Litigating the Accountability of Clinical Genomics Laboratories
Seeking Accountability from Clinical Genomics Laboratories
A wrongful death case pending in a federal court in Columbia, SC—Williams v. Quest Diagnostics, Inc., et al.—demonstrates the very high stakes involved with clinical genomics testing. The case also underscores the ongoing struggle to (1) establish accountability when mistakes happen and (2) establish a reasonable and effective level of governmental oversight for the laboratories performing such tests. The case could have dramatic implications for the future practice of genomic medicine, and regardless of how this specific case ultimately plays out in the courtroom, the allegations should serve as a serious wake up call to those involved in genomics for clinical purposes in any way.
Overview of the Case
Note: This summary is based on allegations in the pleadings or reports in the media. The facts of the case have not yet been adjudicated and are in dispute.
Amy Williams, the plaintiff in this case, is suing defendants Quest Diagnostics, Athena Diagnostics, and ADI Holdings as an individual and as the personal representative of her deceased son’s estate. Athena is a subsidiary of ADI Holdings which is itself a subsidiary of Quest. Williams claims that the Athena lab’s negligence, stemming primarily from noncompliance with the statutory requirements of the Clinical Laboratory Improvement Amendments (CLIA), caused the death of her son, Christian. The case was initially filed in a South Carolina state court in February 2016 but was quickly removed to federal court the following month. (Defendants can often “remove” a case filed in state court to the nearest federal court if, as here, the plaintiff could have filed it in federal court in the first place. The negligence claim will continue to be adjudicated according to South Carolina state law, notwithstanding the removal to federal court.)
According to the complaint, Christian was born in August 2005 and began having seizures when he was about four months old. His doctors treated his seizures with sodium channel blocking medications, but these were not effective. The doctors began to suspect that Christian had Dravet syndrome, also known as SMEI (Severe Myoclonic Epilepsy of Infancy) or at least wanted to rule out that diagnosis.
Some background about this condition and the gene involved
At the time Christian was born, SCN1A was known to be one of the most clinically important genes involved in epilepsy. SCN1A is the gene encoding the voltage-gated sodium channel, neuronal type I, alpha subunit protein (Nav1.1). The gene is located at 2q24.3 (i.e., it is on the long arm of chromosome 2). Its highly conserved structure (which is a sign often indicative of functional importance) consists of 26 exons and encodes a protein of 2000 amino acid residues. Mutations in this gene cause a spectrum of epilepsy phenotypes, the most severe being Dravet syndrome. More than 1250 epilepsy-causing mutations have been reported in this gene, and about 95% of patients with Dravet syndrome have de novo heterozygous mutations (meaning the parents do not carry the mutations). For patients with SCN1A mutations, sodium channel blocking medications (like carbamazepine and lamotrigine) are to be avoided because they can exacerbate the seizures, and other treatment options (such as valproic acid, topiramate and stiripentol and a ketogenic diet) are to be considered. Genetic testing is now be considered a routine part of clinical epilepsy practice. The majority of individuals with Dravet syndrome survive to adulthood, but the mortality rate for individuals with Dravet syndrome is high (estimated at 10-15% or perhaps ~15-20%). Death can occur at any age, although most frequently during childhood.
Suspecting Dravet syndrome, in January 2007 Christian’s treating doctors ordered genetic testing for SCN1A from Athena to get a more specific diagnosis and gain a better understanding of viable treatment options. While still awaiting the results from Athena’s “SCN1A sequencing clinical diagnostic test” as to whether Dravet syndrome could be ruled out, the doctors reached a preliminary diagnosis that the boy suffered from “probable mitochondrial encephalomyopathy” (a disorder caused by mitochondrial DNA, which is inherited only from mother to child). Athena issued its report in June 2007, six months after the test was ordered. The June 2007 report indicated Christian had an SCN1A mutation (1237T>A, Y413N) and classified the mutation as a “variant of unknown significance” (or VUS; in other words, not specifically known to be either pathogenic or benign). Relying on the information in the report (which, to reiterate, did not report any variants as pathogenic or identify any treatments to avoid), Christian’s doctors continued to treat him for an unspecified mitochondrial disorder with increasing doses of sodium channel blocking medications (which are standard for epileptic seizures but are contraindicated for those with Dravet syndrome). It is necessary to note that, in line with industry custom, Athena did not issue a copy of the June 2007 report directly to the patient (or Williams as the patient’s parent). Christian’s condition worsened (even as doses of the sodium channel blocking medications were increased in attempts to treat him), and he died in January 2008. The cause of death was listed as “seizure due to complications of mitochondrial disorder (type not specified).”
