HIPAA’s 2019 penalty tiers made culpability a record to keep
HHS’s April 2019 enforcement discretion cut the annual HIPAA penalty cap for violations an organization could not have known about from $1.5 million to $25,000. What a penalty could cost now turned on what it could prove about knowing and fixing.
Looking Back revisits past weeks one at a time, up to the most recent, and asks what it meant for the banks, health systems and public agencies we serve. The series was written in 2026, with the benefit of hindsight, and each article says which week it looks back on.
Two-minute summaryThe dek and the first sentence of each section, quoted from the article
HHS’s April 2019 enforcement discretion cut the annual HIPAA penalty cap for violations an organization could not have known about from $1.5 million to $25,000. What a penalty could cost now turned on what it could prove about knowing and fixing.
- The week On April 30, 2019, the Department of Health and Human Services published a notice in the Federal Register titled “Notification of Enforcement Discretion Regarding HIPAA Civil Money Penalties.”
- Proving you could not have known: diligence as the first tier’s evidence The lowest tier is not available to an organization simply because it did not know.
- Proving reasonable cause: decisions and their reasons The second tier covers violations due to reasonable cause and not willful neglect.
- Proving when you knew and when you fixed it: the 30-day clock The two willful neglect tiers are separated by correction.
- Avoiding the top tier: a known problem left in place The fourth tier, willful neglect not corrected in time, kept the $1.5 million annual limit under the notice, and its minimum per-violation amount was the full $50,000.
- How the evidence gets asked for The tiers matter only when HHS asks an organization to account for a violation.
- Running the evidence as a system The four tiers describe what an organization may need to prove.
- Where the argument is weakest The strongest objection is that the notice was enforcement discretion, not law.
- How it differs by kind of organization Large systems have the resources to keep the records this argument requires and the complexity that makes it hard.
- With hindsight The rulemaking HHS said it expected did not arrive.
The week
On April 30, 2019, the Department of Health and Human Services published a notice in the Federal Register titled “Notification of Enforcement Discretion Regarding HIPAA Civil Money Penalties.” It was signed on April 23 by the director of the HHS Office for Civil Rights, which enforces the HIPAA Privacy, Security and Breach Notification Rules, and filed for public inspection on April 26. In its own words, it was effective indefinitely.
The notice changed how much HHS would collect, at most, in a calendar year for HIPAA violations. Its background explained why. The original 1996 HIPAA statute allowed a civil money penalty of up to $100 per violation, capped at $25,000 a year for all violations of an identical requirement. In 2009 the HITECH Act strengthened enforcement by creating four categories of violation with rising penalty tiers based on culpability: the person did not know, and with reasonable diligence would not have known, of the violation; the violation was due to reasonable cause and not willful neglect; the violation was due to willful neglect and was corrected in time; and the violation was due to willful neglect and was not corrected in time.
The statute attached a per-violation amount and an annual limit to each tier. The notice set them out: $100 per violation with a $25,000 annual limit; $1,000 with $100,000; $10,000 with $250,000; and $50,000 with $1.5 million. HHS’s 2009 interim final rule, adopted in a 2013 final rule without change to the penalty tiers and annual limits, had read the statute as conflicting and applied the highest annual limit, $1.5 million, to all four tiers. Commenters, as the 2013 final rule recorded, had argued that a single cap ignored the lower limits Congress set. HHS had kept its reading.
In April 2019 HHS changed its mind. The notice said that, after further review by the HHS Office of the General Counsel, the department had determined that the better reading of the HITECH Act applied separate annual limits: $25,000 for no knowledge, $100,000 for reasonable cause, $250,000 for corrected willful neglect and $1.5 million for uncorrected willful neglect. As a matter of enforcement discretion, all HIPAA enforcement actions would be governed by those interim tiers, adjusted for inflation, until further notice. The minimum and maximum amounts per violation did not change.
