Evidence-Informed Milestones for Developmental Surveillance Tools

The AAP Systems of Services for Children and Youth With Special Health Care Needs team identified and convened 8 subject matter experts (SMEs) in different fields of child development. The group included developmental-behavioral, neurodevelopmental, and general pediatricians; child and developmental psychologists; and a professor of special education and early intervention. All SMEs had graduate training and experience in research methodology and medical decision-making and clinical experience in developmental surveillance, screening, and evaluation. One SME was an editor of and 2 contributed to the AAP’s Bright Futures: Guidelines for Health Supervision of Infants, Children, and Adolescents (4th ed).⁹  Two SMEs were lead authors of the AAP’s 2020 clinical report, “Promoting Optimal Development: Identifying Infants and Young Children With Developmental Disorders Through Developmental Surveillance and Screening”¹ ; another had long-time experience as a developer of screening tools.

The SMEs developed 11 criteria for CDC surveillance milestones and tools (Table 1). SMEs nominated the criteria on the basis of their clinical experience and use of CDC surveillance materials. Nominations were discussed, and those that were unanimously agreed upon were included as criteria. Of note, the SMEs agreed that milestones should be easily observed in natural settings and ≥75% of children would be expected to achieve a milestone at a given age. The strategy (≥75%) was chosen to support clinical judgment regarding performing additional developmental screenings, with validated screening tools as a next step to assess a child’s risk for developmental delays. This strategy may also prevent a wait-and-see approach because most children of the same age would be expected to achieve the milestone. To reduce confusion about when to be concerned, developmental warning signs were eliminated because not achieving milestones that most children (≥75%) are expected to achieve similarly warrants more in-depth surveillance and consideration for developmental screening.

Milestones for possible inclusion in CDC surveillance materials were identified by the SMEs. Existing CDC milestones were automatically included for evaluation and were the foundation for the revised checklists. Other milestones were identified by a literature review and commonly used developmental resources.

A broad literature search was conducted in March 2019 using MEDLINE, PsychInfo, and ERIC databases. Search terms were developed in collaboration with the CDC librarian after discussion of the objective of the literature review and comprised the following: (1) milestone, normative (data, table, range, value) or age (range, appropriate) AND (2) child development or infant development AND (3) percentile, psychometrics, predictive values, red flags, warning signs, assessment, monitor, delay, or reference (standard, values).

An article was included for evaluation if it was written in English, contained evidence that supported at least one normed individual developmental milestone or included published clinical opinion (ie, consensus milestones) that children exhibit the milestone by a specific age, and limited to children aged ≤5 years. Articles were excluded if evidence was limited to special populations (eg, preterm infants) or risk factors. Non–peer-reviewed articles, dissertations, and books were also excluded.

Articles were evaluated by the first 2 authors to determine if they met inclusion criteria. Articles chosen by both authors, either on initial review or after additional discussion and agreement between these 2 authors, had milestone data extracted for review by the SME group.

SMEs nominated several additional resources for evaluation, including parent resources, professional teaching resources, and commonly used screening and diagnostic evaluation tools (Table 2). Inclusion criteria for these resources were availability in English and contained evidence that supported at least one individual developmental milestone or included published clinical opinion (ie, consensus milestones) that children exhibit the milestone by a specific age. All nominations were accepted for additional review.

Existing CDC milestones served as a foundation for identifying milestones and were automatically evaluated.

The following data were extracted from existing CDC milestones, articles selected for review, and developmental resources: individual milestones, supporting references, and any normative data or published clinical opinion for the milestone. Milestones with conflicting data (eg, age at which most children should achieve the milestone) were flagged for additional discussion.

4A comprehensive database was created with information extracted from CDC milestones and resources for individual milestones from ages 2 months to 5 years, aligned with AAP HSV ages, and shared with the SME group. Original data sources, such as articles that met inclusion criteria and nominated developmental resources, were also shared with the SME group. SMEs collaborated through an in-person meeting, 6 virtual meetings, and e-mail reviews of decision summaries from January to September 2019.

During meetings, the SMEs discussed categorizing milestones into 4 developmental domains: (1) social emotional, (2) language/communication, (3) cognitive, and (4) motor. These domains were previously used in CDC materials and could help parents to learn about different areas of child development (eg, social-emotional skills in addition to language/communication skills). Because milestones often represent skills across several domains, they were placed in the domain in which the SME group believed that parents would most likely identify them. For example, reciprocal play skills involving other people were categorized as social-emotional, and other play skills were categorized as cognitive.

