Practical Considerations for Reinterpretation of Individual Genetic Variants

Abstract

With the growing use of genetic testing in medicine, the question of when genetic findings should be reinterpreted in light of new data has become inescapable. The generation of population and disease-specific data, development of computational tools, and new understandings of the relationship of specific genes to disorders can all trigger changes in variant classification that may have important implications for patients and the clinicians caring for them. This is a particular concern for patients from groups underrepresented in current reference datasets, since they have higher rates of uncertain findings. Here we identify the challenges to implementing a systematic approach to variant reinterpretation and propose solutions. In particular, we address (a) the infrastructure needed to support implementation of systematic variant reinterpretation, (b) the issues around obtaining consent from patients for reinterpretation, (c) the process for triggering reinterpretation, (d) pathways for the flow of reinterpreted data, (e) considerations for how to cover the costs of reinterpretation, and (f) practical issues related to implementation of processes and policies that address these issues, including the importance of a fixed duration during which there is an expectation that variants will be reinterpreted.

Introduction

The rapid increase in genetic testing in healthcare has created a large volume of genetic variants, some of which can pose challenges in classification. As additional population and disease-specific data become available, computational tools evolve, and functional data are generated, variant classification can change. For example, of 100 participants who underwent genomic sequencing in the MedSeq study, 13% received updated interpretations of originally identified variants after a mean of 13 months from when the original results were generated (1). Similarly, in a study of women who had undergone BRCA1/2 testing, 12.4% of variants were reinterpreted over a 5-year period (3). In another study, reinterpretation of reported variants in 185 pediatric patients with epilepsy two to five years after initial testing resulted in reclassification in 36.2% of cases (2). In these and similar studies, the majority of reinterpretations involve variants of uncertain significance (VUS), most of which are reclassified to likely benign/benign. These examples illustrate the dynamic nature of sequence data interpretation and highlight the importance of tracking, updating and reporting these data.

“Reinterpretation” as used in this article refers to the process of reevaluating genetic variants that have been previously analyzed/interpreted to determine if there is new information that may alter their classification (4). “Reclassification,” which can be an outcome of reinterpretation, denotes the process of changing the designated genetic variant classification per the American College of Medical Genetics and Genomics (ACMG)/ Association for Molecular Pathology (AMP) guidelines to or from pathogenic/likely pathogenic/VUS/likely benign/benign (5). We focus on reinterpretation of previously reported variants from clinical testing, since reanalysis of exome and genome data in their entirety, somatic variants, and data from reproductive screening raise issues beyond those addressed here. Although technical aspects of reinterpretation from the point of view of the laboratory have been considered previously (4), there are other important stakeholders (patients, clinicians, and payers) beyond the laboratories to be considered, along with ethical and legal concerns.

An illustration of why reinterpretation may be necessary comes when a new association, or lack of association, is established between a gene and a disease. When sufficient evidence accumulates to support that association or lack thereof, the interpretation of all variants in the gene may need to be reconsidered. Conversely, rigorous assessment of some genes on gene panels may undermine presumed disease associations. For example, ClinGen groups have contested associations of all but one gene associated with Brugada syndrome (6), necessitating the reinterpretation of many variants.

Reclassification can have different downstream clinical implications depending on the clinical context and whether a disease is preventable or treatable, including whether preventive interventions, such as surgery, which may be irreversible and potentially life-altering, are indicated. This has consequences not only for the individual tested but also for family members whose risk status for diseases such as cancer, cardiomyopathies, and arrhythmias—all of which can be life threatening—may be very different when viewed in light of reinterpretation for their family member.

Reinterpretation of genetic variants disproportionately affects—and is likely to disproportionately benefit—individuals of non-European ancestry. Higher rates of reclassification for these groups are due in part to the increased likelihood of a VUS, a result of inadequate representation of these groups in current genomic data. Although the guidelines for variant interpretation from the ACMG/AMP state that no medical decisions should be made on the basis of a VUS (5), there is evidence that VUS results may lead to unnecessary and potentially harmful medical procedures or impede preventive screening (7). In a large study of 10,000 patients undergoing clinical gene panel testing for hereditary cancer, 22.7% of participants of European ancestry had a VUS reported, compared to 37.3% with Asian ancestry and 39.7% with Black or African American ancestry (8).