The notice was explicit about its limits. It said it created no legal obligations and no legal rights. It did not amend the regulation. HHS said it expected to engage in future rulemaking to revise the penalty tiers in the regulation to better reflect the text of the HITECH Act. Until then, the codified rule at 45 CFR 160.404 would continue to show a single $1.5 million annual limit for every tier, and the notice would govern how HHS actually applied it.
The change applied to covered entities, meaning health plans, health care clearinghouses and health care providers that conduct certain transactions electronically, and to their business associates, which the regulation names alongside covered entities in its penalty provisions. It did not change what the Privacy, Security or Breach Notification Rules require. It changed what failing to meet them could cost, and it tied that cost more tightly to how culpable the organization was.
The headline effect of the April 2019 notice was lower maximum penalties for HIPAA violations. For an organization that had not known of a violation and could not have known with reasonable diligence, the annual cap for violations of an identical requirement fell from $1.5 million to $25,000. For reasonable cause, it fell to $100,000. The notice could be read as relief, and in a narrow sense it was.
For the people who run a health system’s privacy, security and compliance programs, the notice did something more useful than lower the ceiling. It made the ceiling depend on a finding about the organization’s state of mind and conduct, and that finding would rest on evidence. Under the 2013 rule, a hospital that had done everything reasonable and a hospital that had ignored a known problem faced the same $1.5 million annual limit per requirement. After April 2019, the gap between them, in HHS’s application of its own rule, was the gap between $25,000 and $1.5 million. The difference would be settled by what each could show.
Our argument is that the notice turned culpability into a record a covered entity or business associate keeps. The four tiers are the four parts of that record, and they do not overlap: an organization’s exposure for any violation lands in exactly one. First, evidence that it did not and could not have known, which is evidence of reasonable diligence. Second, evidence that a known shortfall had reasonable cause, which is evidence of decisions and their reasons. Third, evidence of when it learned of a problem and when it fixed it, which decides whether willful neglect was corrected in time. Fourth, evidence that it did not leave a known problem in place, because uncorrected willful neglect kept the full $1.5 million limit. We take each in turn, then test the argument and look at how it applied to different kinds of organization.
Two points of scope. The notice concerned civil money penalties that HHS imposes; it did not change the requirements of the HIPAA rules. And it was an exercise of enforcement discretion, not a rule, so it created no rights an organization could enforce against HHS. Organizations should take advice from counsel on how the tiers apply to any specific matter.
Proving you could not have known: diligence as the first tier’s evidence
The test is reasonable diligence, not ignorance
The lowest tier is not available to an organization simply because it did not know. The statute and the regulation describe a person who did not know and, by exercising reasonable diligence, would not have known of the violation. Ignorance that diligence would have cured does not qualify. That turns the first tier into a question about the organization’s monitoring: what did it do that would have surfaced the problem, and why did that not surface it?
The evidence that answers that question is the evidence of a working security and privacy program. An accurate and thorough risk analysis, which the Security Rule requires, kept up to date, is the starting point. Records of the risk management decisions that followed it, of audit log review, of access reviews and of workforce training show that the organization was looking. Business associate agreements and the steps taken to oversee business associates show that it looked beyond its own walls.
The gap between a policy and a record
A health organization can have policies that say all the right things. A policy that says access logs are reviewed is not evidence that they were. For the first tier, what matters is the record that the activity happened: the review report, the ticket, the sign-off, the date. An organization that cannot produce those records may find that reasonable diligence is hard to establish, because it cannot show what it did.
We think the practical test is simple. For each safeguard the organization relies on, could it produce, within days, evidence that the safeguard operated in the period in question? If the answer is no for a safeguard that would have caught the violation, the organization should expect argument about whether reasonable diligence was exercised.
Diligence has to cover the systems people forget
A violation that no one knew about may well sit in a place no one was watching. We would look first at systems outside the central electronic health record: departmental applications, research data stores, legacy systems kept for records retention, shared drives, and vendor-hosted services that hold protected health information. A risk analysis that left those out would be a weak foundation for a claim of diligence about a violation found in one of them.
