|The kid in red is in the same grade and classroom as the other four
In my post ofSeptember 20, 2010, Immaturity in YoungChildren: Officially a Disease, I described two studies published ina very obscure journal, the Journal of Health Economics, thatboth found nearly identical data about the diagnosis of ADHD in schoolchildren. In the these articles, two different research groups (Evans,Morrill, &Parente, 29, 2010 657–673; Elder, 29 2010, 641–656) using fourdifferent data sets in different states came to the same conclusion.
In one, roughly 8.4 percent of children born inthe month prior to their state’s cutoff date for kindergarten eligibility – whotypically become the youngest and most developmentally immature children withina grade – were diagnosed with ADHD, compared to 5.1 percent of children born inthe month immediately afterward. The study also found that the youngestchildren in fifth and eighth grades were nearly twice as likely as theirolder classmates to regularly use stimulants prescribed to treat ADHD! The results of the second study were quite similar.
Translated into numbers nationwide, as Steindór summarized in his comment on myblog, this would mean that between900 thousand (Elder) and 1.1 million (Evans et al. 2010) of those childrenunder age 18 in the US diagnosed with ADHD (at least 4.5 million) aremisdiagnosed.
Now, a year and a half later, another study,published in a more widely read journal and reported widely in the news, came upwith the exact same conclusion. (“Influence of relative age ondiagnosis and treatment of attention-deficit/hyperactivity disorder in children”by Richard L. Morrow, et. al., CanadianMedical Association Journal, published on line March 5, 2012).
Ina cohort study (a study of a group ofindividuals with something in common followed over time) of more than 900,000Canadian children, researchers found that boys born in the month of December(the cutoff birth date for entry to school in British Columbia) were 30% morelikely to be diagnosed with ADHD than boys in their grade who were born theprevious January.
Thisnumber was even more dramatic in the girls, with those born in December 70%more likely to be diagnosed with ADHD than girls born in January.
Inaddition, both boys and girls were at a significantly higher risk of beingprescribed an ADHD treatment medication if they were born in the later monththan in the earlier one.
"It could be that a lack of maturity in theyoungest kids in the class is being misinterpreted as symptoms of a behavioraldisorder," said lead author Richard L. Morrow.
Could be? About about “is?”
Someof these behaviors could include not being able to sit still, not being able tofocus and listen to the teacher, or not following through on a task, he added.
"Youwouldn't expect a 6- and 9-year-old to behave the same way, but we're oftenputting a 6- and 7-year-old in the same class. And we're learning that youcan't expect the same behaviors from them," he added. "We would liketo avoid medicalizing a normal range of childhood behaviors." No sh*t!
This problem has been complicated recently by the fad of "redshirting" children for kindergarten: overachieving parents purposely starting them at age six rather than five in order to give them a competitive advantage academically over their classmates. Now children in the same class may be as much as two years apart in age.
Thestudy authors went on to note thatpotential harms of overtreatment in children include increased risk forcardiovascular events, as well as effects on growth, sleep, and appetite. There was no mention of the harm of makingthis diagnosis and using these potentially toxic medication instead of investigating and addressing possiblepsychosocial reasons for “hyperactivity” such as a chaotic family environment orabusive and/or inconsistent parenting practices.
This brings up the issue of the risk to the heart and the rest of thecardiovascular system posed by stimulant use. There have been several studies recently published that have been reported inboth the medical and lay media that claim that this risk is minimal.
This is in an interesting contrast to the publicity about anarticle, published this week in BMJ Open (theonline version of theBritishMedical Journal
) thatpurported to show that the use of sleeping pills increases the risk of dyingfrom all causes by a factor of 4 over just two and a half years. Sleeping pills are generally regarded as far less dangerous and less likely to be abused than stimulants. The FDA categorizes benzos as "Schedule IV" (lower likelihood of abuse) and stimulants as "Schedule II" (most likely to be abused short of the illegal "Schedule I" drugs).
That study about sleeping pills seemed to me to be a bit hard to believe, especially sinceepidemiological studies are notoriously unreliable. But even if the numbers are valid, the factthat the risk of death from all causesincreases most likely means that there is some other characteristic, or a bunch morecharacteristics, of the population of people who are prescribed sleepers thatare not characteristic of other populations. Those additional factors might explain thefindings.
As for stimulants, inthe February, 2012 issue of the AmericanJournal of Psychiatry, there is an article on methylphenidate(Ritalin and its variations) and risk of heart problems in adults. Using a large medication database, researchers matched about 44000methylphenidate (MPH) users and about 176,000 controls.
They looked at main theincidence of a cardiac event defined as a myocardial infarction, stroke,ventricular arrhythmia, or sudden death. They found a 117% increased risk - orover double the risk – in the Ritalin group. After adjustment for somepotential confounding factors, the risk was still 84% higher.
The news storiesabout the study on the benzo’s seemed to be meant to scare people out of usingthem, while the stories about increased risk in stimulant users seemed to bemeant to reassure people about using them. Of course, both of these studies described relative risk and not absoluterisk (See my postStats.comfromNovember2, 2011).
This means that “doublethe risk” means the risk might go from, say, a tenth of a percent to two tenthsof percent. Double a very small risk isstill a very small risk. The absoluterisk in this example would have gone up just one tenth of one percent. Still, if millions of people are getting the prescriptions,this increased risk can still turn out to apply to a sizeable number of people.
Physicians will not be able to see the increased risk in their clinicalexperience. As Nassir Ghaemi says,“They don't happen in 10-20% ofpatients in our practice; they happen in 1-2% (or 0.1-0.2%), and so, theaverage clinician, faced with a welter of patients, doesn't make the causalconnection.”
The question should be, what are the risks versus the benefits from taking the medication. For sleeping pills, for instance, one mightwant to know if there is a much larger increased risk of death for people whoare sleep deprived. For example, before the practice was stopped, medical interns would routinely work 36 hour shifts. Fatal accidents on the car trip from the hospitalback home were not all that unusual.
Then there is the whole question of other, non-pharmacological treatments,which is relevant for both the use of sedatives and stimulants. Of course, they do not work for everyone either.
An editorialin the same issue of the American Journal of Psychiatry as the study of Ritalinin adults sounded reassuring about stimulant use. Based on that study, I’m not so reassured.