September 1, 2016 (Vol. 36, No. 15)
DeeAnn Visk Ph.D. Founder and Principal Writer DeeAnn Visk Consulting
The Circle That Is Personalized Medicine Consists of More than Just Doctor, Patient, and Patient Data
Technical hurdles to the wide-spread use of genetic data in humans are disappearing. The cost of sequencing an entire human genome is sinking to around $1,000. Commercial companies are offering genetic tests for their clients’ personal use.
Given that changes such as these are occurring, you could easily imagine that you will, one day soon, arrive at your doctor’s office carrying a thumb drive containing your entire genomic sequence. What then? Will your doctor simply download your genomic information, run a computer program, and generate a report about your chances of developing a disease or experiencing an adverse reaction to a particular medication?
Unfortunately, that day is still far off. But patients are already availing themselves of genetic testing services, and bringing their test results to their doctors’ offices. The doctors, however, may be unprepared.
Do you suppose your primary care physician has the time to research each mutation, deletion, frame-shift, or duplication of your genome? Do you see your doctor searching PubMed, reading through reams of papers to determine if the differences in your cytochrome P450 genes might limit your prescription options? Probably not, unless your primary care physician is actually your personal physician, a healthcare provider solely devoted to you and your genome. Most likely you don’t enjoy this sort of “personalized medicine.”
As a practical matter, personalized medicine will be realized through intermediaries that can help clinicians work in accordance with the latest genomic knowledge. One such intermediary is the Clinical Pharmacogenetics Implementation Consortium (CPIC), an organization working to streamline the process of translating genetic test results into gene-based prescription recommendations.
“The goal of the CPIC is to facilitate the process to apply pharmacogenetics to clinical practice by developing guidelines for clinicians,” states Kelly E. Caudle, Pharm.D., Ph.D., CPIC coordinator, St. Jude Children’s Research Hospital. “We do not tell clinicians if genetics testing should be ordered, but rather what to do with the information when it is available.” The CPIC strives to provide clear, concise information for providers, explaining how genetic data applies to their patients.
“A genetic test is unlike a blood glucose test, which is merely a snapshot of a dynamic situation,” Dr. Caudle continues. “Genetic results follow a patient for a lifetime.” Hence, genetic information needs to be entered into a patient’s electronic health record (EHR) in a time-independent way. When a physician orders a drug for a patient 15 years after the genetic test was added, an alert needs to inform the clinician about the results of genetics testing in relation to the treatment.
According to Dr. Caudle, the CPIC not only provides gene-based clinical recommendations, it also supports EHR mechanisms. Such mechanisms can, for example, facilitate the ordering of a pharmacogenomics panel in advance of clinical need (preemptive genotyping) or in response to a specific drug indication.
EHRs may reflect consistent guidelines for the translation of genotype test results and even present recommendations such as “increase the drug dose,” “decrease the drug dose,” or “consider prescribing a different drug.” “By following the Institute of Medicine standards for developing trustworthy clinical practice guidelines,” notes Dr. Caudle, “we are insuring that the information presented by the CPIC is of the highest quality.”
Genetic information is revised and expanded daily. To keep up with these changes, CPIC guidelines are updated on an ongoing basis. This is essential, as what was a good review of genetic information one year, may be outdated the next.
Maintaining a Flexible Interface
In addition to being able to add to the knowledge base of genetic information in humans, the CPIC also strives to maintain the ability to integrate with several EHR forms. Different medical groups often have different applications that they use to access EHRs. The ability of CPIC information to integrate with all EHR types is essential for information to flow freely.
“CPIC guidelines enable the translation of genetic laboratory test results into actionable prescribing decisions for specific drugs,” declares Robert R. Freimuth, Ph.D., assistant professor of biomedical informatics at the Mayo Clinic and co-chair of the CPIC Informatics Working Group. “A key assumption underlying CPIC guidelines is that as clinical genomic testing becomes more common, clinicians will have those results available but may face uncertainty about how to use those results to inform their choice of treatment.”
“One theme that has emerged is the need for standardization within pharmacogenetics,” maintains Dr. Freimuth. “How is clinical genomic data represented? We need to be able to exchange genomic information within and between systems and institutions using specific data standards.
“Next, we need to define the language we use. If different words are used, the interpretation of results can become variable. We need common definitions provided by terminology standards.
“Additionally, standard models are essential to describe how genomic data and clinical decisions are linked to one another. Knowledge must be captured and structured so that computers and people can read it. Genotype to interpreted phenotype tables must be standardized—these tables are really at the core to understanding and communicating information.
“Finally, messaging standards provide the means to exchange data between systems and deliver it to physicians. The first three types of standards (data, terminology, and models) ensure that the information is clearly represented; the messaging standards provide the means for information to be exchanged between computer systems and communicated to clinicians. Together, all four types of standards are necessary to ensure information is clearly communicated and interpreted.”
Thus, CPIC guidelines utilize a standard format, including a standard system for grading levels of evidence, linking genotypes to phenotypes, and assigning prescribing recommendations based on these genotypes and phenotypes.
“We hope,” concludes Dr. Freimuth, “that this standardization effort will reduce confusion among clinicians and provide a resource that is usable both now and in the future.”