Knowledge of one person can be knowledge of the organization
The Breach Notification Rule offers a useful parallel. It treats a breach as discovered on the first day it is known to the covered entity or, by exercising reasonable diligence, would have been known to it, and it deems the covered entity to know of a breach if any workforce member or agent other than the person committing it knows or should have known. Whatever the exact reach of that rule in a penalty case, it shows how HHS thinks about organizational knowledge. A help desk technician who noticed something odd, or a vendor’s engineer who saw a misconfiguration, may be where the organization’s knowledge begins. Diligence includes the routes by which those people raise what they see.
Counting violations
The annual limits apply per calendar year to all violations of an identical requirement or prohibition. An incident can involve more than one requirement, for example a failure to conduct a risk analysis, a failure to implement access controls and an impermissible disclosure. Each can carry its own annual limit. The lower first-tier cap helps only for the requirements where the organization can show the first-tier facts. An organization that met the diligence test for one requirement and not another could see its exposure split across tiers.
Proving reasonable cause: decisions and their reasons
The middle ground
The second tier covers violations due to reasonable cause and not willful neglect. It is the space between an organization that could not have known and one that was indifferent. An organization in this tier knew, or should have known, of the problem, but it was not indifferent to the obligation. Its annual limit under the notice was $100,000.
The regulation defines reasonable cause as an act or omission in which the organization knew, or by exercising reasonable diligence would have known, that it violated a provision, but did not act with willful neglect. The question is therefore whether the organization was indifferent, and the evidence that answers it is the record of how the organization made its decisions. A documented risk assessment that identified the gap, a remediation plan with a realistic timeline, a budget request, a compensating control put in place in the meantime, and an approval by an accountable executive all show an organization that saw a problem and acted on it in a considered way. The same gap without any of that record looks much more like neglect.
Risk acceptance has to be explicit
The Security Rule lets covered entities and business associates choose security measures, taking into account their size, complexity and capabilities, their infrastructure, the costs of security measures and the probability and criticality of risks. For addressable implementation specifications, it lets them decide that a specification is not reasonable and appropriate, provided they document why and implement an equivalent alternative measure where reasonable and appropriate. That flexibility is only as good as the documentation behind it. An organization that decided not to encrypt a class of devices, for instance, needs the written analysis and the alternative safeguards it relied on. Without them, a regulator has little reason to see the decision as reasonable.
Resource constraints and reasonable cause
Smaller providers and rural hospitals face real limits on money and staff. A plan that matched those limits, documented and followed, can be evidence of reasonable cause. A plan that existed only in a slide deck is not. The difference again lies in the record: dates, owners, progress and the compensating steps taken while the larger fix waited.
Vendor delays
Some gaps persist because a vendor has not delivered a fix, a medical device cannot be patched, or a contract cannot be changed until renewal. We think those situations are most likely to support reasonable cause where the organization documented the vendor dependency, pressed for a fix and protected the system in the meantime, for example by segmenting it from the wider network. An unpatchable system with no documented mitigation reads differently from one with a record of each step.
The record of an addressable decision
For each addressable implementation specification, the Security Rule requires an assessment of whether it is a reasonable and appropriate safeguard in the organization’s environment, and, if the organization does not implement it, documentation of why and an equivalent alternative measure where reasonable and appropriate. Those documents are exactly the kind of record that supports reasonable cause. We would keep them current, tie each one to the risk analysis it relied on, and review them when the environment changes, such as when a new remote access method, a new cloud service or a new class of device arrives.
Proving when you knew and when you fixed it: the 30-day clock
Two dates decide the tier
The two willful neglect tiers are separated by correction. Under the regulation, willful neglect is treated as corrected if it was corrected during the 30-day period beginning on the first date the covered entity or business associate knew, or by exercising reasonable diligence would have known, that the violation occurred. Under the notice, corrected willful neglect carried a $250,000 annual limit; uncorrected willful neglect kept $1.5 million.
