Identification of Neuropsychiatric Copy Number Variants in a Health Care System Population

This cohort study evaluates the suitability of including pathogenic copy number variants associated with neuropsychiatric disorders in population screening by determining their prevalence and penetrance and exploring the personal utility of disclosing results.

burden of disease that may be caused by NPD CNVs. Five CM categories were selected based on relevance to known CNV-related phenotypes and highest confidence in the accuracy of EHR-derived ICD-9/10 codes (eTable 2). A genetic diagnosis of a NPD-related CNV in the EHR was identified by the presence of an ICD-9/10 billing code indicating a chromosomal/structural abnormality (eTable 2). Tests of association were performed with logistic regression adjusted for age and sex. Thirteen individuals, missing sex information in their EHR were excluded from these analyses. Reported p-values are Bonferroni corrected for the nine tests of association performed in the study. All statistical analyses were performed using R version 3.4.1. Penetrance was calculated using the formula below: CNV penetrance = Number of cases with CNV Number of controls with CNV

Disclosure Protocol and Assessment of Participant Responses
A subset of nine pathogenic CNVs were prioritized to be returned to participants, based on the frequency in our dataset and the likelihood that the CNV could also cause non-NPD health considerations (e.g., congenital cardiac anomaly) that would be more likely to have implications for medical care. ICD9/10 codes were manually reviewed for individuals with a prioritized CNV to assess for evidence of an NPD phenotype. Individuals with one of these CNVs and a documented NPD, including depression and anxiety, in their EHR were eligible for clinical disclosure of results. DNA from samples maintained in the MyCode® biorepository CLIA environment were sent for orthogonal confirmation in a CAP-and CLIA-certified clinical laboratory and 141 (100%) results were confirmed. The clinical confirmation test report was scanned into the EHR and the disclosure protocol initiated.
We developed a results disclosure process to return prioritized NPD-related CNVs to adult patientparticipants ( Figure 2), based on the existing MyCode® GSC program 10 . The proposed CNV disclosure process and returnable CNVs underwent review by Geisinger's Ethics Advisory Council, Clinical Oversight Committee, Genomics Council, and MyCode® Governing Board prior to implementation.
Three licensed certified genetic counselors experienced with NPD were responsible for contacting and counseling participants about results, facilitating cascade testing for relevant family members, and providing clinical guidance to other healthcare providers. Initiation of result notification occurred from March 2017 to March 2019. A disclosure session outline (eFigure 1) was developed to ensure consistency among genetic counselors and to guide psychosocial assessment of participants' reactions. Particular consideration was given to reactions upon learning about a genetic NPD etiology and the effects of cognitive and psychiatric symptoms on the comprehension of information provided. Discussion included a description of the CNV, inheritance, knowledge of the associated clinical manifestations, the variable expressivity inherent to brain disorders, and the complex interplay between high-impact genomic variants and environmental factors. Personal medical and NPD histories and family histories were elicited, including participants' experience with NPD and treatment history. Participants were asked for their thoughts about communicating with relatives about their result and about whether and how the result was valuable to them. Genetic counseling disclosure sessions were documented in participants' EHR, with detailed three-generation pedigrees that include NPD and other brain disorders; relevant medical articles and healthcare guidelines were also provided to participants' primary healthcare providers.
A mixed-methods approach was utilized to evaluate a sub-set of participants' immediate responses to receiving NPD-related results, including qualitative analyses of written genetic counselor post-disclosure assessments of participant responses and audio recordings of in-person disclosure sessions. Retrospective chart reviews of disclosure session notes were conducted, if needed. The post-disclosure assessments were implemented initially to reduce participant burden and were eventually phased out after this analysis. Audio-recording of inperson disclosure sessions was implemented, with participant written consent (IRB #2017-0273), after results had been reported to the first group of 34 participants and remains on-going. Detailed thematic analysis of a larger dataset of audio-recorded transcripts, which is consistent with this subset, will be reported separately (Wain et al., in preparation), but major themes from preliminary analyses are reported here. Post-disclosure assessments were completed for all participants who also consented for audio-recorded sessions. Datasets (audio-recorded transcripts and genetic counselor post-disclosure assessments) were assessed independently using a grounded theory approach 11 . Genetic counselor post-disclosure assessment data were available for both in-person and telephone disclosures (eFigure 1). For each dataset, two independent coders generated codes for themes and subthemes, with discussion to reach consensus and develop final codebooks. Themes derived from the genetic counselor assessments were also discussed with the entire GSC team.
We took our working definition of "personal utility" from Bunnik et al., 2015 who provide the following: "genomic information has personal utility if and only if it can reasonably be used for decisions, actions or selfunderstanding which are personal in nature." They go on to clarify that "personal utility can be (indirectly) related to health and disease 12 ." This definition informed our interpretation of identified themes which we determined to be related to personal utility.