Factors affecting the age at diagnosis of autism spectrum disorders in Nova Scotia,Canada

Abstract

While early diagnosis of autism spectrum disorders (ASD) is essential for ensuring timely access to early intervention services, there is limited existing literature investigating factors that delay this diagnosis. This population-based cohort study explored the age at which children in Nova Scotia, Canada, are diagnosed with ASDs and the factors associated with this age. Children diagnosed with an ASD between January 1992 and December 2005 were identified from a cohort of live births in the province between 1990 and 2002. Demographic and clinical variables were extracted from population-based perinatal and administrative health databases and evaluated as predictors of age at ASD diagnosis. Of 122,759 live births, 884 cases of ASDs were identified during the study period. The median age at diagnosis within the cohort was 4.6 years. In adjusted linear regression analysis, a one year increase in maternal age at delivery was associated with a 0.06 decrease in age at ASD diagnosis (p = .0007). Children who were residents of Halifax County received their diagnoses 0.52 years later than residents of other counties (p = .0054). A diagnosis of attention-deficit/hyperactivity disorder (ADHD) was associated with a 1.29-year increase in age at diagnosis (p < .0001). These results suggest that potential exists for improving early detection of ASDs in the province. Future research in this field has the potential to contribute to our understanding of the causal pathways linking the demographic and clinical variables we have identified and the age at diagnosis of ASDs.

Keywords

autism age of diagnosis ADHD maternal age at delivery

Autism spectrum disorders (ASDs) are a class of neurodevelopmental conditions characterized by deficits in social functioning and communication skills, and by the presence of repetitive behaviours or limited interests (American Psychiatric Association (APA), 2000). ASDs include autistic disorder, the most profound disorder on the spectrum, pervasive developmental disorder – not otherwise specified, a milder condition, and Asperger syndrome, which is characterized by relatively strong cognitive and language skills. Recent studies have placed the prevalence of ASDs at 6 to 7 per 1000, making them some of the most common disorders of development in the world (Fombonne, 2009). This recent observed increase in ASD prevalence, coupled with evidence suggesting that children benefit from intervention programs that are delivered early in life (Eldevik et al., 2009; Howlin et al., 2009; Reichow and Wolery, 2009; Rogers and Vismara, 2008), has emphasized the importance of early detection. While there is increasing evidence to indicate that children with ASDs can be reliably diagnosed as young as two years of age (Charman and Baird, 2002; Chawarska et al., 2007; Eaves and Ho, 2004; Lord, 1995; Zwaigenbaum et al., 2009), many children are not diagnosed until later in childhood (Latif and Williams, 2007; Mandell et al., 2002, 2005; Ouellette-Kuntz et al., 2009; Shattuck et al, 2009; Wiggins et al., 2006; Williams et al., 2008). A recent Canadian study reported median ages at first ASD diagnosis ranging from 39 months to 55 months (Ouellette-Kuntz et al., 2009). Similar trends have been reported elsewhere (Latif and Williams, 2007; Shattuck et al., 2009; Williams et al., 2008). This delay has the potential to postpone access to early intervention programs, thereby decreasing the chance of more positive outcomes (Bryson et al., 2003). Identifying factors associated with the age at which ASDs are diagnosed could therefore be helpful to understanding barriers to accessing diagnostic and intervention services.

Few studies have examined factors associated with the age at diagnosis of ASDs. Level of impairment and diagnostic subtype have consistently been linked with an earlier age at diagnosis (Mandell et al., 2005; Shattuck et al., 2009; Wiggins et al., 2006), while conflicting findings have been reported with respect to race and sex (Mandell et al., 2002, 2005; Shattuck et al., 2009; Wiggins et al., 2006). To date, there has been little research examining factors affecting the age at diagnosis of ASDs in Canada. This study attempted to address this gap in the literature, using data from a clinical population-based perinatal database and administrative health databases in one Canadian province.