Preemptive Pharmacogenetic Testing
While the CPIC guidelines are shaping up nicely, other institutions are working on integrating genetic data with their EHRs. “Normally, pharmacogenetic testing is done after the fact,” states Stuart Scott, Ph.D., assistant professor of genetics and genomic sciences, Icahn School of Medicine at Mount Sinai. “We see this as less helpful for patients and advocate doing pharmacogenetic testing beforehand—preemptively,”
Before preemptive pharmacogenetic testing can be implemented, certain elements may have to be in place. According to Dr. Scott, these elements include suitable clinical guidelines, duly informed clinicians, demonstrations of cost-effectiveness, and assurances that insurance companies will honor reimbursement plans.
At Mount Sinai, Dr. Scott is participating in programs that preemptively test patients for genes involved in drug response variability. Patient-specific gene-drug information is added to the patient’s EHR, where it sits until the doctor orders a drug for which there is pertinent genetic information. Ordering such a drug essentially pulls an informational trigger—the relevant test result is retrieved and presented to the ordering physician with CPIC therapy recommendations. Certain specialties such as cardiology, psychiatry, oncology, and pain management already have well-known cases of gene-drug pairs, which were the first to be included in the implementation programs at Mount Sinai.
“With approximately 2,000 patients enrolled since 2013, the system now ‘fires’ point-of-care pharmacogenetic results and recommendations about once every two weeks,” notes Dr. Scott. “The physician receives a message through the EHR. For example, the message might say, ‘patient is a poor metabolizer of Drug A—consider X, Y, and Z alternatives.’”
Initially, Dr. Scott was concerned that it would be necessary to genotype 10,000 people before “triggering” would occur. However, activation of the pharmacogenetics functionality in Mount Sinai’s EHR system is occurring with only one-fifth of that number enrolled.
“We think it’s important to give physicians options, not orders, as physicians ultimately know the clinical context of their patients better than anyone,” insists Dr. Scott. “The guidelines can be rejected for any reason. We are carefully measuring provider feedback and experience using the EHR system. This information from clinicians informs how we refine our pharmacogenetic implementation programs without adding more work for physicians.”
Dr. Scott feels that doctors need education on using pharmacogenetics, as most have never had formal training before. To this end, medical students at Mount Sinai have the option of having their own personal pharmacogenetic tests performed. Later, when the students attend Dr. Scott’s lectures, the de-identified clinical results are reviewed.
Dr. Scott envisions patients receiving genetic testing for a drug response panel during initial visits to the doctor’s office. This information can then follow them for the rest of their lives via their EHRs, and potentially also through smart phone apps and other personal electronic devises.
International efforts to incorporate pharmacogenetics into routine medical practice can be seen in European countries. One ambitious undertaking, sponsored by the European Union (EU), is a multicenter study to evaluate the cost-benefit ratio of preemptive pharmacogenetic advice in comparison to conventional treatment considerations. This will be undertaken by the Ubiquitous Pharmacogenetics consortium.
“Six different healthcare centers throughout Europe will enroll patients, says Magnus Ingelman-Sundberg, Ph.D., professor and section head, department of physiology and pharmacology, Karolinska Institute. “At each center, one patient group will be treated based on pharmacogenetic information of the patient’s genome, and another control group will receive conventional medical care. By comparing the two regimens, we can do a cost-benefit analysis to determine the value of using preemptive pharmacogenetics testing.”
Dr. Ingelman-Sundberg concurs with the idea that it is important to keep track of the latest developments in pharmacogenetics and drug-genome interactions. The European Medicines Agency (EMA) fulfills the function of the FDA in the EU.
“Employing DNA to predict a patient’s reaction to a particular drug is easier, safer, and less costly than testing for the phenotype by in vivo studies,” Dr. Ingelman-Sundberg points out. “However, we are still researching exactly what the individual’s DNA sequence means for the expression of drug-metabolizing enzymes, drug transporters, and drug targets in the patient.
“Many rare mutations which are now not present in the commonly used genotyping platforms influence drug response. We still do not know all the mutations out there and how they affect the phenotype. These unknown rare mutations confound current preemptive genotyping for many drugs and have to subsequently be incorporated into the future genotyping platforms.”
Other factors in Europe lend themselves to clinical evaluation of pharmacogenetics. “In the Netherlands, there is a good infra-structure for basic science and a very strong track record of translational and clinical trials,” illuminates Stefan Sleijfer, M.D., Ph.D., head, department of medical oncology, Erasmus MC Cancer Institute, the Netherlands. “This situation is perfect to execute complex clinical studies. Distances are relatively minor. We have a single government-run healthcare system with mandatory health insurance for all residents.”
One interesting study focuses on determining which treatment to use for a particular cancer, especially metastatic cancer. Patients join the study, and tumor samples are obtained. From there, information on treatment response and overall survival is easily captured. Genetic information about the tumors is collected, too. “We now have the resources to do this for 6,000 patients,” articulates Dr. Sleijfer.
Another study conducted in the Netherlands is a drug repurposing study for patients for whom no treatment is available. Patients who are in good condition and willing to be treated will be examined for genetic variants. From there, off-label use of medications will be considered.
Cancer recognizes no geopolitical boundaries. To find new treatments and approaches, we need to consider worldwide research efforts, not only in cancer treatment, but also in pharmacogenetics.