By Turna Ray
Researchers at the Cleveland Clinic are conducting several studies with the goal of informing the integration of genomics into the care of patients at the academic medical center.
At a personalized medicine conference last week hosted by the Center for Business Intelligence, Charis Eng, director of the Cleveland Clinic's Genomic Medicine Institute, discussed three investigations that have added to the body of knowledge about how genomic medicine should be adopted into the existing healthcare infrastructure.
In one study, Eng and Brandi Leach, another Cleveland Clinic geneticist, looked at how genetic risk data gleaned from direct-to-consumer genomics firm Navigenics compares to risk assessments based on patients' family history for diseases. A second survey conducted with DTC genomics firm 23andMe explored whether doctors improved their knowledge of genetics through participatory continuing medical education courses.
Lastly, Cleveland Clinic also tracked the number of BRCA mutation tests and microarray tests that were incorrectly ordered by medical professionals without a background in genetics. The monetary losses that the Cleveland Clinic stood to incur in these cases convinced the institution to reorganize and increase the involvement of genetic professionals in ordering genomic tests.
Family History and Genetic Testing
In its study comparing disease risk assessments from Navigenics' SNP test results to family history, Cleveland Clinic researchers looked at 44 subjects and considered their chances for developing breast cancer, prostate cancer, and colon cancer. Each subject was categorized into one of three risk categories ─ general population, moderate, or high risk ─ based on their family history or genomic test results.
"Overall concordance between family history risk assessments (FHRA) and personal genomic screenings (PGS) was low for all three cancers." Eng and Leach reported in an abstract for a presentation at the American Society of Human Genetics annual meeting last November.
Genome scan scores placed eight subjects at a moderate or high risk for prostate cancer, while by family history these people were in the "general population" category. The study also identified 12 participants as having moderate colon cancer risk by genetic testing scores, but with family history they fell into the "general population" risk category. Finally, none of the three subjects with a high hereditary risk of prostate cancer were reported to be as high risk following genetic testing.
When study participants' cancer risks were considered by FHRA, 10 had hereditary breast cancer risk, but PGS only identified one person as high risk. Of the nine participants with hereditary colon cancer, none were deemed high risk with PGS. In total, 22 people were found to have a hereditary risk of cancer, but PGS only identified one as having a high genetic risk.
"This is especially concerning for patients at hereditary risk that may elect a PGS approach over validated FHRA and testing," the study authors wrote in the abstract.
Overall, for cancer risk assessments, that study found that FHRA and PGS agreed on the subjects' risk categories 46 percent of the time. "Based on our study, FHRA and PGS may be complementary tools for cancer risk assessments," the researchers said.
Eng told PGx Reporter that there is ongoing research to figure out how scores from genome scans can be used with other medical data to inform patient care. As an example, she cited work by Jiangfeng Xu of Wake Forest School of Medicine looking at whether genetic testing can be used with prostate-specific antigen testing to help guide decisions about whether more invasive prostate cancer diagnosis is needed.
"By using prostate cancer-associated SNPs, [Xu] showed that he could adjust borderline PSAs up or down and hence help the biopsy decision," Eng said. However, in talks, Xu has said that using genetics and PSA scores to inform patient treatment is a complex decision and needs further research.
Ultimately, Eng and Leach's findings support family history as the gold standard, and they recommend that FHRA "should be used to clinically evaluate an individual's risk of developing cancer until further research is done to prove that PGS can predict an individual's risk or can be integrated with FHRA to increase sensitivity."
Muin Khoury, director of the Centers for Disease Control and Prevention's Office of Public Health Genomics, has often said that family history is a more informative and cheaper way of determining predisposition to diseases than DTC genome scans. The Office of Public Health Genomics last year issued a video warning against the limited utility of DTC genetic testing, and urged people instead to take a good family history (PGx Reporter 08/11/2010).
[ pagebreak ]
However, proponents of the DTC genomics model maintain that family history is not necessarily a good surrogate for the genetic risk of common diseases. "It accounts for a very small portion of the genetic risk. It has been estimated that 50 percent to 70 percent of risk of most common diseases is genetic versus environment," Jeff Gulcher, chief scientific officer at Decode Genetics, said recently at an advisory committee meeting on DTC genomics hosted by the US Food and Drug Administration. Gulcher added at the meeting that no conventional risk factors exist for many of the most common diseases, such as prostate cancer (PGx Reporter 03/09/2011).
Navigenics issued a letter to the US Surgeon General around the time the Eng and Leach study came out last year urging for a national effort to educate people about new disease prevention tools, such as genetic tests, that can provide information about disease risk when family history doesn't.
In the letter, Navigenics suggested that genetic information and family history should be used together to inform individuals about their health, since genetic data could potentially encourage people to make better healthcare decisions, and because even a thorough family history can't predict whether most people might get many common, complex illnesses, such as diabetes, many types of cancers, and cardiovascular disease.
