William Check, PhD
“Only connect,” wrote E.M. Forster in an oft-quoted passage in Howard’s
End. In the course of the novel, Forster’s characters discover how
difficult it is to achieve true communication, given the complexities
of the human psyche. Add to the complexities of psychology the complexities
of molecular medicine, and you can see that communicating results of medical
genetics lab testing to clinicians can be challenging. This is particularly
true of screening for carriers of cystic fibrosis mutations, which is
considered a prototype for the clinical application of molecular analysis
of inherited diseases.
“The CF gene was linked to the disease in 1985 and isolated in 1989,” says Jean Amos Wilson, PhD, scientific director of human genetics at Focus Diagnostics, Cypress, Calif. “It is one of the first of many that appeared at that same time. So everything we say for CF is true for many other diseases we test for.”
There’s no escaping it—CF carrier screening is complex. “It is interesting that a lot of people perceive that it is easy to report results of CF carrier screening, ” says Margaret Gulley, MD, director of molecular pathology and associate professor of pathology at the University of North Carolina, Chapel Hill. “But this test is actually quite complex to interpret. Not all mutations are equal in terms of their impact on severity of disease. Also, we would like to have our report include as much information as possible that would help the clinician know what to do with the results.”
In Dr. Gulley’s experience, clinicians are looking for an interpretive report so they don’t have to sift through the literature themselves to figure out the clinical import of the findings. “Sometimes,” Dr. Gulley says, “interpretation involves not only the results of the laboratory test but correlating those results with other tests that the patient had, such as sweat chloride, or testing of other family members. For results to be most useful to clinicians, there should be a pathologist or other laboratory professional who can answer questions about what it means to that patient in light of the overall clinical findings.”
Concern about effective communication of molecular genetic testing is
more than theoretical. “There is published evidence that physicians misinterpret
complex test results, especially in the molecular genetic area,” says
Wayne Grody, MD, PhD, professor in the Divisions of Molecular Pathology
and Medical Genetics in the Departments of Pathology and Laboratory Medicine,
Pediatrics, and Human Genetics, UCLA School of Medicine. For example,
a 1997 paper on gene testing for familial adenomatous polyposis concluded
that “Patients who underwent genetic tests for familial adenomatous polyposis
often received inadequate counseling and would have been given incorrectly
interpreted results” (Giardiello FM, et al. N
Engl J Med. 1997;336:823-827).
“I’m not sure the fault lies with our laboratory reports,” Dr. Grody says. “The ones I’ve seen are quite complete and give conditional probabilities, residual risk, and other pertinent information. Physicians are either not reading the report or trying to read it and not understanding it because it is beyond what their background can absorb.”
Making this task even more difficult soon will be the wider dissemination of carrier screening for CF mutations into non-esoteric laboratories that don’t have medical genetics expertise in-house. “I never dreamed that this test would leave the genetics laboratory,” says Dr. Amos Wilson. “But it has.” Dr. Gulley says, “More laboratories are doing it because clinicians are demanding it.” It is recommended that CF carrier screening be offered to all pregnant women, so it has become a high-volume test. Since a sendout test is an expense rather than a revenue producer, many laboratories have decided to bring CF carrier screening in-house.
As a result, this test is now migrating to laboratories that have molecular expertise but no on-site clinical molecular geneticist, medical geneticist, or genetic counselor. “We all worry about that,” Dr. Grody says. “It is a frequent topic of discussion, especially now when more manufacturers have come out with kits and ASRs [analyte-specific reagents] that are robust and technically fairly easy to use.” Smaller hospital and reference laboratories can do the analytic part adequately. But, Dr. Grody says, “We worry how well they will handle the preanalytic part and reporting.”
Ordering of the test is also migrating to nonspecialists. Pediatrician Tracy L. Trotter, MD, of the San Ramon Valley (Calif.) Primary Care Medical Group, says, “More and more this test is ordered by primary care physicians or at least nongeneticists. There is a need to communicate results so that nongeneticists can understand the clinical implications.”
Effectively communicating the results of medical genetic testing was the focus of a Nov. 9, 2005 workshop sponsored by the Centers for Disease Control and Prevention, Association for Molecular Pathology, and Wadsworth Center of the New York State Department of Health. One of the convenors, Ira Lubin, PhD, a geneticist in the CDC’s Laboratory Practice Evaluation and Genomics Branch, summarized the aim of the workshop for CAP TODAY. He contrasted a medical genetic test with a complete blood count. A CBC can be ordered routinely and the clinician will know what to do with the results. A genetic test, on the other hand, tells you whether a mutation is present, but to have meaning it needs some level of expert interpretation.
