DOI: https://doi.org/https://doi.org/10.57187/smw.2023.40131
Autism spectrum disorder
Standard deviation
University Children’s Hospital Basel
Pediatric Institute of Southern Switzerland, EOC, Bellinzona
Autism Diagnostic Observation Schedule
Autism Diagnostic Interview – Revised
Modified Checklist for Autism in toddlers
Autism spectrum disorders (ASD) are pervasive developmental disorders with a heterogeneous phenotype and mixed patient outcomes.
The prevalence of autism has risen worldwide [1], particularly in Switzerland. Epidemiological research shows a prevalence of 1% in the general population [2]. In Switzerland, ASD is diagnosed at a mean age of 5.4 years [3], though it can be reliably diagnosed at 24 months of age or younger [3–5]. To our knowledge, there are no epidemiological studies looking at age of diagnosis in Switzerland.
Despite compulsory health insurance for all Swiss residents, ASD is diagnosed relatively late, often precluding the possibility of early preschool intervention [2, 6]. One reason for the persistence of late diagnosis could be the inconsistent use of screening tools such as the Modified Checklist for Autism in Toddlers (M-CHAT) [7], in addition to a lack of knowledge of early signs of autism among paediatricians [3]. The M-CHAT questionnaire is an easy-to-use and effective screening tool with good specificity in children aged 18 months or older [8–10].
Standard autism diagnostic practices are based on a structured parent interview and an observational assessment, frequently the Autism Diagnostic Interview–Revised (ADI-R) [11] and the Autism Diagnostic Observation Schedule–Second Edition (ADOS-2) [12]. Both tests require specific training for administration and provide complementary information necessary for an ASD diagnosis, either documentation of autistic behaviours present before age 5 as reported by the parent or caregiver (ADI-R) or observation of current autistic behaviours (ADOS-2) by a trained clinician in accordance with the latest version of the ICD-10 and DSM-5 diagnostic manuals.
Early diagnosis is crucial for improving outcomes in children and toddlers with ASD [13–16]. Studies show that early diagnosis facilitates gains in adaptive behaviour and IQ through timely intervention [13, 17]. For example, a recent randomised controlled trial showed important differences in the gains of children who started early intervention at 18 months of age versus 27 months of age [16]. Language abilities, including expressive language, receptive language and daily communication skills can improve through early treatment [18]. If early intervention is practiced with patients younger than 4 years, there is a good chance of decreasing ASD severity and of increasing cognitive and behavioural skills [19], as well as both social and daily living skills [20]. However, the gains available through early intervention appear to be accessible only if a child is diagnosed during preschool years. Later treatment may be less effective [21, 22].
A late diagnosis of ASD can lead to parental stress, delayed intervention and less optimal school placement [23]. However, despite these risks, there are many barriers that delay diagnosis of ASD. Isolating behaviours or slow language development can be falsely interpreted as sluggish development or shyness [24]. Parents often recognise delayed speech development or other non-specific signs of ASD, such as atypical sleep and aspects attributed to temperament, that do not differ from what is seen in other children with delayed development [25].
In Switzerland, services for children with disabilities are attributed differently in each of the 26 cantons, making it difficult to track and compare similar treatments. Outside of Switzerland, therapy and early intervention programmes are implemented with varying outcomes. While some children show great signs of improvement from early intervention, others benefit minimally, if at all [26]. The reasons for these differences in outcome are not yet understood [27]. One possible explanation is the variability inherent in the ASD phenotype, as well as differential cognitive functioning. Children with higher cognitive functioning and less severe symptoms may improve the most from early intervention [28]. However, they are also frequently diagnosed at a later age [28, 29]. More severe autistic disorders tend to be diagnosed earlier.
The objectives of this study were twofold: (1.) to measure age at diagnosis in two major diagnostic centres in Switzerland; (2.) to determine whether age at diagnosis decreased over a specific period during which a specific measure was implemented: an M-CHAT-based screening model in southern Switzerland or sporadic paediatrician training in northwestern Switzerland.
We conducted a retrospective, two-centre study using past clinical records from children and toddlers (aged 0–16 years) diagnosed with ASD in the neuropaediatric departments of two Swiss children’s hospitals between January 2006 and December 2016. We first searched the records of all patients diagnosed with ASD in both hospitals in early 2017. Nineteen different variables were extracted from the clinical records. The study was approved by the local ethics committees from the northwestern and central Swiss university hospitals (EKNZ BASEC 2016-01063) and Ticino (EK TI BASEC 2016-01063) and performed in accordance with the ethical standards defined in the Declaration of Helsinki.