Milestones were reviewed for 2 to 3 ages at a time (eg, 2, 4, and 6 months) so that skill progression could be considered. Milestones were reviewed by the SME group on the basis of the criteria outlined in Table 1. First, SMEs considered milestones with normative data that supported achievement by ≥75% of children at a particular age. Next, they considered milestones from screening and diagnostic tools. Finally, SMEs considered milestones that were based on published clinical opinion. Each of these data sources was cross referenced with the others to get a sense of the evidence base available to support inclusion of a milestone in revised CDC checklists.

If there was disagreement across sources or lack of evidence supporting a milestone for a specific age, additional research was conducted, and additional evidence was then evaluated by the SMEs according to the developed criteria. This included reviewing diagnostic evaluation tools (Table 2) and/or conducting a separate PubMed search for a specific milestone or a related skill. For example, additional evidence was sought for age-specific development of gestures like lifting arms to be picked up, waving bye-bye, and blowing a kiss.

SMEs used the available evidence base and their clinical experience to determine if and at what age a milestone would be included in the revised checklists. Only milestones with unanimous agreement were included.

During the evaluation process, SMEs denoted milestones they included for surveillance but believed that additional research could improve age placement, quality, and/or quantity of supporting evidence or better capture the underlying developmental construct of the milestone. Finally, the SMEs simplified the milestones and added examples to try to improve understanding. The CDC team then reviewed milestones from a cultural and health communication perspective (eg, family friendly, fifth- to sixth-grade reading level), and SMEs reviewed those changes again to ensure that the milestone still represented the developmental construct being assessed.

With the use of back-translation methods, the milestones were translated into Spanish by CDC Multilingual Services and reviewed by 2 native Spanish-speaking pediatricians. Cognitive testing with a diverse sample of parents located in different regions of the United States provided feedback on relatability and clarity, which led the SMEs to make additional changes to wording. Results of cognitive testing will be published separately.

Of the 1027 articles generated from the literature review, 34 met inclusion criteria. Of those, 24 contained normative data, and 10 contained published clinical opinion for ≥1 milestones. Six additional articles were found through independent searches to evaluate individual milestones when SMEs determined that the milestone was appropriate but insufficient or conflicting evidence supported its use at a specific HSV age. Of the articles with normative data, 14 (58.3%) described populations within non–English-speaking countries.

Tables 3 to 6 list the milestones that SMEs included in CDC checklists after critical evaluation and unanimous decision. They are presented by domain with HSV age, references that support the inclusion of the milestone at that age, and whether the milestone was an existing or new CDC milestone. Supporting references are divided into normative data, developmental screening and evaluation tools, and published clinical opinions.

Previously, CDC had 216 milestones across 10 checklists. With the addition of 15- and 30-month checklists and the evidence review process, 159 milestones were included across 12 checklists. This represented a reduction of 57 (26.4%) CDC milestones, with the average number of milestones per checklist decreasing from 22 to 13.

Of the final 159 milestones that met the evaluation criteria, 94 (59.1%) were based on CDC original milestones and 65 (40.9%) were added on the basis of the milestone identification and evaluation process. One third of the 94 retained CDC milestones were moved to a different age on the basis of the criterion that ≥75% of children would be expected to achieve the milestone by that age. When moved, 21 of those 31 milestones were transferred to an older age. More than half (56.5%) of the original 216 milestones were eliminated (Supplemental Table 7) on the basis of SME criteria, including 25 that were duplicated across checklists at different ages. For example, Tries to use things the right way, like a phone, cup, or book” was on both the 12- and 18-month checklists; this milestone was placed only at 15 months on the basis of supporting evidence. Additionally, eliminating vague terms, such as may or begins, resulted in moving and changing milestones; for example, “Begins to pass things from one hand to another” was removed from the list of milestones at age 6 months and was included as “Moves things from one hand to her other hand” at age 9 months.

The 1-, 2-, and 3-year-old checklists had the greatest decrease in the number of milestones, with a ≥50% reduction. Approximately half of the aggregate loss for these ages was due to moving milestones to the new 15- and 30-month checklists. When combining all new and revised checklists for 1 to 3 years, the number of milestones was reduced by 25.7%, consistent with the 26.4% reduction of milestones across the other age ranges.

All 4 domains had a reduction in number of milestones. Cognitive milestones decreased by 34 (50.7%), social emotional decreased by 16 (27.5%), language decreased by 4 (9.1%), and motor decreased by 3 (6.4%). Social-emotional and cognitive domain milestones were the least likely to have normative data available. The social-emotional domain had 25 (59.5%) milestones with 0 to 1 normed references and 8 (19.0%) with ≥3 normed references. The cognitive domain had 19 (57.6%) milestones with 0 to 1 normed references and 8 (24.2%) with ≥3 references. In contrast, language and motor domains had 11 (27.5%) and 10 (22.7%) milestones with 0 to 1 normed references and 21 (52.5%) and 30 (68.1%) milestones, respectively, with ≥3 normed references.

