Taiwan's single-payer National Health Insurance system encompasses its entire population, and it's National Health Insurance Research Database tracks all medical claims in the system. That makes it easy to conduct nationwide population studies.
Two Taiwanese research teams availed themselves of that database to explore in-depth a surprising relationship between the birth month of children and rates of ADHD diagnosis.
In principle, the two should be unrelated. The likelihood of diagnosis should be the same regardless of the month a child is born. But the data are clear that this is not so. Children born late in summer are the most likely to be diagnosed with ADHD, and those in autumn are the least likely.
Using a nationwide database of over 29 million persons, one of the teams (Hsu et al.) found that children born in April were 6% more likely to be diagnosed with ADHD than the year-round mean, those in May 12% more likely, those in June 20% more likely, and those in July and August well over 25% more likely.
Conversely, children born in September were 19% less likely to be diagnosed with ADHD than the year-round mean, followed by a gradual increase in likelihood with each succeeding month until the following September.
The second team (Chen et al.) analyzed some 9.5 million children and adolescents in the same reserch database, and found that those born in August were 67% more likely to be diagnosed with ADHD than those born in September, after adjusting for age, sex, residence, and income. August births were also almost twice as likely (80% more likely) as September births to be on long-term treatment with ADHD medications.
The first team also performed a meta-analysis of eleven studies with a combined total of over 580,000 participants in North America (the U.S. and Canada), Europe (U.K., Germany, Norway, Sweden, Denmark), Asia (China, Taiwan, South Korea), and Oceania (Australia). Children born in the summer (June through August) were 13% more likely to be diagnosed with ADHD than the year-round mean, whereas those born in autumn were 13% less likely to be diagnosed with ADHD. This confirms that this pattern is not confined to Taiwan. It is worldwide.
Note carefully that the sharp discontinuity between August and September corresponds with the break-of point that decides which children get assigned to which school class. Anyone who turns a certain age by the start of the school year in September is included in the class associated with that age, whereas those turning the same age later are held back in the following class. That means that in any given class, those born in September are the oldest children and those born in August the youngest.
As signaled earlier, the likelihood of an ADHD diagnosis should be independent of something as obviously arbitrary as a birth month. That suggests there may be an unconscious bias trending against younger students when it comes to diagnosis.
Chen et al. concluded, "The effect of relative age on diagnoses and prescriptions was determined to last from childhood to adolescence but attenuated with age. Relative age is an indicator of brain maturity in cognition, behavior, and emotion and may thus play a critical role in the likelihood of being diagnosed as having childhood mental disorders and subsequently being prescribed psychotropic medication. Therefore, clinicians should consider the relative age effect in the childhood mental health care context."
Mu-Hong Chen, Kai-Lin Huang, Ju-Wei Hsu, Shih-Jen Tsai, Tung-Ping Su, Tzeng-Ji Chen, Ya-Mei Bai, "Effect of relative age on childhood mental health: A cohort of 9,548,393 children and adolescents," Acta PsychiatricaScandinavica (2021), online ahead of print, https://doi.org/10.1111/acps.13327.
Chih-Wei Hsu, Ping-Tao Tseng, Yu-Kang Tu, Pao-Yen Lin, Chi-Fa Hung, Chih-Sung Liang, Yun-Yu Hsieh, Yao-Hsu Yang, Liang-Jen Wang, Hung-YuKao, "Month of birth and mental disorders: A population-based study and validation using global meta-analysis," Acta Psychiatrica Scandinavica (2021), online ahead of print, https://doi.org/10.1111/acps.13313.
_logo%201.svg)