Method

Participants

Nova Scotia is an eastern Canadian province with an estimated population of 940,000. The population is primarily English speaking and, according to 2006 Canadian census data, just over 4% is made up by visible minorities (Statistics Canada, 2010). A cohort of all live hospital births between 1 January 1990 and 31 December 2002 was identified from the Nova Scotia Atlee Perinatal Database. This province-wide clinical database houses variables related to the prenatal, labour and delivery, and neonatal periods for all Nova Scotia hospital births of at least 500 g in weight or 20 weeks of gestation. Children in the cohort were linked to data in three administrative health databases: the CIHI Hospital Discharge Database, the MSI Physician Billings Database, and the Mental Health Outpatient Information System using methods previously described (Dodds et al., 2010). The final study cohort consisted of those children with at least one record of an ASD diagnosis in any of these databases between 1 January 1992 and 31 December 2005. An ASD diagnosis was defined as a code of International Classification of Diseases (ICD)-9 299 or ICD-10 F84 in either the first or second diagnostic field. Due to concerns regarding the reliability of diagnoses in children under the age of two years (Chawarska et al., 2007; Cox et al., 1999), those whose only ASD diagnostic code appeared before this age were excluded from the study. An earlier study reported the sensitivity and specificity of these data sources for detecting true cases of ASD as 69.3% and 77.3%, respectively (Dodds et al., 2009).

Materials and procedures

Demographic and clinical variables pertaining to the affected child, the mother of the affected child, and the family of the affected child were extracted from the Nova Scotia Atlee Perinatal Database. These included the sex of the child, marital status of the mother at the time of delivery, maternal age at delivery, birth order of the child within the family, county of residence, evidence of psychiatric illness in the mother, birth weight, gestational age, evidence of a diagnosis of the child with any major congenital anomaly or the following co-morbidities: intellectual disability (ICD-9 317-319 or ICD-10 F70-79), cerebral palsy (ICD-9 343 or ICD-10 G80), epilepsy (ICD-9 345 or ICD-10 G40), and attention-deficit/hyperactivity disorder (ADHD) (ICD-9 314 or ICD-10 F90). Variables were selected based on the results of previous studies, the plausibility of their association with the age at diagnosis, and their availability in one of the databases.

The age at which the first health claim for an ASD appeared in any of the three administrative health databases was calculated for each child. To assess the suitability of the age at first ASD health system claim as a proxy for age at ASD diagnosis, a simple validation procedure was performed. Records of children diagnosed with an ASD by the IWK Autism Team between 2001 and 2005 were linked with ASD health claims in the administrative databases. The Autism Team at the IWK Health Centre in Halifax receives referrals for children who are suspected of having an ASD from Halifax County and other areas of the province. Diagnoses are made using the Autism Diagnostic Interview Revised (Lord et al., 1994), the Autism Diagnostic Observation Schedule (Lord et al., 2000), and clinical judgment. These diagnostic protocols are consistent with the current recommended best practices for diagnosing ASDs, and as such, are considered the ‘gold standard’. The time difference between the date of first ASD administrative health claim and the date of the formal gold standard diagnosis was calculated for each child. Additional investigations were undertaken to determine if the time difference between the ASD administrative code and gold standard diagnosis was associated with any of the demographic and clinical variables of interest. Mann–Whitney U tests were performed on variables with cell sizes larger than 5 to determine if the length of time elapsed differed significantly based on these variables.

The median age at diagnosis was calculated for the final study cohort. Age-specific incidence rates were calculated using denominator information derived from the original cohort of live Nova Scotia hospital births. For the purpose of incidence rate calculations, children were considered to be at risk until they were diagnosed with an ASD, died, or moved out of the province.

All children in the study cohort were included in analyses exploring relationships between demographic and clinical variables of interest and the age at diagnosis. Variables were described using frequencies and proportions or with medians and ranges, as appropriate. The median age at ASD diagnosis and associated range was calculated for all levels of each categorical variable. Mann–Whitney U tests were used to determine if the age at diagnosis varied significantly based on the demographic and clinical variables of interest.

