Black Families’ Experiences of Developmental Screening: Review of Well Child Visits to Inform Enhanced ASD Risk Assessment

Abstract

Objective:

Black families face barriers to early diagnosis of autism spectrum disorder (ASD). Most work emphasizes systemic delays to diagnosis rather than how existing screening procedures may impact identification. Our goal was to examine pediatric care visits in which screening was most likely to occur in order to document behaviors of parents and providers.

Method:

We examined 18-36 month primary care visits in our electronic health record system (n = 99) of 39 4- and 8-year-old Black children later diagnosed with ASD. We extracted qualitative and quantitative data and engaged in consensus coding. We captured whether formal screening occurred, the content of concerns of parents and providers, and referral patterns for follow-up care or evaluation.

Results:

Consistent with existing work, we found differences in parent and provider concerns and discrepancies in referral rates. Parents often endorsed concerns about language, sleeping or eating habits, behavior, or motor skills rather than ASD, but specific mention of ASD as a concern increased over time. Referrals for follow-up care were more likely when providers, not parents alone, expressed concerns about patient development.

Conclusion:

Pediatric providers cannot place the burden on families to raise autism concerns. Although some level of developmental risk was noted at most visits for children later diagnosed with ASD, referrals were only made when providers were also concerned, and the majority of these were for speech-language evaluation. Ongoing work is necessary to better understand how existing care systems interact with diverse families to inform the creation of inclusive screening practices that mitigate diagnostic delays.

Despite advances in screening and awareness, the average age of diagnosis for autism spectrum disorder (ASD) in parts of the United States remains later than four years of age.^1,2 This is especially true for children from traditionally underserved communities, who are much less likely to be diagnosed in toddlerhood.^1,3,4 A recent study investigating a large cohort of Black children in the United States found an average 3-year delay between first parental concern and ASD diagnosis; half of parents reported seeing multiple providers before children were diagnosed.¹ Although ASD prevalence rates are now roughly equivalent between Black and White children, disparities in age of first evaluation persist, with Black children less likely than White children to have a diagnostic visit by age 36 months.² Once evaluated, Black children are more likely to be misdiagnosed as having a behavioral or mood disorder than ASD relative to White children.⁵ These diagnostic delays contribute to family stress⁶ and restrict access to ASD intervention services that are critical for optimal functioning,⁷ potentially creating lifetime disparities.¹

Existing early ASD identification models of “wait-and-see” or “screen-and-refer,” in which medical providers assess developmental risk and either wait to until another visit to monitor change or refer out for subsequent evaluation if deemed appropriate,⁸ may exacerbate wait times and create geographic, financial, or other resource burdens for families.⁹ Evidence suggests that multi-stressed, rural, or traditionally underserved families may not be able to overcome these barriers to attend diagnostic evaluations at a tertiary care center.¹⁰ However, that is not the full picture: even when these more easily quantified barriers are not present, race-based disparities persist, with diagnostic delays present regardless of insurance coverage or socioeconomic status.¹ Concerns also exist that currently utilized developmental screening models have been developed based upon the input of primarily White providers and patient families.¹¹ That is, question structure and format may not reflect content most relevant, important, or accessible to diverse groups based not only upon language, literacy, or education level,¹² but also culturally relevant and specific reports of parenting experiences or developmental milestones.

Importantly, even when these barriers are not present, Black families report that professionals dismiss their concerns; they report that racial bias prevents their children from receiving adequate diagnostic care.¹³ They are less likely than White families to feel like providers spent enough time with their children or were sensitive to their family’s values.¹⁴ To better understand the cultural and systemic barriers affecting the diagnostic experiences of marginalized groups in the United States, Singh and Bunyak¹⁵ conducted a qualitative review and noted themes related to a lack of knowledge or awareness of ASD, the stigmatization of disability labels within the larger context of systems of oppression, a distrust of inequitable health systems, and different familial interpretations of symptoms all contributed to diagnostic delays.¹⁵ Constantino et al.¹ documented the “diagnostic odyssey” of over 500 African-American children with ASD and identified barriers to diagnosis related to wait times, evaluation quality, and the number of professionals seen, suggesting that early diagnostic delays have long-term consequences in the form of intellectual disability diagnoses.

