Souped-up EMR: Some interesting new applications
■ Behind the electronic medical record is the emerging field of bioinformatics. But physicians who want to integrate a patient's genetic code into treatment decisions will need the technology and ability to analyze those data.
By Tyler Chin — Posted Oct. 9, 2006
Most electronic medical records enable doctors to store, organize and query clinical information. The system at the Marshfield (Wis.) Clinic digs all the way down to a patient's individual genetic code.
During the next two years, clinic physicians plan to order genetic tests for patients who are prescribed warfarin, an anti-coagulant sold under various brand names. The clinic will be able to access electronically test results and calculate the ideal drug dosage for each patient based on entered knowledge of genetic makeup, and clinical and environmental data.
Marshfield is an early adopter of the growing integration of clinical and genetic data. Bioinformatics involves the use of massive computing power, statistical analysis and computational algorithms to analyze and solve biomedical problems at the molecular level. Research has been going on since the advent of computers, but the 2003 sequencing of the human genome gave bioinformatics tools a boost.
Researchers hope bioinformatics will lead to discoveries of new treatments and diagnoses, as well as enable physicians to identify people at risk for a particular disease. What some term "personalized medicine" could be given at the practice level with the help of EMRs that can store patients' genetic codes.
"The idea of personalized medicine is that in choosing which drugs to give an individual, we will no longer assume that there is a best drug [for a specific disease] but rather that different drugs will be better for different people with different genotypes and different environmental exposures," said Justin Starren, MD, PhD, director of the bioinformatics research center at the 750-physician Marshfield Clinic. "The process of matching the patient to the drug will become exponentially more complex."
Physicians can expect to be enticed to buy a powerful EMR in the next five to 10 years, one that can digest and process an individual's genetic data in a decision-support system.
"The type of bioinformatics that we specialize in here, which is the genotype-phenotype environmental evaluation, you absolutely cannot do without an EMR because manually trying to abstract that much information is impossible," Dr. Starren said.
But there are questions: How much would this cost? Who is going to pay for it? How do you protect patient privacy? And, most important, is bioinformatics really going to result in better care?
A money issue
Several surveys estimate only 20% of U.S. doctors are using an EMR. A key reason: start-up and first-year costs can run $10,000 to $40,000 per physician. Cost is a point of contention for many physicians who believe they are being asked to pay for a system whose benefit flows to insurers and employers.
Insurers are starting to offer technology subsidies or pay-for-performance programs that reward physicians for implementing technology or meeting certain quality measures. Separately, the federal government has begun working to revise Stark laws and anti-kickback regulations to allow hospitals and others parties to donate technology to doctors.
An initiative the National Cancer Institute's Center for Bioinformatics launched in 2003 could potentially offer a lower-cost alternative for those involved in bioinformatics, said Kenneth Buetow, PhD, the center's director. The cancer Biomedical Informatics Grid initiative, or caBIG, is attempting to make it easier for bioinformatics cancer researchers to share information with each other, and their eventual discoveries to physicians. Although caBIG targets only the cancer community, it offers a prototype of a virtual national network that could be used for other diseases, Dr. Buetow said.
Many insurers already reimburse for some genetic tests, including prenatal, breast cancer and leukemia screenings, but they won't pay for thousands of new genetic tests until they see the tests work and are cost effective, said Russ B. Altman, MD, PhD, an internist and professor of genetics, bioengineering and medicine at Stanford University School of Medicine in California.
"There will have to be [peer-review] studies that show that bioinformatics-based technologies like genomic medicine improve patient outcomes," he said.
Meanwhile, the National Institutes of Health is funding projects designed to cut the cost of genome testing.
"At the moment we take genes and we might [test for] several markers on a given gene for research, but if [the cost of testing] gets down to $1,000 per person, that's what's going to get it done," said Catherine McCarty, PhD, co-principal investigator of Marshfield Clinic's Personalized Medicine Research Project.
A lock on information
If the privacy of genetic information isn't protected, it might limit the number of people donating genetic samples and allowing researchers access to their electronic medical records, said Rex L. Chisholm, PhD, director of the Northwestern University Center for Genetic Medicine in Illinois. In 2002, the center began an initiative to collect blood samples from 100,000 patients at three affiliated hospitals. So far 6,000 participants have enrolled.
"We know of about 10 million different places in the human genome where there's genetic variation and what we need to do is be able to ask the question, 'How does that genetic variation correspond to disease?' " Dr. Chisholm said. "You need to have a lot of individual people participate in a project where you can assess what their genetic variation is by [using] standard genotyping approaches."
In recent years, Congress has considered, but not passed, legislation that would prevent insurers from denying coverage on the basis of genetic predisposition for diseases. As of mid-September, 34 states have genetic anti-discrimination laws related to employment, though some include exceptions allowing employers to screen employees for something in their genetic makeup that could threaten workplace safety, said Alissa Johnson, program principal at the National Conference of State Legislatures. Also, 32 states have passed genetic privacy laws, and 48 have some form of law restricting the use of genetic information to charge higher rates, or cancel or deny coverage for individual or group insurance.
Bioinformatics research is still in the discovery process, but researchers are confident bioinformatics dramatically will begin impacting primary care practice within five to 10 years. At that time, thousands of commercial diagnostic tests and tools will be available to help physicians interpret thousands of genetic variations, said Isaac Kohane, MD, PhD, a pediatric endocrinologist at Harvard Medical School in Boston. Dr. Kohane is researching the timing of how genes turn off and on during brain development, and how variation in the regulation of these genes may lead to disorders such as autism.
He expects bioinformatics will produce research highly beneficial for patient care, but he is concerned it could be derailed. In an article in the July 11 Journal of the American Medical Association, he warned that the necessary computer infrastructure must be in place to support doctors in the use and interpretation of genomic-based tests.
"For most of these tests in genetics that are coming down the pike, we're poorly educated," said Dr. Kohane. "Nor could we possibly be educated for when is the right time to use 30,000 different genetic tests. If we don't do this right, the number of false-positives ... will be in the millions.
"That will be costly, may cause morbidity, is going to be unaffordable and therefore is going to shut down important progress in applying genomics to health care if we do not bring to bear sound statistics and automated decision support," he said.
Meanwhile, doctors can wait a little longer.
"Everyone's going to need to have an EMR eventually, but ... [during the transition] physicians who don't have an EMR will still be able to take advantage of the discoveries that bioinformatics and genomic medicine produce that improve their ability to care for their patients," Dr. Chisholm said.