It should be easy for you to find out what proportion of people who do trans change their mind. Some evidence always comes in handy in cases like this. And it seems that you do your own research.
I've seen reported "desistance rates", and they vary widely due some key differences (and different types of methodology flaws) in the studies.
Early studies suggesting desistance rates of near 90+% had the flaws of having practically no vetting, and making the assumption that if someone didn't respond to the 3-year follow-up survey, that means "they must have desisted", which isn't an assumption that could be made, and could've artificially inflated their desistance rate by upwards of 40 percentage points.
Later studies (which were aimed at refuting those earlier studies) were more rigorous in that regard by not making assumptions about the people who were follow-up non-responders, but were much more selective with higher barriers of entry to be considered "a good candidate to study" (like the Sweden studies). So many of the studies that concluded that regret or "detransition" was below 5%, weeded a lot of people out. (I know I've linked the study before, I'll do some digging and see if I can find it again, but it was something like over 60% of the people who identified as trans and volunteered to be part of it, got turned down for not having the proper history of working with a therapist for a period of 2 years prior to embarking on the transition)
So it stands to reason that the widely disparate results emerged.
One study said "anyone who says they're trans, must be trans, no further documentation required, and if they don't answer our call 3 years from now, we'll count that in the desist column"
Other studies said "we're going to exclusively work with people who have claimed this identity for 4 years, and have worked with a therapist for at least 2 of those years"
It stands to reason that working exclusively with a cohort who's been identifying that way for 4 years, and working with a counselor for 2 years and still says "yes, this is who I am, we've covered and ruled out any other psychological factors, I still want to do this" is going to report a much lower "regret rate" than a cohort of people who merely proclaimed it (with no extensive history provided) and who the researchers are making some serious assumptions about when they can't get ahold of them 3 years later.
In fact, I'd suggest that the studies that produce the <5% desist rates, are using criteria for what constitutes "a real trans person" (in order to select their cohort) that many would find "offensive". What those studies are basically suggesting is that the desistance rate is low among what the researchers categorize as "real trans person" (defined by having identified that way for 4 years, and 2 years of working with a counselor who agrees with the assertion) --not surprising though. More due diligence up front, leads to less regret down the road (like with many things)
So, on the spectrum of the two methodologies/cohorts listed, which of the two does the "student goes to the school counselors office, identifies as trans, and wishes to conceal that information from their parents" more closely match up with?
And even the term "desistance" actually comes with some caveats as well due to "desistance" not necessarily meaning "going to back to a cis person"
Per TheConversation (in an interview with a researcher who is, themselves, trans):
One of the more interesting takes I heard on persistence and desistance came from UCSF gender clinician Erica Anderson, who is transgender herself. She views the very notion of measuring persistence/desistance as something of a fool’s errand, because such definitions are mediated by changing cultural norms, the self-perceptions of children and the ways that researchers interpret them.
...doesn’t even like to use the terms persistence and desistance. Those words imply something fixed — a binary state of yes or no. But younger generations of transgender people — and even younger generations in the general population — see gender as more protean, even customizable. Of nearly 28,000 respondents to the 2015 U.S. Transgender Survey, more than one-third said they were some form of nonbinary. That means they may identify as both male and female, neither male nor female, or sometimes male, sometimes female.
Meaning, if one third are suggesting they're "sometimes male/sometimes female", it maybe be a good idea to use caution before address a "sometimes" dilemma with an "always" solution.
If they're basically acknowledging that one-third of an entire generation is identifying as nonbinary (including a "sometimes male/sometimes female" situation, in theory, there's the potential for a person to actually "desist" multiple times, but people tend to not count it when it's desisting to a non-binary identity, or desisting to some or situation that's not "cis-normative")