To further characterize the relationship between the demographic and clinical variables of interest and the age at diagnosis of ASD, a linear regression analysis was undertaken. Initially, each demographic or clinical variable was entered into a univariate model with age of diagnosis as the outcome variable. Unadjusted regression coefficients with associated 95% CIs were calculated for each variable. Demographic and clinical variables were then introduced together into multivariate models to produce adjusted regression coefficients with associated 95% CIs. To identify the most parsimonious combination of variables that would explain the age at diagnosis, a multivariate model was developed through backward elimination regression. All variables were initially included in the model. Variables with p values greater than .05 were removed from the model, beginning with the largest p value. This process was continued until all variables remaining in the model were significant at the level of p < .05. Subjects with missing values for variables in the full model were initially excluded from the analysis, but were reintroduced if the variable was not included in the final model. All analyses were conducted using SAS version 8.2.

Results

The results of the validation procedure are presented in Table 1. A total of 121 children diagnosed with an ASD by the IWK Autism team between 2001 and 2005 were linked with health claims for ASDs. The median length of time between first ASD administrative health claim and gold standard diagnosis was 0.6 years.

Table 1.

Time elapsed between first administrative health claim and gold standard diagnosis.

Variable	n (%)	Median length of timeelapsed in years(range of time)	p value^b
Overall	121	0.6 (0.0–8.6)	–
Maternal marital status^a
Married/common-law	90 (77.6)	0.6 (0.0–8.2)	.7355
Single/widowed/divorced	26 (22.4)	0.6 (0.0–8.6)	.7355
Maternal history of psychiatric illness
Yes	5 (4.1)	0.4 (0.2–2.1)	.6435
No	116 (95.9)	0.6 (0.0–8.6)	.6435
Sex
Male	107 (88.4)	0.6 (0.0–8.2)	.8109
Female	14 (11.6)	0.5 (0.0–8.6)	.8109
Major congenital anomaly
Yes	6 (5.0)	0.8 (0.2–5.1)	.1275
No	115 (95.0)	0.6 (0.0–8.6)	.1275
County of residence
Halifax County	97 (80.2)	0.6 (0.0–8.6)	.7083
Outside Halifax	24 (19.8)	0.5 (0.0–8.2)	.7083

Five missing observations.

Mann–Whitney U test.

The median length of time elapsed varied with demographic and clinical variables, and ranged from 0.4 to 0.8 years. None of the Mann–Whitney U tests comparing time elapsed based on these variables were significant at the level of p < .05. The results of the validation procedure were interpreted as evidence for the suitability of the age at first ASD health system claim as a proxy for the age at diagnosis.

Of 122,759 live Nova Scotia hospital births between 1 January 1990 and 31 December 2002, 884 cases of ASD were identified from administrative health records, for an overall incidence rate of 75.8 cases per 100,000 person-years. Calculation of age-specific incidence rates revealed a peak in the frequency of diagnoses at age three and a subsequent decline across older ages. A graph showing the distribution of age of diagnosis (based on age at the time of first health system claim) has been reported previously (Dodds et al., 2010).

Median maternal age at delivery in this cohort was 28.8 years (range 14.2–43.7 years). Median (and ranges) for gestational age and birth weight were 39 weeks (25–44 weeks) and 3.50 kg (0.68–6.29 kg), respectively. Characteristics of the cohort according to age at ASD diagnosis are presented in Table 2. Median age at diagnosis of the cohort was 4.6 years (ranging from 1.3–15.7 years). Children with intellectual disability, ADHD, a major congenital anomaly, those who were firstborn, and those who were residents in Halifax County had significantly higher ages at diagnosis in the univariate analysis.

Table 2.

Age at ASD diagnosis by exposures of interest.

