Premera’s breach response was judged by what its call center told members

Premera disclosed its breach on March 17, 2015, then mailed millions of letters and staffed phones. Four years later, states cited what agents had told callers. After a breach, every promise made to members is evidence.

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

Premera disclosed its breach on March 17, 2015, then mailed millions of letters and staffed phones. Four years later, states cited what agents had told callers. After a breach, every promise made to members is evidence.

  1. The week On Tuesday, March 17, 2015, Premera Blue Cross said it had been the target of a sophisticated cyberattack.
  2. Hearing about it: the cost of cleaning up first Premera found the attack on January 29 and announced it on March 17.
  3. Being found: reaching people who did not know they were customers Premera had about 1.8 million customers in two states.
  4. Being told: why the call center’s words carried weight The letter, the website and the call center are three channels for the same message.
  5. Being helped: credit monitoring for a health data breach Premera offered two years of free credit monitoring and identity theft protection.
  6. Being kept informed: the long tail after the letter The member experience of a breach does not end when the letter arrives.
  7. With hindsight Washington changed its breach law within months of the Premera announcement.

The week

On Tuesday, March 17, 2015, Premera Blue Cross said it had been the target of a sophisticated cyberattack. The health insurer, based in Mountlake Terrace, Washington, said information on about 11 million people may have been exposed. It had more than 1.8 million customers in Washington and Alaska at the time. The larger figure reflected records going back to 2002 and data from its Vivacity and Connexion Insurance Solutions brands [1].

The company said it had discovered the attack on January 29, 2015. Its investigation indicated that the first intrusion had taken place on May 5, 2014. The exposed information included names, Social Security numbers, bank account information and medical information, among other member details. Premera said it had notified the FBI and hired the security firm Mandiant to investigate and clean its systems [1].

The announcement came about seven weeks after discovery. Premera said it was mailing letters to affected people. It offered two years of free credit monitoring and identity theft protection, and it took questions through a call center.

Washington’s insurance regulator reacted within a day. Insurance Commissioner Mike Kreidler said publicly that he was concerned it had taken Premera about six weeks to inform his office after learning of the attack [1]. By the following week, Washington had announced a market conduct examination of how Premera handled the breach, joined by Alaska and Oregon. It was to look at whether Premera had taken the proper steps to notify the people affected, alongside a separate inquiry into how the attackers got in [2].

Premera explained the timing in the days that followed. As the Spokesman-Review reported, the company said the FBI had not asked it to delay notice. Its security consultant had strongly advised waiting until the systems were cleaned, so the attackers could not cause further damage or make threats when the news broke. During that period the consultant removed implanted malware and improved security. A Premera spokesman said there was no evidence that information had been removed from its systems [2].

The same report gave a sense of scale. Letters were going out at roughly 500,000 a day. By Tuesday, March 24, about 88,000 people had enrolled in the free identity protection [2]. The paper also reported that about 6 million of the current and former customers involved were in Washington.

Two sets of notice rules applied. Under the HIPAA Breach Notification Rule, a covered entity must notify each affected individual without unreasonable delay and no later than 60 calendar days after discovery. The notice must be written plainly. It must describe what happened, the types of information involved and the steps individuals should take. It must say what the entity is doing, and give contact routes that include a toll-free number, email address, website or postal address [4]. Premera’s March 17 announcement fell within the 60-day limit.

Washington’s own breach law, as it stood in March 2015, set no fixed number of days. It required disclosure in the most expedient time possible and without unreasonable delay, allowed delay for a criminal investigation, and did not require notice to the attorney general [5]. The legislature was already considering a bill to change that.

The context mattered too. Anthem had disclosed a much larger attack on its systems in early February. Premera’s disclosure made it the second large Blue Cross Blue Shield plan in six weeks to announce a breach of member data. For many members, the Premera letter was the first time they had thought about what their health insurer held about them, and for how long.

