ABSTRACT
Introduction: The ACMG has recommended returning clinically relevant results for certain genes when identified in research or as secondary findings in diagnostic testing. Research studies have shown that genomic population screening detects patients with previously unrecognized and often actionable health risks or genetic conditions, with acceptably low levels of harm. Cascade testing of relatives at risk is enabled. Screening for recessive disorder carrier status with gene sequencing panels is common in clinical practice. Preventative screenings routinely occur in primary care settings. The cost of reliably sequencing of many genes in a clinically reliable fashion is approaching levels where offering genomic screening tests may be contemplated for entire populations, and the results used for preventative health purposes, including clinical correlation, early screening, and education. In anticipation of universal genome sequence-based screening, integrated with existing health risk screenings, we piloted a novel implementation of clinical genomic population screening in primary care, mostly family medicine clinics. Screening involved clinical sequencing and reporting of 431 genes where variants are associated with personal health risks or recessive disease carrier status. Methods: Interested primary care providers (PCPs) in two Family Medicine practice systems were invited to participate and given onboarding education. Adult patients with any health status were introduced to The Genomic DNA Test and provided test information by their PCPs in the context of preventative health assessment. Patient education materials included paper, online, and video information, a ‘hotline,’ and optional free genetic counseling. Patients completing a bespoke, health system-approved, written clinical consent provided blood or occasionally saliva samples that were NGS sequenced according to validated procedures in a commercial CLIA-certified genetic testing laboratory. Laboratory reports were returned to the PCP and patient after a local genetics professional added a 1-to-3-page messaging document, the Genomic Medicine Action Plan (GMAP). The PDF-format reports and GMAP were placed in the patient’s electronic health record. Only pathogenic (P) and likely pathogenic (LP) variants were reported. Variant classification was according to Sherloc, the performing laboratory’s system. Patients or providers could request free post-test genetic counseling locally, and the performing lab offered free family member testing and limited-cost partner testing for health risk panel genes and recessive disorder panel genes, respectively. Patients with health risk results were defined as being heterozygous for a P/LP variant for a dominant condition or for a recessive condition where some heterozygotes are symptomatic or co-dominant, hemizygous for a P/LP variant for an X-linked recessive condition, or bi-allelic and plausibly in trans for an autosomal (or X-linked in a female) recessive condition. Many such conditions that are common have reduced or low penetrance, and were characterized as increased risk compared to those not having those variants. When increased risk was identified, the GMAP recommended appropriate medical responses and/or patient education. As part of quality assessment of the pilot, the frequencies of reported results and certain events are monitored. Results: Between November 2019 and October 2021, 186 patients with a median age of 58 years were tested by 20 PCPs at no cost to patients or insurance. Testing volumes declined during the COVID-19 pandemic and when other health system events made high demands on PCPs and their staff. Only 13.3% of patients had no reportable variants in any of the 431 genes. Eighty point nine percent were carriers for at least one recessive disease. The most common recessive genes showing carrier status were HFE, SERPINA1, GALT, CFTR, BTD, F5, DHCR7, PC, GAA, GJB2, PMM2, PAH, and PKHD1. Twenty-six percent had at least one potential health risk result identified, 20% if the common thrombophilias are excluded. The most common category was hereditary cancer risk (7.5%), followed by F5, F2, and SERPINC1 thrombophilia variants (6.5%), hereditary hemochromatosis 1 (HFE) (4.3%), cardiovascular disorders, mostly cardiomyopathies (3.8%), alpha-1-antitrypsin deficiency or other pulmonary disorder (3.8%), familial Mediterranean fever heterozygotes (1.6%), G6PD deficiency (1.1%), and lipid disorder (0.5%). Two patients had health risks in two areas, and two in three areas. Interestingly, BRCA1 and BRCA2 variants were only identified in males. Thirteen patients, about 7%, had an amended report issued during the period. This happened when an unreported variant of uncertain significance was reclassified as LP or P, or when LP became P, and the performing laboratory issued an amended report. Surprisingly few patients took advantage of the free genetic counseling. No patient adverse events were reported by the participating PCPs despite ongoing outreach, nor by patients. Conclusion: Genomic population health screening can be successfully implemented in primary care settings with use of limited but essential genetic professional assistance, after careful planning and input from other medical specialties. A significant proportion of adults not selected for health status harbors germline genetic variants associated with increased health risk. In the absence of a culture where routine genomic screening is expected and where patient genomic competency is high, PCP capacity limits are a barrier to universality. Inclusion of genes for both health risk results with variable degrees of penetrance and for recessive carrier status, and multiple simultaneous results, dictates careful messaging of the implications, while doing so in a primary care setting begs a concise and efficient process. Rates of carrier detection were in-line with predictions based on general population frequencies. Rates of health risk detections were higher than earlier research programs because a larger number of genes with a much broader scope of health risk was included, including disorders with low penetrance yet meaningful clinical relevance and carefully-designed care pathways meant to optimize care while avoiding unnecessary additional testing. We conclude that genomic population health screening of primary care patients where large numbers of genes are clinically sequenced is feasible in a real-world health system, and that value exists for some tested patients now. Research to overcome certain technical limitations of current clinical genomic testing methods and to better stratify risk level in the context of incomplete penetrance should enhance the value of universally-offered genomic population health screening in the future.