Variable	n (%)	Median age atdiagnosisin years (range of ages)	p value^c
Overall	884	4.6 (1.3–15.7)	–
Maternal marital status^a
Married/common-law	593 (72.9)	4.4 (1.3–15.7)	.7145
Single/widowed/divorced	221 (27.2)	4.7 (1.3–13.9)	.7145
Maternal history of psychiatric illness
Yes	23 (2.6)	4.0 (1.9–8.6)	.2296
No	861 (97.4)	4.6 (1.3–15.7)	.2296
Sex
Male	714 (80.8)	4.5 (1.3–15.7)	.8232
Female	170 (19.2)	4.7 (1.7–14.6)	.8232
Intellectual disability
Yes	94 (10.6)	5.5 (1.6–13.5)	.0086
No	790 (89.4)	4.4 (1.3–15.7)	.0086
Cerebral palsy
Yes	19 (2.2)	4.6 (1.9–11.3)	.9681
No	865 (97.9)	4.6 (1.3–15.7)	.9681
Epilepsy
Yes	60 (6.8)	4.6 (1.5–12.2)	.8439
No	824 (93.2)	4.6 (1.3–15.7)	.8439
Attention-deficit/hyperactivity disorder
Yes	352 (39.8)	5.6 (1.3–15.7)	<.0001
No	532 (60.2)	4.0 (1.3–14.3)	<.0001
Major congenital anomaly
Yes	52 (5.9)	5.7 (1.7–15.7)	.0137
No	832 (94.1)	4.5 (1.3–14.6)	.0137
County of residence
Halifax County	318 (36.0)	4.9 (1.7–15.7)	.0072
Outside Halifax	566 (64.0)	4.4 (1.3–14.6)	.0072
Birth order^b
Firstborn	415 (47.0)	4.8 (1.3–14.3)	.0042
Later born	468 (53.0)	4.4 (1.3–15.7)	.0042

70 missing observations.

One missing observation.

Mann–Whitney U test.

Unadjusted regression coefficients and associated 95% CIs calculated from univariate linear models with the continuous age at diagnosis outcome variable are presented in Table 3. In total, associations with five demographic and clinical variables were found to be significant at the level of alpha = .05. These included ADHD diagnostic status, birth order, county of residence, major congenital anomaly status, and maternal age at delivery. Three variables remained after the backward elimination regression: ADHD diagnostic status, county of residence, and maternal age at delivery (Table 4). An ADHD diagnosis was associated with an ASD diagnosis occurring 1.29 years later when compared to children without an ADHD diagnosis (95% CI: 0.93–1.64). Being a resident of Halifax County was associated with a 0.52-year later age at ASD diagnosis when compared to residents of other counties (95% CI: 0.15–0.88), while each year increase in maternal age at delivery was associated with a 0.06 year decrease in the age at ASD diagnosis (95% CI: –0.09– –0.02).

Table 3.

Unadjusted linear regression modeling age at diagnosis.

Variable	Unadjusted regressioncoefficient (95% CI)	p value
Maternal age (years)	−0.06 (–0.09– –0.03)	.0004
Gestational age (weeks)	−0.01 (–0.10–0.08)	.8510
Birth weight (kg)	−0.20 (–0.48–0.07)	.1475
Maternal marital status
Married/common-law	0.09 (–0.34–0.52)	.6817
Single/widowed/divorced	–
Maternal history of psychiatric illness
Yes	−0.85 (–1.99–0.28)	.1407
No	–	.1407
Sex
Male	0.04 (–0.42–0.49)	.8798
Female	–	.8798
Intellectual disability
Yes	0.52 (–0.07–1.10)	.0829
No	–	.0829
Cerebral palsy
Yes	0.15 (–1.10–1.40)	.8141
No	–	.8141
Epilepsy
Yes	−0.15 (–0.87–0.57)	.6801
No	–	.6801
Attention-deficit/hyperactivity disorder
Yes	1.34 (0.98–1.70)	<.0001
No	–	<.0001
Major congenital anomaly
Yes	0.88 (0.11–1.64)	.0249
No	–	.0249
County of residence
Halifax County	0.46 (0.08–0.84)	.0166
Outside Halifax	^–	.0166
Birth order
Firstborn	0.53 (0.17–0.89)	.0043
Later born	–	.0043

Table 4.