Participants were children and toddlers diagnosed with ASD between January 2006 and December 2016 at the Neuropaediatric Department of University Children’s Hospital of Basel (UCHB) or the Neuropaediatric Unit of the Paediatric Institute of Southern Switzerland (IPSI), San Giovanni Hospital, Bellinzona. Patients diagnosed with an Asperger-like profile due to an absence of language delays, a risk for higher age at diagnosis [29], as well as diagnoses made before 2006 or after 2016 were excluded from the cohort. All diagnoses were coded according to the International Classification of Diseases, 10th edition. Children and toddlers were diagnosed using the ADOS-2 [12] and ADI-R [11] instruments, which were administered by trained professionals.
Both hospitals are reference centres for autism diagnosis made in their respective areas of Switzerland. The UCHB Neuropaediatric Department covers children living in the northwest of Switzerland. Most ASD diagnoses in this region are made at UCHB directly, which serves a region of approximately 650,000 people (approximately 8% of the total Swiss population). From 2006 to 2016, UCHB organised occasional trainings for regional paediatricians in collaboration with an outpatient clinic (in cooperation with local child psychiatric services) specialised in detecting and diagnosing autism spectrum disorders. The systematic use of screening tools was not part of the referral system in the northwestern region of Switzerland.
In southern Switzerland, use of M-CHAT was strongly encouraged by IPSI at the 2-year developmental health check between 2009 and 2012. By 2013, paediatricians were systematically using this checklist at well-baby appointments [30]. Ramelli et al. previously showed a significant decrease in age at diagnosis using this screening tool [30]. At-risk cases were subsequently referred for evaluation by regional paediatricians to a specialised outpatient clinic in IPSI. IPSI diagnoses most cases of ASD in children in the southern Italian-speaking part of Switzerland serving an area of approximately 350,000 people.
Clinical records for 258 ASD diagnoses were consulted; 112 cases were excluded for the above-mentioned reasons (an Asperger-like profile or missing data), leaving 146 records for analysis. Information on sex, nationality, native language, other languages spoken at home, presence of comorbid cerebral pathology and presence of comorbid epilepsy were available both in the IPSI and the UCHB datasets and were therefore used in the analysis (see table 1).
The dataset was created using SPSS (version 24) and Excel software applications, and was characterised by full information for all variables (no missing data). Descriptive statistics were used to illustrate the variables considered. Categorical variables were presented as percentages, whereas continuous variables were presented using either mean and standard deviation or median and interquartile range according to the distribution type (evaluated through a skewness-kurtosis normality test). Differences between hospitals were tested using the chi-square test (or Fisher’s exact test, where appropriate) and the Mann-Whitney test for categorical and continuous variables, respectively. Period-specific differences in age at ASD diagnosis between hospitals, as well as hospital-specific differences, were also assessed using the Mann-Whitney test. We used linear regression for assessing the relationship between age at ASD diagnosis and a four-category variable of interest created to account for hospital- and period-specific conditions (i.e. category 1 = UCHB, 2006–2012; category 2 = IPSI, 2006–2012; category 3 = UCHB, 2013–2016; category 4 = IPSI, 2013–2016), while controlling for a set of covariates. We fitted three versions of the model with a different reference category for the variable of interest, for the purpose of assessing whether age at ASD diagnosis was lower for category 4 (IPSI, 2013–2016) regardless of the reference category considered. Model 1 (M1) had category 1 (UCHB, 2006–2012) as its reference, while model 2 (M2) had category 2 (IPSI, 2006–2012) and model 3 (M3) had category 3 (UCHB, 2013–2016). We calculated the Variance Influence Factor (VIF) for each model to check for multicollinearity, while we used the Breusch-Pagan and Cook-Weisberg test to check for model heteroskedasticity. The statistical significance threshold was set at p <0.05. All statistical analyses were conducted with Stata/IC 16.0 (StataCorp, 4905 Lakeway Drive, College Station, Texas, USA).