Most breach analysis starts with the intrusion: the phishing email, the malware, the months of undetected access. For the 11 million people Premera wrote to in March 2015, the breach began with an envelope, a website and a phone call. Their experience of the incident was the response. What they were told, when, and by whom decided whether they trusted the company afterward.

The record that followed shows how that experience was later judged. When thirty states settled with Premera in 2019, their allegations included more than the security failures. They cited what the company’s call center agents had told worried callers after the breach. Reassurances given on the phone in 2015 became part of a consumer protection case four years later.

That is the argument of this article. A breach response for members is a series of promises, made at scale under time pressure, and every one is evidence. We follow the member’s experience in five steps, from first notice to the long tail:

  • Hearing about it: when members were told, and the trade-off between telling quickly and cleaning up first.
  • Being found: who had to be reached, including people who had never been Premera customers in any sense they recognized.
  • Being told: what the letter, website and call center said, and why the words carried legal weight.
  • Being helped: what remedy was offered, and how well credit monitoring fit a health data breach.
  • Being kept informed: the months after the letter, when questions, scams and litigation continued.

The lessons apply to health plans and to the health systems, ambulatory networks and business associates that hold the same kinds of data.

Hearing about it: the cost of cleaning up first

Premera found the attack on January 29 and announced it on March 17. In between, its consultant was removing malware and closing the attacker’s access. The company said this sequence was deliberate. Announcing while the attacker was still inside might have prompted the attacker to destroy data, dig in deeper or threaten the company.

The case for waiting

The argument for remediation before notice is real. An attacker who learns that a victim has detected it can change tactics. It can move to systems the defenders have not yet examined, deploy destructive tools or exfiltrate whatever it can reach before access is cut. A public announcement is also a signal to other criminals that a large population of personal data may be for sale. Incident responders often prefer to complete containment before the news breaks.

The HIPAA rule allows for this. Its standard is without unreasonable delay, with an outer limit of 60 days from discovery. A covered entity that spends part of that window on containment, and can show why, is acting within the rule. The rule also has a separate provision for delays requested by law enforcement. Premera said the FBI had not asked for one. Its delay rested on its consultant’s advice.

The case against

The argument against is the member’s exposure. For seven weeks, people whose Social Security numbers and bank details may have been taken did not know to watch their accounts or freeze their credit. Each day of delay was a day in which any misuse would go unnoticed by the person most able to spot it.

Regulators felt the delay too. Washington’s insurance commissioner objected publicly to learning about the attack weeks after Premera did. His deputy noted that the regulator could have been told confidentially. That point is important for any covered entity. Remediation before public notice does not require silence toward regulators. A confidential early notice to the primary regulator preserves the containment window while avoiding the impression that the company was hiding the incident.

The member who calls before the announcement

A containment period creates its own communication problem. During those weeks, some members will notice suspicious activity and call. A few will ask directly whether there has been a breach. Front-line staff either do not know about the incident or know and have been told not to discuss it.

Both cases carry risk. An agent who does not know may give a confident denial that later looks false. An agent who knows may hint at something, and the hint spreads. A sound plan gives a small, briefed team a holding answer for those calls. The answer acknowledges the concern, takes details, and avoids any statement that the organization knows to be untrue. Those calls should be logged, because they may be early signs of misuse that the investigation needs.

Deciding the timing

The timing decision is a member experience decision as much as a security one. It should be made with both views in the room. A practical approach has four elements.

  • A documented rationale, written at the time, for any delay between discovery and notice, tied to specific containment steps and their expected completion.
  • Early confidential notice to the primary regulators, and to any business partners whose members are affected, as soon as the facts are stable enough to share.
  • A notice pack prepared in parallel with containment, so that letters, web content and call center materials are ready the day containment ends.
  • A firm internal date, well inside the legal deadline, after which notice goes out even if containment is incomplete, with a plan for what to say in that case.

The last element matters because containment can drag. Investigations find new compromised systems. A company that ties notice to the end of containment, without a backstop, can find itself near the legal limit with no plan.