Most VUS that are reclassified are determined to be likely benign or benign, although some are upgraded to likely pathogenic or pathogenic (9). However, not all reclassifications involve VUS. By tracking nearly 1500 variants in cancer genes originally classified in a category other than benign and reported over a 20-year period, researchers found an overall reclassification rate of 18.1%, with patients from non-European ancestries having higher rates of reclassification (ranging as high as 42.9% in BRCA1/2 and 24.4% for 40 other cancer genes in patients of African ancestry) compared with patients of non-Hispanic European ancestry (27.9% and 7.2% respectively) (10). Although reclassification of variants initially identified as pathogenic/likely pathogenic is rare, some have been reclassified to VUS, a change that can be associated with regret for patients and their family members, especially those who have undergone irreversible surgical intervention based upon what they believed to be certain information (11). While genetic testing can often provide benefit, the field of human genetics is still evolving, leaving us with a dilemma about what to do with previously reported results as we improve the accuracy of our interpretations.

The continued evolution of our understanding of human genetic variation has prompted calls for routine reinterpretation of individually reported variants, an issue some laboratories have already started to address (12). In the United States, at least one large commercial laboratory has announced that “reclassification alerts to clinicians will be sent when a clinically significant variant classification update occurs.” (13) A report of interviews of genetic clinicians in Europe, Australia, and Canada suggests that laboratories in other countries are also beginning to adopt such practices (14). Moreover, some argue for an ethical duty to periodically reinterpret genetic variants while the state of knowledge is still in flux (15).

We recently conducted a research study funded by the National Human Genome Research Institute (NHGRI), involving a series of focus groups and surveys with stakeholder groups (patients and parents of patients who have undergone genetic testing, clinical genetic providers, laboratory genetic providers, and non-genetic providers in specialties that frequently use genetic tests) as well as workgroups considering the clinical, ethical, legal, and economic aspects of variant reinterpretation (16). Although no courts have imposed liability on laboratories or clinicians for failure to reinterpret variants or issue revised reports (17), these stakeholders expressed strong support for routine reinterpretation of reported variants, at least under some circumstances, with no fixed duration of the obligation (16). Given the likelihood that variant reinterpretation will become increasingly common, there is a need to develop consensus on a practical framework to apply during this period of dynamic change in genetic knowledge. Some aspects of recontacting patients were addressed previously in an ACMG “points to consider” document (18), but here we address in greater detail the specifics for how systemic variant reinterpretation could be implemented; the separate issues of whether such reinterpretation is ethically desirable (15) or legally required (17,19) have been addressed in other publications.

The framework proposed here highlights the issues that laboratories, clinicians, and other stakeholders need to consider with regard to variant reinterpretation and, where possible, suggests potential approaches. The views embodied here reflect perspectives gathered as part of the study referred to above.

In the sections that follow, we address (a) the infrastructure needed to support implementation of systematic variant reinterpretation, (b) the issues around obtaining consent from patients for reinterpretation, (c) the process for triggering reinterpretation, (d) pathways for the flow of reinterpreted data, (e) considerations for how to cover the costs of reinterpretation, and (f) practical issues related to implementation of processes and policies that address these issues, including the importance of a fixed duration during which there is an expectation that variants will be reinterpreted.

Infrastructure Needed to Support Implementation

Variant reinterpretation at scale and at reasonable cost requires affordable, automated methods and databases to compare reported classifications of variants with new data and with classifications by other laboratories. There are growing numbers of patients being tested for an increasing number of genes, and while algorithms to automatically identify high-yield variants for reinterpretation will be crucial, some cases will require resource-intensive manual review by an analyst. The publicly available ClinGen and ClinVar databases could be used to support reinterpretation, especially as evidence accumulates to: 1) assess the strength of association with diseases of specific genes included in gene panels and 2) perform and summarize evidence reviews as groups submit interpretations and supporting evidence about variants to these databases. Laboratory databases will require the ability to cross reference variants across cases to efficiently manage the task and therefore the software that laboratories use must be configured in a way that is variant-centered and not limited to sorting only by cases.

