Early Intervention referral information, transmission, and sources—A survey of state Part C Coordinators and analysis of referral forms

Abstract

Objective:

Early Intervention (EI) referral is a key connector between health care and early childhood systems serving children with developmental risks. This study aimed to describe the US network of EI referrals by answering: “What information is sent to EI?” “Who sends it?” and “How is it sent?”

Method:

This study combined an analysis of national document- and website-based referral forms with a survey of state Part C Coordinators (PCCs). Data on referral forms were systematically collected from state agency websites. PCCs from 52 jurisdictions were surveyed to assess current EI referral practices. Descriptive statistics were used for responses to multiple-choice items; free-text answers were condensed into key study themes.

Results:

EI referral forms came as e-documents (81%) or websites (35%), and 72% were in English only. They emphasized family and referral source contact information and reason for the referral. Survey results indicated that healthcare (45%) sends the most referrals, followed by families (30%). EI Agencies received referrals by phone (38%), electronically (23%), email (17%), and fax (17%), and PCCs valued this diversity of methods. Few states received referral data directly from Electronic Health Records (EHR); however, PCCs hope to eventually receive referrals via websites, mobile devices, and EHRs.

Conclusion:

EI referral data flow is complex with opportunities for loss of children to follow-up. This study described how EI referrals occur and provides examples of how communication and access to information may be improved.

INTRODUCTION

About one in six children in the US has a developmental disability or delay.¹ Early developmental services access for these children is important because neuroplasticity and response to therapeutic interventions is highest in early life. As a result, early identification and treatment of young children at risk for developmental disabilities is critical to improve children’s chances of maximal academic success and contribution to society as adults.^2,3

The Individuals with Disabilities Education Act Part C Early Intervention (EI) Program provides developmental services, often at no direct cost, to US children under age 3 with developmental delays or those with diagnosed conditions likely to affect development. States set their own eligibility threshold of developmental delay and define their own lists of diagnosed conditions; they may also choose whether to include infants and toddlers at risk of substantial delays. However, as many as half of the estimated 0.8 million US children with developmental delays in a given year are not receiving developmental services through EI.^4,5 Furthermore, children facing poverty, racial, and other social disadvantage—those who might benefit the most from EI—are also those less likely to receive it.^6,7 Children who would be eligible for EI must progress through the initial surveillance, screening, referral, evaluation, and treatment plan generation, before they can finally enroll in services. Unfortunately, not all children complete this process.

The EI referral is a uniquely vulnerable step in this EI access pipeline, because it traverses two different service sectors and information ecosystems that are frequently “siloed” from each other. Records of children with developmental disabilities are fragmented between health care, educational, and other organization types that do not exchange information easily. As a result, providers in both healthcare and early childhood systems are frequently unaware which children are not getting needed care, and why.⁸ In addition, most states cannot track the number of children screened or referrals initiated; they can only record the number of referrals their EI agencies receive.⁹

At least three studies confirm that many children drop out of the pipeline once they are referred. Proportions of children lost to follow-up once referred to EI were 69% (McManus et al. 2020⁵, 1,887 out of 2,746 children referred), 49% (Guevara et al. 2013¹⁰, 162/332), and 38% (Atkins et al. 2020¹¹, 29/77). Studies based on the Massachusetts Pregnancy to Early Life Longitudinal (PELL) data system, although valuable, do not appear to count the number of referrals initiated at their source.¹²

Reasons for low rates of Early Intervention use include stigma, mistrust, logistical challenges, limited resources, and communication failures between families, providers, and health/educational systems.^13–15 Of these, communication failures may be particularly important and amenable to intervention. Fortunately in this decade, data sharing policy changes and advances in cloud computing offer new opportunities for solutions. Successful solutions also require an understanding of current practice, but data is limited about what information an EI referral contains, who sends referrals, and how referrals are sent. To address this gap, in this study we analyzed referral forms from all states and surveyed Part C Coordinators (PCCs) to 1) describe information sent when referring a child to Early Intervention, 2) identify the primary sources of these referrals, and 3) identify their transmission methods. Results from this study can inform improvement in EI access and associated information systems for state EI agencies, healthcare organizations, and others.