The overall sample was composed of 70 children assessed at UCHB and 76 assessed at IPSI; table 1 gives their sociodemographic and clinical characteristics. These children did not differ between hospital sites by sex (approximately 80% males), presence of epilepsy (affecting 3–4% of the children) and period of diagnosis (slightly more than half were diagnosed with ASD between 2013 and 2016). Instead, children assessed at UCHB were more likely to be foreign (57% vs 42% at IPSI, p = 0.049), have a mother tongue different to the regional language (27% vs 4% at IPSI, p <0.001), speak more than one language (56% vs 30% at IPSI, p = 0.002) and suffer from cerebral pathology (13% vs 1% at IPSI, p = 0.006). Overall, the median age at ASD diagnosis was significantly higher at UCHB (44 months vs 36 months at IPSI, p <0.001).
Table 1Sociodemographic and clinical characteristics of study participants.
| Regional hospital | p-value of the statistical test | |||
| UCHB (n = 70) | IPSI (n = 76) | |||
| Sex | Male | 81.4% | 81.6% | 0.981 |
| Female | 18.6% | 18.4% | ||
| Nationality | Swiss | 42.9% | 57.9% | 0.049* |
| Foreign | 57.1% | 42.1% | ||
| Mother tongue | Matching regional language | 72.0% | 96.0% | <0.001*** |
| Different to regional language | 27.1% | 4.0% | ||
| Languages spoken | One | 44.3% | 69.7% | 0.002** |
| Two or more | 55.7% | 30.3% | ||
| Organic cerebral pathology? | No | 87.1% | 98.7% | 0.006** |
| Yes | 12.9% | 1.3% | ||
| Epilepsy? | No | 97.1% | 96.1% | 0.539 |
| Yes | 2.9% | 3.9% | ||
| Period of ASD diagnosis | 2006–2012 | 41.4% | 48.7% | 0.379 |
| 2013–2016 | 58.6% | 51.3% | ||
| Age in months at ASD diagnosis: median (IQR) | 44.0 (18.5) | 36.0 (16.0) | <0.001*** | |
ASD: autism spectrum disorders; IQR: interquartile range; UCHB: University Children’s Hospital of Basel; IPSI: Paediatric Institute of Southern Switzerland.
* p <0.05
** p <0.01
*** p <0.001
The detailed analysis of the differences in age at ASD diagnosis between hospitals and periods is presented in table 2. Considering the differences between hospitals, the median age at ASD diagnosis was significantly lower at IPSI both for the period 2006–2012 (39 months vs 47 months at UCHB, p = 0.038) and the period 2013–2016 (30 months vs 43 months at UCHB, p <0.001). If we consider the differences between periods within the same hospital, at UCHB the median age at ASD diagnosis did not change significantly (47 months for the period 2006–2012 vs 43 months for the period 2013–2016, p = 0.680), while it decreased significantly at IPSI (39 months for the period 2006–2012 vs 30 months for the period 2013–2016, p = 0.003).
Table 2Differences in months at ASD diagnosis between hospitals and between periods.
| Regional hospital | p-value of the statistical test | |||
| UCHB (n = 70) | IPSI (n = 76) | |||
| Difference in the age in months at ASD diagnosis between hospitals for the period 2006–2012 | Age: median (IQR) | 47.0 (20.0) | 39.0 (12.0) | 0.038* |
| Difference in the age in months at ASD diagnosis between hospitals for the period 2013–2016 | Age: median (IQR) | 43.0 (15.5) | 30.0 (14.0) | <0.001*** |
| Difference in the age in months at ASD diagnosis between periods for UCHB | Age, 2006–2012: median (IQR) [n = 29] | 47.0 (20.0) | – | 0.680 |
| Age, 2013–2016: median (IQR) [n = 41] | 43.0 (15.5) | – | ||
| Difference in the age in months at ASD diagnosis between periods for IPSI | Age, 2006–2012: median (IQR) [n = 37] | – | 39.0 (12.0) | |
| Age, 2013–2016: median (IQR) [n = 39] | – | 30.0 (14.0) | 0.003** | |
ASD: autism spectrum disorders; IQR: interquartile range; UCHB: University Children’s Hospital of Basel; IPSI: Paediatric Institute of Southern Switzerland.
* p <0.05
** p <0.01
*** p <0.001
The results of the estimated linear regression models are in table 3. The VIF (maximum mean value of 1.53) did not point to multicollinearity problems, but the Breusch-Pagan and Cook-Weisberg test indicated clear model heteroskedasticity (p = 0.000). To fix this problem, we calculated robust standard errors for assessment of statistical significance.