Linear regression modeling age at diagnosis (final model).

Variable	Adjusted regression coefficient (95% CI)	p value
Maternal age (years)	−0.06 (–0.09– –0.02)	.0007
County of residence
Halifax County	0.52 (0.15–0.88)	.0054
Outside Halifax	–	.0054
Attention-deficit/hyperactivity disorder
Yes	1.29 (0.93–1.64)	<.0001
No	–	<.0001

Discussion

In this cohort, affected males appeared more often than affected females in a ratio of 4.4:1, which is consistent with the sex ratio of 3–4:1 reported in the literature (Bryson and Smith, 1998; Ouellette-Kuntz et al., 2006). The rate of co-morbidity with intellectual disability was 10.6%, which is substantially lower than many existing estimates. Recent studies of ASDs have reported the proportion of children meeting the criteria for intellectual disability to be in the range of 14.7–46.7% (Chakrabarti and Fombonne, 2001; Shattuck et al., 2009; Wiggins et al., 2006; Williams et al., 2008). The present study differs from previous investigations in that we relied on administrative health databases to ascertain cases of both ASDs and intellectual disability. One possible explanation for the low rate of co-morbidity is that diagnostic codes from health records are poor measures of intellectual disability. Previous studies have demonstrated that overall prevalence estimates of intellectual disability are lower when based on health sources alone (Roeleveld et al., 1997). This may be especially true in cases where children meet the diagnostic criteria for both intellectual disability and ASDs. Professionals may be reluctant to assign a diagnosis of intellectual disability for fear of further stigmatizing the child (First and Tasman, 2006). Another factor that might have influenced our results is the fact that many children are not diagnosed with intellectual disabilities until they enter the school system (Larson et al., 2001). In our cohort, children born in later years were not followed long enough to reach school age, and therefore may not yet have received their diagnoses of intellectual disability.

This study revealed that age at ASD diagnosis in Nova Scotia during the study period was best explained by ADHD diagnostic status, maternal age, and county of residence. The results of our study indicate that children with ADHD diagnostic codes received their ASD diagnoses over a year later than children without this code. One possible explanation is that level of impairment acted as a confounder in this relationship. Specifically, children who received diagnostic codes of ADHD may have been less severely impaired than those without ADHD codes, and it is this level of impairment that predicted age at diagnosis. This is consistent with earlier studies showing greater impairments to be associated with earlier diagnoses (Mandell et al., 2005; Shattuck et al., 2009; Wiggins et al., 2006). Although both the Diagnostic and Statistical Manual of Mental Disorders (DSM)-IV and ICD-10 diagnostic criteria exclude the diagnosis of ADHD when an ASD is present (APA, 2000; World Health Organization, 1993), recent studies have reported high degrees of symptom overlap between children with ASDs and those with ADHD (Hartley and Sikora, 2009; Simonoff et al., 2008; Sinzig et al., 2009). This overlap has been suggested to be a contributing factor to later diagnoses among children with milder impairments (Hartley and Sikora, 2009). In the present cohort, it is possible that symptoms of ASD were initially incorrectly attributed to or masked by symptoms of ADHD, leading to a delay in assigning a formal diagnosis of ASD. The timing of the first ADHD code with respect to the first ASD code seems to support this scenario. In almost 60% of cases when an ADHD diagnostic code was present, it appeared sometime before the first ASD diagnostic code.

In this study, an increased maternal age at time of delivery was associated with a decreased age at diagnosis, after controlling for birth order. One possible explanation for the association is that older mothers may be more educated. Women who have attained a higher level of education may be more familiar with ASDs and their symptoms.

