Genomics, Personalized Medicine and Electronic Medical Records
Upcoming SlideShare
Loading in...5

Genomics, Personalized Medicine and Electronic Medical Records



We are now unlocking the secrets of health at a molecular level – which includes not only why some people get diseases, but also how to prevent or cure them. However, as Osler points out, knowing ...

We are now unlocking the secrets of health at a molecular level – which includes not only why some people get diseases, but also how to prevent or cure them. However, as Osler points out, knowing this information is only valuable in the context of making it available for the right patient at the right time.

This presentation provides a basic introduction to genomic or personalized medicine, and discusses how this information can and should be integrated into our electronic medical record systems.

These slides were originally presented at the HIMSS Annual Conference in February of 2007.



Total Views
Views on SlideShare
Embed Views



3 Embeds 206 185 17 4


Upload Details

Uploaded via as Microsoft PowerPoint

Usage Rights

© All Rights Reserved

Report content

Flagged as inappropriate Flag as inappropriate
Flag as inappropriate

Select your reason for flagging this presentation as inappropriate.

  • Full Name Full Name Comment goes here.
    Are you sure you want to
    Your message goes here
Post Comment
Edit your comment
  • Lyle Berkowitz, MD is a practicing internal medicine physician who has researched and consulted in the field of applied medical informatics throughout his professional career.  Dr. Berkowitz is an attending physician at Northwestern Memorial Hospital and is the Medical Director of Clinical Information Systems for the largest primary care medical group in Chicago.  In this latter capacity, he has helped his group successfully implement an enterprise-wide electronic medical record system, and has been paperless in his own clinical practice since the Fall of 2002.  Dr. Berkowitz is on the Advisory Board for the Association of Medical Directors of Information Systems and the Editorial Board of Healthcare Informatics. Mr. Strier is the National Innovation Leader for Deloitte Consulting's Healthcare Provider Practice, specializing in the convergence between healthcare and technology as well as disruptive market and product development strategies ranging from personalized medicine to telehealthcare.  He is a Guest Faculty at the joint Harvard Medical School/MIT Health Science & Technology Program, a member of Global eHealth Advisory Board at the World Bank, member of the Conference Advisory Board at Healthcare Research & Innovations Congress and member of the National Advisory Board at the Emerging Technologies and Healthcare Innovations Congress.
  • The ambitions of healthcare providers have not changed in the past hundred years, but the capabilities certainly have. We are now unlocking the secrets of health at a molecular level – which includes not only why some people get diseases, but also how to prevent or cure them. However, as Osler points out, knowing this information is only valuable in the context of making it available for the right patient at the right time. There are 3 billion letters in the human DNA code. While more than 99.9 percent of DNA is identical between any two humans, about on in every 1200 amino acid pairs varies from one person to another. These single nucleotide polymophisms (SNPs) are what differentiates us from each other. Furthermore, haplotypes are a set of SNPs on a single chromatid that are statistically associated. It is thought that these associations, and the identification of a few alleles of a haplotype block, can unambiguously identify all other polymorphic sites in its region. The causes of common disease are very complex. We know that both environment and genetics play important roles. Furthermore, most diseases, as well as responses to medicines, involve the interaction of multiple genes.
  • Learning Objective 1: Understand the importance of integrating clinical IT systems with molecular science and consider how your organizations will take advantage of this synergy. Learning Objective 2: Provide examples of successful integrations between molecular science and clinical IT systems, and discuss the potential for future projects that unite these two fields. Learning Objective 3: Review and improve your organization's strategic IT plan to fully incorporate the emerging needs of a more personalized, molecular-based care delivery system that is well integrated with your current and future plans for clinical information systems. Learning Objective 4: Relate the implications of emerging models of care, based on molecular medicine, to clinical staff to build consensus around the acquisition and adoption of new IT systems and applications.
