Mental Health, Genomic Medicine and Patient Care for our Veterans Gerry Higgins, M.D., Ph.D.; Facilitator Chair, Genomics Advisory Panel, OSHERA TOPICS 1. Mental Health Services – Challenges: → Suicide → Post-Traumatic Stress → Treatment-resistant depression → Poly-pharmacy and comorbidity 2. The Importance of the Million Veteran Program and Genomic Medicine at the VA 3. Integration of Genome Data into the VA’s EHR: → Why it is important for the individual Veteran → Update, challenges and risks Rank order, Top Health Care Challenges in the Veterans Health Administration; Survey of 67 VA Clinicians, All Specialties at 24 Sites:1 1. Treatment-Resistant Depression 2. Treatment-Resistant Anxiety 3. Traumatic Injury 4. Addiction / Alcoholism 5. Posttraumatic stress 6. Cardiovascular Disease 7. Oncology 8. Gastroenterology 9. Dementia 10. Sleep Disorders 1. Conducted by Liberty Mutual Patient Safety Institute under contract with the VHA. Challenges - Suicide →“Suicide rates among Veterans treated in the VA System are about 50% higher than in the general population, and the rate of suicide among active duty service personnel has recently exceeded the rate in the general population.” →From DoD Suicide Event Report Program (2008-2010): Previous diagnosis of Mood Disorder including Bipolar Previous diagnosis of Anxiety Disorder Previous diagnosis of Major Depressive Disorder Previous diagnosis of Posttraumatic Stress 22% 19% 14% 6% Outpatient behavioral health care visit within past 30 days 21% Outpatient health care consultation visit within past 30 days Taking or previously taken psychotropic medications? 44% 23% Challenges - PTS → Using the strictest PCL criterion for OIF/OEF Veterans PRE-DEPLOYMENT Grouped: 3.0% (95% CI: 2.9-3.1) POST-DEPLOYMENT Support Unit 5.0% (95% CI: 4.9-5.2) Operational Unit 19.6% (95% CI: 19.1-20.2) Percentage diagnosed → Time-course using DSM-5 ‘Experiencing’ criteria 35 30 25 20 15 Unintentional Exposure to trauma Intentional Exposure to Trauma Israeli War Veterans 10 5 0 1 3 6 12 10 20 month months months months years years* *5.2% after 20 years Challenges – Treatment-Resistant Depression →Major depressive disorder (MDD) strikes 14% - 17% of Americans during their lifetime. → >30% of individuals with MDD do not achieve remission after 4 long antidepressant trials. → 22% of Americans view MDD as a ‘personal weakness.’ → MDD costs the U.S. economy an estimated $150 B each year. →Co-morbidities: CARDIO MDD SUICIDE PTS Poly-Pharmacy Goals of VA Genomic Medicine Program • Collect and link genetic information with VA Electronic Health Record and thereby: o o o Discover genetic predispositions, causes and mechanisms of disease Better define treatments Pharmacogenomic & interventional customization Via research, advance understanding in all these areas • Establish how genetic information will be used in clinical medicine o o Translational research to link genotype to phenotype Complex, adult, multi-gene diseases possibly with strong environmental influences • The Million Veteran Program o o ‘Whole genome sequencing & analysis’ to provide largest study of disease associations and human genetics Will help all studies examining how genome variants lead to disease and help understand gene x environment interactions Breakthroughs in Epigenomics may lead to understanding of how combat stress leads to PTS Epigenomics: A genomic approach to studying environmental effects, primarily DNA methylation, on gene function. Repetitive combat stress causes Epigenomic changes to genes in the stress response Breakthroughs in Epigenomics may lead to understanding of how combat stress leads to PTS Combat stress → Prolonged stress response Hypermethylation of glucocorticoid receptor PTS Breakthroughs in Epigenomics may lead to understanding of how combat stress leads to PTS → Patients with PTSD have increased plasma levels of cortisol, show a dysfunctional cortisol rhythm, and an exaggerated stress response. Plasma Cortisol Architecture for Pharmacogenomic Decision Support1: Infrastructure Overview of Required Elements Pharmacogenomic Decision Support 1.Computer processable medical knowledge 2.Computer-interpretable patient data 3.Generation of patientspecific advice using knowledge and patient data CDS Pre-Requisites Resources 1.Centrally managed repositories of medical knowledge 2. Standardization of CDS information for genomic medicine 3. Standard approach for representation and locating patient data → Rules Engine → MLM → Machine Learning component such as Support Vector Machine (SVM) → Remote web access input 1Based on: Kawamoto K et al. A national clinical decision support infrastructure to enable the widespread and consistent practice of genomic and personalized medicine. BMC Medical Informatics and Decision Making. 2009; 9:17 doi:10.1186/1472-6947-9-17. Architecture for Advanced Pharmacogenomic Decision Support: What’s Missing? Drug – Gene Interactions (DGIs) → Here are some reasons that the DGI domain has not received the focus it deserves: A. Drug-drug interactions are difficult to model, except in a generic manner, and once the algorithm has to handle many drug pairs (i.e., >2) it becomes an impossible combinatorial problem – especially as polypharmacy seems to be on the rise. To add a patient's variable metabolizer phenotype into the math, and the problem becomes intractable. B. The advent of next generation sequencing (NGS) has changed everything. Every vendor is very concerned that maintenance and upgrades will require tremendous resources to keep up with the very rapidly world of ‘omics. C. Physicians are already upset about medication alert fatigue, and since they were never thoroughly educated in genomics or pharmacogenomics, vendors are uncertain about how to proceed. In addition, both professional medical organizations and payors are still somewhat skeptical about gene testing. Preemptive Genotyping Store pharmacogenomic data in EHR High Risk Patients in Medical Home Obtain known actionable pharmacogenomic variants for each patient using multiplexing / NGS panels Trigger EHR system automatically provides physician with optimal therapeutic regimen for patient Advanced pharmacogenomic decision support Find possible risk variant Search for pharmacogenomic data in patient’s EHR Inpatient Gets New Prescription Possible integration of omic data directly into the VistA electronic health record 1. How could this be achieved? 2. Since epigenomic data, which has been highly replicated, shows a correlation between hyper-methylation of the glucocorticoid receptor and PTS (as well as MMD and BPD) – how can we integrate these data? 3. What concerns do Veterans have about protection of personal privacy in this rapidly moving domain? END OF PART ONE: Questions? Discussion?