Influence of relative age on diagnosis and treatment of attention-

deficit/hyperactivity disorder in children

Running head: Relative age in ADHD diagnosis

RL Morrow, EJ Garland, JM Wright, M Maclure, S Taylor, CR Dormuth

4 tables; 1 figure; 2,483 words

Author information:

Richard L. Morrow, MA, is a health research analyst in Anesthesiology, Pharmacology and

Therapeutics at the University of British Columbia, 210-1110 Government St, Victoria, BC,

V8W 1Y2, Canada, tel: 250-590-5955; fax: 250-590-5954, richard.morrow@ti.ubc.ca.

E. Jane Garland, MD, FRCPC, is clinical professor of Psychiatry at the University of British

Columbia and clinical head of the Mood and Anxiety Disorders Clinic, BC's Children's Hospital,

BC Mental Health and Addictions Services, P4-215 Mental Health Building, 4500 Oak Street

Vancouver V6H 3N1, Canada, tel:604-875-2737, fax: 604-875-2099, jgarland@cw.bc.ca.

James M. Wright, MD, PhD, is professor of Anesthesiology, Pharmacology & Therapeutics and

Medicine at the University of British Columbia, 317-2176 Health Sci Mall, Vancouver, BC, V6T

1Z3, Canada, tel: 604 822-4270, fax: 604 822-0701, jim.wright@ti.ubc.ca.

Malcolm Maclure, ScD, is professor of Anesthesiology, Pharmacology & Therapeutics at the

University of British Columbia (Rm 3201 JPPN 910 West 10th Ave Vancouver, BC, Canada)

and a co-director of research and evidence development in Pharmaceutical Services Division of

1  

 
the BC Ministry of Health Services, Victoria, BC, tel: 250-952-1010, fax: 250 952-1584 ,

malcolm.maclure@gov.bc.ca.

Suzanne Taylor, PharmD, is executive director of Drug Use Optimization branch of

Pharmaceutical Services Division of the BC Ministry of Health Services and clinical professor at

the UBC Faculty of Pharmaceutical Sciences, Vancouver, BC, tel: 604 660-1217, fax: 604 660-

2108, suzanne.taylor@gov.bc.ca.

Colin R. Dormuth, ScD, is assistant professor of Anesthesiology, Pharmacology and

Therapeutics at the University of British Columbia, 210-1110 Government St, Victoria, BC

V8W 1Y2, Canada, tel: 250-388-9912, fax: 250-590-5955, colin.dormuth@ti.ubc.ca.

Corresponding author: Richard L. Morrow, University of British Columbia, 210-1110

Government St., Victoria, BC V8W 1Y2 Canada. Tele: 250-388-9912, Fax: 250-590-5954,

email: richard.morrow@ti.ubc.ca. (This email address can be published.)

Funding: This work was funded by a Catalyst Post-Market Drug Safety and Effectiveness grant

from the Canadian Institutes of Health Research (grant DSA-103525) and from a 5-year

renewable grant to the University of British Columbia from the British Columbia Ministry of

Health Services.

Conflict of interest: None to declare.

Abbreviations:

ADHD attention-deficit/hyperactivity disorder

BC British Columbia
2  

 
CI confidence interval

Key words: attention deficit, pharmacotherapy, diagnosis, stimulant medications, cohort study,

relative age effect

Author contributions:
All authors participated in the conception and design of the study and in analysis and

interpretation of the results. Richard Morrow and Colin Dormuth drafted the manuscript, and all

authors critically revised it and approved the final version for publication.

3  

 
Abstract

Background: The annual cut-off birth date for entry into school is December 31 in British

Columbia, Canada, so children born in December are typically the youngest in their grade. We

conducted a cohort study to examine the influence of relative age on diagnosis and drug

treatment for attention-deficit/hyperactivity disorder (ADHD).

Methods: The study cohort consisted of 937,943 children in British Columbia who were 6 to 12

years of age at any time from December 1, 1997, to November 30, 2008. We calculated the

absolute and relative risk of being diagnosed with ADHD and of receiving a prescription for

methylphenidate, dextroamphetamine, amphetamine or atomoxetine for children born in

December compared to children born in January.

Results: Boys who were born in December were 30% (relative risk 1.30; 95% CI: 1.23-1.37)

more likely to be diagnosed with ADHD than boys born in January. Girls were 70% (relative risk

1.70 (CI: 1.53-1.88) more likely to be diagnosed with ADHD if born in December compared to

January. Similarly, boys were 41% more likely (relative risk 1.41; CI: 1.33-1.50) and girls 77%

(relative risk 1.77; CI: 1.57-2.00) more likely to be prescribed an ADHD medication if born in

December compared to January.