That makes two dates central. The first is the date the organization knew or should have known. The second is the date it corrected the violation. Both need evidence. An organization that cannot show when it learned of a problem leaves the first date open to argument, and the “should have known” standard can move that date earlier than the organization would like. An organization that cannot show when the fix was complete leaves the second date open too.
Building the clock into incident handling
The practical step is to treat the date of discovery as a recorded fact in every privacy and security incident, not a detail filled in later. Incident tickets should capture when the issue was first reported, by whom (by role), when it was escalated to privacy or security, and when each corrective action was completed and verified. Those records already help with breach notification timelines; after April 2019 they also marked a wider gap in penalty exposure, because correction in time now separated a $250,000 annual limit from a $1.5 million one.
What counts as correction
Correction should mean the violation stopped and the cause was addressed, not that a plan was approved. If a misconfigured system was exposing records, correction is the configuration change and the verification that it worked. If a policy failure caused repeated impermissible disclosures, correction is the changed process and evidence it is being followed. We would document both the fix and its verification, with dates, so the organization can show that correction happened inside the window.
When 30 days is not enough
Some problems cannot be fully fixed in 30 days. The regulation’s affirmative defense allows for correction within such additional period as HHS determines appropriate, but an organization cannot count on that judgment in advance. A system replacement or a network redesign takes longer. In those cases we think the organization’s best position is to stop the immediate violation within the window through interim measures, such as disabling access, isolating a system or suspending a process, and to document the longer remediation separately. That record shows correction of the violation even while the underlying improvement continues.
One discovery date, two clocks
The discovery date does double duty. For a breach of unsecured protected health information, the Breach Notification Rule requires notice to individuals without unreasonable delay and in no case later than 60 calendar days after discovery. For any underlying violation found to be willful neglect, the penalty regulation’s 30-day correction period runs from the date the organization knew or should have known. An organization that fixes the discovery date carefully and consistently serves both purposes; one that records different dates in different places invites questions about both.
Avoiding the top tier: a known problem left in place
The limit that did not move
The fourth tier, willful neglect not corrected in time, kept the $1.5 million annual limit under the notice, and its minimum per-violation amount was the full $50,000. The notice made the top tier relatively more expensive by lowering the others. The distance between doing the work and not doing it grew.
Where uncorrected neglect comes from
Willful neglect is not a matter of intent to harm. The regulation defines it as conscious, intentional failure or reckless indifference to the obligation to comply. In a large organization, that can arise from a known finding that sits unresolved in an audit report, a risk analysis that was done once and never updated, or a security exception that was granted and never revisited. Each of those leaves a record showing the organization knew. The defense is a record showing it acted.
Tracking findings to closure
We think every finding from an internal audit, an external assessment, a risk analysis or an incident should go into one register with an owner, a due date and a status, and that the register should be reviewed by a senior committee on a schedule. Overdue items should require a documented decision, either a new date with reasons or a risk acceptance by an accountable executive. That register is the evidence that separates an organization working through known problems from one that is indifferent to them.
Repeat findings weigh twice
Among the factors HHS considers in setting a penalty amount is the organization’s history of prior compliance, including whether a current violation is the same as or similar to previous indications of noncompliance and whether, and to what extent, the organization tried to correct them. A finding that recurs is therefore costly in two ways. It is evidence that the organization knew of the problem, which bears on the tier, and it is an aggravating factor in setting the amount within the tier. Closing findings properly, and checking later that the fix held, protects against both.
Board and executive oversight
Because the top tier turns on indifference, the attention of senior leadership is part of the evidence. Minutes showing that leadership received reports on privacy and security risk, asked questions and approved resources support the case that the organization took its obligations seriously. Silence at the top does the opposite. The notice did not mention governance, but the culpability tiers made it relevant.
How the evidence gets asked for
The tiers matter only when HHS asks an organization to account for a violation. The procedural rules in the same part of the regulation show when that happens and how much room an organization has to present its evidence. Read alongside the notice, they make the case for keeping evidence ready rather than assembling it under pressure.