How the clock differs by type of organization

The HIPAA clock runs the same way for every covered entity, but the parties around it differ. A health plan like Premera notifies its own members. Where it administers self-funded employer plans, the employer’s group health plan may be the covered entity and the insurer its business associate. Who sends the letters then depends on the contracts, and the employer’s benefits team needs to be told early.

A business associate, such as a claims processor or billing vendor, must tell the covered entity without unreasonable delay and within 60 days of discovering a breach. The covered entity then notifies individuals. When the vendor takes most of the window, the covered entity has little time left for its own notice. Health systems that rely on many vendors should write shorter reporting periods into their business associate agreements.

Large breaches carry two further notices under the rule. A breach affecting more than 500 residents of a state or jurisdiction requires notice to prominent media outlets there. A breach affecting 500 or more people requires notice to HHS at the same time as the individual notices. Those media and HHS notices make the incident public on the covered entity’s timetable, so the member-facing channels must be ready when they go out.

Telling the right people first

The order of notice also shapes the experience. In a breach of this size, employees, brokers, employer groups and provider offices all receive questions from members who have seen the news. Each of those groups needs a briefing before the public announcement, so that the first person a member asked could give an accurate answer. A broker who learns about a breach from a client’s phone call cannot help that client.

For a health system, the same applies to clinicians and front-desk staff. Patients ask the people they see. If those people have not been briefed, they improvise, and improvised answers become part of the record.

Being found: reaching people who did not know they were customers

Premera had about 1.8 million customers in two states. It wrote to about 11 million people. The gap between those numbers is the second part of the member experience problem.

Who was in the data

The affected population included former members whose records dated back to 2002. It included people covered through Premera’s other brands. Some had moved away years earlier. Some had died. Some had been children on a parent’s policy and were now adults with different addresses. Many would not recognize the sender of a letter about a company they had not dealt with in a decade.

That pattern recurs in health care. Health plans and health systems keep records for long periods, for legal, clinical and administrative reasons. A breach of a long-lived data store reaches people the organization has no current relationship with. Their contact information is stale, and their trust has to be built from nothing in one letter.

Retention decided the size of the mailing

The number of people an organization must notify is set long before a breach, by what it chose to keep. Premera’s affected population was several times its membership because the compromised systems held records going back more than a decade [1]. Every former member in those systems was a future letter, a future call and a future identity protection enrollment.

Health organizations have real reasons to keep data. Medical record retention laws, claims disputes, audits and litigation holds all require it. Those reasons rarely require full Social Security numbers and bank details for people who left a decade ago to sit in systems reachable from the corporate network. Moving old records to restricted archives, removing identifiers that are no longer needed and enforcing retention schedules reduces the population a future breach can reach. Retention is usually discussed as a compliance topic. After Premera, it was also a member experience decision with a measurable cost per record.

The mechanics of mass notice

The HIPAA rule sets the method. Notice goes by first-class mail to the last known address, or by email if the individual has agreed. For a deceased person whose next of kin is known, notice goes to the next of kin or personal representative. Where contact information is insufficient for ten or more people, substitute notice is required. That means a conspicuous web posting for 90 days or notice in major print or broadcast media, plus a toll-free number active for at least 90 days [4].

At Premera’s scale, those rules became a logistics program. Mailing 500,000 letters a day required an outside mail house, address hygiene against change-of-address databases, and tracking of returned mail. Returned mail had to feed the substitute notice process. Each letter needed a unique code to enroll in identity protection, which meant a data handoff to the protection vendor before mailing began.

Each of those steps involved personal data leaving the covered entity. The mail house and the identity protection vendor were business associates handling the same data that had just been breached. Their security, and the contracts covering them, became part of the response.

Recognition and trust

A letter from an unfamiliar company, asking the reader to call a number or visit a website and enter personal details, looks like a scam. Fraudsters know this and often copy real breach letters. The legitimate notice has to establish that it is real without training recipients to trust any letter of that kind.