  • Understanding molecular medicine, through both laboratory and imaging techniques, deepens our ability to detect, diagnose and treat disease. Genomics, the study of genes, is the most common area of study since it involves a stable, albeit large, data set. Genomic data is particularly useful in identifying certain diseases, unveiling risk factors for other diseases, and predicting how well certain drugs will work in humans. This last area, called pharmacogenomics, is an increasingly popular area of study involving both drug effectiveness (efficacy) and drug side effects. This is a critical field since many medications are only effective in 50% of the population and cause side effects in another large percentage, but we don’t know ahead of time how individuals will react. Being able to test for gene differences ahead of time will both improve effectiveness and decrease side effects for patients. Other important areas of study include proteinomics and metabolomics. While genetic markers are stable, these biomarkers are constantly changing as a function of both genetic and environmental exposures. They are particularly important in diagnosing diseases and calibrating treatment regimens. Finally, molecular imaging (e.g. PET Scans) is an increasingly important area in the field of cancer diagnosis and treatment calibration.
  • Current day medicine involves disease prevention (e.g. guidance on diet and exercise habits), early detection of problems (e.g. Colonoscopy, Mammograms), and treating a problem that has occurred. However, we are limited in each of these areas. For preventive care, our guidance could be more specific and could carry more weight if we had genetic testing to help a patient understand their risk. For early detection, we must rely on “macro” level events, such as visualizing polyps, feeling masses, imaging structural distortions, and tracking non-specific biomarkers (e.g. PSA). The field of molecular detection will help us to more quickly identify diseases and thus have more success in treating them before they spread or cause excessive damage. For disease treatment, we know that not all medications or other treatment options work for all patients. Molecular medicine has the ability to help us choose the best treatment regimens, as well as more easily follow the progression of disease over time.
  • BRCA1 gene mutations are associated with increased breast, ovarian, and possibly prostate, and colon cancers; while BRCA2 gene mutations are associated with breast, pancreatic, gallbladder, and stomach cancers. BRCA positive patients should get more aggressive monitoring, and some opt for even more aggressive measures including bilateral mastectomies and oopherectomies to minimize their future risk of breast and ovarian cancer. Several genes affecting the Cytochrome P450 pathway determine a patient’s ability to metabolize a large variety of medications. Depending on results, a patient is usually put into four categories of metabolism: Normal, Ultrarapid, Intermediate, and Poor. The ultrarapid group may therefore not respond well to normal dosages, while the Intermediate and Poor groups may suffer side effects related to the drug levels getting too high. About 20-30% of women with breast cancer are HER-2 positive (meaning they have too many copies of the HER-2 gene, and thus too much HER2 protein). These breast cancers are often more aggressive and harder to treat. Herceptin is a monoclonal antibody that specifically targets the HER-2 genes and thus is only effective for HER-2 positive cancers. The FDA therefore requires that molecular testing confirm elevated HER-2 levels before allowing physicians to prescribe Herceptin.
  • The future of molecular medicine will track the pace of information liquidity. With social networking, advances in communications and a broad social compact amongst health researcher, the most powerful tools for personalizing medicine may not come from proprietary informatics vendors, but rather through open source networks, shared collaboration – especially given the scale of the informational challenges at hand. Some current examples include: The cancer Biomedical Informatics Grid™, or caBIG™, is a voluntary network or grid connecting individuals and institutions to enable the sharing of data and tools, creating a World Wide Web of cancer research. The goal is to speed the delivery of innovative approaches for the prevention and treatment of cancer. The infrastructure and tools created by caBIG™ also have broad utility outside the cancer community. caBIG™ is being developed under the leadership of the National Cancer Institute's Center for Bioinformatics . Adjuvant! Online is an application designed to help health professionals and patients with early cancer discuss the risks and benefits of getting additional therapy (adjuvant therapy: usually chemotherapy, hormone therapy, or both) after surgery. The goal is to help health professionals make estimates of the risk of negative outcome (cancer related mortality or relapse) without systemic adjuvant therapy, estimates of the reduction of these risks afforded by therapy, and risks of side effects of the therapy. These estimates are based on information entered about individual patients and their tumors (for example, patient age, tumor size, nodal involvement, histologic grade, etc.) These estimates are then provided on printed sheets in simple graphical and text formats to be used in consultations. It is available online
