Clinically Actionable Findings Derived From Predictive Genomic Testing Offered in a Medical Practice Setting.

OBJECTIVE: To assess the presence of clinically actionable results and other genetic findings in an otherwise healthy population of adults seen in a medical practice setting and offered "predictive" genomic testing. PATIENTS AND METHODS: In 2014, a predictive genomics clinic for generally healthy adults was launched through the Mayo Clinic Executive Health Program. Self-identified interested patients met with a genomic nurse and genetic counselor for pretest advice and education. Two genome sequencing platforms and one gene panel-based health screen were offered. Posttest genetic counseling was available for patients who elected testing. From March 1, 2014, through June 1, 2019, 1281 patients were seen and 301 (23.5%) chose testing. Uptake rates increased to 36.3% [70 of 193]) in 2019 from 11.8% [2 of 17] in 2014. Clinically actionable results and genetic findings were analyzed using descriptive statistics. RESULTS: Clinically actionable results were detected in 11.6% of patients (35 of 301), and of those, 51.7% (15 of 29) with a cancer or cardiovascular result = did not have a personal or family history concerning for a hereditary disorder. The most common actionable results were in the BCHE, BRCA2, CHEK2, LDLR, MUTYH, and MYH7 genes. A carrier of at least one recessive condition was found in 53.8% of patients (162 of 301). At least one variant associated with multifactorial disease was found in 44.5% (134 of 301) (eg, 25 patients were heterozygous for the F5 factor V Leiden variant associated with thrombophilia risk). CONCLUSION: Our predictive screening revealed that 11.6% of individuals will test positive for a clinically actionable, likely pathogenic/pathogenic variant. This finding suggests that wider knowledge and adoption of predictive genomic services could be beneficial in medical practice, although additional studies are needed.

Assessing the stability of biobank donor preferences regarding sample use: evidence supporting the value of dynamic consent.

Dynamic consent has been proposed as a strategy for addressing the limitations of traditional, broad consent for biobank participation. Although the argument for dynamic consent has been made on theoretical grounds, empirical studies evaluating the potential utility of dynamic consent are needed to enhance deliberations about the merits of dynamic consent. Few studies have assessed such considerations as whether donor preferences may change over time or if participants would use a dynamic consent mechanism to modify preferences when they change. We administered a 66-item survey to participants in a large DNA biobank. The survey sought to gauge the stability of donor preferences specified at the time of biobank enrollment, specifically the stability of donors' preference regarding posthumous availability of biospecimens to next-of-kin. We received 1164 completed surveys for a response rate of 72%. Forty percent of respondents indicated a preference regarding sample availability on the survey (T2) that was inconsistent with the preference they had expressed when they enrolled in the biobank (T1). Most (94%) individuals with inconsistent preferences regarding sample availability had initially restricted sample availability at T1 but were comfortable with broader availability when asked at the time of the survey (T2). Our findings demonstrate that preferences regarding sample use expressed at the time of enrollment in a DNA biobank may not be reliable indicators of donor preferences over time. These findings lend empirical support to the case for a dynamic consent model in which biobank participants are approached over time to clarify their views regarding sample use.

Acceptability of Electronic Visits for Return of Research Results in the Mayo Clinic Biobank.

OBJECTIVE: To understand patient characteristics related to acceptability of returning individual research results via various modalities, focusing on electronic visits (e-visits). PATIENTS AND METHODS: Twelve hundred participants from the Mayo Clinic Biobank were selected using a stratified random sampling approach based on sex, age, and education level. Mailed surveys ascertained return of results preferences for 2 disease vignettes (cystic fibrosis and hereditary breast cancer) and a pharmacogenomics vignette. The study was conducted from October 1, 2013, through March 31, 2014. RESULTS: In all, 685 patients (57%) responded, and 60% reported liking e-visits, although the option of receiving results in an office visit was liked most frequently. Multivariable logistic models showed that the odds of liking the use of e-visits for returning results for cystic fibrosis and hereditary breast cancer were higher among those with higher education and better genetic knowledge and among those not living in proximity to the Mayo Clinic (Rochester, Minnesota). Level of genetic knowledge was not considerably associated with accepting e-visits, whereas education level remained important. For all vignettes, those who are divorced were less likely to accept e-visits. CONCLUSION: Researchers are faced with a difficult challenge of returning results with a method that is both acceptable to recipients and logistically feasible. This study implies that the use of e-visits may be a viable option for return of results to stratify the chasm between in-person genetic counseling and online portal receipt of results.

Variability in assigning pathogenicity to incidental findings: insights from LDLR sequence linked to the electronic health record in 1013 individuals.

Knowledge of variant pathogenicity is key to implementing genomic medicine. We describe variability between expert reviewers in assigning pathogenicity to sequence variants in LDLR, the causal gene in the majority of cases of familial hypercholesterolemia. LDLR was sequenced on the Illumina HiSeq platform (average read depth >200 × ) in 1013 Mayo Biobank participants recruited from 2012 to 2013. Variants with a minor allele frequency (MAF) <5% predicted to be functional or referenced in HGMD (Human Gene Mutation Database) or NCBI-ClinVar databases were reviewed. To assign pathogenicity, variant frequency in population data sets, computational predictions, reported observations and patient-level data including electronic health record-based post hoc phenotyping were leveraged. Of 178 LDLR variants passing quality control, 25 were selected for independent review using either an in-house protocol or a disease/gene-specific semi-quantitative framework based on the American College of Medical Genetics and Genomics-recommended lines of evidence. NCBI-ClinVar included interpretations for all queried variants with 74% (14/19) of variants with >1 submitter showing inconsistency in classification and 26% (5/19) appearing with conflicting clinical actionability. The discordance rate (one-step level of disagreement out of five classes in variant interpretation) between the reviewers was 40% (10/25). Two LDLR variants were independently deemed clinically actionable and returnable. Interpretation of LDLR variants was often discordant among ClinVar submitters and between expert reviewers. A quantitative approach based on strength of each predefined criterion in the context of specific genes and phenotypes may yield greater consistency between different reviewers.

Health behaviors and quality of life predictors for risk of hospitalization in an electronic health record-linked biobank.

BACKGROUND: Hospital risk stratification models using electronic health records (EHRs) often use age and comorbid health burden. Our primary aim was to determine if quality of life or health behaviors captured in an EHR-linked biobank can predict future risk of hospitalization. METHODS: Participants in the Mayo Clinic Biobank completed self-administered questionnaires at enrollment that included quality of life and health behaviors. Participants enrolled as of December 31, 2010 were followed for one year to ascertain hospitalization. Data on comorbidities and hospitalization were derived from the Mayo Clinic EHR. Hazard ratios (HR) and 95% confidence interval (CI) were used, adjusted for age and sex. We used gradient boosting machines models to integrate multiple factors. Different models were compared using C-statistic. RESULTS: Of the 8,927 eligible Mayo Clinic Biobank participants, 834 (9.3%) were hospitalized. Self-perceived health status and alcohol use had the strongest associations with risk of hospitalization. Compared to participants with excellent self-perceived health, those reporting poor/fair health had higher risk of hospitalization (HR =3.66, 95% CI 2.74-4.88). Alcohol use was inversely associated with hospitalization (HR =0.57 95% CI 0.45-0.72). The gradient boosting machines model estimated self-perceived health as the most influential factor (relative influence =16%). The predictive ability of the model based on comorbidities was slightly higher than the one based on the self-perceived health (C-statistic =0.67 vs 0.65). CONCLUSION: This study demonstrates that self-perceived health may be an important piece of information to add to the EHR. It may be another method to determine hospitalization risk.