Several design choices help. The letter can name the brand or employer group through which the person was covered, so a former member recognizes the connection. It can direct readers to a web address they can verify on the organization’s main site, rather than only to a new domain. It can say clearly what the organization will never ask for by phone or email. It can offer a way to enroll in identity protection without calling a number printed in the letter.

Networks and shared members

Health plans often process claims for each other’s members, and health systems share records across affiliated practices and exchanges. When data about another organization’s members or patients is breached, the people affected may have no direct relationship with the breached party at all. Their own plan or provider will get the calls.

That raises questions that are better settled before an incident. Who writes to those people? Whose name goes on the letter? Who pays for the identity protection? Contracts among plans, and agreements within clinically integrated networks, often say little about this. A short clause on breach notice roles and costs avoids a negotiation in the middle of a crisis.

Accessibility and language

The rule requires plain language. A large affected population also includes people who read other languages, people with low vision and people without internet access. A response that relies on a website excludes some of them. A health plan with members in many communities should have translated letters and multilingual call center capacity ready before an incident, because there is no time to build them during one.

Health systems face the same issue with patients. Their notice obligations under HIPAA are the same as a plan’s. Their patient population may be more diverse and less used to dealing with the organization by mail. A patient who received care once in an emergency department may not recognize the hospital’s billing name.

Being told: why the call center’s words carried weight

The letter, the website and the call center are three channels for the same message. They are written and run by different teams, often with outside vendors, under deadline. If they say different things, the member notices, and later so do regulators.

The allegation

Washington’s attorney general announced a $10 million settlement with Premera in July 2019, on behalf of a coalition of thirty states. The complaint alleged that Premera had failed to address known security weaknesses. It also alleged that Premera misled consumers about its privacy practices before and after the breach. The announcement described what call center agents had told callers. In substance, agents said there was no reason to believe callers’ information had been accessed or misused. They also said significant security measures were already in place [3].

The states’ view was that those statements were misleading, given that experts and Premera’s own auditors had warned for years of inadequate security. Premera settled without the claims being tried. For our purposes, the merits matter less than the mechanism. A reassurance written into a call center script, and repeated thousands of times, was treated as a statement by the company.

How scripts get written

Call center scripts after a breach are usually written fast. The communications team drafts them. Legal reviews them. The call center vendor turns them into a knowledge base and trains agents, many of them temporary staff hired for the surge. The goal is to calm callers and keep calls short.

Each of those steps can introduce risk. Communications staff tend toward reassurance. Legal review may focus on admissions of fault rather than on the accuracy of comforting statements. Agents under pressure paraphrase, and a paraphrase of a careful statement can become an inaccurate one. A script line saying the company had found no evidence of misuse so far can turn, in an agent’s mouth, into a promise that nothing was misused.

Governing the words

A health plan or health system can treat breach communications as controlled statements, with the same discipline as a public filing. That approach has several components.

  • One source of facts. A single, versioned fact sheet that every channel draws from, updated as the investigation learns more, with a named owner.
  • Accuracy review. A reviewer who checks each reassuring statement against what the investigation actually knows. Statements about security should describe what has been done, not characterize it as strong or significant.
  • Agent guardrails. Clear lists of what agents may and may not say, with approved answers to the hardest questions and an escalation path for anything outside them.
  • Quality monitoring. Call recording and sampling during the surge, with daily review of what agents are actually saying and fast correction of drift.
  • Change control. When the facts change, every channel changes the same day, and the old versions are archived with dates.

The archive matters later. A company asked years afterward what it told members on a given date needs the script version, the web content and the training materials from that date. Without them, it relies on memory and on whatever callers recall.

The website as the record

The incident website usually outlives the call center surge. It is where journalists, regulators, plaintiffs’ lawyers and members return for the organization’s current account. Its frequently asked questions often contain the most specific statements the organization makes: what data was involved, whether it was encrypted, whether misuse has been seen and what the organization has done.

Those answers deserve the same accuracy review as the letter. They should carry a date of last update. When an answer changes, the page should say so rather than silently replacing it. A member who read one version in March and another in May should be able to see that the facts moved, and why.