According to the complaint, Williams spent years grieving for her son’s death and obtaining counseling to help her cope with her false belief that she was responsible for his disorder (since a child inherits mitochondrial DNA from his/her mother). Several years later, Williams requested a copy of the 2007 report directly from Athena but was not given access. A copy was finally obtained in September 2014 with the help of a genetic counselor. In January 2015, Quest/Athena jointly issued a “revised report” that listed Christian’s SCN1A variant as a “known disease-associated mutation” and directed the recipient to “disregard previous report.” The 2015 report did not list any new references or information supporting a reclassification, did not identify the date when the reclassification occurred or who directed that decision, and did not document any attempts by the lab to notify either the medical provider who had ordered the original test or the patient(s) affected by the reclassification. The report also did not indicate whether a new analysis was performed on a retained DNA sample or whether Athena merely prepared a new report using the original variant call file retained from the 2007 analysis. That the revised January 2015 report provided no new references that were unavailable at the time the June 2007 report was issued is striking: if the references cited in the 2015 report were available at the time of the original report’s issuance in 2007, the original classification should have been different. All of these factors tend to support Williams’ position that the 2015 report corrected an error in the original classification, rather than providing an update or reinterpretation necessary to reflect how scientific understanding had developed since the original test was performed.
Williams alleges that Athena (along with its corporate affiliates) was negligent and breached the applicable standard of care by (1) failing to provide a genetic confirmation that Christian had Dravet syndrome, (2) failing to adhere to its own post-analytic classification system, (3) failing to notify anyone (the medical provider or the patient) of the reclassification of the variant the lab had identified, and (4) failing to identify the contraindicated treatments (i.e., the sodium channel blocking medications). The complaint also alleges extensive violations of the Clinical Laboratory Improvement Amendments (CLIA) as well as a violation of the Health Information Portability and Accountability Act (HIPAA) right to access personal health information (45 CFR §164.524), when Williams requested and was denied direct access to the 2007 Report. In addition, Williams brings claims for fraudulent or negligent misrepresentation, civil conspiracy to cover up the mistake, and engaging in an unfair and deceptive trade practice.
This case draws attention to several unanswered legal questions, including
1. What responsibility do laboratories have to make sure that they adhere to their own variant classification schemes and, when they make a mistake, that they identify and correct it appropriately?
2. Can ordering physicians and health providers rely on the reports issued by labs conducting genomic testing, or do they individually have a professional obligation to be competent in genomics?
3. Whose responsibility is it to stay abreast of the developments in the scientific literature and ensure that accurate information is used in generating results reports and, ultimately, in making medical decisions on the basis of such reports?
4. Whose responsibility is it to revisit medical decisions (such as ruling out one disease or diagnosing another) based on interpretations of results that subsequently become outdated?
5. Can laboratories be held accountable for regulatory noncompliance and injuries from lab mistakes using an ordinary negligence theory, or must injured individuals (end-users, patients, etc.) base their causes of action in medical malpractice theory?
6. What does “constructive notice” mean in an industry that has not been known for its transparency or willingness to provide individuals with direct access to the data and results of the genomic analyses performed?
7. Overall, what responsibility do laboratories owe to each individual patient whose DNA is analyzed, and how much of that burden may laboratories shift onto (a) the medical providers ordering the tests and making diagnostic and treatment decisions in reliance upon the reported results or (b) the patients themselves.
Moreover, this case has serious implications for the sufficiency of regulatory oversight of the laboratories and specific tests.
