Genetic testing information offered directly to patients does not appreciably increase or decrease their use of follow-up healthcare services or interventions, according to results from a two-year National Institutes of Health study.
Beginning enrollment in 2007, the study, led by the NIH's Intramural Research Program, the Group Health Cooperative in Seattle, and the Henry Ford Health System in Detroit, offered genetic susceptibility testing for eight common health conditions to nearly 1,600 patients in the HFHS. Researchers then tracked the use of healthcare services in subjects' automated records, comparing their use of health services before and after the testing period.
The researchers published some final results from the project, called the Multiplex Initiative, this month in the early online edition of the journal Genetics in Medicine, showing that physician visits and laboratory services in the post-test period did not change significantly for either the 217 subjects who chose to be tested or the remaining participants who declined testing.
According to Colleen McBride, chief of the National Human Genome Research Institute's Social and Behavioral Research Branch and an author of the study, the group's evaluation is the first to use electronic health records, instead of self-reported behavior, to measure the impact of genetic testing on future consumption of health care services.
These data represent the latest findings from the Multiplex Initiative, one of the first studies to look at people's attitudes toward genetic risk information for common, preventable conditions. According to McBride, based on the totality of data emerging from the initiative, she would not recommend consumers pay online, direct-to-consumer genomic testing services, such as 23andMe, to learn their genetic risk for common diseases, since the information customers learn from these services does not seem to markedly influence their behavior in any way.
"We took on a lot of what have been [main] suppositions about these tests," McBride told PGx Reporter this week. "So [now, overall] would we endorse DTC genetic testing for common preventable health conditions? … Probably not at this point."
"Would [we recommend] spending $100 on it? Probably not," she said.
"But is it going to harm you? Is it going to create such demand on the healthcare system that we can't manage it? Is it going to cause people to be upset and worried? No, probably not."
McBride noted that the growth of DTC genetic testing has raised fears that patients may misunderstand test results and demand increased services like diagnostic testing for predicted illnesses, which in turn would drive up healthcare costs.
"We know patient demand is a big driver for what gets done, so it’s a legitimate concern," she said. "This is the first data to really bring to light that it might not be as bad as they thought."
The group's results support other investigations, like a study by Scripps last year which found that DTC genetic testing does not lead to significant changes in behavior or anxiety (PGx Reporter 12/7/2011).
In the Multiplex Initiative study McBride and her colleagues offered patients from the HFHS a multiplex genotyping test comprising 15 SNPs associated with susceptibility to eight common, adult-onset health conditions. The final study cohort, after excluding ineligible subjects, was 1,599 subjects who were between 25 and 40 years old.
Of the patients enrolled in the trial, 1,098 bowed out of the genetic testing after a baseline survey. Another 284 went as far as to visit the study website at least once, but in the end, decided not to be tested. Meanwhile, 217 people did choose to be tested and gave blood samples for evaluation.
According to McBride, the multiplex test administered in the study is designed by the researchers and comprises a panel of SNPs intended to cover variants associated with increased susceptibility for diseases like diabetes, high blood cholesterol, high blood pressure, osteoporosis, and certain cancers.
"At the time, we were looking for variants that had a sizeable enough evidence base that we could actually say to a patient, this is what it tells you," McBride said. "This was before 23andMe or Navigenics, or any of the other [DTC genomic] companies came out. We anticipated that this was coming."
The group found that all patients who elected to undergo the multiplex test carried at least one genetic risk marker and most had an average of nine risk-associated variants.
Participants received results through the mail, and a "research educator" phoned each subject to explain the results and answer questions, the study authors wrote. A follow-up interview with participants followed three months after they received their results.
According to McBride, while physicians knew that they had a patient in the trial, there was no communication between the study group and these providers and few participants reported sharing results with their doctor.
"I think that’s because we did a fairly good job talking about the limits to the information before they made a decision to test, in terms of the medical actionability of these test results," McBride said.
"Most of the recommendations [associated with these risks] are generic, like exercise more, eat more fruits and vegetables, lose weight, don't smoke," she explained. "Again, that's mostly what these DTC tests can do as well – they can't really give any specific genomic-based recommendations [yet]."
The researchers then mined subjects' health records to track whether having genetic susceptibility information about four of the eight conditions — colorectal cancer, lung cancer, type 2 diabetes, and atherosclerotic coronary heart disease ─ impacted how frequently they visited their doctor and whether they got laboratory tests or procedures to obtain more information about their health.
The team compared use of services for each of the three cohort subgroups—those that took the test, those that investigated the test online but did not take the test, and those that didn't move past the initial survey—for a year pre- and post-testing, or before and after participation in the trial.
The researchers found that participants who opted for genetic testing did not change their overall use of health care services compared with those not tested. In fact, none of the three groups showed increased use of healthcare services, though overall, those who sought testing showed higher use than the other two groups both before and after they were tested.
"It looks like people who took up the genetic tests were slightly higher users to begin with, but their use didn't increase [after testing] … and we didn't see a decline either," said McBride.
The team's latest publication follows several others highlighting data from the Multiplex Initiative that focused on additional questions about testing adoption and effects. "This is probably the last paper, and in our minds maybe the biggest impact because there wasn’t data out there about this impact on the healthcare delivery system and there has been a lot of handwringing," McBride said.
The researchers have published other findings from the project discussing a reduced rate of testing adoption among African Americans, as well as results indicating that those tested were able to understand results and make informed decisions.
One previous report found that consumers who have received their genetic risk assessments through the web may improve their health outcomes through a more informed interaction with their doctors (PGx Reporter 7/29/2009).
While McBride said the group's results seem to indicate that, with the right education, genetically-tested patients are unlikely to make inappropriate demands of healthcare systems, the study has also highlighted the limitations of DTC testing to really affect health.
While patients in the study were unlikely to increase their pursuit of healthcare services, neither were they likely to make changes in their lifestyles or health practices that may affect their disease risk or outcomes, according to McBride.
"As a behavioral scientist I have to tell you," she said, "we throw everything in the book at [people] to get them to change behavior, and we get very modest change and low maintenance of change over time.
"So telling someone they have a slight increased risk of a very common disease, the likelihood that that would change their behavior in and of itself is pretty low."
According to McBride, the researchers asked subjects about motivation to change behavior as part of the study, and found those who actually showed up for testing were the most motivated overall.
"My instinct says these are people who tried a lot of things to change [their] health habits and have not been successful, because it's hard," she said.
"They're thinking genomics is going to bring them something new, hoping some health professional is going to say, 'Based on your genomics you should do this kind of exercise, you should eat this kind of food,' and we just can't do that yet."
McBride believes that researchers and healthcare providers should focus their energies on identifying genetic links to more specific and actionable behavioral changes. "I don't think its risk communication [alone] that’s going to get people to improve their health outcomes," she added.