The privacy notice problem

The states also cited Premera’s pre-breach privacy notice, which told members the company took steps to secure its buildings and electronic systems from unauthorized access [3]. Most organizations have similar language. It reads as a routine statement until a breach makes it a claim to be tested.

The fix is not to remove such statements, since HIPAA notices of privacy practices have required content. It is to make sure the organization can support what its notices say, with evidence of the controls they describe. Compliance teams that review privacy notices annually can ask the security team to confirm each factual claim, and record the answer.

Testing the script before it is needed

Most incident response exercises test the technical side: detection, containment and recovery. Fewer test the member side. A useful exercise gives the call center vendor a realistic scenario and a draft script, then has staff play difficult callers. A caller whose child’s details were exposed, a caller who was never a member, and a caller who reports fraud already happening are all predictable. Each reveals gaps in the script and in the escalation path.

The exercise also tests capacity. A vendor that promises to scale to hundreds of agents within days should show how it would recruit, train and supervise them. Temporary agents trained in a few hours are the ones most likely to paraphrase a careful statement into an inaccurate one.

Who owns the member’s question

Inside a health plan, member services owns the call center. Security owns the investigation. Legal owns the regulatory exposure. Communications owns the message. During a breach, the member’s question cuts across all four. Someone needs authority to decide what the member is told, with input from each. In many organizations that role is undefined until the incident starts, and the first days are spent working it out.

Being helped: credit monitoring for a health data breach

Premera offered two years of free credit monitoring and identity theft protection. That was the standard remedy of the time, and it addressed real risks. The stolen data included Social Security numbers and bank account information, which can be used to open accounts or commit fraud.

What credit monitoring covers

Credit monitoring watches credit files for new accounts, inquiries and changes. It alerts the person when something appears. Identity theft protection services typically add help restoring a stolen identity and some insurance against costs. They are useful against financial identity theft.

What it does not cover

A health data breach creates other risks that credit monitoring does not see. Medical identity theft, in which someone uses another person’s identity to obtain care or file claims, shows up in claims records and medical records, not credit files. Insurance fraud using a member ID number shows up in explanation of benefits statements. The exposure of claims information about diagnoses or treatments carries privacy harms that no monitoring service can reverse.

A health plan is unusually well placed to address some of these. It can flag accounts for unusual claims activity. It can encourage members to review their explanation of benefits statements and make it easy to report claims they do not recognize. It can reissue member ID numbers where the risk justifies the cost. Those measures fit the data that was taken in a way that generic credit monitoring does not.

Enrollment and its limits

By March 24, about 88,000 people had enrolled in Premera’s free protection, a week into a mailing to 11 million [2]. Early enrollment is often low because letters are still arriving. Enrollment in breach monitoring offers has been widely reported as low relative to the number eligible. People who never enroll get no protection from the offer.

The design of enrollment matters. A process that requires a code from the letter, a long online form and a second round of personal data entry loses people at every step. Simpler enrollment, reminders and a clear explanation of what the service does all raise uptake. They also cost more, because the vendor is paid per enrollee. That is a trade-off between budget and member protection, and it should be made consciously.

Matching the remedy to the person

A single offer for everyone is simple to run, but the affected population is not uniform. A current member with a bank account on file faces different risks from a former dependent whose only exposed data is a name and member number. Tailoring the remedy by data type, such as additional bank account guidance for those whose account details were taken, puts help where the risk is. It also makes the letter more credible, because it describes the reader’s actual situation.

Buying the remedy before the breach

The identity protection service is usually bought in a hurry. The organization chooses a vendor in the first days, negotiates price and terms under pressure, and hands over a file of affected people. That file is a new transfer of the data that was just breached.

