ACMG Backs Down a Bit

57 sauceA year ago, the American College of Medical Genetics and Genomics (ACMG) released its Recommendations for Reporting of Incidental Findings in Clinical Exome and Genome Sequencing. As I reported in a July 2013 post, the core recommendation was this: “The ACMG recommends that for any evaluation of clinical sequencing results, all of the genes and types of variants in the Table should be examined and the results reported to the ordering physician.” Specifically, the ACMG recommended that whenever a lab does whole genome or whole exome sequencing on a patient, it should examine all 57 [now 56] genes on the list included in the Recommendations and report any clinically significant findings to the ordering physician. It would then be the duty of that physician “to provide comprehensive pre- and post-test counseling to the patient.” Most controversially, the ACMG recommended that the test findings “be reported without seeking preferences from the patient and family and without limitation due to the patient’s age.” As I characterized it in the July post, “patients should be given the 57-gene screening whether they want it or not and told the results even if they say they don’t want them.”

A huge controversy erupted. Critics stressed the apparent affront to the ethical principle of patient autonomy, the limited state of medical knowledge about the pathway from genetic mutations to disease, the problems in deciding what conditions are “amenable to medical intervention,” and the special concerns presented by applying the Recommendations to child patients (note “without limitation due to the patient’s age,” above). (There was inconsistency in the ACMG’s application of the Recommendations to children. As GLR Contributing Editor Jen Wagner has written elsewhere, before it issued the Recommendations the ACMG had issued a joint policy statement with the American Academy of Pediatrics that stressed the need for parental involvement in the genetic testing of newborns and children.)

As the controversy over the Recommendations unfolded, journals published special issues devoted to the topic. The ACMG published a “Clarification” in the journal it sponsors, Genetics in Medicine, that dealt with the major issues raised by the critics, but it did not modify the Recommendations. Now the ACMG is backing down, at least a bit, and showing greater deference to patient autonomy. In an April 1, 2014, press release, the ACMG announced that its board of directors had voted to approve an “update” to the recommendations. The revised version recommends that patients be offered “an opportunity to opt out of the analysis of medically actionable genes when undergoing whole exome or genome sequencing.” The opt out discussion between doctor and patient should take place at “the point where the sample is sent [to the lab] rather than at the time when results are received by the clinician, as was originally recommended.” The update is said to have followed the expression of “divergent and valuable opinions” by the ACMG membership, and to reflect an apparent “consensus among ACMG members.”

As I understand the updated approach, the right to opt out would involve the analysis and interpretation of the sequencing results. The sequencing of the whole genome or exome would be done in any event. If a patient opted out of part of the 56-gene panel, the patient’s sequence at the relevant loci would not be analyzed, meaning that there would be no interpretive findings for the lab to return to the clinician and patient. The opt-out right would not put the lab and doctor in the position of having findings that they would then have to keep from the patient—that information would not be generated in the first place.

My UNC colleague Jim Evans, a medical geneticist who is a leading participant in genomic policy debates, had recommended just such a change—he called it a “compromise”—last fall. In an October 2013 editorial accompanying a special issue of Genetics in Medicine, which he edits (and where the ACMG will publish the current and any future updates), Evans wrote: “let us routinely examine the ‘ACMG 56’ for clearly deleterious mutations when genome-scale sequencing is performed. But let us allow patients to opt out of such analysis of their specimen.” He went on to stress that he was “uncomfortable” with his recommendation, because it “affords patients an illusory degree of autonomy and that, in reality, discussions about an unlikely and hypothetical scenario that they choose to opt out of that does not really boil down to an informed choice.” I think the point is that it would be difficult for a patient to grasp what it is like to be in the unlikely situation of receiving a positive report about a potentially dangerous mutation; hence, any choice based on an evaluation of such a situation would be ill-informed. Nonetheless, Jim concluded, this compromise, with discomfort on both sides, represents the best currently-available approach.

I agree with everything Jim Evans said, and thus with the ACMG’s new approach. I sympathize with doctors who are concerned about having to practice with one metaphorical hand tied, unable to obtain and use potentially significant information about their patients. But I think that patient autonomy—even if “illusory”—is the higher value in this case. This is particularly true because of the relatively attenuated clinical value of at least some of the information that is at stake here: patient choices may be ill-informed, as Jim said, but nobody is fully informed, as critics of the original ACMG Recommendations argued. Patients should not have autonomy to make doctors do things that contradict their best medical judgment, but they should be able to make them refrain from doing things.

A final question is the one I raised in my original post: what is the practical effect of the Recommendations, as originally promulgated and now as updated? Do they now constitute a mandate to physicians and a standard of care for genetic medicine?