Black families are traditionally underrepresented in ASD research, including development of screening methodologies. Growing evidence points to racial, ethnic, or cultural differences in how parents identify and communicate about developmental risk.¹⁶ Donohue et al. surveyed 174 White and Black parents of toddlers later diagnosed with ASD prior to a comprehensive developmental evaluation.¹⁷ They analyzed qualitative data and discovered that relative to White parents, Black parents were more likely to endorse concerns related to aspects of developmental delay not related to ASD. Specifically, both Black and White parents endorsed speech as a primary initial concern, but whereas White families endorsed social concerns as the second more frequent category, Black families instead endorsed motor and then “other medical” (e.g., sleeping, toileting) concerns. White families were almost twice as likely to endorse concerns aligning with social or restrictive and repetitive behavior categories than Black families. They noted that this has important implications for how medical professionals approach screening for ASD risk in a manner that is sensitive to racial, ethnic, or cultural differences in conceptualization and reporting of developmental concerns. These findings have been replicated in other work showing that Black or Latino families may have different conceptualizations of “typical” child development.¹⁸

Zuckerman et al.¹⁸ recommend that standardized surveillance procedures occur to reduce the chance of providers choosing not to ask minority families screening questions due to implicit or explicit bias. They specifically note that modifying screening procedures based on cultural norms is an important step, in addition to training providers in cultural variations in developmental and behavioral expectations in early childhood. Race-based care disparities emerge when examining not only developmental screening, but also provider levels of concern about developmental risk and provision of additional care.¹⁵ As evidence of these differences and disparities continues to build, prominent recent work has embodied calls for action and change within pediatric research and practice. Broder-Fingert et al.¹⁹ identified multifold contributions to, and byproducts of, structural racism within the ASD diagnostic pathway that likely contribute to delayed diagnosis for Black children. They call for an action plan to address inequities through actions such as understanding racial bias in our diagnostic conceptualization of ASD and embedding cultural humility into diagnostic practice.

Addressing the systemic structural racism that exacerbates these care disparities will require novel identification processes that specifically target underserved groups at risk for tremendous diagnostic delays, including Black families.^1,3,19 However, before new tools or processes can be developed, we must first understand the experiences of Black families within existing systems of care. Specifically, it is important to understand what Black parents of young children eventually diagnosed with ASD report to medical providers, and how those medical providers respond based upon existing care systems and structures.

As a first step in understanding these dynamics, we investigated two cohorts of Black children later diagnosed with ASD that received primary care at recommended ages for ASD screening at a large medical center. Our primary goals were to: 1) Understand what primary care screening procedures were used, 2) Gather data on parent- and provider-reported concerns, 3) Describe referral patterns, and 4) Analyze any differences across age cohorts. Importantly, this work did not include a White comparison group because its focus is on describing the documented screening experiences of Black families as a first step toward refining screening procedures, not demonstrating the presence of the care disparities described in the literature review above. This work contributes to the body of knowledge by engaging in retrospective review, informed by extant literature on healthcare disparities, of electronic health record data to describe patterns of early parent- and provider-reported developmental concerns for a sample of Black children later diagnosed with ASD.

METHODS

Creation of standardized data extraction template

Our first step was to identify those families across time whose children were ascertained as having ASD, that also had medical visits within our electronic record system (i.e., those children with documented care experiences that could be examined to determine areas of clinical growth and improvement). The next step was to review quantitative (e.g., number of visits) and qualitative (e.g., text-based provider notes) data from electronic medical records related to any visit with a primary care pediatrician in which developmental concerns were discussed between the ages of 12-40 months.

To do this, a team of clinical experts (including clinical psychologists, data extraction experts, Black stakeholders and developmental medical experts, and research staff) convened and engaged in an iterative process of reviewing well visit record structure and content and then identifying codable qualitative and quantitative data. These meetings resulted in a standardized coding form capturing 1) whether standardized aspects of screening procedures were administered, including the Modified Checklist for Autism in Toddlers-Revised [MCHAT-R],²⁰ the Parents’ Evaluation of Developmental Status [PEDS]²¹ and screening questions related to language, social, self-care, and motor skills; 2) the results of such screening (screen negative/screen positive/concerns noted); 3) whether referrals were made, and to where; 4) text-derived information about parent and provider concerns, and 5) information about DSM-5 ASD symptoms from any source associated with the visit.