“While a laboratory is not in the business of direct patient care,” Dr. Lubin says, “knowledge and expertise of professionals within the medical genetics laboratory provide an interpretive role that’s important to the clinician for presenting the patient with appropriate clinical management options.” The lab interpretation integrates the analytic test result with an understanding of the test method and its relevance to the clinical condition for which the test was ordered, and pertinent patient and family information, he notes.
The question is how best to communicate that understanding to clinicians.
Sometimes the best way to make a point is by telling a story. Here are four stories told by workshop attendees that illustrate the potential for faulty interactions between laboratories and clinicians. Dr. Amos Wilson’s experience demonstrates that misinterpretation can have real-life human consequences. “If the report states that the patient’s screen was negative for mutation analysis and is not interpreted correctly, the patient may understand from that that they are definitely not a carrier,” she says. But that may not be true. Because the CF screen is 90 percent sensitive in non-Hispanic Caucasians, in a person with no family history, or a family history in which the mutation is not known, a negative screen reduces but does not eliminate the possibility that the subject is a carrier. “I have two patients in my practice who were negative for the mutation panel who were indeed carriers and went on to have affected children,” Dr. Amos Wilson says. Both of these couples were counseled appropriately and understood that the risk was low but not zero. In each case one member of the couple was at high risk, but there was no way to get an unambiguous answer for the partner. They made the decision to get pregnant. What if a couple in the same situation was incorrectly told that they had no risk and, with that understanding, went ahead and had a child—and it had CF?
Elaine Lyon, PhD, medical director of molecular genetics at ARUP Laboratories and assistant professor of pathology at the University of Utah, made a more subtle point. A child with symptoms suggestive of CF was tested at another laboratory for CF mutation status by complete gene sequencing, which identified two mutations that were not on the standard panel. A request came from the treating physician to Dr. Lyon to test the siblings, who also had minor symptoms. She requested the original report, which stated that both mutations had been seen before on the same chromosome. If that was true, the proband was a carrier but did not have CF, or, if he did have CF, the mutation on the other chromosome could not be identified even by state-of-the-art full gene sequencing. (If both mutations are on the same chromosome, the gene on the other chromosome would be intact. Since CF is an autosomal recessive disease, the proband would not have the disease.) “We worked with the doctor to get the parents in for testing,” Dr. Lyon says. She found that both mutations were from the same healthy parent, showing that they were on the same chromosome. Dr. Lyon says this result surprised the clinician, who had the mutational confirmation of his diagnosis “taken away,” leaving the children’s symptoms unexplained.
“If we had sequenced both sibs directly, they could have received a diagnosis of CF as well,” she says. “Having that report in our hands was essential, because we understood it, even though the clinician didn’t.” In this case the medical genetic expertise in the laboratory saved the children from a wrong diagnosis, with its considerable psychological and medical consequences. Avoiding an incorrect diagnosis also required that the original laboratory report have all pertinent information, not just a simplified conclusion. “We have seen the same situation twice and have discussed the possibility several times when two rare mutations were identified,” Dr. Lyon says. Resolving these situations requires parent samples. “Unfortunately,” she adds, “we don’t always get them.”
Dr. Trotter’s experience was not with CF but long QT syndrome, or LQTS. A family that had recently moved to his area came to him because the 10-year-old son had two episodes of syncope at basketball camp. While taking a history he found out that the boy’s 15-month-old sister was born premature and had problems requiring ventilation shortly after birth. The infant had been evaluated for LQTS, including a genetic analysis, which discovered no mutations. “That piece of information, unfortunately, was handled incorrectly,” Dr. Trotter says. “The neonatologist had poor understanding that there are many mutations causing this disease that the current test cannot find. So the fact that you don’t find a mutation doesn’t mean the child does not have it.” Through several steps the laboratory’s report ended up in the pediatrician’s summary as “normal test.”
To obtain the complete chart, Dr. Trotter consulted the hospital where the baby girl was born. After several requests, it finally came. Sure enough, the laboratory report had a footnote saying that a mutation is found in only 30 to 40 percent of LQTS cases and recommending ECGs on all family members, a step that became lost in the crisis. Dr. Trotter ordered ECGs, which showed moderately prolonged QT intervals (420-470 msec) for the two children as well as the mother, who had never had symptoms. All three are now doing well on beta-blockers.