Three routes in
Under the regulation, any person who believes a covered entity or business associate is not complying may file a complaint with HHS, generally within 180 days of when the complainant knew or should have known of the act or omission. HHS may also conduct compliance reviews on its own initiative. And a covered entity that discovers a breach of unsecured protected health information must notify HHS: for breaches involving 500 or more individuals, at the same time as it notifies the individuals, and for smaller breaches through a log it keeps and reports to HHS no later than 60 days after the end of each calendar year. Each route can lead to an investigation, and each starts the evidence request at a different point in the organization’s own timeline.
Willful neglect changes the procedure as well as the cap
The regulation says HHS will investigate a complaint, and will conduct a compliance review, when a preliminary review of the facts indicates a possible violation due to willful neglect. In other cases it may do so. So the same culpability question that sets the annual limit can also decide whether an investigation happens at all. An organization whose early response to a complaint or breach report shows a working program, with records to match, gives HHS less reason to see possible willful neglect at the preliminary stage.
A short window for mitigating evidence
If a matter is not resolved by informal means, the regulation requires HHS to tell the organization and give it an opportunity to submit written evidence of mitigating factors or affirmative defenses, within 30 days of receiving that notice. Thirty days is not long to reconstruct a year of security operations. The factors HHS considers include the nature and extent of the violation and the harm, the organization’s history of prior compliance, including whether it corrected previous indications of noncompliance and how it responded to technical assistance and prior complaints, its financial condition and other matters as justice may require. Each of those is easier to show from records kept as a matter of course.
Timely correction can remove the penalty altogether
The regulation contains an affirmative defense that the notice did not change and that sharpens the argument. For violations on or after February 18, 2009, HHS may not impose a civil money penalty if the organization establishes, to HHS’s satisfaction, that the violation was not due to willful neglect and was corrected within the 30-day period beginning when it knew or should have known of it, or within such additional period as HHS determines appropriate. For a violation in the first or second tier, then, timely correction with proof can mean no penalty at all. The discovery date and the correction date matter for every tier, including the two that have nothing to do with willful neglect.
How long the exposure lasts
The regulation also limits how long HHS has to act: no penalty action may be entertained unless HHS commences it within six years from the date of the occurrence of the violation. That period lines up with the Security Rule’s six-year retention requirement for its documentation. We read the two together as a planning horizon: the records that would decide a tier should be kept at least as long as the organization could be asked to answer for the violation they concern.
Running the evidence as a system
The four tiers describe what an organization may need to prove. The work of proving it falls on whoever keeps the records, and in a health system those records are spread across security tools, ticketing systems, compliance software, email and shared drives. The notice made it worth treating that evidence as a system with an owner, rather than as something assembled after an investigation begins.
Retention that matches the exposure
The Security Rule requires covered entities and business associates to retain the documentation it requires for six years from the date of its creation or the date when it last was in effect, whichever is later. That rule covers policies, procedures and required records of actions, activities and assessments. It also sets a sensible floor for the operational evidence that supports the tiers: risk analyses, review records, incident tickets and remediation logs. An organization that deletes its ticket history after a short period may lose the dates that decide whether willful neglect was corrected in time.
Evidence for each requirement, not each incident
Incidents are how violations come to light, but the annual limits apply requirement by requirement. An evidence system organized only by incident will struggle to answer the question HHS will ask about each requirement in play: what did the organization do, over the relevant period, to meet this one? We would index the evidence both ways. Each incident record should note which requirements it touched, and each requirement, such as risk analysis, access control, audit controls, training or business associate agreements, should have its own trail of activity over time. When a single incident raises questions about several requirements, the organization can then present a separate, coherent account for each, which is how the tiers will be applied.
Consistent dates and times
Evidence of timing is only as good as the clocks behind it. Security tools, ticketing systems and email each stamp their own times. When an investigation compares them, inconsistencies become arguments. Synchronized time sources, a documented time zone convention and a single field in the incident record for the discovery date reduce that risk. These are small engineering decisions with direct effects on how a regulator reads the record.