Identification of visits for review

Our site has been a Centers for Disease Control Autism and Developmental Disabilities Monitoring site (CDC ADDM; NU53DD000010) since 2015 and has participated in ongoing and increasing efforts to understand race-based disparities in diagnostic identification within this data set.^2,22 Using the new CDC ADDM methodology for autism case definition, we identified 204 Black children (ages 4 and 8 years) with ASD in our surveillance area for study year 2018, which represent approximately 20% of the total children with ASD in our ADDM catchment. Children were identified as Black or African-American based upon parent-provided data in the electronic health record. Our CDC ADDM catchment area includes 11 counties surrounding a large university-affiliated medical center in the southeastern United States. This area encompasses a mix of primarily metropolitan counties with substantial variation in socioeconomic status (median household income range: $43,819 – $100,140).

The ADDM methods identify children in both the health and educational records. We first identified how many children had medical records available at the participant medical center, and then examined how many of them had primary care visits within outpatient pediatric clinics documented and available for review prior to the point of diagnosis or age 40 months, whichever occurred first. Please see Figure 1 for information about how many visits were reviewed and excluded to identify our final sample. The sample included 39 children who attended 99 separate well child visits or other primary care encounters (mean = 2.54 visits, sd = .99; see Table 1). Visits were conducted by 45 separate providers who conducted an average of 2.2 visits each (sd = 1.65, range: 1-7).

Figure 1.

Record and visit identification pathway.

Table 1.

Number of visits by age, screeners administered (n / % positive), presence of concern (n / % of total visits), and visits resulting in referrals (n / %).

Age of visit	Total visits (n)	M-CHAT Given? Total n (pos; %)	PEDS Given? Total n (n pos; %)	Parent concern n (% of visits)	Provider concern n (% of visits)	Visits with referrals n (% of visits)
18-24 months	31	27 (17; 63%)	26 (11; 35.5%)	17 (54.8%)	16 (51.6%)	17 (54.8%)
24-29 months	28	28 (10; 35.7%)	23 (10; 35.7%)	15 (53.6%)	18 (64.3%)	15 (53.6%)
30-35 months	13	4 (2; 50%)	12 (5; 38.5%)	10 (76.9%)	11 (84.6%)	7 (53.8%)
36 months	20	1 (0; 0%)	19 (6; 30.0%)	14 (70.0%)	14 (70.0%)	9 (45.0%)
Other	7	1 (1; 100%)	2 (0; 0%)	5 (71.4%)	7 (100%)	5 (71.4%)
Total	99	61 (30; 49.2%)	82 (32; 32.3%)	61 (62%)	66 (66.7%)	53 (53.5%)

Note: “Pos” = “positive screening result,” or scoring as showing developmental difference or concern.

Coding procedures

Each well visit was reviewed by two clinicians with expertise in ASD according to the standardized coding form described above and then consensus coded. The coding rubric used for ASD symptoms was based upon CDC ADDM abstract methodology. This included qualitative record data (such as text notes) reflecting provider or parent concerns or recommendations. Additional targets of coding related to standardized aspects of visits (including types of developing screening used) were identified via initial shared consensus coding of five visits. Content identified in this way and then coded going forward included: mention of concerns from either the provider or parent, the category of concern, results of developmental screeners (such as the MCHAT or PEDS), whether the child spoke 10 words or less, and referrals from the provider to developmental or diagnostic services. We also documented specific language in the notes, provider communication or forms that described developmental concerns (e.g., a parent report of “he won’t talk”), characteristics or behaviors that could be mapped onto DSM-5 criteria for ASD (e.g., poor eye contact, response to name, sensory interests), as well as specific associated features of ASD as aligned with ADDM methodology (e.g., abnormalities in eating, sleep disturbance, tantrums).