A bit of context: oversight of genetic testing has been a regulatory hot potato
The scientific understanding of genetic variants and their roles in health and disease is constantly changing. As genomic analysis becomes routinized, best practices have begun to emerge and early variant classification schemes (i.e., standards and guidelines issued by professional organizations for determining whether a variant is pathogenic, benign, or something in between) have already undergone revisions. Moreover, the secrecy that has dominated the industry since the 1990s (driven by the desire to maintain proprietary databases) is—albeit slowly and not without resistance—being challenged and displaced by advocates and grassroots efforts in favor of the illuminating properties of openness and the public benefits of a medical data commons. In the last 25 years, so much in genomics has changed, and yet so much in the regulatory space remains the same.
The Clinical Laboratory Improvement Amendments Act of 1988 (CLIA) was intended to provide government oversight of clinical lab protocols and processes. There are now more than 250,000 laboratories registered. Fifteen years ago, the Center for Medicare & Medicaid Services (CMS), the agency responsible for enforcing CLIA, was advised by the working group of its advisory committee (CLIAC) and later petitioned by three non-profit organizations to add genomics as a designated specialty area for CLIA enforcement. No such specialty area has been created, and the recommendation was formally rejected by CMS in 2007 on grounds that it would be too expensive to implement. Ten years ago the Government Accountability Office (GAO) concluded, “CMS oversight is not adequate to enforce CLIA requirements.” According to the GAO report, CMS was not striking the right balance between enforcement of the statutory requirements through penalties and education of laboratories to improve future compliance. CMS was reluctant to sanction even those labs with repeat deficiencies. CMS also was unwilling to conduct unannounced CLIA inspections (called “surveys”), and its practice of announcing the inspections perhaps undermines the ability of inspections to provide an accurate picture of lab quality. As recently as last year, this policy of announced inspections was reiterated by CMS. Nevertheless, the statutory requirements of CLIA for lab practices are clear, and CMS is the agency tasked with ensuring compliance with them. A few of the CLIA regulations are the following:
• 42 CFR 493.1249(a), 42 CFR 493.1289(a), and 42 CFR 493.1291(a) require laboratories to “establish and follow written policies and procedures for an ongoing mechanism to monitor, assess, and when indicated, correct problems identified” in the pre-analytic, analytic, and post-analytic systems, respectively.
• 42 CFR 493.1241(c) requires laboratories to ensure that a request for testing includes the identification and contact information for the ordering person/provider and that the request for testing includes “information relevant and necessary for a specific test to ensure accurate and timely testing and reporting of results, including interpretation”
• 42 CFR 493.1283 requires laboratories to maintain proper records of tests, including specific information such as identifying who performed the test, the date the specimen was received, the dates of all tests performed, etc.
• 42 CFR 493.1291(k) requires laboratories to take action when errors in the reported test results are detected, including specifically that laboratories must “(1) Promptly notify the authorized person ordering the test and, if applicable, the individual using the test results of reporting errors. (2) Issue corrected reports promptly to the authorized person ordering the test and, if applicable, the individual using the test results. (3) Maintain duplicates of the original report, as well as the corrected report.”
If CMS identifies deficiencies during its inspections of a lab’s procedures, the agency has the authority to issue sanctions. Moreover, anyone can file a complaint with CMS regarding concerns about a lab’s operations. Two states, Washington and New York, are exempt from CLIA and are considered to have state regulatory schemes that meet or exceed CLIA requirements. CMS has maintained that its oversight authority is limited to lab processes and not the clinical validity of the tests that labs use, passing that responsibility off to the Food and Drug Administration (FDA).
Over the past decade, the FDA has attempted to exert authority over all laboratory-developed tests (LDTs), a subject covered extensively on the Genomics Law Report. The FDA released a proposed framework in 2014 and a year later released a controversial report of 20 cases that, it argues, demonstrate why the FDA (and not CMS) is the appropriate agency to handle the bulk of genomics lab testing oversight responsibility. Several legislative alternatives to the FDA’s proposed framework have also been under consideration. Regulating LDTs is on the FDA’s “A-List” for FY2016 and an upcoming National Human Genome Research Institute workshop suggests that an FDA policy could be announced before the next president takes office. Assuming the FDA has the authority to regulate all genetic analyses as medical devices, the agency would assess their analytical validity and clinical validity and could, when problems are identified with specific tests, issue medical device recalls to notify affected individuals and require corrective actions (similar to the post-analytic requirements already imposed by CLIA at the laboratory level).