Who owns the record
Privacy officers, security officers, compliance and internal audit each hold part of the evidence. We would name one owner for the evidence system as a whole, placed in compliance or the security office, with a documented map of where each type of record lives, how long it is kept and who can retrieve it. When OCR sends a data request, the owner should be able to produce a coherent set of records for the period in question without a scramble.
Counsel and privilege
Some investigations of incidents are run at the direction of counsel, and organizations may seek to protect certain analyses. That is a legal judgment for counsel. Operationally, we would keep the factual records of what happened and when, such as tickets, logs and change records, in the ordinary course, because those are the records that establish discovery and correction dates. Organizations should agree with counsel in advance how investigative reports and the underlying operational records will be handled, so that the facts the tiers depend on remain available.
Testing the system before it is needed
An evidence system that has never been tested may fail when it matters. We would run an exercise once a year: pick a past incident or a sample control, and ask the owner to produce, within a set time, the records showing what the organization knew, when, what it did and when it finished. Gaps found in an exercise cost little to fix. Gaps found during an investigation can move a violation into a higher tier.
Where the argument is weakest
Discretion can be withdrawn
The strongest objection is that the notice was enforcement discretion, not law. HHS said it created no rights, and the codified rule still applied a $1.5 million limit to every tier. A future administration could withdraw the notice without rulemaking. We accept that. Our argument does not depend on the lower caps lasting. It depends on the tiers, which come from the statute, and those would continue to sort penalties by culpability whatever annual limits applied.
Settlements happen outside the penalty process
A second objection is that the regulation lets HHS resolve a matter by informal means, which may include demonstrated compliance or a completed corrective action plan or other agreement, so a formal penalty may never be imposed. Even so, the penalty framework sets the terms of negotiation. An organization that can show first-tier or second-tier facts negotiates from a different position than one facing a credible willful neglect finding.
Per-violation counts can still add up
A third objection is that the per-violation amounts did not change, and a large incident can involve many violations. The annual limits still cap the total per identical requirement per year, but an incident touching several requirements across more than one year can produce large totals. That reinforces the argument: evidence that moves each requirement into a lower tier matters requirement by requirement and year by year.
Lower caps could weaken the case for investment
A further concern is that lower caps in the lower tiers might reduce the pressure on organizations to invest in security and privacy. If the worst outcome for an unknowable violation was $25,000 a year per requirement, a budget committee might discount the risk. We think that reading misses where the money is. The penalty is only one cost of a violation; the incident behind it brings response costs, operational disruption and reputational harm that the notice did not touch. And the lower caps were available only to organizations that could prove diligence or reasonable cause, which itself requires investment.
Other enforcers
A fifth point is that HHS is not the only body that can act on a health data failure. State laws on data security and breach notification, and other federal regulators for some activities, carry their own penalties under their own rules, which the HHS notice did not touch. The evidence that supports a lower HIPAA tier may help in those settings too, but the tiers themselves apply only to HHS’s HIPAA penalties.
How it differs by kind of organization
Large health systems
Large systems have the resources to keep the records this argument requires and the complexity that makes it hard. Their risk lies in scale: many applications, many vendors and many departments, each of which can hold a known problem that never reached the central register. For them, consolidating findings and dates into one system is the main task.
Small and rural providers
Smaller providers have fewer systems but thinner staff. The notice helped them in cases where they could show diligence or reasonable cause, which depends on simple, consistent documentation more than on expensive tools. A dated risk analysis, a short remediation plan and a log of what was done are within reach of almost any practice.
Health plans
Health plans handle large volumes of member data and many transactions with providers and vendors. Their exposure can lie in disclosures through mailings, portals and vendor processing. Evidence of oversight of those processes, including quality checks and vendor monitoring, supports the lower tiers when an error occurs.