The Scripps Translational Science Institute is also looking at whether Navigenics risk scores match risk assessments ascertained through family history as a follow-on to a study it published in the New England Journal of Medicine in January that examined the behavioral and psychological impact of DTC genomic testing (PGx Reporter 1/12/2011).
That data is currently unpublished, but at the Future of Genomic Medicine Conference hosted by Scripps earlier this month, Cinnamon Bloss, a Scripps researcher and a co-author on the NEJM study, said that the team found that the Navigenics estimates "correctly classified people with and without family history" of various conditions. Furthermore, in conditions with high heritability and high prevalence, the Navigenics estimates "were statistically significant in terms of their ability to classify," she said.
Bloss said that the results are "preliminary evidence for the clinical validity" of DTC genomics testing, but stressed that the effect sizes for the study were very small.
Eng noted that the Cleveland Clinic research was conducted separately from the Scripps analysis. "It would be nice to have it replicated by an independent party," she said.
When it comes to driving adoption of genomic medicine, physician education is often identified as a major barrier. At an advisory committee meeting hosted by the FDA earlier this month, 23andMe expressed an interest in developing CME and other types of educational programs to advance doctors' understanding of genomics.
Eng said that Cleveland Clinic collaborated with 23andMe to figure out what doctors were learning from interactive CME courses. Hypothesizing that doctors learned more through participatory CME, the researchers submitted questionnaires to doctors after they took these courses to gauge what they got out of them.
The results of the survey suggest that the CME courses didn't necessarily increase doctors' understanding of genomic topics, but they "certainly learned empathy and sympathy for their patients," Eng observed.
The questionnaires uncovered that while more than 90 percent of physicians felt that genetics was important to their practice of medicine, less than 25 percent said they were comfortable discussing genomics results with their patients. "What was resoundingly said was the experience made them empathize more with their patients about genetic testing," Eng noted. "So, they felt they would make better doctors in advising their patients in this regard."
This study is under consideration for publication, so Eng couldn't divulge further details.
The conclusions of the Cleveland Clinic/23andMe assessment are similar to a nationwide survey conducted three years ago by Medco looking at how genomics knowledge impacts adoption of pharmacogenomics. Out of more than 10,000 physicians from Medco's provider database that completed the questionnaire, 98 percent of doctors felt that patient genetic profiles influenced their response to drugs, but only 10 percent felt "adequately informed" about PGx testing.
Changing Care, Lowering Costs
In the post-healthcare reform environment, before implementing genomically-guided personalized medicine strategies, payors and providers are paying close attention to whether such interventions would improve patient outcomes and cut unnecessary costs. As a research institution, Cleveland Clinic is trying to show through a pilot project that genetic testing can save money when administered to those who truly need it based on their medical information and family history.
[ pagebreak ]
Cleveland Clinic assessed how many of its employees received BRCA1/2 testing under its health benefits plan during 2008, and found that the plan paid for around 500 tests costing more than $380,000. Of the tests ordered, 70 genetic tests, costing around $53,000, were ordered by Cleveland Clinic's Center for Personalized Genetic Healthcare.
The rest, around 430 tests costing more than $330,000, were ordered by non-genetics professionals. These patients, according to Eng, likely received no evaluation from a genetics expert and didn't receive genetic counseling.
"This was a rough gauge for cost-effectiveness," Eng said of the analysis. "At the time it was done, we were just changing our practice model. So, when we increased access and upped education on genetics, of course, many more non-genetics healthcare providers were pleased to have us see their cancer genetics patients."
Several insurers have tracked testing patters with Myriad's BRACAnalysis Test, and estimated that around 20 percent of tests ordered do not meet professional society guidelines or meet their own coverage requirements. As such many payors have placed the test under prior authorization schemes to ensure that those getting the test really need it based on family history and medical data. Additionally, several insurers also provide genetic counseling as a condition for reimbursement (PGx Reporter 11/04/2009).
Continuing its cost-effectiveness estimations, Cleveland Clinic also looked into how much it paid for microarray testing ordered in 2010. The number was around $1 million, and "almost all" the tests ordered by non-genetics professionals were deemed to be ordered incorrectly, while only a minority of tests were ordered by genetics professionals. These findings led Cleveland Clinic to institute a policy that under the pilot project only genetics professional could order microarray tests and other genetic tests.
It is an oft-cited fact that demand for certified genetics medical professionals is outpacing their availability, and this need is felt at the Cleveland Clinic as well.
"What we need is to increase our presence by embedding genetics professionals, mainly counselors, with non-genetics clinicians through our main campus clinics, as well as in our regional practice," Eng said. Genetic counselors at the institution are already stretched thin, however, and the demand for their expertise shows no signs of decreasing.
Eng said that last year, around six full-time genetic counselors at the clinic saw 3,000 outpatient visits, in addition to their in-patient consultations.
Have topics you'd like to see covered in Pharmacogenomics Reporter? Contact the editor at tray [at] genomeweb [.] com.