If laboratories have the primary responsibility of triggering reinterpretation (as we suggest in Triggering Reinterpretation below), it will be important for laboratories to build and maintain databases of their own variants and interpretations and contribute their data to ClinVar to allow candidates for reinterpretation to be identified as new information becomes available. They will also need to establish parameters that trigger variant reinterpretation (e.g., thresholds for allele and/or homozygous genotype frequency data in ancestrally diverse cohorts to consider reclassifying VUS to likely benign, or recurrences of rare missense variants exclusively in affected individuals to reclassify VUS to likely pathogenic). Functional data assessing a large allelic series in a gene could also trigger reinterpretation of variants for that gene. With a standardized data model of variant-level information, including both variants and accompanying variant metadata, it should be possible to automate identification and reinterpretation of variants with discrepant interpretations in ClinVar (i.e., after completion of a ClinGen expert review) and to aggregate evidence to support reinterpretation, using parameters that can be specified by the laboratory. Changes in ACMG/AMP variant classification guidelines could also trigger reinterpretation of variants. Specialized software for this process will be an essential component of the infrastructure for variant reinterpretation and should be a priority for the genetics community. Its development could be funded by the National Institutes of Health with the stipulation that it remain open source. This software could be used by laboratories to identify variants and cases in their internal databases for reinterpretation based upon adjustable parameters they set. Establishment of standardized variant databases within each laboratory would be required and can be time- and resource-intensive and requires harmonization of the variant nomenclature, genome build or transcript ID with ClinVar or other gene-specific databases and genomic resources. This will ensure that variants are annotated similarly across data sets so that the most accurate and up-to-date data are applied to each variant for interpretation. As new genome builds are established, this process will also require infrastructure and resources to lift over data to the new reference genome. Collaboration across genomic resources will ensure there is robust interoperability that can scale over time.

It is inevitable that some laboratories that perform genetic testing will go out of business, given rapid changes in technology and market structure, making it impossible for them to sustain variant reinterpretation. Potential approaches to preserve their data and to allow reinterpretation to occur when needed are important to consider. At a minimum, ensuring that all laboratories submit their variant data to ClinVar will at least allow data generated in defunct laboratories to continue to inform the interpretation process.

Obtaining Consent for Reinterpretation

Variant reclassification can lead to changes in medical care—including screening, prevention, and treatment—and may have implications for cascade testing of family members or reproductive decisions. This may lead to changes in perceived vulnerability to serious medical conditions and psychological distress. Hence, patients should be made aware of the possibility that their genetic test results might be reinterpreted at a future time. We suggest that patients who consent to genetic testing will want the most accurate possible results from their tests and therefore can be assumed to want results updated. The laboratory workflow is more readily scalable if all patients are included in reinterpretations; otherwise, a lab could have an incongruity in the interpretation of the same variant across patients—even potentially patients within the same family. The clinical workflow is also most effective when the clinical laboratory report reflects the most current interpretation of a genetic variant for all patients, lest clinical decisions be based on outdated information. The most efficient way to inform patients of the potential for changes in variant interpretation over time is at the time the original consent for testing is obtained. Patients should be told concisely, both in writing and in the oral discussion, about the possibility that new information could change the interpretation of their results, the timeframe during which this might happen (see Covering the Costs of Reinterpretation below), and that, should there be a reclassification, the clinician who ordered the initial test will be contacted with an updated report that should become part of the patient’s health record. They should also be told how the information will be shared with them (see Pathways for the Flow of Reinterpreted Data below). Consent and requisition forms for genetic testing should be designed to include information about reinterpretation, including the parameters under which an updated report may be issued and any fixed duration during which reinterpretation would occur.

The consent process also offers the opportunity to inform patients about the importance of keeping their clinicians, especially the clinician who ordered their genetic testing, informed about changes in their contact information, so that they can be notified about reinterpreted results in the future. If patients do not maintain current contact information with their clinicians, they are essentially opting out of receiving reclassifications. Patients should be offered the opportunity to identify a family member who can receive the results in the event that the patient is deceased, incapacitated, or unable to be reached when a new variant classification is available. This option can be particularly important when findings indicate familial risk for cancer and other serious but preventable conditions. This information could be retained by the ordering clinician to guide the return of updated results when a new classification is received in the future. When the patient is a minor, permission will need to be obtained from a parent. Minors who become adults during the period when reinterpretation may occur (see Covering the Costs of Reinterpretation below) could be informed that, in the event an updated report is received, their parents will be contacted to find out how the now-adult patient can be reached so the information can be offered directly to the patient.

Triggering Reinterpretation

Many factors can trigger the need for reinterpretation. These include changes in the patient’s health status, pregnancy/plans for pregnancy, availability of updated family history, variant segregation studies, new knowledge about the clinical significance of the gene/variant based upon internal laboratory data or publicly available data, and the availability of newly approved therapies or clinical trials. In general, clinicians are best positioned to appreciate the relevance of changes in health status or family history, while laboratories are most aware of changes in the classification of variants they have reported and have access to the informatics tools to monitor public databases efficiently for new information. We suspect that reinterpretation will most often be initiated by laboratories, but it will remain important for the ordering clinician or current treating clinician to be able to request a reinterpretation when clinically indicated, and there may be circumstances in which patients themselves will seek to initiate the process in collaboration with their provider. Such requests for reinterpretation may be infrequent, especially if the laboratory will trigger reinterpretation in a systematic way and so informs ordering clinicians. There may be repeated requests by the same party, although limits could be considered to contain costs, with requests exceeding the limit resulting in additional charges to the patient.