The child’s region of residence was found to be a significant predictor of the age at diagnosis. Regional differences in the age at diagnosis were reported in two studies (Mandell et al., 2005; Shattuck et al., 2009). In this study, children living in the Halifax region were diagnosed with ASDs later than those living in other counties. One possible explanation for this discrepancy is that diagnostic services in the early years of the study period may have been largely unavailable outside of Halifax. During this time, children living in other areas may have been more likely to have a diagnostic code immediately assigned by a family physician while children in Halifax County were referred for formal ASD diagnostic services. Another possible explanation for the discrepancy is the size difference between the populations served by the diagnostic services. The Autism Team at the IWK Health Centre is responsible for evaluating children referred within the heavily populated Halifax region. It is possible that children living in Halifax face greater competition for the diagnostic services available in their community than children living in more rural areas.

Sex of the child was not a significant predictor of the age at diagnosis in this study. This is consistent with findings from all but one previous study (Mandell et al., 2002, 2005; Shattuck et al., 2009; Wiggins et al., 2006). Gestational age and birth weight were also not significantly associated with age at ASD diagnosis, which is consistent with earlier findings (Shattuck et al., 2009). Birth order of the child was not retained in the final linear model. Therefore, our hypothesis that a lack of familiarity with normal child development by first-time parents would result in later ages at diagnosis among firstborn children was not supported. Maternal history of psychiatric illness was also not associated with the age at diagnosis. To our knowledge, no previous studies have investigated birth order or maternal history of psychiatric illness as they relate to age at diagnosis.

This study had several key strengths. First, it employed a large sample size (n = 884), which is among the largest samples from a single geographic region to be included in an investigation of factors affecting the age at ASD diagnosis. The results of this study are also the first to be reported from Nova Scotia. Additionally, this project utilized a previously validated data source for all subjects. The use of administrative data eliminated the possibility of recall biases, while the use of a consistent data source throughout the study ensured that there would be no chance of confounding due to differing data collection methods. Finally, this study investigated several new variables as predictors of age at diagnosis.

This study has also several important limitations that warrant consideration. First, information on the severity of the condition and diagnostic subtype was not available in the administrative health databases used to identify cases of ASD. These factors have been identified as predictors of age at diagnosis in virtually all previous studies (Mandell et al., 2005; Shattuck et al., 2009; Wiggins et al., 2006, Williams et al., 2008). Our failure to control for these factors could have biased our results in either direction. In particular, level of impairment and/or diagnostic subtype may have confounded the relationship between the ADHD diagnostic status and the age at diagnosis. Another limitation of the administrative data usage is that the three health claim databases almost certainly did not capture all individuals who were diagnosed with ASDs during the study period. The sensitivity of the three databases has been previously reported as approximately 70% (Dodds et al., 2009). However, this estimate was not found to vary significantly based on several of the variables evaluated in the present study, including sex, birth order, birth weight, county of residence, maternal age, and the diagnosis of a major congenital anomaly. This suggests that systematic differences did not exist between the individuals we captured in our study and those we missed.

Our decision to use the age at first ASD health claim as a proxy for age at diagnosis is another potential limitation of our study. Although our validation procedure indicated that the deviation from the true date of diagnosis was minimal in most cases and that the time elapsed between diagnostic code and gold standard diagnosis did not vary based on the variables of interest, there may have been some misclassifications. Finally, the use of the Atlee Perinatal Database limited our cohort to affected children born in Nova Scotia hospitals. Therefore, children diagnosed with ASDs who immigrated to the province sometime after birth were not included in this study.

Conclusion

This study revealed valuable information regarding the age at diagnosis of ASDs in Nova Scotia and factors affecting it. Future research in this field has the potential to contribute to our understanding of the causal pathways linking the demographic and clinical variables we identified and the age at diagnosis. Specifically, detailed investigations of medical records of affected children could provide valuable information regarding diagnoses of ADHD as well as details regarding the child’s level of impairment and diagnostic subtype, which are likely significant predictors of the age at diagnosis. Identified discrepancies in the age at diagnosis based on certain demographic and clinical variables highlight the need for coordinated strategies for early detection in Nova Scotia.