“Lucky for the patients none had sudden cardiac death while we were figuring it out,” Dr. Trotter says. His conclusion: “Medical genetic tests are frequently inconclusive and frequently misunderstood by us in primary care.”
Joy B. Redman, MS, CGC, encounters clinician misunderstanding related to ordering rather than reporting. “I spend a lot of time talking to doctors who order CF sequencing rather than screening,” says Redman, who is a clinical genetics specialist at Quest Diagnostics/Nichols Institute in San Juan Capistrano, Calif., and who represented the National Society of Genetic Counselors at the workshop. One responsibility of Redman and the other genetic counselors at Quest Diagnostics is to ensure appropriate test ordering. “I call about every CF sequencing request,” Redman says. “We get lots of sequencing tests ordered by test code that obstetricians didn’t mean to order.” Sequencing is much more expensive and takes six weeks, while the 23-mutation screening panel takes five days. Sometimes a physician from a CF clinic really meant to order sequencing. When she tells the physicians there is a 90 percent chance to detect a mutation with the screening panel, she says, “Most are okay with that.”
Once in a while the client will insist on sequencing. This happened with
a physician who was the mother of a CF patient. “She said, ’I want the
best test,’” Redman says. Whether sequencing is the best initial test
is debatable. “Whenever you do a sequence there is a chance to see a mutation
that has not been reported,” Redman says. “How do you interpret that—as
a disease mutation or a benign polymorphism?” Using sequencing in place
of screening is especially problematic. But Redman points out that doing
sequencing after a negative screening panel, in a person whose partner
screened positive, for instance, is not recommended as routine practice
in the American College of Medical Genetics’ guideline (Grody W, et al.
Genet Med. 2001;3:149-154).
It’s bad enough that many physicians don’t have medical genetic expertise. What’s worse is the possibility that, in the private practice setting, it may not even be a physician who orders the tests and reviews results and reports them to patients. A pilot survey of 84 OB-GYN practices in New York City that was presented at the workshop revealed that in many practices nonmedical personnel were responsible for filling out genetic test requisition forms (44 percent), reviewing genetic test results (59 percent), and reporting results to the patient (86 percent).
“Personally I wasn’t surprised by these findings,” says Margaret McGovern, MD, professor of human genetics and pediatrics and co-director of the molecular diagnostics laboratory in the Department of Human Genetics at Mount Sinai School of Medicine, New York, who conducted the survey. “In a busy OB-GYN practice it’s not at all surprising that secretaries are filling out paperwork.” However, Dr. McGovern notes, genetic test requisitions are more detailed than typical tests. “Similarly,” she says, “when you get a result back, there is a distinct possibility that a secretary may not understand the subtleties of the report. So this may be the way they run their offices, but genetic testing may pose some problems.”
A second part of the study was for the physicians to indicate on a scale from one to five how well they were satisfied with statements in reports. Almost all scores were less than three, which one workshop participant called “appallingly low.” Such scores are typically four or five. “In general, they didn’t like more complicated reports,” Dr. McGovern says. Reports that contained more numbers or technical terminology (that is, Bayesian calculation) got lower scores. “Clinicians want a simple report,” Dr. McGovern concludes. “I understand why laboratory directors put all this information in reports,” she says. “I also understand why clinicians don’t want it or find it not helpful.”
One solution might be what Dr. Lyon calls multi-level reporting, with a simplified interpretation for the clinician and complex underlying data provided for genetic counselors and laboratory directors in case additional testing is needed. “In our situation, it was very important to have the original results,” she says.
Dr. McGovern is repeating the survey in a larger population to see if the results generalize.
In a prior survey of genetic counselors, Dr. McGovern found that many had difficulty with reports and called for clarification. “It’s a no-brainer that a clinician who is not doing genetics full time will have even more difficulty,” she says.
Dr. Gulley found “very interesting” the data about who is reading the test results and who is telling the patient. “That’s an argument that results should be as comprehensible as possible. There must be something that can be understood when a secretary is reading the results over the phone to a patient and that could be faxed to other people involved in care of the patient.” Reports should be written at the level of the generalist physician, in Dr. Gulley’s view, and include an interpretation about what the result means to that patient.