We note that individuals in historically marginalized groups, who on average have lower health literacy (20) and higher rates of economic disadvantage (21), are likely less well-situated to request reinterpretation on their own. Moreover, they also have less consistent health care utilization with the same clinician (22), reducing the likelihood that someone who is following them closely will be in a position to trigger requests for reclassification based on changes in their clinical state or family history. Thus, systematic laboratory-centered processes to trigger reinterpretation that include allele and/or genotype frequencies from large cohorts of individuals not of European ancestry may be more beneficial.

Triggers for a laboratory to initiate reinterpretation may be driven by newly generated data from ClinGen and ClinVar. However, other data sources that may be useful as indicators of the need for variant reinterpretation include expansion of reference population data sets, especially those from individuals not of European ancestry, functional studies for variants in specific genes, gene-specific databases, and large disease-specific gene discovery programs that validate new disease-gene associations. Methods are continuously being developed to enhance interpretation of noncoding variants, splicing, missense, and in-frame insertion/deletions, and improvement in these prediction tools could also provide support for reassessment of some variant types. Laboratories may elect automatically to incorporate data feeds from FDA-recognized sources, as well as newer sources not yet FDA-approved. Additionally, there will likely be ongoing revisions to the ACMG/AMP and other professional guidelines for variant interpretation that would lead to reclassification of some variants.

Triggering variant reinterpretation should be done judiciously to minimize the burden on the laboratory, since human involvement in the reassessment of the variant will be required even if much of the process is automated. Therefore, default thresholds should be set to ensure that reinterpretation has a high probability of leading to reclassification: e.g., a new ClinGen evidence review; publication of a robust, large, relevant data set; and/or if multiple laboratories issue a consistent interpretation that differs from the original interpretation. Although laboratories might set different parameters for variant reinterpretation, experience across laboratories should result in consensus criteria to trigger reinterpretation over time. Professional societies may also contribute to the development of such criteria. For certain diagnoses that are medically actionable, a lower threshold for variant reinterpretation might be considered given the greater importance of accurate classification. As the established default threshold might be considered insufficient or overly restrictive by the patient, the laboratory could offer a fee-for-service alternative to the standard process of variant reinterpretation that would be included with the initial testing charge (see Covering the Costs of Reinterpretation below). This alternative would have broader indications for reinterpretation.

Pathways for the Flow of Reinterpreted Data

The default flow of information is typically from the laboratory to the ordering clinician and from the ordering clinician to the patient. Since patients move and/or change clinicians over time and generally do not have unified health records, mechanisms may be needed for laboratories to be notified directly by patients of changes in the clinician who should receive updated interpretations. If an ordering clinician is no longer in practice or leaves the facility where the test was ordered and a reinterpretation arrives, a system will be needed for the facility to designate an alternative clinician to receive these results and notify the patient. Differences in the structure of clinical practices (from small private practice groups to large academic medical centers) may require different processes for support and oversight of clinician notification of reinterpretation. With the 21^st Century Cures Act (23), laboratory results in the electronic health record (EHR) are automatically pushed out to patients, providing infrastructure to automate communication of reclassifications if reports from genetic testing flow into the EHR. All of these complexities underscore the importance of emphasizing to the patient their responsibility for maintaining current contact information, particularly in cases in which the likelihood of variant reinterpretation is high or would be clinically impactful. The initial consent discussion provides an opportunity to discuss how patients will be notified about reclassifications, especially since patients can access their electronic medical records and test reports, which may include reinterpretations, through patient portals (18).

When an ordering clinician (or a designated alternate) receives a revised genetic test report, he or she should attempt to contact the patient using the last known contact information in the medical record. Given the indeterminacy regarding when a doctor-patient relationship ends for purposes like this, clinicians should attempt to contact the patient (e.g., by registered mail) even if they have not treated the patient for an extended period of time (24). However, it is not reasonable to burden the clinician with affirmative obligations to attempt to track down a patient who fails to respond to notification via their preferred means of contact. If the clinician is able to contact the patient, the clinician should offer to review the revised report with them—either remotely or in person—in the context of the patient’s medical and family history. Patients, of course, retain the right to decline to receive this information for reasons of their own, but should be told if the updated results will be added to their EHR. If the EHR automatically pushes all laboratory reports to the patient, as is increasingly the case, that redundancy in the system will help ensure that the patient is notified of a revised result, but may also make it more difficult for them to remain unaware of the result. If the patient is deceased or unable to be reached and has identified a designated recipient, the clinician should convey the revised report to that designated person if they can be contacted.

