Transtorno do Espectro do Autismo: uma abordagem multifatorial

Abstract

Autism Spectrum Disorder (ASD) is considered one of the most serious disorders in child development due to its impact in the communication skills, language, social interaction, social cognition and patterns of behavior. The identification of factors associated with ASD, especially the modifiable ones, is important since it can guide public health policies for prevention. The early diagnosis and intervention are key factors for a better prognosis for someone with ASD. The diagnosis of cases with suspicion of ASD must be done based in historical data and behavior observation. There is a scarcity of validated instruments in Brazil to aid the clinical observation in the assessment of children with ASD symptoms. In this context, this thesis was carried out through the production of two studies. The first study aimed to develop an Instrument to assist the diagnostic evaluation of Autism Spectrum Disorder – IADASD and verify its psychometric properties. It is a methodological study performed in two stages: (1) Definition of the instrument objectives, population and establishing of conceptual structure; (2) Items elaboration and response scale; Items selection, organization and instrument structuring, (3) Content and apparent validity and (4) Construct validity, reliability and sensitivity. It was established that ASDDAI is an instrument to be adopted by professionals of health care, with the goal to aid in the assessment of the impairments that are characteristic of ASD in children between 2 to 12 years old. The construction of ASDDAI was guided by the conceptual structure of the Diagnostic and Statistical Manual of Mental Disorder (DSM-V). 50 items were organized in two domains: Social Communication (SC) (35 items) and Restrictive and Stereotyped Behavior (RSB)(15 items). The domain SC was constituted by three subdomains: the ability to communicate and language (18 items), social abilities (12 items) and inflexible behavior (5 items), while the domain RSB, was constituted by two subdomains: restrictive and repetitive behavior patterns (6 items) and sensorial disfunctions (9 items). The assessment of the instrument content validity was performed by 13 judges through the Content Validity Index (CVI). The results suggest evidence of validity in general scale, domains, subdomains and individual items, with CVI values over 80%. The items of ASDDAI were submitted to apparent validity, performed by 35 health care professionals that used Percentage of Agreement with values above 80%. After this stage, ASDDAI was tested in a sample of 82 children, 41 with an ASD diagnosis (ASD group) and 41 neurotypical children (neurotypical group). Scores of general scale, domains and subdomains were calculated through the sum of the number of items whpse answers indicated the presence of signs characteristic of ASD. All scores of the ASD group were significantly superior to the neurotypical group, demonstrating the capacity of the instrument in capturing the differences between the groups. As for the reliability, the general scale presented Kuder-Richardson Coefficient KR20 = 0.94 and the values obtained for the domains and subdomains were over 0.70, characterizing internal adequate consistency. To assess the sensitivity of the instrument a curve ROC (Receiver Operation Characteristic) was used, which demonstrated results that ASDDAI is capable to correctly classifying the presence of impairments that are characteristic of ASD. The following areas under the curve (AUC) were obtained: General Score (AUC =0.99; CI95% = 0.94 -1.00); domain SC (AUC = 0.98; CI 95% = 0.94 – 1.00) and subdomain RSB (AUC= 0.95; CI95% = 0.87 – 0.98). The second study aimed to investigate the interrelations among ASD and sociodemographic factors, health conditions, Family ASD history, and Attention Deficit Hyperactivity Disorder (ADHD) history in a population of Brazilian children and teenagers. Data from an epidemiological case control study performed with subjects aged 2 to 15 years old over the years 2015 to 2016 were used. The sample was constituted of 1,134 children and adolescents, 248 with ASD diagnosis (case group) and 886 neurotypical (control group). A theoretical model was elaborated based on the correlations or associations, described in the literature, among the analyzed variables. Structural equation modeling was used to estimate the strength of the correlations. The results suggest that the male gender (β = 0.33); p<0.001), older maternal age when giving birth (β = 0.15; p<0.001), occurrence of fetal distress (β = 0.19; p<0.001), presence of stress or sadness of the mother in pregnancy (β = 0.22; p<0.001), ASD in family (β = 0.19; p<0.001) and ADHD in family (β=0.12; p<0.001) had a positive effect over ASD. The socioeconomical condition correlated positively with mother age when giving birth (β = 0.34; p<0.001), as well as prematurity with fetal distress (β = 0.15; p<0.05).