RESULTS

Developmental Screening Results

In this sample of Black children later diagnosed with ASD, most children received formal developmental screening using the MCHAT or PEDS (see Table 1). These tools were administered to parents just before the well child visit and then reviewed by providers as the visit occurred, in alignment with standard of care procedures at our institution at the time these visits occurred. More children screened positive for ASD risk on the MCHAT at 18 months than 24 months (Table 1), with 30 positive MCHATs total across 21 unique children. As seen in Table 2, children were more likely to be referred for ASD evaluation if they failed more than one MCHAT screening.

Table 2.

Parent only, provider only, and shared (parents / providers) concern within visits, by concern type, across timepoints.

Overall Categories of Expressed Concern
Visit and reporter	Speech	Behavioral Concern	ASD	Developmental Delay	Visits where a referral was made
18-24 month visit
Parents only (n = 7)	4	3	0	1	0
Providers only (n = 6)	5	0	2	5	7
Parent / Provider (n = 10)	9 / 9	4 /1	0 / 4	2 / 8	10
24-29 month visit
Parents only (n = 2)	2	2	0	1	0*
Providers only (n = 5)	5	0	2	3	4
Parent / Provider (n = 13)	11 / 13	8 / 2	0 / 4	1 / 7	11
30-35 month visit
Parents only (n = 1)	1	0	0	0	0
Providers only (n = 2)	2	0	1	2	2
Parent / Provider (n = 9)	7 /8	9 / 3	0 / 4	0 / 3	5
36-month visit
Parents only (n = 2)	1	2	1	0	0
Providers only (n = 2)	2	0	1	0	0
Parent / Provider (n = 12)	10 / 10	6 / 6	3 / 4	2 / 4	7
Other visit
Parents only (n = 0)	0	0	0	0	0
Providers only (n = 2)	1	1	1	2	2
Parent / Provider (n = 5)	4 / 3	2 / 2	1 / 2	2 / 2	2

Note: “Parent / Provider” indicates visits where both parents and providers expressed developmental concern, but not necessarily about the same developmental area.

^*= At both of these visits, although no referrals were made, providers had previously referred children at the 18-month timepoint.

Parental and Provider Concerns

Looking at the entirety of visit documentation, parents expressed some aspect of developmental or behavioral concern in 62% of visits (n = 61). Examining this in more detail (Table 3), the most commonly noted parental concerns related to language or behavior. In 5% of visits (n = 5), parents specifically mentioned ASD. Providers expressed some aspect of developmental or behavioral concerns in 67% of visits (n = 66; Table 1). Again, speech or language was the most common concern (see Table 2). In contrast to parents, providers noted concerns for ASD in around 25% of visits (n = 25).

Table 3.

Associated features of ASD documented by visit timepoint.

Concerns documented across all sources Visits n (%)
Visit timepoint	Eating	Sleep	Aggression	Motor	Hyperactivity	Tantrums	Language Delay	Play Delay
18-24 month (n = 31)	11 (35.5%)	0 (0.0%)	2 (6.5%)	7 (22.6%)	1 (3.2%)	1 (3.2%)	18 (58.1%)	5 (16.1%)
24-29 month (n = 28)	6 (21.4%)	5 (17.9%)	3 (10.7%)	3 (10.7%)	4 (14.3%)	1 (3.6%)	18 (64.3%)	3 (10.7%)
30-35 month (n = 13)	6 (46.2%)	3 (23.1%)	2 (15.4%)	1 (7.7%)	0 (0.0%)	3 (23.1%)	10 (76.9%)	5 (38.5%)
36-month (n = 20)	2 (10%)	2 (10%)	1 (5%)	4 (20%)	4 (20%)	5 (25%)	22 (55%)	1 (5%)
Other (n = 7)	4 (57.1%)	0 (0.0%)	2 (28.6%)	2 (28.6%)	1 (14.3%)	1 (14.3%)	6 (85.7%)	1 (14.3%)
All visits (n = 99)	29 (29.3%)	10 (10.1%)	10 (10.1%)	16 (16.2%)	10 (10.1%)	11 (11.1%)	74 (74.5%)	15 (15.2%)