Covering the Costs of Reinterpretation

The simplest mechanism for covering the cost of single-variant reinterpretation and issuance of a reclassification report by the laboratory would be to build the charge for reinterpretation into the original charge for the test, as at least one large genetic testing laboratory is already doing (13). The charge would include any reinterpretations that are needed over a fixed duration of reinterpretation, such as five years. The cost of reinterpretation could then be spread across everyone being tested, minimizing the financial burden on the laboratory and the incremental cost for any particular patient. Although the incremental cost per patient is likely to be modest, payers may resist paying an extra, up-front charge for re-interpretation when there is uncertainty whether a specific patient will be among those who ultimately need reinterpretation services. In addition, payers would need to see evidence of the impact of reinterpretation, including its clinical utility (in terms of net patient health outcomes or impact on clinical decision making) (25, 26). Including reinterpretation services in the CPT code for the test itself thus poses a number of reimbursement challenges. Alternative approaches could also be considered – for example, laboratories could undertake to review all patients’ genetic results at least once, even without a specific trigger, or reinterpretation could be billed separately when it is performed using a reanalysis CPT code, with all the complexities that would entail.

A fixed duration for reinterpretation would offer laboratories and third-party payers (who frequently reimburse genetic testing charges) a clear basis for estimating costs involved (25). This also insulates laboratories from excessive burden and cost due to open-ended and potentially repeated reinterpretation. It reflects the reality that over time, as new genetic testing technologies are developed, the value of updating the original test results may diminish, as new versions of tests supersede previous testing methods. In the absence of robust data on the frequency of revised interpretations over time, any period chosen for the duration of the commitment to reinterpret will necessarily be somewhat arbitrary. Taking the above considerations into account, we suggest that a five-year period may be reasonable, however, laboratories should have the discretion to offer reinterpretation over a longer period of time. Importantly, whatever timeframe is adopted should be clearly communicated in the initial consent process and the test report. As data become available about the frequency and cost of reinterpretation, the period of time and/or the initial charge for the test may need to be adjusted accordingly.

Of course, it is not only the laboratory that will bear the costs of generating a reclassification report. The ordering clinician to whom the report is provided will face costs associated with transmitting the reclassification information to the patient and could bill for a clinical encounter to explain the reclassification to the patient/family. In some cases, the implications of the reclassification for the patient’s medical care will be obvious, for example, if a variant is downgraded from a variant of uncertain significance to likely benign. However, other cases that are less common may have more significant implications, such as a change in classification from variant of uncertain significance to likely pathogenic, for which the clinician may need to spend time exploring the medical literature or consulting with a genetics or other content expert to better understand the management implications of the change. The clinician will then need to reach out to the patient or, if the patient is unavailable or deceased, to the patient’s designee, as described in Pathways for the Flow of Reinterpreted Data above. Although there may be existing codes that can be used for billing purposes for these tasks, it may be preferable for insurers to adopt a specific code for this purpose. Among other benefits, a unique code for communicating reclassified genetic test results would allow system-level tracking of the frequency and cost of the process.

Practical Issues Related to Implementation

Implementation of any policy of routine variant reinterpretation should occur only after laboratories and clinical facilities have established the necessary infrastructure for this process. Moreover, although there would be value in reinterpreting results from previous testing and laboratories may choose to do so, there are clear advantages to phasing in this policy for newly ordered genetic tests on a prospective basis. This will allow patients to be informed of the possibility of reinterpretations prospectively and ensure a gradual ramp-up of reinterpretation activities by laboratories and clinicians. Medical centers and large outpatient practices, including genetic clinicians, non-genetic clinicians, and primary care clinicians, will be able to use that time to develop policies and procedures for their facilities, including for contacting patients when reclassified results are received and for disseminating reclassification reports to patients if their clinician has left the facility. If there are significant changes to the ACMG/AMP guidelines for variant interpretation in the future, including but not limited to changes to the current five categories, the guidelines should specify whether previously reported variants will be reinterpreted using the new categories.