Dr. Grody calls the survey results “the single biggest surprise of the meeting” and says he was “kind of stunned.” They emphasize the discrepancy between laboratory directors who try to report complex test results comprehensively and clinicians who are trying to run a busy practice. “Those two factions keep diverging,” Dr. Grody says. “Our tests get more complex and they get busier with fewer resources. Clinicians have a tough job and a short time to do it. It’s not fair to ask them to do a lot more.”
Dr. Lyon, too, didn’t realize nonmedical personnel were handling requests and results. “The majority of tests have no family history, no symptoms, and are negative for the mutation panel. I hope those are the results that are being given over the phone. But if a positive family history, positive result, or a symptomatic patient was involved, I’m hoping that would be handled differently,” she says.
Many say the solution to this situation lies in educating future physicians about medical genetics in medical school. The goal is for molecular genetic testing to become part of a physician’s knowledge base. Says Dr. Trotter, “I find the medical students that we occasionally have in our office are much better informed [about genetic testing] than people my age.”
Unfortunately, that will take a generation. What can be done now?
Dr. Gulley, chair of the CAP Molecular Pathology Committee, heads the CAP’s effort to develop reporting guidelines for molecular pathology tests. “Existing guidelines in the CAP checklist document how laboratories should do test reporting,” she says. “We are now compiling the reporting aspect into one centralized document that will cover all aspects of molecular testing—oncology, infectious disease, inherited disease, and identity testing.” A draft report will appear in about a year.
She thinks it is important that there be a “model report.”
“People don’t need to use it verbatim, just as a guide for the format and content expected by clinicians. If we had a recommended content and format it would help nonspecialists to look at genetic laboratory reports and quickly find the information they need.” A professional organization or governmental body with broad representation would have to oversee such a model—that is, periodically review and modify it. “Ira Lubin’s group is getting input from a wide group of players,” Dr. Gulley says.
In her view, molecular genetics has many of the same issues as anatomic pathology with synoptic reports for cancer. “Some places have menus, and as the pathologist signs out a cancer case, he or she works through the checklist to ensure a complete report.” Another parallel between anatomic pathology reports and reporting of molecular genetic test results is the synthetic aspect. “Pathologists are trained and feel comfortable synthesizing data created beyond the walls of their own laboratory,” Dr. Gulley says, “including clinical presentation, physical examination, and other findings.”
A model format is “worth exploring,” Dr. Grody agrees. “How far we will be able to get on that remains to be seen. Laboratory directors in these high-complexity laboratories, who are board certified and highly trained, view this as both an art and a science and like to retain discretion for how they express things. Individual reports are often edited extensively by the director.” He acknowledges that there are a number of elements that probably should be included in every report, such as residual risk if the patient could still carry a mutation despite a negative screen, a recommendation for genetic counseling if appropriate, and clinical implications. “We may get to where we list essentials that should be in every report without having a complete text that laboratories have to follow.”
In his experience, the guidelines of the American College of Medical Genetics and the CAP checklist “come closest” to a model format that he has seen. In an appendix to the 2001 ACMG recommendations is a set of model reports with wordings for all combinations of mutations. “Laboratories could have used those word-for-word,” Dr. Grody says.
Some have objected that a too-tailored report amounts to practicing medicine. “I think we are practicing medicine as laboratory directors,” Dr. Grody says. A laboratory director should have the same latitude in crafting a laboratory report as a surgeon might have in crafting a surgical approach, he says. “Not everyone at the workshop agreed with that,” he admits. “But I have argued that is why we have board certification.” Since many directors of medical genetics laboratories are PhDs rather than MDs, Dr. Grody’s approach raises a philosophical question: Can non-MDs practice medicine? “That is why we have board certification open to PhDs,” Dr. Grody asserts.
Dr. Amos Wilson takes this position: “Language in the report does not have to be uniform. We need to have interpretations accurate, concise, and at the appropriate level. We don’t need to have a specific way to say a person is a carrier.” She thinks that many laboratory reports meet these criteria now. “My impression is that laboratories with geneticists are doing a pretty good job overall. I collect laboratory reports, and when I look at a competitor’s report I often find something I might want to use.” On the other hand, she concedes that she has a biased sample. “Laboratories with no geneticist—that is where my concern is much higher,” she says.
“We often ask ourselves to practice medicine,” she adds. “And we may go too far with our recommendations [in reports]. But we realize how complex this area is.”