Combined concerns

We next tallied concerns across both parents and providers. As seen in Table 3, considering all sources of information (provider observations, screening measures, parent reports), visits documented a variety of associated features of ASD over time, including sleep problems, delayed motor milestones (“will walk with his hands but not stand alone”), delayed play skills (“not playing well with toys”), and eating habits (“picky”). We then looked at each visit, identified the presence of developmental concerns, and then determined whether that concern was expressed by a parent only, a provider only, or both (see Table 4). When looking at “both,” the most common shared concern related to speech and language delays. In a closer look at concerns flagged during visits, we found that 48.7% of the sample (n = 19 children) at 18-months and 53.8% of the sample (21 children) at 24-to-36 months were saying less than 10 words.

Table 4.

Screen positives and referral patterns over time, per patient.

	Referred at any time point for:
	Patients (n)	ASD Eval	ST	Part C	OT	Not Referred
No positive screen ever
	16	5	1	7	1	11
Positive MCHAT at any visit
No MCHAT	2	0	0	1	0	1
0 positive	16	8	5	4	1	5
1 positive	12	8	7	5	2	1
2 positive	9	7	4	3	0	0
Positive PEDS at any visit
No PEDS	2	2	0	0	0	0
0 positive	16	6	9	5	2	4
1 positive	13	9	9	5	1	2
2 positive	5	2	5	3	1	0
3 positive	3	2	2	3	0	0
Referrals for children with “dual positive” MCHAT and PEDS at same visit
1 dual positive	9	7	6	7	0	2
2 dual positives	3	1	3	2	0	0

Note:

ASD Eval = Autism Spectrum Disorder Evaluation. ST = Speech-Language Therapy Evaluation. Part C = State Part C early intervention services. OT = Occupational Therapy Evaluation.

Each row within a category represents a separate group of patients (i.e., MCHAT 2 positive patients are not included in the MCHAT 1 positive row).

Patterns of referrals

We next wanted to understand provider behavior based upon the types of referrals made. As seen in Table 3, providers made referrals for follow-up evaluation or care in 53.5% (n = 53) of visits. Of note, when the provider alone expressed some aspect of developmental or behavioral concern, referrals were made 88% of the time (15 out of 17 visits). When both the provider and parent expressed concerns, referrals were made 75% of the time (35 out of 49 visits). In 10% of visits (n = 10), in which the parent alone expressed concerns but the provider did not have concerns, no referrals were made (see Table 2). These were children for whom previous referrals had not already been placed. Providers placed referrals at the next visit for all but one child, who was never referred across three encounters.

Differences in diagnostic age and parental concern between 4- and 8-year-olds

Within our total catchment group of 177 children, 54% (n = 95) were diagnosed with ASD after 36 months of age. Out of the 39 children in our sample, 18 were born in 2010 and 21 were born in 2014. When comparing the age of diagnosis between groups, 26% of children born in 2010 were diagnosed before 36 months (mean diagnostic age: 57 months) and 30% of children born in 2014 were diagnosed before 36 months (mean diagnostic age: 35 months). An Independent Sample T-Test indicated that this is a statistically significant difference (p < .004) in mean diagnostic age across groups. The frequency with which parents shared developmental concerns with providers also differed across groups. Children born in 2010 completed 46 visits, in which 46% of parents (n = 21) expressed concerns. In contrast, children born in 2014 completed 53 visits, in which 75% of parents (n = 40) expressed concerns. Referrals made showed a more modest increase; 50% of well-visits for children born in 2010 (23 / 46) resulted in a referral, whereas 57% of visits for children born in 2014 (30/53) led to a referral being placed.

DISCUSSION

In the current work, we utilized a population-based surveillance system (i.e. CDC ADDM) to identify Black children with ASD that had primary care visits prior to age 40 months available for review, in order to better understand the early developmental screening experiences of Black families. Many parents of young children later diagnosed with ASD expressed concerns about language delay, tantrums, sleeping, or motor skills as part of developmental monitoring. However, although some level of developmental risk was noted at most visits, referrals were only made when providers were also concerned about children, and the majority of these were for speech-language evaluation.

