Most children with autism spectrum disorder (ASD) show symptoms by 2 years of age, but many are not diagnosed until age 4 or older. Because earlier diagnosis of ASD is associated with positive outcomes, the American Academy of Pediatrics recommends universal screening for ASD. Conducted by primary care providers, universal screening would evaluate all children between ages 18 and 24 months for signs of ASD to promote earlier detection of the disorder and to ensure that all children, especially those from underrepresented groups (including rural and lower-income children), receive adequate supports and services. However, the US Preventive Services Task Force concluded that there is currently insufficient evidence to recommend universal screening. Existing data typically underrepresent children of color and children from low-income families. Additionally, there is limited data about the long-term outcomes of children who screen negative for ASD and subsequently do not return for follow-up.

In this study, researchers examined the accuracy of universal screening in a real-world setting by using a screening tool and long-term follow-up among a diverse population of toddlers in a large pediatric primary care health network. A total of 23,634 children were screened in the study and 50.4% were screened more than once. The researchers used the Modified Checklist for Autism in Toddlers with Follow-Up (M-CHAT/F), the most widely used screening tool for this age group. Primary care providers administered the M-CHAT/F to all children aged 16 to 26 months who received a well-child check-up from a Children’s Hospital of Pennsylvania primary care provider. All children in the study received follow-up screening at or after age 4. The researchers used the children’s electronic health records to confirm an ASD diagnosis, considering a child to have ASD if the diagnosis appeared more than once in the health record or if the diagnosis was provided by a specialist.

The researchers screened 91% of children who had eligible visits. Of these children, 9.5% screened positive for ASD on the first administration of the M-CHAT/F. After a second screening, the final number of children who screened positive for ASD dropped to 6.2%. The researchers determined that the M-CHAT/F is less accurate in detecting ASD in this “real-world” cohort than previously assumed. The M-CHAT/F positively identified only 38.8% of children who were later diagnosed with ASD (similar to what has been reported in other studies), and children who did screen positive for ASD only had a 14.6% probability of actually having the disorder.

There were differences in the sensitivity (percentage of true positives among those who screened positive) of the M-CHAT/F across sex, age, race, and income. The M-CHAT/F was less sensitive for girls than for boys. The M-CHAT/F was more sensitive for older toddlers than for younger toddlers at the initial screening, and combining results from repeated follow-up screenings yielded greater sensitivity than the first screening or any single follow-up screening. Children of color and children from lower income households were more likely to screen positive for ASD than white children or children from higher income families. However, the researchers found that the M-CHAT/F more accurately diagnosed white children with ASD than children of other racial or ethnic groups. Children from urban areas, low-income families, or families who used Medicaid were also less likely to receive an accurate diagnosis. These children were less likely to receive repeated screenings.

Despite demographic disparities, children who screened positive received an ASD diagnosis on average 7 months earlier than children who initially screened negative but did have ASD (as determined at a later date). This finding indicates that, in many cases, accurate early screening can accelerate a child’s diagnosis and potentially lead to earlier intervention. The researchers concluded that universal screening in primary care is possible when administered electronically and fully integrated with electronic health records. Although new tools are needed in addition to the M-CHAT/F to detect a greater proportion of children with ASD and reduce racial and socioeconomic disparities in the screening process, universal screening remains a promising method for early ASD detection among young children.

Although most children with ASD are diagnosed between the ages of 3 and 4 years, many children show ASD-related symptoms as early as 12 months of age. Because toddlerhood (12 to 36 months) is a stage of rapid brain development, late diagnosis often represents a missed opportunity for early intervention during this important developmental period.

Despite the potential benefits associated with early detection, the US Preventive Services Task Force believes there is currently insufficient evidence to recommend universal screening for ASD. Some opponents of universal screening for 18-month-old children have expressed concerns about diagnostic stability, the likelihood that an initial ASD diagnosis is confirmed at subsequent evaluations. Previous studies that have addressed diagnostic stability have only used small groups of 18-month-old children who were not representative of the general population, and no such studies have included a control group of children without ASD.

The goal of this study was therefore to determine diagnostic stability of ASD among children 12 to 36 months old who were randomly selected from the general population. The researchers also aimed to evaluate the diagnostic stability of ASD compared to other diagnostic categories (e.g., language or developmental delay) among children of this age group.