Table 3Results from the linear regression models. Robust standard errors in brackets.
| Variables | Dependent variable: age in months at ASD diagnosis | ||
| Coefficient (robust std. error) | 95% confidence interval | ||
| Lower | Upper | ||
| Model 1 (reference category: UCHB 2006–2012) | |||
| IPSI 2006–2012 | –7.356 | –14.810 | 0.097 |
| (3.769) | |||
| UCHB 2013–2016 | 1.756 | –7.092 | 10.604 |
| (4.474) | |||
| IPSI 2013–2016 | –13.102** | –20.827 | –5.376 |
| (3.907) | |||
| Model 2 (reference category: IPSI 2006–2012) | |||
| UCHB 2006–2012 | 7.356 | –0.097 | 14.810 |
| (3.769) | |||
| UCHB 2013–2016 | 9.112* | 1.809 | 16.416 |
| (3.693) | |||
| IPSI 2013–2016 | –5.746* | –10.311 | –1.180 |
| (2.308) | |||
| Model 3 (reference category: UCHB 2013–2016) | |||
| UCHB 2006–2012 | –1.756 | –10.604 | 7.092 |
| (4.474) | |||
| IPSI 2006–2012 | –9.112* | –16.416 | –1.809 |
| (3.693) | |||
| IPSI 2013–2016 | –14.86*** | –22.662 | –7.054 |
| (3.946) | |||
| Observationsb | 146 | ||
| R-squared | 0.191 | ||
* p <0.05
** p <0.01
*** p <0.001
a All models include the following control variables: sex (male vs female), nationality (Swiss vs foreign), native language (commensurate with regional language vs different to regional language), languages spoken (one vs two or more), organic cerebral palsy (absence vs presence) and epilepsy (absence vs presence).
b All models had the same number of observations and R-squared.
The three models indicated that, when controlling for sociodemographic and clinical covariates included in the model, age at ASD diagnosis was significantly lower at IPSI for the period 2013–2016 compared to all the reference categories considered. More specifically, we found greater differences compared to the UCHB reference categories, with an average reduction in age at ASD diagnosis at IPSI of 13–15 months, regardless of the period considered (p <0.01). The IPSI difference between periods was smaller, but still significant, with an average reduction in age at ASD diagnosis of approximately 6 months between the two periods (p <0.05).
We observed that systematic implementation of the M-CHAT screening questionnaire at the year 2 well-baby visit was associated with a decrease in age at diagnosis in southern Switzerland (from 39 to 30 months). In northwestern Switzerland, we did not observe a significant decrease in age during the same period, despite comparable diagnostic methods and an effort to include infantile autism as a subject in trainings and continued education for paediatricians. Moreover, in southern Switzerland, more children and toddlers were diagnosed with ASD, even though IPSI serves a much smaller portion of the population than northwestern Switzerland.
In Switzerland, everyone has compulsory healthcare insurance and access to healthcare. Therefore, both cohorts should have had access to healthcare providers, such as paediatricians, child psychiatrists and neuropaediatricians (developmental paediatricians). A difference in healthcare resources in the regions would most likely not explain the difference in age at diagnosis. In Switzerland, a lower age at diagnosis facilitates preschool access to therapeutic intervention programmes that can improve patient functioning and school placement [31].
There presently exists a wide choice of screening questionnaires that assess autism risk in young children [5, 10, 32]. In addition to good sensitivity and specificity, M-CHAT has the distinct advantage of being simple for parents to complete and simple for the clinician to score, allowing for quick and efficient administration during the short time that parents are waiting for an appointment [9]. The simplicity of the M-CHAT questionnaire surely contributed to the ease of implementation of the project, allowing for widespread adherence by paediatricians and a possible relationship with the decrease in age at diagnosis.
While further investigation is needed, our retrospective study supports the idea that implementation of M-CHAT at paediatrician well-baby visits may be a promising approach to lowering age at diagnosis in other geographic regions where paediatricians are the main source of referrals for specialised evaluations. During the period we studied, mean age at diagnosis in southern Switzerland fell from 39 months to 30 months, making some participants eligible for early intervention when they otherwise would not have been. The demonstration of such a decrease is encouraging, but not yet optimal. It will be important to continue to decrease the age at which autism is diagnosed in southern Switzerland to take full advantage of the neurological plasticity associated with the toddler years [16].