Buying in advance changes the terms. A pre-arranged agreement can fix per-enrollee pricing, set capacity commitments for the call center and enrollment site, and define how the vendor protects and deletes the data it receives. It can require enrollment reporting by day, so the organization knows how many people are protected. It can require translated materials and accessible enrollment options. Many cyber insurance policies cover notification and monitoring costs, sometimes through a panel of approved vendors. Checking the policy terms against the response plan avoids learning during an incident that the preferred vendor is not covered.

Children and the deceased

Records dating back to 2002 included children who had been dependents. Children’s identities are valuable to fraudsters because misuse can go undetected until the child applies for credit years later. Adult credit monitoring does not fit them. Some providers offer minor identity monitoring. A 2018 federal law later made credit freezes free nationwide and let parents place them for children under 16.

For deceased members, the risk runs through their estate and next of kin. Notice to next of kin is required where the address is known. Offering a meaningful remedy is harder, and few breach responses address it well.

Two years and after

Two years of monitoring was common in 2015. The data taken, including Social Security numbers and dates of birth, does not expire. A member whose details were taken in 2014 faces risk long after a two-year service ends. Later settlements and regulatory expectations moved toward longer terms in some cases, but no fixed period matches the life of a Social Security number. The honest message to members is that the service covers a period and that some precautions, such as credit freezes, are permanent and free.

Being kept informed: the long tail after the letter

The member experience of a breach does not end when the letter arrives. In Premera’s case it ran for years, through regulatory examinations, class actions and settlements. Members received further notices, questions arose about whether new fraud was connected to the breach, and the company’s statements continued to be tested.

The second wave of questions

After a large breach, a second wave of contact follows the first. Members who experience fraud months later call to ask whether it is connected. Members who receive suspicious calls or emails, sometimes impersonating the breached company, ask whether they are real. Each of those contacts needs an answer consistent with what the company has said before.

A response plan that staffs the call center for the first month and then scales it down leaves those members with nowhere to go. Keeping a smaller, trained team in place for a longer period, with updated materials, costs less than the reputational and legal cost of inconsistent answers.

Updates as facts change

Investigations change their findings. The initial estimate of affected people at Premera was about 11 million. The later figure in the attorneys general’s settlement was more than 10.4 million [3]. Dates move too. The states described unauthorized access from May 5, 2014 to March 6, 2015, which means access continued for weeks after discovery while the clean-up was under way [3].

Members deserve to know when material facts change. An organization that updates its public information as findings change, with dates on each update, builds credibility. One that leaves its first statement unchanged invites later challenge when the record shows something different.

Regulators read the letters too

Washington’s market conduct examination looked at how Premera notified people as well as how it was breached [2]. State insurance regulators, attorneys general and HHS all receive copies of notices or complaints about them. Members who are unhappy with a call often complain to the regulator next.

That makes complaint handling part of the breach response. A plan or health system should route breach-related complaints to a dedicated team, answer them on a set timetable and keep a log that can be shared with regulators. Patterns in those complaints show where the letter or script is failing. They also show a regulator that the organization was listening, which matters when the regulator later decides how to characterize the response.

Litigation as a member experience

Class action litigation followed the Premera disclosure quickly. For members, it meant further mail: notices of the case, of proposed settlements and of claim procedures. Many members could not tell those notices from the company’s own communications, or from scams.

The organization has limited control over class notices. It can make sure its own channels explain the difference, and that its call center knows what to say about the litigation. That is one more script, with the same governance needs as the first.

Closing the incident for members

At some point the organization has to decide that the response is over for members. The call center returns to normal staffing, the incident website is updated for the last time and the monitoring service runs out. Each step is a communication in its own right.

A final update that says what was learned, what changed and where to go with future questions closes the loop. It also creates a dated record of the organization’s final account. Without it, the last public word may be the first estimate, issued weeks into an investigation that later changed its findings.

Measuring the member experience of a breach

Few organizations measure how members experienced a breach response. Useful measures are straightforward to collect. They include time from discovery to notice, the share of letters delivered and returned, and call center wait times and abandonment. They also include enrollment rates in offered protection, complaint volumes to regulators and the results of call quality reviews. Reviewing them after the incident, and comparing them with plan targets, turns a crisis into data for the next plan.