Interpretation: Our analyses provide evidence of a relative age effect in the diagnosis and

treatment of ADHD in children in British Columbia. These findings raise concerns about

potential harms from over-diagnosis and over-prescribing, which include adverse effects on

sleep, appetite and growth, and increased risk of cardiovascular events.

4  

 
Attention-deficit/hyperactivity disorder (ADHD) is the most commonly diagnosed

neurobehavioural disorder in children and is widely treated with stimulant and non-stimulant

medication. [1, 2] While the specific causes of the disorder are unknown, both genetic and

environmental factors have been hypothesized to contribute and may interact. [3, 4] Several

studies have investigated the potential association between season of birth and the incidence of

ADHD. [5-9] Season of birth has been studied for a number of mental and behavioural disorders,

including schizophrenia for which an association is well-established [10] and autism for which

the evidence of an association is weaker. [11-13] Among studies hypothesizing an etiological

relationship between season of birth and ADHD, a consistent seasonal pattern has not been

found. [5-9]

While previous research on birth date hypothesized an etiological relationship to the

disorder, two recent studies using survey and private health plan data in the United States

reported that birth date was a predictor of diagnosis and treatment of ADHD based school

starting age. [14, 15] Due to the cut-off birth date for entry into kindergarten or grade one,

children within the same grade may be almost one year younger and less mature than their peers.

These US studies found that children who were relatively younger than their peers within the

same grade were more likely to be diagnosed or prescribed medication for ADHD. This pattern

persisted across states with school entry dates at different times of the year and depended on the

cut-off date rather than the season. This has been termed the “relative age effect” and has been

found to affect outcomes in education and athletics. [16, 17]

The potential effect of children’s relative age within grade on treatment and diagnosis of

ADHD has been little studied and has not been analyzed in populations outside the United States,

where reported prevalence of ADHD is high relative to international comparisons and where

5  

 
ADHD medication usage exceeds levels in other developed countries including Canada [18, 19].

We were interested to know whether the effect of relative age was inherent in the US context

(due to underlying factors such as differences in the health system, cultural attitudes or

pharmaceutical industry marketing), or whether it would be present in the different context of

Canada. We conducted a cohort study to explore the influence of relative age on diagnosis and

treatment of ADHD in a large sample of children with data from a nearly universal public health

plan in the setting of British Columbia (BC), Canada. In BC the cut-off birth date that allows

entry into kindergarten or grade one is December 31. Consequently, children born in December

are typically the youngest in their grade. We hypothesized that the younger children within each

grade would be more likely to be diagnosed with ADHD and to be prescribed ADHD

medications.

Methods

Data sources

We used linked data from provincial administrative health databases for prescription

drugs (PharmaNet), physician services (Medical Services Plan) and hospital admissions

(Canadian Institute for Health Information Discharge Abstracts Database). PharmaNet contains

records of all medications dispensed at community pharmacies in British Columbia, and rates of

underreporting and misclassification were expected to be minimal because data quality checks

are performed when claims are transmitted.[20] Similarly, physician services and hospital

admissions data were expected to be reliable on the basis of studies comparing patient charts

with administrative data in Canada.[21, 22] Our data was representative of most of the BC

population and excluded only the approximately 4% of residents who are federally insured and

not covered by the BC Medical Services Plan (aboriginal people, prisoners and military).

6  

 
Study design

The study period was December 1, 1997, to November 30, 2008. The study cohort

consisted of 937,943 children who were ≥6 and ≤12 years of age at any time during the study

period and who had been registered in the BC Medical Services Plan (MSP) for at least one year

before their entry into the cohort. Children entered the cohort at the latest of their sixth birthday,

the start of the study period or their having been covered by MSP for one year, and follow-up

continued until the earliest of the day before their thirteenth birthday, the end of the study period

or the end of their MSP coverage. Medications included in the analysis of prescribing were

methylphenidate, dextroamphetamine, amphetamine and atomoxetine.

We categorized patients according to their calendar month of birth and calculated the

proportion of patients prescribed an ADHD medication and the proportion of patients diagnosed

with ADHD by each month. To estimate the influence of relative age, we calculated the absolute

risk difference and relative risk of receiving an ADHD diagnosis for children who were born in

December compared to those born in January based on these proportions. We similarly estimated

the absolute and relative risks of receiving an ADHD prescription. A Cochran-Armitage trend

test was used to test the influence of relative age on diagnosis and prescribing across all months

of birth from January to December.

In addition to our main analysis, we conducted analyses stratified by year and by age and

a series of sensitivity analyses. In the analysis stratified by year, we analyzed data in one-year

categories from December to November of the following year (e.g., 1997-98 refers to 1

December 1997 to 30 November 1998). In the analysis stratified by age, we used one-year age

groups from age 6 to 12 years. As sensitivity analyses, we conducted the same risk calculations

with cohorts of children born in other periods in close proximity to the school entry cut-off date.