METHODS

Our approach consisted of two parts: 1) analysis of states’ standardized Early Intervention (EI) referral forms, and 2) a survey of state EI agency Part C Coordinators (PCCs) on their agency’s EI referral methods.

EI referral forms are readily available from many state agency websites. They reflect the information priorities of EI agencies during the transition between the referring person and EI staff receiving the referral.

State Part C Coordinators are key parts of each state’s EI program leadership, managing the program, facilitating the work of the interagency state leadership team, and coordinating professional development staff and projects. Often having risen through the ranks of their own agencies, they represent their state’s EI program to the US Office of Special Education Programs, which administers funding and technical support to states under Part C of the Individuals with Disabilities Education Act. Our survey of state PCCs allowed an overview of referral processes in a national sample. The Early Childhood Technical Assistance Center provided PCC and EI agency contact information for all fifty states, the District of Columbia, and Puerto Rico.¹⁶ State EI eligibility criteria categories, detailed in Supplemental Digital Content, were established by the IDEA Infant & Toddler Coordinators Association 2010 Data Committee.¹⁷ Lead agency type was established using the methodology of Twardzik and colleagues.¹⁸

Both the referral form analysis and the survey results contributed answers to each of the three study aims, as detailed below. See Figure 1 for a schematic of study activities. The Institutional Review Board of Oregon Health & Science University (OHSU) approved this study. IRB-approved information sheets were sent with survey invitations and embedded in the survey, and informed consent was inferred from completion of the survey.

Figure 1. Early Intervention referral information, transmission, and sources Study Flow Diagram.

a. In a few states, more than one person shares the position of Part C Coordinator

b. A 37^th state indicated a state-wide form, but unable to obtain for analysis

Referral Form Data Collection

We manually examined all 52 agency websites for EI referral forms. We collected both downloadable, printable electronic documents (hereafter referred to as “e-documents,” e.g. Portable Document Format, PDF) and website-based forms with information entered directly online (hereafter referred to as “web forms”). We collected forms from every state with statewide forms and, where states offered both, one each of e-document- and website-based formats. We categorized states as having no statewide forms if their website instructed visitors to contact their local EI agency to access EI and no referral form was available or mentioned on the site. We noted the date of last form revision, the presence of document uploading capacity for web forms, the presence of fields for preferred language and the child’s race, and the existence of forms in languages other than English.

PCC Survey Development and Data Collection

Survey Population.

With consent of the IDEA Infant Toddler Coordinators Association,¹⁷ we surveyed PCCs representing their state’s EI agency in all fifty states, the District of Columbia, and Puerto Rico. In a few states, more than one person shares the position of Part C Coordinator; we invited all of them to maximize likelihood of response.

Survey Development.

We developed survey items to address our three study aims, to assess capacity for use of electronic referral methods, to allow unstructured comments, and to characterize respondents. We explained technologic terms in plain language, displayed progress throughout, and allowed respondents to leave questions unanswered. Study staff revised survey drafts with feedback from two primary care pediatricians, two informaticians, two speech pathologists, a physical therapist, a family medicine physician at a community health center, and a subject matter expert at the Early Childhood Technical Assistance Center.¹⁶ They reviewed the survey for clarity, ease of completion, and feasibility of information requested. We pilot-tested the survey (see Supplemental Digital Content 1: Survey Questions) with four state and regional EI leaders in Oregon.

Survey Content.

The survey contained two items asking whether a statewide referral form was used and how often. Four items asked about EI referral source type (e.g., parent/family, doctor/clinic, Childcare/Head Start, Child Protective Services, Others), proportion of referrals from each source, available means of referral transmission (e.g. fax, telephone, email, other electronic means, postal, in person) and estimated proportion of referrals received by each means. Another item clarified the type of electronic referral receipt. Four survey items asked about statewide EI data systems, what types of information they held, and whether they use the Common Education Data Standards¹⁹ supported by the US Department of Education, and Health Level Seven messaging standards such as HL7 and Fast Health Interoperability Resources (FHIR).²⁰

The survey included open-ended questions about state EI system processes (“What about your state’s EI referral system works well?” “What doesn’t work well?” and “How would you make it better?”), and an optional comment section. We also collected survey respondent characteristics, including professional background and years in position.