A California-based early detection screening program, Get SET Early, referred 2,241 toddlers to the study for ASD evaluation. To establish a control group for comparison, clinics participating in Get SET Early also referred typically developing (TD) toddlers. The study included 1,269 toddlers who received at least two evaluations from licensed psychologists every 12 months until 3 years of age. At each evaluation, the children received a designation of ASD, ASD features (showing signs, but not enough to meet DSM diagnostic criteria), developmental delay, language delay, other issue, TD, or TD sibling of an ASD child.

The researchers determined that overall stability for an ASD diagnosis was .84, meaning that 84% of children retained their original ASD diagnosis at follow-up evaluation. ASD diagnostic stability increased steadily as children aged—from .5 at ages 12 to 13 months, to .79 at 14 months, and to .83 at 16 months. In general, diagnostic stability was lower among children who were diagnosed with other delays, likely because non-ASD developmental delays often self-correct within the first few years of life. Overall stability for a TD diagnosis was .79.

Most of the children who transitioned from one diagnosis to another were initially determined to have language or developmental delays and then received an ASD diagnosis at later evaluation. The least common transition was from initial ASD designation to later TD designation, representing only 1.8% of the 400 children who were initially diagnosed as having ASD. The majority (71%) of these children who received initial ASD diagnoses were first evaluated at 12 to 13 months old, indicating that immediate referral for early intervention may have improved their long-term outcomes.

The researchers concluded that ASD diagnosis is stable by 14 months of age, and toddlers who receive an early ASD diagnosis are very unlikely to be considered TD later in life. An initial ASD diagnosis was found to be more stable than any other diagnosis, including TD. While the American Academy of Pediatrics recommends initial screening for ASD at 18 months of age, the results of this study suggest that ASD is reliably detectable and diagnosable in children as young as 14 months of age.

Identifying ASD in children can be a complex, time-intensive process that involves observing the child and obtaining a detailed developmental history from the caregiver. As a result, children whose families experience systemic barriers to accessing quality pediatric services may never receive a clinical ASD diagnosis, even when ASD symptoms are present. Current research on disparities in ASD diagnosis indicate that factors such as a child’s race and ethnicity, symptom severity, and co-occurring conditions can influence timing of diagnosis. White children are generally diagnosed 1–2 years earlier than non-white children. Unfortunately, children with ASD symptoms who receive a late ASD diagnosis—or never receive one at all—have less opportunity to access interventions and other services that may improve their long-term outcomes.

The goal of this study was to determine how many children have ASD indicators but no ASD diagnosis documented in health or education records. Additionally, the study aimed to identify demographic, individual, and service-related factors that might influence the presence of an ASD diagnosis in a large and diverse group of children. The study used data from the Autism and Developmental Disabilities Monitoring (ADDM) network, funded by the Centers for Disease Control and Prevention (CDC). During surveillance year 2014, the ADDM network conducted ASD surveillance in 11 sites located in AZ, AK, CO, GA, MD, MN, MO, NJ, NC, TN, and WI. Expert clinicians reviewed health and education records of eight-year-old children for social deficits that indicate symptoms of ASD (e.g., limited interest in other children or reduced eye contact). They then applied a standardized coding scheme to determine if children met ASD criteria outlined in the Diagnostic and Statistical Manual of Mental Disorders—Fifth Edition (DSM-5). They also coded information about behavioral characteristics, intellectual functioning, co-occurring conditions, and the presence of an existing ASD diagnosis documented in service records.

The researchers determined that 4,498 of the surveyed children met ADDM criteria for ASD surveillance. One-quarter of these children had documented ASD indicators comparable to diagnostic criteria (at least three social criteria and at least two of four behavioral criteria) but no formal ASD diagnosis documented in service records. Of these children, more than half (55%) were not receiving any ASD interventions or services in public school. Factors associated with not having a clinical diagnosis of ASD were non-White race, no intellectual disability, first developmental concern after three years of age, first developmental evaluation after three years of age, special education eligibility other than ASD, and need for fewer supports.

These results highlight the importance of reducing disparities in the diagnosis of children with ASD so that appropriate interventions can be promoted across communities. One strategy known to reduce these disparities is the use of patient navigators, who provide comprehensive guidance from diagnosis to intervention and help families overcome barriers to care. Other strategies that show promise at reducing disparities include the use of picture screens to overcome language barriers, presumptive eligibility approaches that promote early intervention for children who screen as at-risk, and extended screening efforts beyond preschool age to identify older at-risk children.