In each of our cohorts, fewer than half of children were identified with ASD prior to age 36 months. This may be a function of lengthy wait times at the university-affiliated tertiary care center. Encouragingly, when comparing the 8-year-old to the 4-year-old cohort, more parents of the younger cohort mentioned developmental concerns as part of well child visits. However, in a subset of visits, parents voiced concerns but, in absence of provider concerns, referrals were not made. Other work has found differences in provider referral behaviors based upon patient demographic data, with differences found based upon White vs. Black race, sex (male versus female), insurance type, and household income.²³ Although all but one of these children was eventually referred, the months of waiting lost could make a significant and negative impact. This “wait-and-see” approach,²⁴ in which provider waiting for developmental concerns to become more evident to make referrals, may have delayed access to early intervention and diagnostic care, particularly given the upper age limits of Part C (36 months) and the lengthy waitlists for most tertiary care centers.

Parents in this sample often expressed concerns about expressive language delays, sleeping or eating habits, behavior, or motor skills. Each of these is commonly associated with ASD, but may not represent a “classic” observable symptom aligned with diagnostic criteria or screening tools. This underscores the importance of a primary care workforce trained to recognize and query around possible ASD in the context of early developmental monitoring.^25,26 Providers cannot assume that ASD is not present just because parents do not use that term, particularly given documented differences in diverse communities related to familiarity with ASD, appropriate parenting or child behavior, or stigma associated with disability. The American Academy of Pediatrics encourages more thorough evaluation of developmental risk at any time a parent expresses concern, regardless of provider impression.⁸

Our results reflect the need for prompt, easily accessible referral pathways for diagnostic confirmation and subsequent intervention, regardless of whether ASD is mentioned. A recent qualitative study of diverse parents whose children received false positive results on ASD screening noted many benefits of the screening process, including increased understanding of typical child development and identification of other delays.²⁷ In future work, it will be important to determine how culture and race impact the concerns deemed most relevant or pressing across diverse parent groups, including Black families in the United States, and how to train providers to elicit these concerns. This may become particularly salient as children get older and differential diagnosis becomes more complex, with children mistakenly flagged as having attention- or behavior-related disorders as opposed to ASD.²⁸ It will also be important to evaluate whether specific clusters of non-specific developmental concerns, such as the trio of expressive language delay, tantrums, and sleep difficulties identified here, are more or less predictive of future diagnoses. Understanding such pathways may assist primary care providers in identifying those children most likely to benefit from comprehensive evaluation.

This work has several limitations. It is based on a small sample of children from a single medical center across a time span of several years, during which screening and referral practices may have changed, the number of children identified with ASD risk may have increased at large, and providers themselves transitioned to different roles. Our data was extracted from the medical record and based upon a survey created with stakeholder input; although standardized in this way, it was also limited in its ability to capture more subtle aspects of experiences that influence pediatrician behavior in real-time. Data reflects what was captured in visit notes and not the full breadth of the provider-patient interactions that occurred. Because of the nature of this record review, data was not available on provider characteristics that might have impacted patient-provider racial or ethnic concordance.²⁹ Additionally, because of the large number of providers included, we were unable to examine patterns of provider-specific behavior or the impact of repeated visits over time, with most children seen by different providers at each visit. We also did not have additional demographic information about patients to better characterize the sample. Finally, we did not compare data to a sample of White children, choosing instead to focus on Black families as a traditionally understudied and underrepresented group of research participants. However, broader understanding patterns of developmental screening across diverse groups from different racial, ethnic, and linguistic backgrounds will be critical for future work.

Acknowledging that structural racism impacts not only developmental screening but also our conceptualization of ASD¹⁹ is the first step in creating inclusive systems of care. Identifying ASD in toddlerhood can be complex, with care decisions related to multiple factors and inhibited by numerous barriers even when aggressively sought.³⁰ Our work highlights the critical necessity of identifying and eliminating barriers to equitable care, starting with increased understanding of how racially, ethnically, and linguistically diverse parents recognize and describe developmental risk to the primary care providers on the frontlines of developmental assessment. It emphasizes the need to gather systematic stakeholder input from families that will enhance screening tools and methodologies with input from those outside of the majority groups, mitigating existing biases in diagnostic care to improve outcomes for all children.