  • Much of molecular medicine is still a basic science that has not yet been integrated into clinical care. This is related to multiple factors, including the still early nature of the science, the cost of testing, and the immense amount of information generated. It is clear that the use of information technology is critical to enabling the true integration of molecular medicine with clinical care. The first area of integration will involve “Understanding” how these molecular findings actually affect the human population. We need to answer questions such as “Which gene differences cause what diseases”, “Which gene differences affect therapy” and “which dynamic biomarkers and imaging changes are associated with what diseases and other processes”. Much of this will involve brute force of searching billions of data elements, looking for differences that are associated with clinical findings, and then testing to ensure reliability. The second area of integration involves the ability to “Apply” this information in a normal clinical workflow. The use of electronic medical records will greatly facilitate this process by allowing a platform that enables appropriate storage of molecular data and real-time clinical decision support during preventive care, diagnostic and treatment workflows.
  • The NUgene Project ( ), a long-term research study sponsored by the Center for Genetic Medicine at Northwestern University, is one of the first genetic banking studies in the country. Patients at Northwestern-affiliated hospitals and clinics are able to anonymously donate DNA from their blood samples and health information from their electronic medical records. This information will be used by researchers to examine the role genes play in the development and treatment of common diseases. There are currently over 5000 people enrolled, with a goal of 100,000 people by 2010. The first example of this integration involves a study on the possible causes of abdominal aortic aneurysms. Researchers will isolate and compare DNA from a group of participants with a history of abdominal aortic aneurysms, a disease control group consisting of participants with vascular diseases other than aortic aneurysms, and an ethnically matched control samples from NUgene participants without a history of vascular disease or chronic inflammation. DNA sequences of relevant pro-inflammatory cytokines will be compared between these groups of participants and controls to help determine whether certain gene sequences are associated with abdominal aortic aneurysms or other vascular diseases. The International HapMap ( www. hapmap .org ) is a similar project.
  • Assuring anonymity and security of donated patient data is of extreme importance for biobanks such as the NUGene Project. Some of the participant protection mechanisms include the following: NUgene has been granted an NIH Certificate of Confidentiality that protects the privacy of research study participants All study materials are coded with unique serial numbers to minimize the likelihood that non-study personnel will learn the identity of NUgene participants All study information routed over the Internet uses secure websites with advanced encryption technologies The data repository itself consists of a dual database system where primary identifiers, such as names, are housed in a separate database from all NUgene data NOTIS = Northwestern Oncology Trial Information System. It tracks participants on protocols for all the cancer trials and studies at Northwestern.  NUGene uses components of this infrastructure to track its participants and their identifiable information separate from the NUgene system.