Finally, not to be forgotten, the Federal Trade Commission (FTC) also has some important oversight authority in the area of clinical genomics. The FTC protects the public from false and deceptive advertising and also “unfair or deceptive trade practices” under Section 5(a) of the FTC Act, 15 U.S.C. 45(a)(1). The FTC along with state attorneys general enforcing similar state acts could play a role in protecting the public (i.e., individuals as consumers) but have played little role to date.
Athena’s classification scheme for the 2007 Report and what was known in 2007
According to the lab’s stated seven classification options listed on the 2007 report (1. known disease-associated mutations; 2. predicted disease-associated mutations; 3. amino acid changes of unknown significance; 4. variant of unknown significance; 5. inconclusive; 6. indeterminate; and 7. benign), the original classification of the mutation (4. variant of unknown significance) was incorrect and should have been reported as “1. Known disease-associated mutation.” Athena apparently got the DNA analysis right (identifying the variant) but the review of the literature wrong, as there were already two publications that associated this specific mutation with Dravet syndrome (Berkovic et al., 2006 and Harkin et al., 2007). According to Athena’s own classification scheme and glossary, these available publications required the variant identified to be classified as a “known disease-associated mutation.” Moreover, while the classification scheme of the 2007 report described a VUS as one “similar to those observed in disease-associated mutations and benign polymorphisms,” no benign observations (i.e., observations of this variant in individuals without Dravet syndrome) had been reported to the reliable databases, according to the plaintiff.
While it is undoubtedly challenging for laboratories to stay atop the literature for all conditions in real-time, the 2007 report lists as “pending” a manuscript that had already been published in 2005, suggesting that Athena’s integration of biomedical literature was at least two years behind. In addition, the lab appeared to be unaware of a publication coauthored by one of its own directors: Dr. Sat Dev Batish was a co-author of the Harkin et al. 2007 paper that was published prior to the issuance of the June 2007 report and which associates this specific mutation with Dravet’s syndrome. Moreover, the two relevant publications (Berkovic et al. 2006 and Harkin et al., 2007) apparently came out of work subject to a patent owned by Bionomics (U.S. Patent #7,078,515) for the identification of SCN1A mutations that was licensed and used by Athena, and Christian’s specific mutation (1237T>A, Y413N) was cited in Patents #8,129,142 and #7,709,225, which were exclusively licensed to Athena for SCN1A testing.
Early development of the case
In April 2016, the defense attorneys filed a motion to dismiss the lawsuit and memorandum of law in support of their motion. In addition to denying the allegations and arguing that the pleadings lacked specificity, the motion to dismiss was a strategic attempt to reframe the plaintiff’s claim of ordinary negligence as a claim of medical malpractice. This switch is important, because in South Carolina medical malpractice claims, in addition to caps on damages (e.g., S.C. Code Ann. §15-32-220), are subject to a 6-year statute of repose (S.C. Code Ann. §15-3-545(A)) in addition to a 3-year statute of limitations (explained below). In reframing this case as one of medical malpractice, the defense attorneys argued that the statutes of limitations and repose bar the case from proceeding, as it has been nearly nine years since the June 2007 Report and over eight years since the boy died. The defense tries to position Quest and Athena as “licensed health care providers,” arguing that Athena’s CLIA accreditation and the collaborative role labs play with medical doctors justify Athena’s meeting the definition. Importantly, as has been reported elsewhere, in addition to CLIA accreditation, Athena is accredited by the College of American Pathologists (CAP) and has a clinical laboratory permit in NY. The plaintiff’s attorneys filed a response to the motion to dismiss as well as an amended complaint in May 2016.