7  

 
We analyzed cohorts of children born during December 17-31 in relation to children born during

Jan 1-15 (in which a relative age effect was predicted) and compared the result to control

analyses comparing the birth date cohorts of December 2-16 versus December 17-31 and January

1-15 versus January 16-30 (in which no effect was expected). We also analyzed a cohort of

children born during December 28-31 relative to a cohort born during January 1-3. These

analyses included children based on the same criteria as our main analyses, except for the birth

dates to define the cohorts.

Results

The percentage of children diagnosed or prescribed medication for ADHD is summarized

by month of birth in Table 1. The January and December cohorts were similar in size and

children in both cohorts had a mean age of 7.8 years at cohort entry. The proportion of children

receiving an ADHD diagnosis or treatment increased in birth months following January for both

boys and girls until September with a plateau until December (although the percentage of boys

diagnosed decreased slightly in December). These rising trends in diagnosis and treatment by

birth month were confirmed with Cochran-Armitage tests (all one sided p-values, p<.0001). The

variation in ADHD treatment by birth month is depicted in Figure 1, showing an increase in the

percentage of children treated after January until the later months of the year.

The risk of diagnosis and treatment for ADHD in the December cohorts as compared to

the January cohorts is reported in terms of absolute risk difference and relative risk in Table 1.

Boys who were born in December were 30% more likely to be diagnosed with ADHD than boys

born in January. Girls were 70% more likely to be diagnosed with ADHD if born in December

compared to January. Similarly, boys were 41% more likely and girls 77% more likely to be

prescribed an ADHD medication if born in December compared to January.

8  

 
 The one-year period prevalence of diagnosis and treatment for ADHD increased

gradually from 1997-98 to 2007-08 for both boys and girls, peaking in the most recent years

(Table 2). The relative risk of diagnosis and treatment for ADHD in the December cohorts as

compared to the January cohorts remained relatively stable over time (with no statistically

significant heterogeneity in relative risk across yearly strata). An elevated relative risk of

diagnosis and treatment for ADHD among children born in December compared to January was

present in all age groups from 6 through 12 years (Table 3). The relative risk for girls born in

December diminished at the higher end of this age range for both diagnosis and treatment (chi-

square tests for heterogeneity of relative risk, p<.05 and p<.001 respectively).

Sensitivity analyses

In the sensitivity analyses, children born during the last two weeks of the year were at

higher risk of receiving diagnosis and treatment for ADHD than children born in the first two

weeks of the year in results similar to our main analyses in analysis 2, whereas our control

analysis of periods within December and January produced results close to the null in analyses 1

and 3 (Table 4). When we considered cohorts of three days’ duration before and after the school

entry cut-off in sensitivity analysis 4, results were similar to our main analyses. Risk of diagnosis

and treatment was comparable to the analysis by birth month for boys, and for girls the reported

risks were in the expected direction although the sample size was small and the risk of treatment

was not statistically significant.

9  

 
Interpretation

Main findings

Our study reports a pattern of prescribing and diagnosis of ADHD that is consistent with

the hypothesis of a relative age effect. Children born in the month before the province’s school

entry cut-off date would typically be the youngest and least mature within their grade and are at a

higher risk of treatment and diagnosis. Our sensitivity analyses support this interpretation and

discourage the alternative interpretation that an underlying seasonality related to the etiology of

the disorder is producing this effect. A sudden change in risk apparently occurs at the school

entry cut-off date at the end of the year. Even children born in the three-day period prior to the

end of the calendar year were observed to experience a higher risk than children born in the

following three days in the new year, but they must be considered to be born in the same season

(within a six-day period). The attenuation in rising risk for the birth months from October to

December, as shown in Figure 1 and Table 1, likely reflects that children predisposed to ADHD

have early behavior problems that cause them to be held back from school for a year to allow

them more time to develop sociable behavior.

Explanation and comparison with other studies

The relative age effect we report based on population-wide public data in the Canadian

province of BC is consistent with analyses of relative age in two studies using survey and private

health plan data in the United States [14-15]. In our study a sudden change in the percentage of

children diagnosed or treated for ADHD occurs around the BC school entry cut-off date of

December 31, while in the US studies a discontinuity occurs around school entrance dates

occurring at different times of year for different states. In interpreting their findings, the authors

of the US studies considered that the relative immaturity of younger students within grade may

10  

 
lead to inappropriate diagnosis of ADHD. While the alternative interpretation that older children

are under-diagnosed or that the social pressures on younger children amplify the symptoms of

the disorder must also be considered, the evidence of a relative age effect reported in our study

raises the concern that children who are younger within grade level are over-diagnosed and over-

treated. The potential harms of over-diagnosis and unnecessary drug treatments are important to

consider. Children who are treated with ADHD medications are exposed to adverse effects on

sleep, appetite and growth, and are exposed to an increased risk of cardiovascular events [23].