  • The Clinical Proteomics initiative is a joint initiative between the NCI and the FDA to correlate protein and gene expression patterns for early detection and cancer screening, to establish therapeutic response endpoints, and to monitor drug toxicity during treatment. An initial study using computer-assisted detection of proteomic patterns in ovarian cancer was completed in 2004, marking the first application of proteomic technology to patient diagnosis. The study analyzed blood proteins in 50 women with known ovarian cancer, and 50 without. A computer-based artificial intelligence algorithm identified diagnostic proteinomic patterns that distinguished cancer from non-cancer. Ongoing research is trying to validate these findings. This early detection method has the potential to make a substantial difference in the mortality rate of ovarian cancer as more than 80% of ovarian cancer patients are diagnosed at a late stage and have a 20% or less chance of survival at five years. In contrast, the 20% of women diagnosed with early stage disease have a 95% survival rate at five years. There is also the exciting potential of using this technique to diagnose additional types of diseases.
  • Correlagen Diagnostics, Inc develops and commercializes genetic testing services using a high throughput automated approach that incorporates sequencing, variant analysis, and results reporting. They focus on Genes that Matter ™, where variation is correlated with disease and where testing would impact diagnosis and treatment. Integrated results reporting is the concept that the test findings can be reported in the context of a specific patient’s other data, including demographics, known diagnoses, vital signs and other lab results. Some typical examples of Genes that Matter ™ include those for Maturity-Onset Diabetes of the Young (MODY), Early-Onset Coronary Heart Disease and Severe Combined Immunodeficiency (SCID).
  • Heterogeneity in patient response to chemotherapy is consistently observed across patient populations, including both efficacy and toxicity. Much of this variance is likely due to genetic differences that affect drug metabolism. One of the most common of these situations involves the use of Purinethol (mercaptopurine) to treat leukemia in children. One in 300 children (0.3%) has a gene defect for the enzyme thiopurine methyltransferase (TPMT) which can result in mercaptopurine toxicity and associated bone marrow failure. If this defect is identified, the drug can still be safely used at a lower dose. New labeling now includes information about the risk of severe bone marrow suppression in TPMT activity-deficient genotypes and TMPT testing has now become standard practice for children who may receive this medication.
  • In the ideal scenario, a patient with a TPMT deficiency has that genetic information stored in their electronic medical record. So when mercaptopurine is ordered, the physician will get an alert explaining that they need to lower the dose of the medication or consider a separate treatment. Ideally, the alert would actually allow them to modify the order at the same time. If they ignore this alert, they will need to explain their reasoning.
  • Primary care physicians diagnose diabetes in adults on a regular basis and reasonably assume adult onset diabetics are “Type 2 diabetics”. They initiate treatment in a consistent manner and adjust treatment regimens if they do not respond to the standard medications. However, 2-5% of diabetics are actually in new category often called “Maturity Onset Diabetes of the Young” (MODY). MODY is distinct from both Type 1 and 2 diabetics because it has components of both decreased insulin production and decreased insulin sensitivity. Additionally, there are 6 main sub-types of MODY, each of which have their own genetic variations and implications for treatment. This is a good example of the increasing information and complexity physicians have to comprehend on a regular basis. Accomplishing this task in a paper-based world would be extremely hard and inconsistent. Additionally, new medical information often takes 10-15 years to travel from bench research to clinical care. Consider how the use of information technologies like electronic medical records integrated with molecular medicine can facilitate and accelerate this process.
  • A physician using an EMR enters a new diagnosis of Diabetes for a patient. An alert can be created that asks the physician to consider ordering genetic testing for MODY. The sensitivity of this alert can be improved by additionally relating it to demographics (e.g. patient under 50), vital signs (e.g. high weight/obesity), and previous blood results (e.g. glucose, ketones in urine, insulin antibodies). In addition to adding a new diagnosis to a patient’s problem list, other EMR actions could also be used as the trigger. For example, an alert for diagnostic testing could be triggered by the report of a specific lab (e.g. elevated HbA1C for diabetes), or pathology finding (e.g. ordering HER-2 testing if a breast biopsy comes back positive), or input of family history (e.g. early CAD testing). If one of these genetic variations are present, the system might also prompt the physician to inform the patient to get family members tested, and facilitate documentation of that discussion.