Practices that rely on outside IT support
A physician practice or clinic may depend on a managed service provider or an electronic health record vendor for much of its technology. That arrangement does not move its obligations to the provider. For the tiers, it means much of the evidence of diligence sits with someone else: patch records, backup tests, access logs and incident tickets. A practice should make sure its contracts require the provider to keep those records and hand them over on request, and it should ask for a periodic summary it can file. Without that, the practice may be unable to show the diligence its provider actually exercised.
Hybrid entities and academic medical centers
Some organizations are hybrid entities: single legal entities that are covered entities, whose activities include covered and non-covered functions, and that designate health care components. A university with a medical center is one example. For them, the evidence question includes boundaries: showing which components are covered, which systems hold protected health information and how data moves between covered and non-covered parts. A violation found at the boundary, for example in a research system that received clinical data, will turn on whether the organization knew the data was there.
Business associates
Business associates, such as billing companies, cloud providers and analytics firms, are named in the penalty provisions alongside covered entities. For them, the tiers create a second audience for their evidence: their covered entity customers, who will want to know whether a business associate can show diligence and timely correction. We think business associate agreements should require that evidence to be kept and shared on request.
The common thread
Across every kind of organization, the same records decided the tier: a current risk analysis, dated evidence that safeguards operated, documented decisions with reasons, a recorded discovery date, a verified correction date and a register of findings tracked to closure. The April 2019 notice did not require any of them to be kept in a particular way. It made each of them worth more, because each could move a violation down a tier or out of penalty altogether.
With hindsight
The rulemaking HHS said it expected did not arrive. As of the electronic Code of Federal Regulations current through November 12, 2026, 45 CFR 160.404 still states a limit of $1,500,000 for identical violations during a calendar year in each of the four culpability tiers for violations on or after February 18, 2009. A Federal Register search on November 14, 2026, for documents affecting 45 CFR part 160 published since May 2019 returned 16 documents, among them COVID-19 enforcement discretion notices, privacy rule proposals and rules, the 2025 Security Rule proposal, an inflation adjustment rule and a 2026 rule on health care attachments standards. None of them changed those four annual limits.
The inflation-adjusted table HHS keeps at 45 CFR 102.3 reflects the same position. In the current text, the 2025 maximum adjusted calendar-year cap is $2,190,294 for each of the four HIPAA tiers. The lower limits HHS applies come from the 2019 notice, which said its tiers would be adjusted for inflation until further notice, not from the codified table. An organization reading the regulation alone would still see one cap for all four tiers.
Congress later added a separate consideration. Public Law 116-321, enacted January 5, 2021, requires HHS, when making determinations about fines, audits and remedies agreed in resolving potential Security Rule violations, to consider whether a covered entity or business associate adequately demonstrated recognized security practices in place for not less than the previous 12 months. The law states that it gives HHS no authority to increase fines for a lack of those practices. Its text does not mention the culpability tiers or the annual limits; it adds a factor that may mitigate a fine, while the question this article turns on, what the organization knew and when it corrected the violation, still decides which tier applies.
The requirements themselves may change. On January 6, 2025, HHS published a proposed rule to modify the HIPAA Security Rule, citing changes in the health care environment, increases in breaches and cyberattacks, and common deficiencies OCR had observed in investigations. Comments closed on March 7, 2025. As of a Federal Register search on November 14, 2026, no final rule had been published: the search for documents affecting part 164 since January 1, 2025 returned only two, the Security Rule proposal and a January 2026 tribal consultation on proposed Privacy Rule changes. Organizations should check its status before relying on any part of it.
Neither of those later steps removed the need for the evidence the 2019 notice rewarded. The 2021 law adds a twelve-month look-back on recognized practices; the notice still sorts violations by what an organization knew and when it corrected them.
What held up is the structure: four tiers sorted by culpability, with evidence of diligence and timely correction deciding where a violation lands. What did not hold up is the expectation that the regulation would be amended to match. The discretion has now lasted from April 2019 through November 2026, under a notice that called its tiers interim, and it still rests on a notice that grants no rights. In our reading, organizations that built their records around the tiers are better placed whichever caps apply.
Spotted an error? Tell us and we will check it and correct the article where needed.