Inappropriate diagnosis of a child with late-year birth month might lead parents and teachers to

treat children differently or adversely change children’s self-perceptions. Our analyses add

weight to concerns about medicalization of the normal range of childhood behaviors, especially

in boys.

These data underscore the dimensional and developmental nature of ADHD symptoms

[24], and impact of contextual expectations on the likelihood of the diagnosis being made. While

this suggests that age-corrected rating scales and developmentally appropriate evaluation are

essential, even this strategy may not be enough to fully eliminate the relative age effect, as there

may be confounding influences of teacher and parent expectations and the child’s self-perception

in a classroom. For example, inadvertent behavioral reinforcers may magnify the apparently

inattentive, distracting or impulsive behaviors of the youngest children in the class, such as

escaping from a difficult academic task (negative reinforcement) or receiving attention from

teachers and peers for disruptive behavior (positive reinforcement). A previous study found that

teacher perceptions of child behaviour were more strongly related to a child’s age within grade

than parental perceptions, suggesting that “teachers’ opinions of children are the key

mechanisms driving the relationship between school starting age and ADHD diagnoses.” [14] It

11  

 
is possible that closer consideration of a child’s behaviour in multiple contexts, including those

outside the setting of school, in the process of assessing children for ADHD may lessen the risk

of unnecessary diagnosis.

Limitations

We aimed to estimate and explain the influence of birth month on diagnosis and

prescribing for ADHD, but our study faced certain limitations. To interpret the increased risk of

diagnosis and treatment among children younger within grade as over-diagnosis or over-

prescribing implies making an assumption about appropriate levels of diagnosis and treatment,

such as that children born in January were diagnosed and treated at appropriate levels. However,

the true incidence of ADHD is unknown. The proportion of children born in January who were

diagnosed with ADHD might underestimate incidence of the disorder if more mature children

within each grade are better able to cope with an underlying disorder, although it is also possible

that the January proportion overestimates the incidence if a larger issue of overmedicalization of

childhood behaviour exists.

Conclusion and implications for further research

Our analyses provide evidence of a relative age effect in the diagnosis and treatment of

ADHD in children aged 6 to 12 years. The strength of the relative age effect remained relatively

stable over the period of our 11-year study. Although the prevalence of diagnosis and treatment

for ADHD remains considerably higher among boys, it increased for children of both genders in

this age range. While the influence of relative age on diagnosis and treatment may lessen in older

age groups of children, we found the effect of relative age was present at all ages from 6 to 12

years for girls and boys. The potential harms of over-diagnosis and over-prescribing and the lack

of an objective test for ADHD strongly suggest caution be taken in assessment and treatment of

12  

 
children for this disorder. Greater emphasis on a child’s behaviour outside the setting of school

may be warranted in the process of assessing children for ADHD to lessen the risk of

unnecessary diagnosis. Further research into the determinants of ADHD and approaches to its

assessment and treatment should consider a child’s age within grade.

13  

 
References

1. Centers for Disease C, Prevention. Increasing prevalence of parent-reported attention-

deficit/hyperactivity disorder among children --- United States, 2003 and 2007.1439-43,

2010 Nov 12.

2. Pastor PN, Reuben CA. Diagnosed attention deficit hyperactivity disorder and learning

disability: United States, 2004-2006.Data From the National Health Survey. (237):1-14, 2008

Jul.

3. Nigg J, Nikolas M, Burt SA. Measured gene-by-environment interaction in relation to

attention-deficit/hyperactivity disorder. J Am Acad Child Adolesc Psychiatry. 2010

Sep;49(9):863-73.

4. Thapar A, Langley K, Asherson P, Gill M. Gene-environment interplay in attention-deficit

hyperactivity disorder and the importance of a developmental perspective. Br J Psychiatry.

2007 Jan;190:1-3.

5. Liederman J, Flannery KA. Fall conception increases the risk of neurodevelopmental

disorder in offspring. J Clin Exp Neuropsychol. 1994 Oct;16(5):754-68.

6. Mick E, Biederman J, Faraone SV. Is season of birth a risk factor for attention-deficit

hyperactivity disorder? J Am Acad Child Adolesc Psychiatry. 1996 Nov;35(11):1470-6.

7. Seeger G, Schloss P, Schmidt MH, Ruter-Jungfleisch A, Henn FA. Gene-environment

interaction in hyperkinetic conduct disorder (HD + CD) as indicated by season of birth

variations in dopamine receptor (DRD4) gene polymorphism. Neurosci Lett. 2004 Aug

19;366(3):282-6.

8. Brookes KJ, Neale B, Xu X, Thapar A, Gill M, Langley K, et al. Differential dopamine

receptor D4 allele association with ADHD dependent of proband season of birth. Am J Med

14  

 
 Genet B Neuropsychiatr Genet. 2008 Jan 5;147B(1):94-9.