Taken together, the five steps describe a member journey that most incident response plans do not map. Plans are written by security and legal teams around containment and legal deadlines. The member’s journey runs alongside, handled by member services and vendors. Premera’s case shows that the second track is examined as closely as the first, and on a longer timeline.

The strongest objection

A reasonable objection is that companies cannot be blamed for reassuring frightened people. Callers want comfort, and agents are trained to provide it. Holding every phrase to the standard of a securities filing would make breach communications cold and defensive.

The answer is that accuracy and warmth are compatible. An agent can say that the company is sorry and is working to protect the caller. The agent can explain what has been done and what the caller can do, without asserting things the company does not know. The risk comes from statements of fact that turn out to be wrong, not from empathy. Good scripts separate the two.

With hindsight

Washington changed its breach law within months of the Premera announcement. The 2015 bill set a deadline of 45 days for notice to consumers and required notice to the attorney general when a breach affected more than 500 Washington residents [5]. In 2019 the legislature shortened both deadlines to 30 days and broadened the definition of personal information, effective March 1, 2020 [6]. The idea that a company could choose its own timing within a general standard of promptness gave way, in Washington, to fixed clocks.

On July 11, 2019, Washington’s attorney general announced the $10 million multistate settlement. Washington received $5.4 million and 29 other states shared about $4.6 million. The consent decree required Premera to map where HIPAA-protected information was located on its network and to have a third-party expert, approved by the states, prepare data security reports. It required a chief information security officer separate from the chief information officer, with meetings with the CEO every two months and reporting of any intrusion to the CEO within 48 hours. It also required a compliance officer with HIPAA experience and security training for employees who handled personal and health information [3].

The states said the breach affected more than 10.4 million people, including more than 6.4 million Washingtonians, and that unauthorized access lasted from May 5, 2014 to March 6, 2015 [3]. The complaint’s focus on what the call center told consumers, alongside the security failures, is the part of the case most relevant to this article.

A separate class action in federal court in Oregon settled for $32 million paid into a fund for class members. Premera also committed to spend $42 million on information security over three years. The court gave final approval in 2020. The HHS Office for Civil Rights announced its own resolution with Premera in September 2020: a $6.85 million payment and a corrective action plan with two years of monitoring. OCR described it as the second-largest HIPAA settlement payment it had obtained to that point, and said its investigation found failures including the lack of an enterprise-wide risk analysis. OCR also said the attackers had gained access through a phishing email.

What held up from the March 2015 response is the general shape. Mailed notice, a call center, a website and two years of monitoring remained the standard package for large health data breaches for years. What changed is the scrutiny of each element. Regulators and plaintiffs increasingly examined what companies said to affected people, as well as whether they said it on time. The Premera decree itself shows the pattern, with governance and compliance terms set alongside the technical ones.

The timing debate did not go away. Fixed state deadlines tightened the window for remediation before notice. Federal rules for other sectors moved the same way: a 2021 rule from the federal bank regulators, for example, required banks to notify their primary regulator within 36 hours of certain significant incidents. Health care organizations now operate under a patchwork of state deadlines that can be shorter than HIPAA’s 60 days. A plan that relies on using the full HIPAA window for containment can breach a state law along the way.

The member side of breach response also had to absorb a change in scale. In the years after 2015, breaches at health care business associates and technology vendors affected larger numbers of people across many organizations at once. In those cases, the organization the patient knew was often not the one that was breached, and the problem of reaching people who did not recognize the sender grew larger.

The weakest part of the 2015 model was the remedy. Credit monitoring addressed financial identity theft. It did little for medical identity theft or for the privacy harm of exposed health information. Health plans and providers had better tools, such as claims monitoring and record review, but those were rarely offered as part of a breach remedy. That gap remains open for many organizations today.

Spotted an error? Tell us and we will check it and correct the article where needed.

  • Premera
  • breach notification
  • HIPAA
  • call center
  • identity protection
  • health plans