  • A physician using an EMR prescribes a new diabetic medication for a patient with known Type 3 MODY. A sophisticated alert could use both this genetic information and knowledge about whether other drugs have been tried first, and then offer advice if the physician has not chosen the best established protocol. In this case, the physician is attempting to first prescribe Glucophage for a MODY Type 1 patient, and the EMR system is recommending they consider a sulfanurea instead, which is the recommended first line drug for this type of MODY. If the physician ignores this alert, they will be asked to explain their reasoning. Assume the patient also had been found to have a abnormality in the genes controlling her cytochrome P450 pathway. When the physician then changes the medication to glipizide, the EMR can remind them of this abnormality and suggest a modification to the usual dosing, or to choose a sulfanurea that is not affected by that pathway. In a well integrated and relatively painless way, the EMR can thus improve both the efficacy and safety of the prescribing process.
  • Molecular data storage Once identified, how will molecular data be stored and accessed in your EMR system? Ideally it needs to be stored in a structured way for future queries related to reporting and alerts (via CDS). Consider the EMR’s Problem List, possibly using Snomed, an organization’s unique nomenclature, and/or a new Human Genome nomenclature. Consider creating a unique EMR form that holds the genomic data elements. Perhaps most ideal would be if the genomic results could be stored in a structured manner in the EMR’s results database so that they can automatically interact with a CDS system without the need for manual population into another vehicle. Clinical Decision Support (CDS) Diagnostic Workflow: When/Where/How should you alert physicians to consider ordering a predictive or diagnostic test? Treatment Workflows: When/Where/How should you use CDS to help physicians manage treatment options? Personnel Issues Bioinformatics Manager: Who in your organization is responsible for understanding the use of molecular data within your IT systems? Genetic Counselors: Will you have enough genetic counselors to deal with the increase of genetic information expected on the scene. How will you bill for these services.
  • Hospitals Importance of integrating their molecular lab resulting system with their clinical IT systems More specialization of healthcare providers as we identify more sub-types of diseases Rise of Genetic Counseling and related programs Researchers Increased money and attention to the molecular sciences Improve importance and ability to perform translational research that brings bench-top findings into the clinic. The NIH has made it clear that future funding will be biased towards academic centers that are performing more clinical and translational research ( http:// ). Pharmaceutical Companies Pharmacogenomics: More targeting of drugs in R&D and then clinical trials. This may save some drugs that only treat 10% of the population, but will also likely mean the demise of the “Blockbuster Drug” philosophy of “one size fits all”. Marketing shift to help educate patients and physicians about the importance of genetic testing. FDA approval process that might include more genetic testing (e.g. Herceptin) Insurance Companies Cost of tests can potentially hurt profits, or the costs will be shifted to the employers and individuals. Discrimination concerns
  • Technology Advancements (Better, Faster, Cheaper) The $1000 Complete Genome would allow every individual to have their human genome fully mapped and available electronically to interact with preventive, diagnostic and treatment decision. Data such as genetic markers, proteomics, metabolomics and molecular imaging will become more commonplace in helping with early detection and treatment guidance, especially with cancers. Ethical, Financial and Regulatory Issues Ethic debates may likely slow down this process. On one side will be patients or family members who could greatly benefit from this knowledge, while the other side will be privacy advocates who fear the use of these tests for discrimination. Fortunately, the government has already passed the “Genetic Information Non-Discrimination Act of 2005”. Financial issues will limit some tests at first, but eventually insurance companies will realize the cost-benefit of doing at least some selected testing. The use of HSAs will mean patients may find themselves having to decide whether or not to get a very expensive test that might affect their life greatly. Regulatory issues will involve the government’s plan to quickly create standards about the storage of molecular level data.