9. Atladottir HO, Parner ET, Schendel D, Dalsgaard S, Thomsen PH, Thorsen P. Variation in

incidence of neurodevelopmental disorders with season of birth. Epidemiology. 2007

Mar;18(2):240-5.

10. Tochigi M, Okazaki Y, Kato N, Sasaki T. What causes seasonality of birth in schizophrenia?

:1-11, 2004 Jan.

11. Hebert KJ, Miller LL, Joinson CJ. Association of autistic spectrum disorder with season of

birth and conception in a UK cohort.Official Journal of the International Society for Autism

Research. 3(4):185-90, 2010 Aug.

12. Kolevzon A, Weiser M, Gross R, Lubin G, Knobler HY, Schmeidler J, et al. Effects of

season of birth on autism spectrum disorders: fact or fiction? :1288-90, 2006 Jul.

13. Hauschild KM, Mouridsen SE, Nielsen S. Season of birth in Danish children with language

disorder born in the 1958-1976 period.93-9, 2005.

14. Elder TE. The importance of relative standards in ADHD diagnoses: evidence based on exact

birth dates. J Health Econ. 2010 Sep;29(5):641-56.

15. Evans WN, Morrill MS, Parente ST. Measuring inappropriate medical diagnosis and

treatment in survey data: The case of ADHD among school-age children. J Health Econ.

2010 Sep;29(5):657-73.

16. Bedard K, Dhuey E. The persistence of early childhood maturity: International evidence of

long-run age effects. Quarterly Journal of Economics. 2006;121(4): 1437-1472.

17. Cobley S, Baker J, Wattie N, McKenna J. Annual age-grouping and athlete development: a

meta-analytical review of relative age effects in sport. Sports Med. 2009;39(3):235-56.

18. Polanczyk G, de Lima MS, Horta BL, Biederman J, Rohde LA. The worldwide prevalence of

15  

 
 ADHD: a systematic review and metaregression analysis.942-8, 2007 Jun.

19. Scheffler RM, Hinshaw SP, Modrek S, Levine P. The global market for ADHD

medications.450-7, 2007 Mar-Apr.

20. Dormuth CR, Glynn RJ, Neumann P, Maclure M, Brookhart AM, Schneeweiss S. Impact of

two sequential drug cost-sharing policies on the use of inhaled medications in older patients

with chronic obstructive pulmonary disease or asthma. Clin Ther. 2006 Jun;28(6):964-78;

discussion 2-3.

21. Humphries KH, Rankin JM, Carere RG, Buller CE, Kiely FM, Spinelli JJ. Co-morbidity data

in outcomes research: are clinical data derived from administrative databases a reliable

alternative to chart review? J Clin Epidemiol. 2000 Apr;53(4):343-9.

22. Steele LS, Glazier RH, Lin E, Evans M. Using administrative data to measure ambulatory

mental health service provision in primary care. Med Care. 2004 Oct;42(10):960-5.

23. Gould MS, Walsh BT, Munfakh JL, Kleinman M, Duan N, Olfson M, et al. Sudden death

and use of stimulant medications in youths. Am J Psychiatry. 2009 Sep;166(9):992-1001.

24. Levy F, Hay DA, McStephen M, Wood C, Waldman I. Attention-deficit hyperactivity

disorder: a category or a continuum? Genetic analysis of a large-scale twin study. J Am Acad

Child Adolesc Psychiatry. 1997 Jun;36(6):737-44.

16  

 
Table 1. Percentage of children aged 6 to 12 years receiving ADHD diagnosis or treatment,

by birth month. †Cochrane-Armitage test for trend, one-sided p-value. ‡Risk of receiving

ADHD diagnosis or treatment for children born in December in comparison to children born in

January, expressed as risk difference (RD) and relative risk (RR). CI=confidence interval

Figure 1. Percentage of children aged 6 to 12 years treated with ADHD medications, by

birth month.

Table 2. Risk of receiving ADHD diagnosis or treatment for children born in December in

comparison to children born in January, by year. Data is grouped by year from December to

November of the following year, e.g., 1997-98 refers to 1 Dec 1997 - 30 Nov 1998. †Percentage

of children born in any month of the year who received ADHD diagnosis or treatment. ‡Chi-

square tests for heterogeneity of relative risk across yearly strata: χ2=8.3 (p=0.60) for diagnosis

and χ2=7.6 (p=0.67) for treatment among boys; χ2=3.7 (p=0.96) for diagnosis and χ2=5.1

(p=0.88) for treatment among girls. RD=risk difference RR=relative risk CI=confidence

interval

Table 3. Risk of receiving ADHD diagnosis or treatment for children born in December in

comparison to children born in January, by age. †Percentage of children born in any month

of the year who received ADHD diagnosis or treatment. ‡Chi-square tests for heterogeneity of

relative risk across age strata: χ2=12.2 (p=0.06) for diagnosis and χ2=7.7 (p=0.26) for treatment

among boys; χ2=15.3 (p=0.02) for diagnosis and χ2=26.5 (p<.001) for treatment among girls.