The basic procedural issues to consider include the determination of which date(s) start the clock for the applicable statute of limitation, the discovery rule, and constructive notice. Generally speaking, the statute of limitations and statute of repose are affirmative defenses that are to be raised early, as the defendants did in this case. A statute of limitation bars plaintiffs from bringing claims after a specified period of time passes after an injury. A statute of repose is stricter than a statute of limitation in that it bars plaintiffs from bringing claims after a specified period of time from the defendant’s actions, even if the injury has not yet occurred or been discovered. The discovery rule tolls (or suspends) the statute of limitation until the date when the person discovers or until a person with common knowledge and experience through reasonable diligence should have discovered the facts and circumstances giving rise to the cause of action; however, the discovery rule does not toll the statute of repose that applies to medical malpractice actions. When a plaintiff has had actual or constructive notice (that is, a situation in which a person will be considered by the courts to have had knowledge even if he/she did not actually know about the matter), the extended time provided by the discovery rule ends and the statute of limitations starts running. Notice is a difficult argument to make in the context of clinical genomic testing: it has not been customary for patients to obtain lab reports directly, and it is not settled whether patients who do not receive an original copy of the lab’s report should be deemed to have notice if one of their many medical providers was sent report. Moreover, the reasonable inquiry a patient can perform once put on notice is questionable if the information necessary to answer questions (such as access to variants and literature) is closed off in proprietary databases or locked behind paywalls.
Here, there is a debatable legal question as to which date starts the clock. Is it the date of the first report in June 2007; the date of the boy’s death in January 2008; or the date the revised report was given to the plaintiff in January 2015 (which allegedly alerted her for the first time to the original 2007 variant classification as well as the reclassification)? Is it possible that a new, separate clock (rather than a delay of the original clock) would start each day following the issuance of the 2007 report given the apparent post-analytic system failure is a daily recurring failure at the laboratory level to catch erroneous classifications and notify affected parties (the medical provider who ordered the test and the patient to whom it relates)? It is reasonable to argue, as a matter of law, that facts demonstrating a clinical lab’s failure to comply with post-analytic system requirements imposed by CLIA exemplify an ongoing injury, a “fresh injury each day” that starts a new clock for bringing a cause of action in negligence. When an action is continuing and abatable, meaning capable of being stopped or mitigated (like catching a mistake or issuing a required update), each separate instance gives rise to a new claim with a new clock. The plaintiff’s attorneys argue the important date is that of the revised January 2015 report and that, ultimately, the question of when exactly Williams knew or should have known (January 2015 or earlier) is for the jury to decide. The defense argues that the statutes of limitations and repose began running in June 2007 when the original report was issued and that various technical statements, disclaimers, and what the defense refers to as “emphatic warnings” in the June 2007 report as having put Williams on actual or constructive notice. The defense attorneys argue in their memorandum of law that a reasonable person would have been on notice that the June 2007 report was a “tentative conclusion” requiring some sort of additional inquiry and, therefore, the statute of limitations started running then.
Of errors and updates: are industry-wide practices safe enough?
This case is based on an alleged lab error in the original variant classification, rather than in an update or reinterpretation. Yet a major reason this case is worthy of attention is the issue of what follow-up duties a clinical genomics lab has. It has not been industry custom for clinical labs to follow up with the doctors who have ordered genetic tests or the patients for whom the tests were ordered. Moreover, labs (and some medical providers) have balked at requests that patients have full access to their genomic data and results. Yet as Judge Holmes explained in Texas & Pac. Ry. v. Beymer, 189 U.S. 468, 470 (1903), “What usually is done may be evidence of what ought to be done, but what ought to be done is fixed by a standard of reasonable prudence, whether it usually is complied with or not.” (citing Wabash Ry. Co v. McDaniels, 107 U.S. 454 (1883)). Genomics Law Report editor John Conley noted previously that TJ Hooper, 60 F.2d 737 (1932), is a “good model” for a case like this. That case, finding tugboats without reliable radios to be unseaworthy even though equipping tugboats with such safety devices was not industry custom, is known for the legal rule that industry does not determine the standard of care. According to the opinion, by Judge Learned Hand, “…in most cases reasonable prudence is in fact common prudence; but strictly it is never its measure; a whole calling may have unduly lagged in the adoption of new and available devices. It never may set its own tests, however persuasive be its usages. Courts must in the end say what is required; there are precautions so imperative that even their universal disregard will not excuse their omission.”
Whether it is industry custom to do so (or not) or whether CMS is holding them accountable (or not), clinical genomics labs have a statutory obligation under CLIA—for all tests performed in the lab and for the benefit of all individuals whose samples are analyzed or whose judgment is relying upon the reports the lab issued—(1) to have systems in place to ensure proper contact and contextual information is on file and to ensure that protocols (including variant classification schemes, regardless of the clinical validity of those schemes) have been followed, and (2) have post-analytic systems in place to catch deviations from those classification schemes that might have occurred for whatever reason, perform updated literature scans on variants that have been classified, and take reasonable steps to notify those individuals affected by a specific changed classification (whether an identified mistake, a system-wide adoption of new classification schemes, or an update discovery from the literature).