Of the final 159 milestones, 127 (79.9%) were retained or added on the basis of normative data, whereas 32 (20.1%) milestones were included on the basis of screening and evaluation tools, published clinical opinion, and SME opinion. Of milestones with normative data, 32 had normative data from only 1 resource. The SMEs believed that 22 (13.8%) of the final milestones were candidates for additional research, as noted in Tables 3–6; this subset includes the milestone “Sings, dances, or acts for you,” the only milestone included or retained without supporting evidence on the basis of SME opinion.

Of the 77 developmental warning signs listed on the old CDC checklists, 59 (76.6%) had a corresponding milestone on the new checklists.

The CDC checklists support developmental surveillance and other important components of the early identification process by pediatricians and other ECPs, including developmental promotion, parent education and engagement, communication of developmental progress and concerns, and developmental screening.^{1,5,9,72–75}  However, variability across surveillance resources, including the CDC’s, can create confusion regarding what constitutes a concern and when developmental screening between recommended ages might be warranted.^3,4,38,76 

An expert working group convened by AAP sought to improve CDC surveillance tools by enhancing conversations among pediatricians, ECPs, and families regarding childhood development and guiding clinical judgment on when to conduct developmental screening between recommended ages. These tools are not intended to replace validated screening tools but instead to promote optimal child development and encourage professionals to act early through surveillance and screening as outlined in the AAP’s clinical report.¹  Results indicate substantial changes were made developing and applying criteria for surveillance milestones and tools, adding 15- and 30-month checklists, and incorporating evidence-informed milestones.

Clinicians have used attainment of developmental milestones for almost a century⁷⁷  to determine if a child is developing typically. Pediatricians have reported increasing use of milestone checklists from 53.0% in 2002 to 89.6% in 2016, in addition to their report of increased developmental screening.⁷⁸  However, these surveillance milestone checklists are likely based on published clinical opinion due to the lack of published normative milestone data and the lack of citations of original sources on checklists. The domain tables generated from this work linking developmental milestones to empirically informed evidence and published clinical opinion could improve training of professionals and methods for surveillance,^4,7,8  such as incorporation of the CDC’s open access milestones/checklists into electronic health records.

Milestone checklists used in surveillance are intended to prompt conversations, review developmental history and progress, and elicit concerns. The CDC checklists should not replace universal developmental screening, provide a risk categorization, or diagnose DDs. To determine if concerns about milestones should prompt a more in-depth developmental history, observation, and examination along with consideration for assessing actual risk by screening, the SMEs recommended that most children (≥75%) should be expected to achieve milestones by a given age.

SMEs believed that using 50th percentile milestones for surveillance would not support clinical decision-making for developmental screening because only half of children would be expected to achieve an individual milestone by a given age. Using milestones that 85% or 90% of children would be expected to achieve may limit opportunities for additional screening for too many children at risk for developmental delays. A ≥75% criterion was thus agreed upon to balance informed clinical decision-making regarding developmental screening and provide opportunities to identify children at risk for delays as soon as possible. Pediatricians are encouraged to follow AAP recommendations to screen a child, using validated screening tools, when surveillance reveals a concern or anytime a concern is raised, to identify and refine the risk for developmental delays.¹ 

The criterion (Table 1) for milestones representing those that most children (≥75%) would be expected to achieve eliminated the need for the CDC’s previous warning signs because most are now represented as milestones. Using this strategy, SMEs agreed that a child not meeting a milestone should be considered for screening similar to children demonstrating warning signs. Less than a quarter of warning signs could not be replaced with an evidence-informed milestone. Examples not represented as new milestones include subjective items such as “shows extreme behavior (unusually fearful, aggressive, shy or sad)” and items that may be better recognized during a physical examination, such as “stiff or tight muscles.” Additional features of the surveillance tools may identify those types of concerns. For example, the act early message was retained, and new open-ended questions were added to encourage parents to ask about other concerns.

Use of this same strategy is intended to discourage the wait-and-see approach and could prevent worry for children older than the average age of attainment of a milestone but not likely to be at risk for delays.^1,5  It could also eliminate the confusion families had with milestones and warning signs on the same checklist and repetition of milestones across checklists.