RD=risk difference RR=relative risk CI=confidence interval

17  

 
 Table 4. Sensitivity analysis: risk of receiving ADHD diagnosis or treatment for children

(aged 6 to 12 years) born in different periods. RD=risk difference RR=relative risk

CI=confidence interval

18  

 
  
 

             
             
             
             
             
 
Birth   month    
Boys            
Girls          
    Population   Diagnosed  (%)   Treated  (%)   Population   Diagnosed  (%)   Treated  (%)  
                               
January   39,136     5.67   4.38   37,448     1.56   1.09  
                               
February   36,586     5.86   4.64   34,791     1.75   1.30  
                               
March   41,512     6.04   4.72   38,658     1.88   1.38  
                               
April   40,605     6.07   4.98   38,529     1.94   1.37  
                               
May   42,724     6.54   5.26   40,621     1.86   1.36  
                               
June   40,720     6.73   5.32   38,808     2.18   1.72  
                               
July   41,829     7.32   5.84   40,149     2.34   1.69  
                               
August   40,859     7.29   5.92   38,670     2.41   1.77  
                               
September   41,111     7.58   6.22   38,739     2.62   1.92  
                               
October   39,773     7.92   6.20   37,967     2.58   1.86  
                               
November   37,409     7.83   6.23   35,535     2.64   1.91  
                               
December   38,977     7.37   6.17   36,787     2.66   1.93  
                       
All  months   481,241     6.85   5.49   456,702     2.20   1.61  

 Trend  test  (p)†     -­‐19.7  (<     .0001)   -­‐19.3  (<     .0001)     -­‐16.5  (<     .0001)   -­‐14.1  (<     .0001)  
   
             
  (95%  CI)  
RD      
1.71  (1.36,2.05)    
1.80  (1.48,2.11)      
1.09  (0.88,1.30)    
0.84  (0.66,1.01)  
Dec  vs  Jan‡      
           
  (95%  CI)  
RR      
1.30  (1.23,1.37)    
1.41  (1.33,1.50)      
1.70  (1.53,1.88)    
1.77  (1.57,2.00)  
Dec  vs  Jan‡      
                                     

19  

 
 20  

 
 (a)  Boys,  aged  6  to  12  years  
       
   
ADHD        
ADHD      
Year   diagnosis           treatment          
Diagnosed  (%)†   Risk,  December  vs  January   Treated  (%)†   Risk,  December  vs  January  
        RD  (95%  CI)   RR  (95%  CI)‡       RD  (95%  CI)   RR  (95%  CI)‡  
1997-­‐98   2.63   0.62  (0.30,  0.95)   1.30  (1.13,  1.48)   3.37   1.04  (0.67,  1.41)   1.39  (1.23,  1.56)  
1998-­‐99   2.64   0.91  (0.58,  1.23)   1.46  (1.28,  1.68)   3.16   0.97  (0.61,  1.32)   1.40  (1.24,  1.59)  
1999-­‐00   2.58   0.72  (0.40,  1.04)   1.36  (1.19,  1.57)   3.12   1.07  (0.72,  1.42)   1.47  (1.29,  1.66)  
2000-­‐01   2.52   0.45  (0.13,  0.78)   1.21  (1.06,  1.39)   3.09   0.88  (0.53,  1.23)   1.37  (1.21,  1.55)  
2001-­‐02   2.71   0.78  (0.44,  1.11)   1.36  (1.19,  1.56)   3.15   1.28  (0.92,  1.64)   1.56  (1.38,  1.77)  
2002-­‐03   2.95   0.77  (0.42,  1.13)   1.32  (1.16,  1.49)   3.30   1.25  (0.87,  1.62)   1.50  (1.33,  1.70)  
2003-­‐04   3.23   0.75  (0.37,  1.12)   1.27  (1.13,  1.43)   3.51   1.02  (0.63,  1.41)   1.36  (1.21,  1.53)  
2004-­‐05   3.38   1.01  (0.62,  1.40)   1.36  (1.21,  1.54)   3.62   1.22  (0.82,  1.62)   1.42  (1.27,  1.60)  
2005-­‐06   3.52   1.04  (0.64,  1.44)   1.35  (1.20,  1.52)   3.64   1.05  (0.64,  1.45)   1.35  (1.20,  1.51)  
2006-­‐07   3.67   0.94  (0.54,  1.35)   1.31  (1.17,  1.47)   3.72   1.01  (0.59,  1.42)   1.32  (1.17,  1.47)  
2007-­‐08   3.65   0.62  (0.20,  1.03)   1.19  (1.06,  1.34)   3.68   1.02  (0.61,  1.44)   1.33  (1.18,  1.49)  