The authors thank the families who participated in the study, as well as Ms Rosanna Perlini for her precious help with data compilation.
The research was not supported by a grant; rather, the project was undertaken in the context of Dr Sommer’s post-graduate studies.
All authors have completed and submitted the International Committee of Medical Journal Editors form for disclosure of potential conflicts of interest. No potential conflict of interest related to the content of this manuscript was disclosed.
1. Prevalence of Mental Disorders in Children (Boat TF, Wu JT, editors). Mental Disorders and Disabilities Among Low-Income Children. Washington (DC); 2015.
2. Fernell E, Eriksson MA, Gillberg C. Early diagnosis of autism and impact on prognosis: a narrative review. Clin Epidemiol. 2013;5:33–43. 10.2147/CLEP.S41714
3. Studer N, Gundelfinger R, Schenker T, Steinhausen HC. Implementation of early intensive behavioural intervention for children with autism in Switzerland. BMC Psychiatry. 2017 Jan;17(1):34. 10.1186/s12888-017-1195-4
4. Shattuck PT, Durkin M, Maenner M, Newschaffer C, Mandell DS, Wiggins L, et al. Timing of identification among children with an autism spectrum disorder: findings from a population-based surveillance study. J Am Acad Child Adolesc Psychiatry. 2009 May;48(5):474–83. 10.1097/CHI.0b013e31819b3848
5. Zwaigenbaum L, Bauman ML, et al; Early Screening of Autism Spectrum Disorder: Recommendations for Practice and Research. Pediatrics. 2015; 136: Suppl. 1.
6. Rogers SJ, Vismara LA. Evidence-based comprehensive treatments for early autism. J Clin Child Adolesc Psychol. 2008 Jan;37(1):8–38. 10.1080/15374410701817808
7. Robins D, Fein D, Barton M. Modified Checklist for Autism in Toddlers, Revised. Italian Translation: Salomone E. Cecil C & Muratori F; 2014.
8. Kamio Y, Inada N, Koyama T, Inokuchi E, Tsuchiya K, Kuroda M. Effectiveness of using the Modified Checklist for Autism in Toddlers in two-stage screening of autism spectrum disorder at the 18-month health check-up in Japan. J Autism Dev Disord. 2014 Jan;44(1):194–203. 10.1007/s10803-013-1864-1
9. Sunita, Bilszta JL. Early identification of autism: a comparison of the Checklist for Autism in Toddlers and the Modified Checklist for Autism in Toddlers. J Paediatr Child Health. 2013 Jun;49(6):438–44. 10.1111/j.1440-1754.2012.02558.x
10. Petrocchi S, Levante A, Lecciso F. Systematic Review of Level 1 and Level 2 Screening Tools for Autism Spectrum Disorders in Toddlers. Brain Sci. 2020 Mar;10(3):180. 10.3390/brainsci10030180
11. Le Couteur A, Lord C, Rutter ML. Autism Diagnostic Interview – Revised. Italian Translation: Faggioli R, Saccani M. Persico (AM): Tancredi R, Parrini B & Igliozzi R. Giuntipsy Italy; 2005.
12. Lord C, Rutter ML, Dilavore PC, Risi S Luyster RJ, Gotham K, Bishop SL, Guthrie W; ADOS-2: Autism Diagnostic Observation Schedule-Second Version. Italian Translation: Colombi C, Tancredi R, Persico A & Faggioli R. Hogrefe Italy. 2013.