At Focus Diagnostics, Dr. Amos Wilson has instituted a service called GenomEx for interpreting laboratory reports of CF screening tests. Clients—routine clinical laboratories that do not have a clinical molecular geneticist on staff—determine their results and transmit them electronically to Focus, along with patient preanalytic data, such as family history and ethnic background, via a Web-based ordering system. “We then send the performing laboratory an interpretive report tailored to its testing platform,” Dr. Amos Wilson says. GenomEx can provide clinical consulting services to client laboratories as well. “Vendors have made it simple to do data reduction, so the performing laboratory’s job has become easier,” she says. “But the level of complexity and the nuances of interpretation of the results have not changed.”
Focus Diagnostics launched the GenomEx service in early November 2005 and is closing agreements with clients now. “We found that clients have been willing to pay for this service,” Dr. Amos Wilson says. “It helps them from a liability perspective and puts them on a level playing field with esoteric laboratories. Most important, it helps them do the right thing, which they want to do.” She herself feels “good about this,” she says, because she has “moved from being part of the problem to being part of the solution.”
The CDC’s Dr. Lubin observes that opinions vary as to how far laboratory reports should go in making patient recommendations. “Typically, but not always, laboratories will provide an interpretive component with recommendations,” he says. Recommendations vary from referral to a genetic or other specialist to advising the clinician to order additional tests for the patient, family members, or both. “Recommendations are usually made to further clarify the meaning of the current test result and to alert the clinician that others in the family may be affected, at risk for disease, or a carrier of a disease-associated gene,” Dr. Lubin says.
That the medical genetics board is the only medical board that allows PhDs to take the examination is significant, in his view. “It requires PhDs to take a clinical rotation to understand the clinical features to which genetic tests relate,” he says. Some laboratory directors just provide a result, others relate the result to the medical condition. Though boarded professionals have the knowledge and experience to provide the interpretive component, there is a well-recognized line between the lab’s role and that of the clinician who sees the patient, Dr. Lubin says. He adds, “Ideally, I think, the laboratory report should provide sufficient information to integrate the result into the clinician’s overall recommendations to the patient, whether it be for treatment, counseling, or referral.”
Even if agreement can be reached on a prototype reporting format, however, it will still represent only expert opinion as to what works, what Dr. Gulley calls GOBSAT—good ol’ boys sittin’ around a table. Says Dr. Lubin: “We have virtually no data on how test reports are being used for clinical management. Whether they are being used to make proper decisions is not something that has been explored. We have measures based on professional perception of best practices, but we don’t have [outcomes] measures to evaluate our perceptions.”
Dr. McGovern of Mount Sinai School of Medicine agrees that outcomes studies are needed. She has proposed to do such a study in a joint project with the CDC and New York’s Wadsworth Center. In what she calls a 360-degree evaluation, information will be tracked to the laboratory, back to the clinician’s office, then to the patient interview to determine clinicians’ understanding of the report and how accurately results get back to patients. “It will be a complicated study to do,” Dr. McGovern says, “but very important. We hope to carry it out in the next year.”
Even with standardized, comprehensible reporting, in Dr. Trotter’s view, it will still be important for laboratories to have someone on staff who is available to answer questions about ordering or to help in understanding test results. In this way, he, too, sees medical genetics laboratories as being analogous to anatomic pathology. “If you call someone about a biopsy, you can get somebody to talk to you almost all the time,” he says, “which is not true for serum sodium.”
Dr. Grody agrees with the need for experts to be available to interpret genetic test results. “That is one reason why all our genetic test reports say that genetic counseling is recommended to explain these results to the patient and to give other family members the opportunity to be tested when appropriate,” he says. “It gives a hint that this is complex information that the clinician may not have the time, inclination, or knowledge to address, but there are experts out there who can help.” One conclusion of the 1997 paper on familial adenomatous polyposis is that “Physicians should be prepared to offer genetic counseling if they order genetic tests.”
For now, genetics specialist Redman takes a positive view. “General laboratories may very well be aware that they need to contract for genetic expertise,” she says. “Certainly there are many people qualified to act as advisors or consultants.”
Dr. Trotter believes there will have to be an oversight body to tie all these measures into one package and to define them as standard of care. He says, “Standard of care should include analytical expertise, a clear method of transmitting information, and a reference consultant to help clients like me.” Doing medical genetic testing without a consultant has become, he says, like running an MRI machine without a radiologist: “It takes good pictures, but they need interpretation.”
William Check is a medical writer in Wilmette, Ill.