ABSTRACT
Integration of genomics into health practice depends on successful implementation in non-research settings. We describe a medical home-centered implementation at the intersection of genomic medicine and population health in the UVM Health Network. In this clinical implementation, the hospital laboratory orchestrates a collaboration involving primary care providers (PCPs), patient and family advisors, health system administrators, clinical genetics services, oncologists and cardiologists, Vermont’s accountable care organization, and a commercial CLIA genomic testing laboratory. Phenotypically unselected adult primary care patients are offered “The Genomic DNATest” at no cost as part of their regular care. Testing is introduced by primary care providers and their staff using a brief animated video and printed decision aids with graded detail. Question resolution and pre- and post-test genetic counseling is offered at no cost using telephone, video, or in-person visits, and is coordinated bya single phone and email contact point, the Genomic Medicine Resource Center. 431 genes are sequenced for germline health risk and recessive carrier variants;only pathogenic and likely-pathogenic variants are reported. New reports are issued when reported and unreported variants are later reclassified. Test reports are reviewed by a clinical geneticist and genetic counselor. Two brief "action plans" are developed with PCP and patient focus in a single messaging document. This is prepended to the lab reports before release to the PCP, who reviews and then conveys them to the patient. PCPs and their staff receive initial training on the test and process and are invited to participate in an online community with monthly video case discussions. Among the first 72 patients tested, 17% had a health risk identified. This included dominantly inherited disorders and bi-allelic or hemizygous variants for common recessive disorders. Care pathways created in advance using multi-disciplinary expertise were activated for those. Free testing for blood relatives was made available. 76% of tested patients had at least one heterozygous recessive disease variant identified, and low-cost partner testingwas made available. Frequency of positive test results was in line with population frequency predictions. Pre- and post-test genetic counseling uptakewas lower than expected. This raised the question of unmet informational needs. A 2-page anonymous process quality survey mailed twice to the first 61 tested patients had a 31% return rate. Key findings included (1) pre-test engagement methods and decision aids were helpful;(2) the testing decision was influenced equally by value for the individual’s health, for their family’s health, and for researchers;(3) emotions during the ∼4-week time to results were neutral or excited, with none experiencing anxious feelings, and none reported the wait time as too long;(4) 21% reported contacting the Genomic Medicine Resource Center;(5) 16% reported referral to a specialist due to their result;(6) about half reported sharing the results with family members, but none reported any family members getting tested;(7) none indicated they were dissatisfied with the testing and result process, and only one responded they would not recommend others get the test;and (8) all agreed or somewhat agreed that the PCPs officewas the right place to do this testing.While this implementation was designed with scalability and a low management profile in mind, several systems-level barriers were encountered that contributed to lower engagement efforts and slower expansion than planned. This included lack of institutional information technology resources to surmount paper-based systems for requisitions, sample-routing, and consent forms;dependency of the patient engagement process during PCP visits on rooming and nursing staff during times of staffing shortages;susceptibility to practice model disruptions and priorities caused by the Covid-19 pandemic;and PCP time distraction resulting from user interface and polic changes in our EHR during the pilot. These barriers are targets for study and continuous process improvement activities. In summary, an example of clinical genomic population health testing using a medical-home focus has been successfully implemented in a non-research setting, supported by multi-disciplinary collaboration. This implementation depends on minimal staff, avoids financial barriers to access and genetic counseling, and offers a short, defined, test turnaround time as compared to similar biobank-based research programs. Tested patients find the program satisfactory, and meaningful test results are at least as common as in existing population health risk screening archetypes.