45 CFR 164.306, Security standards, general rules (no date shown; amendment history through 78 FR 5693, January 25, 2013). https://www.law.cornell.edu/cfr/text/45/164.306