Survey Sampling.

We sent email and postal invitations to each state’s PCC, with follow up by email and telephone. Invitations contained respondent-specific links, and postal invitations included a $2 bill as an incentive. We used Qualtrics (Provo, UT) and Microsoft Access 2016 to collect data. We resolved discrepancies between survey responses and/or referral form data through emails to respondents.

Data Analysis

We used Microsoft Excel 2016 with Real Statistics Resource Pack (Release 6.8, 2020, www.real-statistics.com) for quantitative analysis including descriptive statistics, tests of difference between groups, and an evaluation of non-response bias using logistic regression. To compensate for variations in the number of forms available for each state, we weighted state-based measures—such as average number of fields—against the number of forms per state.

Aim 1: Information sent when referring a child to EI.

To determine what information is sent, we grouped referral form data fields into categories (e.g. “Parent phone number” placed under Child-Parent Contact) used on the Early Intervention Program Referral Form from the American Academy of Pediatrics (AAP).²¹ We considered this a reference document because it was developed by a group of leaders in the field and we have found references to it across our research—including citations on current state EI referral forms. We computed the total number of fields per referral form, median number of fields in each category, distribution of field number within each category across states, and statistical tests of difference in field number between e-document and web forms.

Aim 2: Sources of EI referrals.

To determine intended users of referral forms, we reviewed the referral form title, statements about who can use the form, the presence of separate forms for multiple user roles, and the presence of fields requesting medical history or diagnosed conditions. We then assigned an intended user category (Healthcare provider, Anyone can use, No user clearly stated, Other) to each form. Keeping in mind that families often place the referral, we calculated the proportion of forms translated into non-English languages.

We used the PCC survey responses to enumerate referral source type (e.g., doctor/clinic, parent/family, etc.) and proportion of referrals from each source type, by state/jurisdiction.

Aim 3: Transmission methods of EI referrals.

To assess types of forms available, we tabulated the number and type of forms on each state agency website. To illustrate the role of referrals that were not based on statewide forms, we assessed alternate means of initiating referrals by the presence of a state-wide referral telephone number, absence of a statewide referral form, and instructions for contacting your local EI provider on agency websites.

To assess methods of sending and receiving referrals using survey data, we tabulated which methods of referral transmission and receipt were available and were most often used, by state, including types of electronic referral receipt. To assess statewide EI data system capacity using survey data, we assessed what types of information data systems held, including use of Common Education Data Standards,¹⁹ FHIR,²⁰ and HL7.

Free text survey responses.

We used NVivo 12 (QSR International) to organize free-text survey responses to the three open-ended questions and comment section. The free-text survey questions were not designed to support qualitative analysis, but rather to allow respondents to express additional thoughts and offer clarifications. BS initially summarized responses based on relevance to study objectives and potential to enrich interpretation of quantitative results. Discussions between BS, JA and KZ then helped to integrate them with other study results. Further feedback was provided by discussion with two practicing Speech Language Pathologists and The DaSy Center for IDEA Early Childhood Data Systems.²²

RESULTS

Referral Form General Results

We found and collected statewide Early Intervention (EI) referral forms from 36 of the 52 state EI agency websites. A 37^th state’s website indicated use of a statewide form but its coordinator did not respond to requests for a copy. We analyzed a total of 59 forms, as some states offered various versions for multiple languages or user roles.

Survey General Results

PCCs from 41 states and jurisdictions (79%) responded to the survey. These leaders most often held a Master’s (51%) or Bachelor’s (32%) degree and had held their positions for 6 years or less (Table 1). Ninety-eight percent of respondents considered themselves either “Extremely” or “Very” familiar with their state’s referral process. We achieved good representation of US states with respect to region, EI eligibility criteria, lead EI agency type (no difference for each with Chi-square, p>>0.05), and the proportion of the population that identify as People of Color, Black, or Asian American (no difference with t-tests for each, p>>0.05).

Table 1.