The defense has suggested that the boy’s physicians may be responsible rather than the lab. Yet the causes of action are based on deviation from or non-compliance with the CLIA regulatory requirements for pre-analytic, analytic, and post-analytic mechanisms which had a foreseeable and direct effect on the boy’s health and safety. This potential significance of this case is that it might bring into focus the risk to public health and safety when laboratory-level mechanisms for quality control (distinct from specific test-level mechanisms that ensure clinical validity and utility) fail or are inadequate. There are multiple potential paths for liability in clinical genomics when something goes wrong (e.g., regulatory noncompliance by the laboratories, medical malpractice by doctors, etc.). It is not yet clear whether a court would consider the actions of doctors after the genetic analysis was performed and a report was issued to be sufficient interruptions of the chain of causation to effectively release labs from legal responsibility for mistakes made in the report’s generation. It also is not yet clear whether a court would find that the standard of care for the industry, regardless of custom, requires that labs deposit genomic variants into a medical data commons and take extra safety measures (like reasonable follow-up and notification of updates or changes).
In terms of regulatory oversight, this case reminds us of how grey this legal area remains. Where does a basic lab function end and the practice of medicine begin? If laboratories are licensed health care providers, at what point have they fulfilled their duties to the individuals served—upon the issuance of a report, or is follow-up notification of some sort required to ensure public health and safety?
Regardless of which party the legal system determines is ultimately responsible (if it reaches such a decision), there are proactive steps that each actor in genomic medicine can take to improve the process.
Patients who have had any genetic testing for medical purposes through any lab might want to (1) request copies of their data and results reports if those have not yet been provided to them; (2) seek a second, updated opinion that (a) verifies the absence or presence of variants and (b) rescans the biomedical literature to get the classification of that variant correct according to both the original and currently used classification schemes; and (3), if there are continued medical mysteries, ask themselves whether they should revisit their doctor(s) to determine if a genetic test upon which the doctor relied in rejection or pursuit of other possible diagnoses or treatments can be relied upon.
Medical doctors who rely on genetic testing reports might also want to seek training to become competent in reviewing genetic reports; however, it is not yet clear whether continuing education courses would provide meaningful competence or simply heighten the standard of care expected of those who have completed such training. It is perhaps more prudent for doctors relying on genomic test reports to develop trusted relationships with close colleagues who have the appropriate genomics training to provide consultations.
Regardless of whether the lab’s tests are FDA-approved or not, each lab must ensure its operations adhere strictly to its own classification schemes, keep its literature scans up-to-date, and as a matter of practice confirm that the individual (patient) as well as the ordering medical doctor has received a copy of the full report. Laboratories also would be wise to take meaningful steps to fulfill their CLIA obligations to notify those previous patients/customers whose reports were done under previous classification schemes when a reclassification for current/future patients/customers has been made for a particular variant. Even a mass notification on a public-facing website would be preferable to silence.
Finally, those responsible for CLIA inspections must devote sufficient resources to the task of ensuring that pre-analytic, analytic, and post-analytic protocols are being followed. Oversight authorities must do a better job of focusing on these three very serious questions: (1) How many people who have received genetic testing from a particular lab harbor a variant that was identified but has since the initial report changed classification? (2) What is being done by the lab to identify these individuals? and (3) What is being done to notify those people and their medical providers?
• March 14, 2016, Turna Ray, “Mother’s Negligence Suit Against Quest’s Athena Could Broadly Impact Genetic Testing Labs” GenomeWeb.
• April 1, 2016, Turna Ray, “Quest, Athena Make Moves in Wrongful Death Lawsuit,” GenomeWeb.
• April 5, 2016, “Quest, Athena File Motion to Dismiss Negligence Suit,” GenomeWeb.
• May 16, 2016, “Plaintiffs Argue Against Quest, Athena’s Motion to Dismiss Negligence Case,” GenomeWeb.