(b)  Girls,  aged  6  to  12  years  

       
 
Year    
ADHD   diagnosis            ADHD  treatment          
Diagnosed  (%)†   Risk,  December  vs  January   Treated  (%)†   Risk,  December  vs  January  
        RD  (95%  CI)   RR  (95%  CI)‡       RD  (95%  CI)   RR  (95%  CI)‡  
1997-­‐98   0.73   0.31  (0.13,  0.48)   1.58  (1.21,  2.06)   0.80   0.36  (0.18,  0.55)   1.65  (1.27,  2.13)  
1998-­‐99   0.67   0.36  (0.19,  0.53)   1.78  (1.35,  2.36)   0.75   0.33  (0.16,  0.50)   1.72  (1.30,  2.28)  
1999-­‐00   0.70   0.29  (0.12,  0.47)   1.57  (1.19,  2.06)   0.79   0.38  (0.19,  0.56)   1.70  (1.31,  2.21)  
2000-­‐01   0.69   0.32  (0.14,  0.50)   1.61  (1.23,  2.10)   0.79   0.37  (0.19,  0.56)   1.69  (1.30,  2.19)  
2001-­‐02   0.76   0.36  (0.18,  0.55)   1.69  (1.30,  2.21)   0.84   0.49  (0.30,  0.68)   1.94  (1.49,  2.52)  
2002-­‐03   0.89   0.35  (0.15,  0.55)   1.55  (1.21,  1.99)   0.92   0.50  (0.29,  0.70)   1.79  (1.41,  2.29)  
2003-­‐04   0.99   0.51  (0.30,  0.72)   1.80  (1.41,  2.29)   1.01   0.43  (0.21,  0.64)   1.57  (1.25,  1.98)  
2004-­‐05   1.05   0.57  (0.35,  0.78)   1.83  (1.44,  2.31)   1.07   0.63  (0.41,  0.85)   1.93  (1.53,  2.44)  
2005-­‐06   1.06   0.48  (0.26,  0.70)   1.64  (1.30,  2.06)   1.05   0.61  (0.39,  0.84)   1.86  (1.48,  2.36)  
2006-­‐07   1.13   0.69  (0.45,  0.92)   1.97  (1.56,  2.48)   1.10   0.67  (0.44,  0.90)   1.96  (1.55,  2.48)  
2007-­‐08   1.19   0.63  (0.38,  0.87)   1.74  (1.40,  2.18)   1.10   0.75  (0.51,  0.98)   2.06  (1.63,  2.60)  

21  

 
 (a)  Boys,  aged  6  to  12  years  
         
 Age    ADHD  diagnosis            ADHD  treatment          
Diagnosed  (%)†   Risk,  December  vs  January   Treated  (%)†   Risk,  December  vs  January  
          RD  (95%  CI)   RR  (95%  CI)‡       RD  (95%  CI)   RR  (95%  CI)‡  
6   2.32   0.82  (0.56,  1.08)   1.48  (1.31,  1.68)   1.64   0.71  (0.48,  0.93)   1.59  (1.37,  1.85)  
7   3.06   0.96  (0.65,  1.26)   1.38  (1.25,  1.53)   2.61   1.02  (0.75,  1.30)   1.53  (1.36,  1.72)  
8   3.39   1.12  (0.81,  1.43)   1.42  (1.28,  1.56)   3.42   1.30  (0.99,  1.62)   1.50  (1.36,  1.65)  
9   3.56   0.90  (0.58,  1.22)   1.30  (1.18,  1.43)   4.01   1.24  (0.91,  1.57)   1.41  (1.29,  1.54)  
10   3.48   0.60  (0.29,  0.92)   1.20  (1.09,  1.31)   4.27   1.24  (0.90,  1.58)   1.37  (1.26,  1.49)  
11   3.28   0.73  (0.43,  1.03)   1.27  (1.15,  1.40)   4.31   1.32  (0.97,  1.66)   1.38  (1.27,  1.50)  
12   3.04   0.63  (0.35,  0.92)   1.25  (1.13,  1.38)   4.18   1.14  (0.81,  1.48)   1.34  (1.23,  1.46)  

  Girls,  aged  6  to  12  years  

(b)            
         