The application of additional criteria (Table 1) to improve clarity in the revised milestones may help parents, pediatricians, and other ECPs to recognize when missing milestones might indicate the need for developmental screening. Although milestones typically span several developmental domains, SMEs believed that keeping domain labels could improve awareness of the less commonly known social-emotional and cognitive development milestones of young children. Showing progression of a skill, like walking, could demonstrate how earlier milestones lay the foundation for later ones. Adding the open-ended question, “Is there anything your child is doing or is not doing that concerns you?” may encourage parents to discuss the quality with which a child exhibits a milestone, milestones not listed within the sample, and atypical behaviors difficult to capture through a list.⁴²  Finally, including information on developmental promotion and acting early may empower families to support their child’s development and to ask about screening if they have concerns.

Surveillance of social-emotional and cognitive milestones supports early identification of children with intellectual and social-emotional disorders, including autism spectrum disorder. In our review, milestones included in the social-emotional domain had the fewest normed references per milestone, followed by the cognitive domain, language domain, and motor domain. When social-emotional milestones were available, they were often self-help rather than social engagement and emotional regulation skills. These results highlight the need for additional research on social-emotional milestones. For example, can social-emotional function be assessed using milestones during developmental surveillance at HSVs, and what factors influence the timing and development of social-emotional skills? Despite relatively more evidence in the language domain, SMEs believed that more research is needed in the intersection of the social-emotional and language domains to recognize and monitor the development of social language in young children.

There were limitations with the revision process. SMEs’ opinions determined which HSV was most appropriate for a milestone, given existing evidence and the goal that most children (≥75%) would be expected to achieve it by that age. For example, if there was evidence that 50% of children reach a milestone at 11 months and 90% reach it at 16 months, it was placed on the 15-month checklist. Eighty percent of the revised milestones had normative data to support their inclusion on a specific HSV age checklist. However, it was necessary to use data such as published clinical opinion to represent milestones in all domains across all ages. Regardless of supporting data type, all revised milestones had unanimous SME support for their inclusion in a surveillance tool. When normative data existed, limitations remained in the availability of milestones that met other criteria. For example, sometimes it was difficult to find naturally observable milestones compared with milestones demonstrated during standardized testing. Although the milestones are evidence-informed using international sources and cognitive testing with parents was done, there is no certainty that these new milestones will resonate with all families or that they are the most relevant milestones for developmental surveillance. These checklists and others have not been tested to see whether they indeed lead to appropriate developmental screening between recommended ages and improve early identification.

There are other gaps in developmental surveillance research. Best practices for conducting the recommended 6 components¹  of surveillance are not available. There are limitations in the use of milestones, even evidence-informed ones, as the sole component of developmental surveillance. Milestones/milestone checklists can support some components, such as taking a developmental history and eliciting concerns, but surveillance also involves observation, examination, education, communication, and clinical decision-making. Moreover, trusting relationships may develop through the longitudinal process of surveillance and improve parents’ confidence in sharing concerns and pediatricians’ confidence to assess concerns raised.

Early identification and intervention for the 1 in 6 children with DDs have been shown to improve outcomes.^79–83  However, less than a quarter of children with DDs receive early intervention services before age 3 years,^84,85  and most children with emotional, behavioral, and developmental conditions, other than autism spectrum disorder, do not receive services before age 5 years.⁸⁶  Developmental surveillance is an important part of early identification^1,87  and facilitates education, communication, and relationship building among parents, pediatricians, and ECPs. Research in early identification has focused primarily on developmental screening, not surveillance. Improvements in surveillance tools and processes could help to identify concerns and support clinical judgment regarding developmental screening to allow more timely referral to early intervention services and additional evaluation. The methods described herein led to substantial revisions of CDC resources to better support developmental surveillance. Best practices for surveillance and improvements in surveillance tools could be supported by additional research on individual normed milestones, particularly social-emotional and cognitive milestones; how parents, pediatricians, and ECPs learn and understand milestones; which milestones are most likely to lead to appropriate screening; whether the use of milestones that most children (≥75%) would be expected to achieve is appropriate for surveillance; whether categorizing milestones into developmental domains is helpful; and whether cultural differences exist in surveillance milestones and processes.^1,3–5 ,7  Nevertheless, based on review of milestone data and clinical experience, the SMEs agreed that most typically developing children would achieve the developmental constructs represented. To our knowledge, this attempt is the first to align empirically informed milestones on parent-completed surveillance tools with objectively defined criteria agreed upon by SMEs. The CDC milestones and checklists can be used in continued efforts to improve developmental surveillance.

We thank Mary Cogswell, RN, DrPH, for her review of earlier versions of the manuscript and pediatricians Natalia Benza, MD, and Jose O. Rodriquez, MD, MBA, for their review of the Spanish milestone translations.

AAP

American Academy of Pediatrics

CDC

Centers for Disease Control and Prevention

DD

developmental delay or disability

ECP

early childhood professional

HSV

health supervision visit

SME

subject matter expert