How likely are the approaches outlined in this article to be adopted? Laboratories, clinicians, and patients alike are motivated to have the most accurate information possible to guide patient management and reproductive decisions for patients and their family members. Genetic testing laboratories may also be incentivized in part by the chance to enhance profits and market share; factors that differentiate them from competing laboratories, such as generous reinterpretation policies, may be viewed as an opportunity to make themselves more attractive to clients. Even so, clinicians may be reluctant to take on additional responsibilities that extend over time, especially in light of patient turnover. Payers are also motivated in part by financial concerns, especially in the near term, and want to ensure that genetic test results will drive appropriate health resource utilization. Ultimately, the costs and benefits of potential solutions and the parties who bear those costs will need to be aligned to motivate and drive an optimal solution for patients.

Conclusion

Accurate genetic variant interpretation, which can change in light of new knowledge, is becoming increasingly important for management and treatment of genetic conditions. It is critical that genetic information be equally available and medically useful to all patients, regardless of ancestry, and as current as possible. Great strides are being made to close current gaps in interpretation. Given the rapid advances in our understanding of the genome, establishment of systematic approaches to accommodate this new knowledge is crucial to the successful implementation of genetic medicine.

Table 1:

Components of the framework for routine reinterpretation of individual genetic variants.

Category	Details:
Infrastructure Needed to Support Implementation	• Development of open-source database and software supported by ClinGen and ClinVar to allow for affordable and automated ways of identifying variants with new data and with classifications by other laboratories  • Laboratories’ building and maintenance of databases of their own variants and interpretations, contributing data to ClinVar, and establishing parameters that trigger variant reinterpretation  • Standardized data model of variant-level information, including variants and variant metadata, to automate identification of variants with discrepant interpretations and aggregate evidence to support reinterpretation
Obtaining Consent for Reinterpretation	• Information for patients at initial consent for genetic testing about possibility of reinterpretation, time frame during which this may occur, process of recontact and how information will be shared with them; emphasis on importance of keeping contact information up-to-date  • Option to identify a family member who can receive results if the patient is deceased, incapacitated, or unable to be reached  • Minors who become adults during the period when reinterpretation may occur informed that if an updated report is received, their parents will be contacted to obtain contact information to offer them the results
Triggering Reinterpretation	• Reinterpretation most often initiated by laboratories, since they are most aware of changes in the classification of variants and have access to the informatics tools to monitor public databases for new information  • Requests from the ordering clinician, current treating clinician or patient possible, although limits could be considered to contain costs  • Initiation of reinterpretation triggered by data from ClinGen and ClinVar and potentially other sources  • Reinterpretation initiated judiciously to minimize burden on laboratories; default thresholds set to ensure that if reinterpretation triggered, high probability exists of a reclassification  • Lower threshold for reinterpretation considered for diagnoses that are medically actionable  • Experience across laboratories and guidelines from professional societies result in consensus criteria for initiation
Pathways for the Flow of Reinterpreted Data	• Default flow of information from the laboratory to the ordering clinician to the patient; mechanisms may be needed to allow for laboratories to be notified directly by the patient’s new clinician who should receive updated reports  • Systems will be created for facilities to designate an alternative clinician to receive results if a clinician leaves the practice  • Clinician attempt to contact the patient using the last known contact information on receipt of revised report  • Clinicians offer to review the revised report in the context of the patient’s medical and family history  • Clinicians attempt to contact designated recipient for deceased or unreachable patients  • Patients retain the right to decline to receive updated information for reasons of their own but should be told if the updated results will be in their EHR  • Integration of laboratory reports for genetic testing in the EHR and patient portal to ensure the patient notified of updated result
Covering the Costs of Reinterpretation	• Costs of reinterpretation and potential reclassification reports built into the original genetic testing charge  • Fixed duration for obligation to reinterpret genetic test results, perhaps 5 years  • Laboratories should have the discretion to offer reinterpretation over a longer period of time, but whatever time frame is adopted should be clearly communicated in the initial consent process and the test report  • Ordering clinician responsible for the costs to recontact patient; can bill for clinical encounter to explain results to patient
Practical Issues Related to Implementation	• Lead time needed to establish necessary infrastructure for this process  • Policy phased in prospectively, not applied retrospectively  • Development of policies and procedures by facilities for when clinicians leave the practice  • Mitigation of concerns about liability by systematic approach to variant reinterpretation with adequate infrastructure  • Laboratories, clinicians, payers, and patients motivated to drive optimal implementation