 Age    ADHD  diagnosis            ADHD  treatment          
Diagnosed  (%)†   Risk,  December  vs  January   Treated  (%)†   Risk,  December  vs  January  
          RD  (95%  CI)   RR  (95%  CI)‡       RD  (95%  CI)   RR  (95%  CI)‡  
6   0.68   0.39  (0.24,  0.54)   1.83  (1.44,  2.32)   0.44   0.21  (0.10,  0.33)   1.74  (1.28,  2.37)  
7   0.91   0.41  (0.24,  0.57)   1.68  (1.36,  2.08)   0.72   0.45  (0.30,  0.59)   2.04  (1.60,  2.60)  
8   1.04   0.68  (0.50,  0.86)   2.11  (1.73,  2.57)   0.94   0.62  (0.45,  0.79)   2.19  (1.76,  2.71)  
9   1.09   0.67  (0.49,  0.86)   1.90  (1.59,  2.28)   1.15   0.87  (0.68,  1.06)   2.25  (1.88,  2.70)  
10   1.03   0.43  (0.26,  0.60)   1.60  (1.32,  1.93)   1.20   0.74  (0.55,  0.93)   2.01  (1.68,  2.41)  
11   0.94   0.38  (0.21,  0.54)   1.54  (1.27,  1.87)   1.17   0.55  (0.37,  0.74)   1.69  (1.41,  2.01)  
12   0.83   0.20  (0.04,  0.36)   1.29  (1.06,  1.57)   1.09   0.25  (0.08,  0.43)   1.29  (1.08,  1.54)  

22  

 
     "Exposed"  cohort   "Unexposed"  cohort   RD  (95%  CI)   RR  (95%  CI)  
                   
Analysis  1   Dec  2  -­‐  16:   Dec  17-­‐31:  
Boys      
   Population    
18,957    
14,853   -­‐     -­‐    
   Diagnosed  (%)   1,420  (7.49)   1,085  (7.30)   0.19  (-­‐0.38,  0.75)   1.03  (0.95,  1.11)  
   Treated  (%)   1,190  (6.28)   884  (5.95)   0.33  (-­‐0.19,  0.84)   1.05  (0.97,  1.15)  

 
Girls          
   Population    
17,924    
13,960   -­‐     -­‐    
   Diagnosed  (%)   469  (2.62)   376  (2.69)   -­‐0.08  (-­‐0.43,  0.28)   0.97  (0.85,  1.11)  
   Treated  (%)   347  (1.94)   261  (1.87)   0.07  (-­‐0.24,  0.37)   1.04  (0.88,  1.21)  

 Analysis  2     Dec  17-­‐31:     Jan  1-­‐15:      

Boys      
   Population    
14,853    
15,068   -­‐     -­‐    
   Diagnosed  (%)   1,085  (7.30)   856  (5.68)   1.62  (1.07,  2.18)   1.29  (1.18,  1.40)  
   Treated  (%)   884  (5.95)   666  (4.42)   1.53  (1.03,  2.03)   1.35  (1.22,  1.49)  

 
Girls          
   Population    
13,960    
14,529   -­‐     -­‐    
   Diagnosed  (%)   376  (2.69)   231  (1.59)   1.10  (0.77,  1.44)   1.69  (1.44,  1.99)  
   Treated  (%)   261  (1.87)   150  (1.03)   0.84  (0.56,  1.12)   1.81  (1.48,  2.21)  

 
Analysis   3   Jan  1  -­‐15:    
Jan  16-­‐30:        
Boys    
   Population    
15,068    
19,154   -­‐     -­‐    
   Diagnosed  (%)   856  (5.68)   1,103  (5.76)   -­‐0.08  (-­‐0.57,  0.42)   0.99  (0.90,  1.08)  
   Treated  (%)   666  (4.42)   837  (4.37)   0.05  (-­‐0.39,  0.49)   1.01  (0.92,  1.12)  

 
Girls          
   Population    
14,529    
18,179   -­‐     -­‐    
   Diagnosed  (%)   231  (1.59)   276  (1.52)   0.07  (-­‐0.20,  0.34)   1.05  (0.88,  1.25)  
   Treated  (%)   150  (1.03)   194  (1.07)   -­‐0.03  (-­‐0.26,  0.19)   0.97  (0.78,  1.20)  

 
Analysis   4   Dec  2  9-­‐31:   Jan  1   -­‐3:      
Boys      
   Population    
3,907    
3,609   -­‐     -­‐    
   Diagnosed  (%)   272  (6.96)   190  (5.26)   1.70  (0.62,  2.78)   1.32  (1.10,  1.58)  
   Treated  (%)   252  (6.45)   155  (4.29)   2.16  (1.14,  3.17)   1.50  (1.24,  1.82)  

 
Girls          
   Population    
3,710    
3,491   -­‐     -­‐    
   Diagnosed  (%)   93  (2.51)   59  (1.69)   0.82  (0.16,  1.48)   1.48  (1.07,  2.05)  
   Treated  (%)   74  (1.99)   49  (1.40)   0.59  (0.00,  1.19)   1.42  (0.99,  2.03)  

23