2/2-Testing the impact of early screening on the long-term outcomes of children with ASD

自闭症谱系障碍(ASD)影响着今天出生的近2%的儿童，但人们对如何治疗知之甚少 积极改变受影响儿童的结局。一方面，包括美国科学院在内的许多人 儿科认为，在婴儿健康检查中进行早期普遍筛查是朝着积极的方向迈出的重要一步 结果，因为它可以导致早期治疗。相比之下，美国预防服务工作组 (USPSTF)最近未能支持早期普遍的ASD筛查，指出这样做的好处很小 明白了。需要为这场辩论提供信息的是检查一大批被检测到的儿童的结果 通过在健康婴儿检查中进行普遍筛查，很早就将他们与没有参与调查的儿童进行比较 一项早期检测计划。在这里，我们建议这样做：我们将检查学龄结果(6-10岁 在圣地亚哥和凤凰城很早就发现了242名患有自闭症的儿童，这是史无前例的样本规模(即， 通过我们的GET SET EARE计划，该计划包括在Well Baby Check时与CSBS进行筛查- UPS(平均年龄17.7个月，范围12-24个月)和立即转介进行全面评估和 如果有必要的话，可以接受治疗。然后，他们将与242名年龄匹配的自闭症社区儿童进行比较， 没有参加我们的早期检测计划的性别、种族、民族和自闭症患者(总计N=484)。给定 由于我们队列的稀有性和独特性，我们计划不仅根据传统的认知指标来描述结果， 社会行为等，但也取决于学校成绩的结果(例如，完全纳入主流和/或失去他们的 诊断)和家庭功能(家庭承受的压力是否较小？)。自全国ASD诊断平均年龄以来 大约4岁，我们预计社区队列中的儿童将有更晚的诊断年龄和更贫穷的年龄 结果比通过GET EARTY计划及早确定的结果要好。考虑到科学的严谨性和重复性， 我们将主动测试在波士顿对103名幼儿进行筛查的独立队列中的研究结果的重复性。 费城和纽黑文(即“东海岸队列”)通过早期项目和来自 相同地区(总N=120)。我们的具体目标是：目标1a：我们将确定临床上有意义的结果亚型 在我们的西海岸队列中使用无偏网络聚类方法的ASD。目标1b：具有这种独特的纵向 在队列中，我们将检查幼儿和学龄儿童之间症状特征、智商和适应功能的变化。 目标1c：我们将通过比较我们早期发现的儿童和 社区队列。目标1D：我们将检查在我们的东海岸样本中复制结果的情况。目标2：使用 免费的回归和机器学习技术，我们的总样本在西方和 东海岸队列(N=602)，我们将测试我们的模型，即在识别和治疗时年龄较早会导致改善 结果。为了检查与良好结果相关的其他因素，调节变量，如SES和水平 治疗参与也将包括在我们的模型中。目标3：由于国家背景(例如，政策、准则)可以 我们还会在结果中发挥作用，因为我们计划收集关键的州一级信息，以将我们的发现放在背景中。

Autism spectrum disorder (ASD) impacts almost 2% of children born today, yet very little is known regarding how to positively alter the outcomes of affected children. On the one hand, many, including the American Academy of Pediatrics, believe that early universal screening at well baby check-ups is an important step towards a positive outcome because it can lead to early treatment. In contrast, the United States Preventive Services Task Force (USPSTF) recently failed to endorse early universal ASD screening, noting that the benefits of doing so are poorly understood. What is needed to inform the debate is to examine the outcomes of a large cohort of children detected very early via universal screening at well-baby checks, and to compare them to children who did not participate in an early detection program. Here we propose to do just that: We will examine the school age outcomes (age 6-10 years) of an unprecedented sample size of 242 children with ASD detected very early in San Diego and Phoenix (i.e., “west coast cohort”) through our Get SET Early program, which involved screening with the CSBS at well baby check- ups (mean age 17.7 months, range 12-24 months), and immediate referral for comprehensive evaluation and treatment if warranted. They will then be compared to a cohort of 242 ASD community children matched on age, gender, race, ethnicity, and SES who did not participate in our early detection program (Total N = 484). Given the rarity and uniqueness of our cohort, we plan to characterize outcomes not only on traditional measures of cognition, social behavior, etc., but also on outcome on school achievement (e.g., fully mainstreamed and/or lost their diagnosis) and family functioning (do families experience less stress?). Since the national mean age of ASD diagnosis is around 4 years, we expect that children in the community cohort will have later ages of diagnosis and poorer outcomes than those identified early via the Get SET Early program. With scientific rigor and reproducibility in mind, we will proactively test the replication of findings in an independent cohort (N=103) of toddlers screened in Boston, Philadelphia, and New Haven (i.e., “east coast cohort”) through Project Early and a matched community cohort from the same region (Total N=120). Our specific aims are: AIM 1a: we will identify clinically meaningful outcome subtypes of ASD in our west coast cohort using unbiased network clustering approaches. AIM 1b: with this unique longitudinal cohort, we will examine changes in symptom profile, IQ and adaptive functioning between toddler and school ages. AIM 1c: we will evaluate program impact by comparing the outcomes children in our early-detected to the community cohort. AIM 1d: we will examine how well findings are replicated in our East coast sample. AIM 2: using complimentary regression and machine learning techniques, with our total sample collapsed across both west and east coast cohorts (N=602), we will test our model that earlier age at identification and treatment leads to improved outcomes. To examine other factors relating to a good outcome, moderating variables such as SES and level of treatment participation will also be included in our models. AIM 3: since state context (e.g., policies, guidelines) could also play a role in outcomes, in we plan to collect key state-level information to place our findings in context.