Genomics, Personalized Medicine and Electronic Medical Records Genomics, Personalized Medicine and Electronic Medical Records Presentation Transcript

  • Enabling the Future of Care Delivery: IT-Driven, Molecular Medicine HIMSS Annual Conference February 27, 2007 Lyle Berkowitz, MD Keith Strier, JD, PAHM
  • Our Ambitions “ To wrest from nature the secrets which have perplexed philosophers in all ages, to track to their sources the causes of disease, to correlate the vast stores of knowledge, that they may be quickly available for the prevention and cure of disease — these are our ambitions.” — Sir William Osler (1849 – 1919)
  • Agenda
    • Background
    • The Great Integration
    • Implications and Future Thinking
  • The Promise Imagine when doctors can…
    • Predict Disease pre-symptomatically with simple testing
    • Prevent Disease by identifying risks, early interventions
    • Diagnose Conditions less invasively, more accurately
    • Select Drugs that maximize benefits and minimize risks
    • Calibrate Treatments to heighten efficacy and recovery
    • Treat/Cure Disease using our own genes
  • Disease Burden Time Cost 1/reversibility Decision Support Tools: Baseline Risk Preclinical Progression Disease Initiation and Progression Assess Risk Refine Assessment Predict Diagnose Track Progression Predict Events Inform Therapeutics Sources of New Biomarkers: Stable Genomics: Single Nucleotide Polymorphisms Haplotype Mapping Gene Sequencing Dynamic Genomics: Gene Expression Proteomics Metabolomics Molecular Imaging Therapeutic Decision Support Source: “Personalized Medicine: Current and Future Perspectives,” Patricia Deverka, MD, Duke University, Institute for Genome Sciences and Policy; and Rick J. Carlson, JD, University of Washington Changing Paradigm of Care Typical Current Intervention Earliest Clinical Detection Earliest Molecular Detection Initiating Events Baseline Risk
  • Real World Examples
    • Stable Genomics ( Inherited Genes )
      • BRCA 1 & 2 predictor of breast and ovarian cancer risks
      • LDLR and APOB predictor of developing early coronary artery disease
      • MODY 1-6 predictor of MODY diabetes; subtypes affect treatment choice
      • CYP2D6/C19 Main cytochrome P450 genes that affects drug metabolism  dosing
      • CYP2C9/VKORC1 variants in these cP450 genes affect warfarin metabolism
      • TPMT guides adjustment of Purinethol dosing in Acute Leukemia patients
    • Dynamic Genomics ( Gene Expression, Biomarkers …)
      • Estrogen Receptor predicts response to Tamoxifen in breast cancer
      • HER-2 Receptor predicts response to Herceptin in breast cancer
      • PSA predicts risk of prostate cancer
      • Cholesterol predicts risk of heart disease and strokes
      • HIV Genotyping to guide selection of therapy
      • PET Scans to diagnose and help manage treatment options for various cancers
  • Evolving Applications
    • Open Source Tools/Networks
  • The Great Integration Understand Apply
    • Genetic Banking
    • Clinical Collaboration
    NUGene Project Understanding Genomics
  • NUGene: Data Flow & Privacy NUgene Database Coded Data Phenotypic Engine Encryption Decryption De-identification Process Medical Record Participant Enrollment Materials Patient Identifiers NOTIS Phenotypic Data Warehouse
  • Clinical Proteomics Initiative Understanding Proteomics
    • Background
      • Joint initiative between the NCI and the FDA
      • Correlate protein and gene expression patterns
        • Early detection and cancer screening
        • Establish therapeutic response endpoints
        • Monitor drug toxicity during treatment
    • Ovarian Cancer Project
      • Phase 1: Identify diagnostic patterns
      • Phase 2: Confirm diagnostic ability
      • Phase 3: Test in real world
  • Correlagen Applying Genomics in the Real World
    • Genes That Matter™
    • Integrated Results Reporting
    • Examples
      • Maturity-Onset Diabetes of the Young (MODY)
      • Early-Onset Coronary Heart Disease
      • Severe Combined Immunodeficiency (SCID)
  • Case Scenario #1
    • 3yo male
    • Acute lymphoblastic leukemia
    • Being treating with mercaptopurine
    How do you currently manage this scenario? How will you manage this in the world of molecular medicine integrated with EMRs?
  • Pharmacogenomic Alert Pharmacogenomic Alert This patient has a TMPT gene defect which indicates a high sensitivity to standard doses of mercaptopurine. [] Cancel Drug Order [] Lower the standard dose (75mg/m2) by 85% to a modified dose (11.25 mg/m2) [] Ignore this Alert
  • Case Scenario #2
    • 32yo Female
    • New onset diabetes (non-ketotic)
    • Non-obese
    How do you currently manage this scenario? How will you manage this in the world of molecular medicine integrated with EMRs?
  • Diagnostic Alert Dx Alert This patient fits the profile for MODY. Consider checking for MODY genetics [] Order MODY Genetic Screen [] Ignore Alert [] Learn more about MODY
  • Treatment Alert Treatment Alert Based on known genomic data and phenotype expression in this patient, the best treatment for their Type 1 MODY diabetes is to start with a sulfanurea. [] Change Order to a Sulfanurea [] Ignore Alert
  • Implications Things to start thinking about Personnel Storage Diagnose Treat Predict CDS
  • Impact on Other Players
  • The Future
    • Technology Advancements
      • The $1000 Complete Genome
      • BioMarker Testing: POC/Continuous
      • Superior Molecular Imaging
    • Ethical, Financial and Regulatory Issues
      • Who should get these tests
      • Who should pay for these tests
      • How should this data be stored
      • How will this data be used
    Better, Faster, Cheaper
  • Resources
    • The Human Genome Project :
    • The NUGene Project :
    • Clinical Proteomics Project :
    • The FDA and Genomics:
    • The CDC and Genomics:
    • NIH & Pharmacogenetics :
    • Non-Profit Organizations
    • News and Updates
  • Thank You