Footnotes

Acknowledgements

The authors thank the Reproductive Care Program of Nova Scotia and the Population Health Research Unit at Dalhousie University for facilitating access to the data and Anne Spencer for help with data analysis. This study was partially funded by a grant from the Cure Autism Now Foundation (now Autism Speaks) and P. Frenette received graduate student funding from the Nova Scotia Health Research Foundation. Although this research is based in part on data obtained from the Population Health Research Unit, the observations and opinions expressed are those of the authors and do not represent those of the Population Health Research Unit.

References

American Psychiatric Association. (2000) Diagnostic and Statistical Manual of Mental Disorders, 4th edn. Text Revised. Washington, DC: American Psychiatric Association.

Bryson

Rogers

Fombonne

(2003) Autism spectrum disorders: early detection, intervention, education, and psychopharmacological management. Canadian Journal of Psychiatry 48(8): 506–516.

Bryson

Smith

(1998) Epidemiology of autism: prevalence, associated characteristics, and implications for research and service delivery. Mental Retardation and Developmental Disabilities Research Reviews 4(2): 97–103.

Chakrabarti

Fombonne

(2001) Pervasive developmental disorders in preschool children. Journal of the American Medical Association 285(24): 3093–3099.

Charman

Baird

(2002) Practitioner review: Diagnosis of autism spectrum disorder in 2- and 3-year-old children. Journal of Child Psychology and Psychiatry, and Allied Disciplines 43(3): 289–305.

Chawarska

Klin

Paul

Volkmar

(2007) Autism spectrum disorder in the second year: stability and change in syndrome expression. Journal of Child Psychology and Psychiatry, and Allied Disciplines 48(2): 128–138.

Cox

Klein

Charman

Baird

Baron-Cohen

Swettenham

. (1999) Autism spectrum disorders at 20 and 42 months of age: stability of clinical and ADI-R diagnosis. Journal of Child Psychology and Psychiatry, and Allied Disciplines 40(5): 719–732.

Dodds

Fell

Shea

Armson

Allen

Bryson

(2010) The role of prenatal, obstetric, and neonatal factors in the development of autism in the offspring. Journal of Autism and Developmental Disorders 5, October 2010 [Epub ahead of print].

Dodds

Spencer

Shea

Fell

Armson

Allen

. (2009) Validity of autism diagnoses using administrative health data. Chronic Diseases in Canada 29(3): 102–107.

10.

Eaves

(2004) The very early identification of autism: outcome to age 4 1/2-5. Journal of Autism and Developmental Disorders 34(4): 367–378.

11.

Eldevik

Hastings

Hughes

Jahr

Eikeseth

Cross

(2009) Meta-analysis of Early Intensive Behavioral Intervention for children with autism. Journal of Clinical Child and Adolescent Psychology 38(3): 439–450.

12.

First

Tasman

(2006) Clinical Guide to the Diagnosis and Treatment of Mental Disorders. Chichester, UK: John Wiley & Sons, Ltd.

13.

Fombonne

(2009) Epidemiology of pervasive developmental disorders. Pediatric Research 65(6): 591–598.

14.

Hartley

Sikora

(2009) Which DSM-IV-TR criteria best differentiate high-functioning autism spectrum disorder from ADHD and anxiety disorders in older children? Autism: The International Journal of Research and Practice 13(5): 485–509.

15.

Howlin

Magiati

Charman

(2009) Systematic review of early intensive behavioral interventions for children with autism. American Journal on Intellectual and Developmental Disabilities 114(1): 23–41.

16.

Larson

Lakin

Anderson

Kwak

Lee

Anderson

(2001) Prevalence of mental retardation and developmental disabilities: estimates from the 1994/1995 National Health Interview Survey Disability Supplements. American Journal of Mental Retardation 106(3): 231–252.

17.

Latif

Williams

(2007) Diagnostic trends in autistic spectrum disorders in the South Wales Valleys. Autism: The International Journal of Research and Practice 11(6): 479–487.

18.

Lord

(1995) Follow-up of two-year-olds referred for possible autism. Journal of Child Psychology and Psychiatry, and Allied Disciplines 36(8): 1365–1382.