13. No authors listed. Early intervention for children with autism. Paediatr Child Health. 2004 Apr;9(4):267–77. 10.1093/pch/9.4.267
14. Rotholz DA, Kinsman AM, Lacy KK, Charles J. Improving Early Identification and Intervention for Children at Risk for Autism Spectrum Disorder. Pediatrics. 2017 Feb;139(2):e20161061. 10.1542/peds.2016-1061
15. Volkmar FR. Editorial: the importance of early intervention. J Autism Dev Disord. 2014 Dec;44(12):2979–80. 10.1007/s10803-014-2265-9
16. Guthrie W, Wetherby AM, Woods J, Schatschneider C, Holland RD, Morgan L, et al. The earlier the better: an RCT of treatment timing effects for toddlers on the autism spectrum. Autism. 2023 Mar;27(8):13623613231159153. 10.1177/13623613231159153
17. Reichow B. Overview of meta-analyses on early intensive behavioral intervention for young children with autism spectrum disorders. J Autism Dev Disord. 2012 Apr;42(4):512–20. 10.1007/s10803-011-1218-9
18. Reichow B, Barton EE, Boyd BA, Hume K. Early intensive behavioral intervention (EIBI) for young children with autism spectrum disorders (ASD). Cochrane Database Syst Rev. 2012 Oct;10:CD009260. 10.1002/14651858.CD009260.pub2
19. Dawson G, Rogers S, Munson J, Smith M, Winter J, Greenson J, et al. Randomized, controlled trial of an intervention for toddlers with autism: the Early Start Denver Model. Pediatrics. 2010 Jan;125(1):e17–23. 10.1542/peds.2009-0958
20. Remington B, Hastings RP, Kovshoff H, degli Espinosa F, Jahr E, Brown T, et al. Early intensive behavioral intervention: outcomes for children with autism and their parents after two years. Am J Ment Retard. 2007 Nov;112(6):418–38. 10.1352/0895-8017(2007)112[418:EIBIOF]2.0.CO;2
21. Fenske EC, Zalenski S, Krantz PJ, McClannahan LE. Age at intervention and treatment outcome for autistic children in a comprehensive intervention program. Anal Intervent Dev Disabil. 1985;5(1-2):49–58. 10.1016/S0270-4684(85)80005-7
22. Smith DP, Hayward DW, Gale CM, Eikeseth S, Klintwall L. Treatment Gains from Early and Intensive Behavioral Intervention (EIBI) are Maintained 10 Years Later. Behav Modif. 2021 Jul;45(4):581–601. 10.1177/0145445519882895
23. Elder JH, Kreider CM, Brasher SN, Ansell M. Clinical impact of early diagnosis of autism on the prognosis and parent-child relationships. Psychol Res Behav Manag. 2017 Aug;10:283–92. 10.2147/PRBM.S117499
24. Elder JH, Brasher S, Alexander B. Identifying the Barriers to Early Diagnosis and Treatment in Underserved Individuals with Autism Spectrum Disorders (ASD) and Their Families: A Qualitative Study. Issues Ment Health Nurs. 2016 Jun;37(6):412–20. 10.3109/01612840.2016.1153174
25. Ozonoff S, Heung K, Byrd R, Hansen R, Hertz-Picciotto I. The onset of autism: patterns of symptom emergence in the first years of life. Autism Res. 2008 Dec;1(6):320–8. 10.1002/aur.53
26. Vivanti G, Prior M, Williams K, Dissanayake C. Predictors of outcomes in autism early intervention: why don’t we know more? Front Pediatr. 2014 Jun;2:58. 10.3389/fped.2014.00058
27. Stahmer AC, Schreibman L, Cunningham AB. Toward a technology of treatment individualization for young children with autism spectrum disorders. Brain Res. 2011 Mar;1380:229–39. 10.1016/j.brainres.2010.09.043
28. Paynter J, Trembath D, Lane A. Differential outcome subgroups in children with autism spectrum disorder attending early intervention. J Intellect Disabil Res. 2018 Jul;62(7):650–9. 10.1111/jir.12504
29. Howlin P, Asgharian A. The diagnosis of autism and Asperger syndrome: findings from a survey of 770 families. Dev Med Child Neurol. 1999 Dec;41(12):834–9. 10.1111/j.1469-8749.1999.tb00550.x
30. Ramelli V, Perlini R, Zanda N, Mascetti G, Rizzi E, Ramelli GP. Early identification of autism spectrum disorders using the two-step Modified Checklist for Autism: experience in Southern Switzerland. Eur J Pediatr. 2018 Apr;177(4):477–8. 10.1007/s00431-018-3097-y
31. Myers SM, Johnson CP; American Academy of Pediatrics Council on Children With Disabilities. Management of children with autism spectrum disorders. Pediatrics. 2007 Nov;120(5):1162–82. 10.1542/peds.2007-2362
32. Sánchez-García AB, Galindo-Villardón P, Nieto-Librero AB, Martín-Rodero H, Robins DL. Toddler Screening for Autism Spectrum Disorder: A Meta-Analysis of Diagnostic Accuracy. J Autism Dev Disord. 2019 May;49(5):1837–52. 10.1007/s10803-018-03865-2