Gender

Suspect Statistics From Plastic Surgeon Society Say 'Gender Confirmation Surgeries' Up 19 Percent Last Year

There were 3,256 such surgeries in 2016, says the American Society of Plastic Surgeons. But how it got this number is anyone's guess.

|

ASPS

A host of U.S. media outlets are reporting that 2016 saw a 19 percent spike in "gender confirmation surgeries," a term used by the American Society of Plastic Surgeons (ASPS) to describe an array of elective facial, chest, and genital procedures performed on transgender patients. The statistic comes from an annual ASPS report, which for the first time includes information on such procedures. But are the numbers legit? A day after publishing an article touting the ASPS findings, Reuters wound up pulling the piece, citing the Society's inability to "vouch for the data which [it] originally supplied."

Yet as of Tuesday evening, the ASPS website still contains the statistics which Reuters published and pulled. It says there were 3,256 "gender confirmation surgeries" in 2016 and 2,740 in 2015. Some 1,759 of last year's procedures involved transfeminine procedures (i.e., those performed on trans women) and 1,497 involved transmasculine procedures; the previous year, it was 1,380 and 1,360, respectively. Using these stats, we get a 19 percent increase in gender-related surgeries overall; a 10 percent increase in surgeries for transgender women; and a 27 percent increase on surgeries for transgender men.

A media relations coordinator for MediaSource, which handles ASPS public relations, says these figures are correct. "There's no problem with the data," Drew Schaar told me Tuesday afternoon, noting that the methodology used for gender-related procedures was the same process used to tally all cosmetic and reconstructive procedures in the ASPS report—some 17.1 million overall, according to ASPS, though 15.1 of these were the minimally invasive sort (e.g., Botox injections, chemical peels, laser hair removal).

According to Schaar, Reuters pulled the article after seeing more information than is publicly available about the data and deciding it didn't meet the news agency's own in-house standards. But he could give no more information about what, precisely, failed to pass muster, and he said more detailed information about survey methods and findings could not currently be provided. Alas, we're left to guess what may have made Reuters lose confidence in the ASPS conclusions. But I'll offer a few of my own observations, based on what limited info is available—the most important bit of which is probably an ASPS fine-print statement that "all values are projected."

The underlying methodology of ASPS's annual survey may be flawed. Each year, ASPS solicits survey responses from all of its 7,000-some members, a group that includes 94 percent of all board-certified plastic surgeons in America, according to ASPS. With the 2016 survey, 703 ASPS-member physicians responded, down from last year (when 733 doctors responded) and 2013 (801 responses). Using answers from this member-surgeon sample, ASPS estimates the total number of cosmetic- and reconstructive-surgery procedure—from chin implants to Botox injections to labiaplasty—performed by American Board of Medical Specialties doctors in a given year. The raw response numbers from ASPS members are thus "aggregated and extrapolated to the entire population of more than 24,500 board-certified physicians mostly likely to perform cosmetic and reconstructive plastic surgery procedures."

This means that only about 2.8 percent of the population allegedly reflected are actual ASPS members who took last year's survey. So that's potential limitation number one—although not necessarily a disqualifying one. Political polls extrapolated to all of American voters are often based on much, much less. But these are, theoretically, run through researchers' weighting magic to make them plausibly representative of the population in question.

In this case, it seems suspect to extrapolate info from physicians who have actively elected to belong to a professional plastic surgeon society to all American physicians who are certified or "likely" to do any sort of plastic surgery procedures. Yet because ASPS is mum on its methodology, we don't know if or how they controlled for differences in the practices of ASPS doctors and outside doctors. If not done properly, this would have major potential to inflate plastic surgery rates across the board.

We don't really know what "gender confirmation surgery" means here. In its main report on plastic surgery in 2016, ASPS breaks data down by procedure. There's one line for "Buttock Implant" and another for "Buttock Lift." Liposuction is separate from "Tummy Tucks," and injectable collagen-based fillers from Hyaluronic acid-filled ones. Yet when it comes to the procedures lumped together as "gender confirmation surgery," we aren't given any specifics. Right now, ASPS merely tells us that the category includes an array of "transmasculine" and "transfeminine" procedures, from genital reconstructive surgery to procedures that help make the face more masculine or feminine. (The now-withdrawn Reuters article reported that only 0.9 percent of transfeminine procedures and no transmasculine procedures in 2016 involved genital surgery. For transgender women, 92 percent of procedures were breast or chest operations and 7 percent facial operations. For trans men, 95 percent were chest or breast operations and 5 percent facial operations.)

How does ASPS decide which surgeries and procedures fall in this category? Did this list fluctuate at all between 2015 and 2016? And how did ASPS determine whether patients obtaining various procedures (many of which aren't exclusive to transgender patients) were doing so for reasons related to gender identity or to vanity or back pain? Again, here is an area where ASPS methods could be perfectly sound but its refusal to relate them leaves room for doubt. If the ASPS methodology is legit, why aren't specific gender-related surgeries listed like all other procedures? Why the need for such secrecy?

We don't know how many doctors, or what types, responded. Of the 703 doctors who took ASPS' 2016 survey, how many conducted any one of the procedures labeled as gender confirmation? We have no idea based on what ASPS has made public. Perhaps only a very small number of survey responses contained any information about transgender patients, making larger extrapolations statistically insignificant. Or perhaps a significant number of members reported a few, but the vast majority came from only a few physicians or practices.

The ASPS report leaves us in the dark about the average number of procedures per physician who did see transgender patients, as well as the types of doctors in the survey sample. But if a few doctors in the survey reported most procedures, and most of the surveyed docs reported none, then extrapolating from the sample-population here could still be flawed. The same thing goes if survey respondents included a larger number of gender-identify specializing doctors than are found in the plastic-surgeon population overall.

We don't know how doctors in the 2015 survey compare to doctors in the 2016 survey. As I noted above, the ASPS survey is sent out to all of the group's members but only elicits around a 10 percent response rate, on average, over the past few years. 2015 was the first year that ASPS posed gender-surgery-specific questions to its members, doing so again last year. Could the inclusion of the new survey question—and this inclusion becoming better known by the 2016 survey—have influenced more plastic-surgeons who specialize in transgender patients to participate? It's certainly possible.

The bottom line is, we have no idea about the makeup of 2015 respondents versus 2016 respondents overall. The entire 19 percent leap reported here—which, remember, is premised on we-don't-know-many actual surgeries reported by actual doctors—could come down to a difference of just a few more patients when we look at raw data, and/or be accounted for entirely based on the differing doctor samples from year to year.