19.

Lord

Risi

Lambrecht

Cook

Jr Leventhal

DiLavore

. (2000) The autism diagnostic observation schedule-generic: A standard measure of social and communication deficits associated with the spectrum of autism. Journal of Autism and Developmental Disorders 30(3): 205–223.

20.

Lord

Rutter

Le Couteur

(1994) Autism diagnostic interview-revised: a revised version of a diagnostic interview for caregivers of individuals with possible pervasive developmental disorders. Journal of Autism and Developmental Disorders 24(5): 659–685.

21.

Mandell

Listerud

Levy

Pinto-Martin

(2002) Race differences in the age at diagnosis among medicaid-eligible children with autism. Journal of the American Academy of Child and Adolescent Psychiatry 41(12): 1447–1453.

22.

Mandell

Novak

Zubritsky

(2005) Factors associated with age of diagnosis among children with autism spectrum disorders. Pediatrics 116(6): 1480–1486.

23.

Ouellette-Kuntz

Coo

Chudley

Noonan

Breitenbach

. (2006) Prevalence of pervasive developmental disorders in two Canadian provinces. Journal of Policy and Practice in Intellectual Disabilities 3(3): 164–172.

24.

Ouellette-Kuntz

Coo

Lam

Breitenbach

Hennessey

. (2009) Age at diagnosis of autism spectrum disorders in four regions of Canada. Canadian Journal of Public Health 100(4): 268–273.

25.

Reichow

Wolery

(2009) Comprehensive synthesis of early intensive behavioral interventions for young children with autism based on the UCLA young autism project model. Journal of Autism and Developmental Disorders 39(1): 23–41.

26.

Roeleveld

Zielhuis

Gabreels

(1997) The prevalence of mental retardation: a critical review of recent literature. Developmental Medicine and Child Neurology 39(2): 125–132.

27.

Rogers

Vismara

(2008) Evidence-based comprehensive treatments for early autism. Journal of Clinical Child and Adolescent Psychology 37(1): 8–38.

28.

Shattuck

Durkin

Maenner

Newschaffer

Mandell

Wiggins

. (2009) Timing of identification among children with an autism spectrum disorder: findings from a population-based surveillance study. Journal of the American Academy of Child and Adolescent Psychiatry 48(5): 474–483.

29.

Simonoff

Pickles

Charman

Chandler

Loucas

Baird

(2008) Psychiatric disorders in children with autism spectrum disorders: prevalence, comorbidity, and associated factors in a population-derived sample. Journal of the American Academy of Child and Adolescent Psychiatry 47(8): 921–929.

30.

Sinzig

Walter

Doepfner

(2009) Attention deficit/hyperactivity disorder in children and adolescents with autism spectrum disorder: symptom or syndrome?. Journal of Attention Disorders 13(2): 117–126.

31.

Statistics Canada (2010) Visual census: Ethnic origin and visible minorities, Nova Scotia. Available at: www12.statcan.ca/census-recensement/2006/dp-pd/fs-fi/index.cfm?Lang=ENG&TOPIC_ID=11&PRCODE=12 (accessed 6 April 2011).

32.

Wiggins

Baio

Rice

(2006) Examination of the time between first evaluation and first autism spectrum diagnosis in a population-based sample. Journal of Developmental and Behavioral Pediatrics 27(2 Suppl): S79–S87.

33.

Williams

Thomas

Sidebotham

Emond

(2008) Prevalence and characteristics of autistic spectrum disorders in the ALSPAC cohort. Developmental Medicine and Child Neurology 50(9): 672–677.

34.

World Health Organization. (1993) The ICD-10 Classification of Mental and Behavioral Disorders: Diagnostic Criteria for Research. Geneva: World Health Organization.

35.

Zwaigenbaum

Bryson

Lord

Rogers

Carter

Carver

. (2009) Clinical assessment and management of toddlers with suspected autism spectrum disorder: insights from studies of high-risk infants. Pediatrics 123(5): 1383–1391.