Part C Coordinator Survey: Respondents & Their States

Respondents		Value
Years in current position		Median = 4.0 (Range 0.6 – 29.0)
Education		Assoc Degree or less: 7%, Bachelor’s: 32%, Master’s: 51%, Doctorate/prof: 10%
Years since degree		Median = 20 (Range 2 – 46)
Familiarity with state’s referral process		Extremely: 71%, Very: 27%, Moderately: 2%, Slightly or Not: 0%
State Attributea		States in Survey	United States
U.S. Census Region (NS, p=0.93)	Northeastern:	7	9
	Midwestern:	11	12
	Southern:	14	17
	Western:	8	13
EI Eligibility Criteriab (NS, p=0.77)	Broad:	21	24
	Medium:	7	12
	Narrow:	12	15
Lead EI Agency Type (NS, p=0.84)	Education:	11	12
	Health:	12	18
	Other:	18	22
		Survey State Mean %	US %
Race (NS, p>>0.05)	All People of Color:	21.7	23.5
	Black:	12.2	13.4
	Asian American:	4.4	5.9

^a.No significant difference between survey states and all US states

^b.EI Eligibility Criteria definitions presented in Supplemental Digital Content 2: State EI Eligibility Categories

Abbreviations: EI – Early Intervention, NS – Not Significant

Free-text responses were submitted for all three questions by 88% (37/42) of respondents, and respondents answered 92% (116/126) of these questions. Comments were made in the optional comment section by 52% of respondents. In all, free-text responses totalled 6,072 words.

When testing for non-response bias against data obtained for all 52 states, survey response did not correlate with any of the following: having a statewide referral phone number (AOR 2.54, 95% CI 0.38 – 16.98), having a statewide referral form (AOR 4.68, 0.79 – 27.75), having an online statewide referral form (AOR 0.11, 0.01 – 1.41), whether the EI agency is part of the state’s health, education, or other government branch (AOR 0.28, CI NS and AOR 4.30, CI NS respectively with Other as reference),¹⁸ or whether states classify their own eligibility criteria as narrow, medium, or broad (AOR 1.03, 0.44 – 2.41, see Supplemental Digital Content 2: State EI Eligibility Categories).¹⁷

Aim 1: Information sent when referring a child to EI

The referral forms—reflecting states’ information priorities for EI referrals—described what data are sent. Figure 2 shows the AAP form categories and how many data fields in each were found across state-wide referral forms. All statewide forms solicited Child-Parent contact information and Reason for Referral, and most asked for information about the Referral Source. Reason for Referral held the most fields on many forms, with specific qualifying diagnoses and risk condition fields numbering as high as 98, exceeding the thirteen Reason for Referral fields on the AAP EI Referral Form.²¹ At the other extreme, 25% of states with statewide forms (not shown in figure) did not request potential eligibility data on their forms, often using only a free-text field labeled “Reason For Referral.” E-documents held more data fields for four out of the eight categories.

Figure 2.

EI Referral Form Analysis: Box plot of data fields found on state-wide EI referral forms, by data field categories appearing on the AAP EI Referral Form.

Data Field Categories: Child-Parent: Child & Parent Contact info; Reason for Referral: Developmental concerns, screening scores, diagnoses; Referral Source: Referral Source Contact info; EI Program: Early Intervention Program Contact info; Feedback: Result info EI staff returns to referral source; Release of Info: Consent by parent to share child’s info; PCP: Primary Care Provider info; Agency Only: For use only by EI staff receiving referral.

Note 1: For states with multiple forms, number of fields represents an average.

Note 2: Data field categories in bold denote statistical difference between the 2 form types (t-test, p<0.05).

Abbreviations: AAP – American Academy of Pediatrics; EI – Early Intervention; IQR – Inter Quartile Range; Q1 – First Quartile

The remainder of information categories on referral forms (EI program information, Feedback Requested, Release of Information, Primary Care Provider, and Agency Use Only) were absent on at least half of the forms analyzed. We identified 133 unique field types across the eight categories. Giving equal weight to each state (some have more forms than others), a median of 49 (range 11 – 142) fields appeared on each form. Although last revision dates were not displayed on the web forms, the average age of revision for e-documents with such information was 3.6 years (weighted mean, n=34).

Fields for parent signature were found almost universally on the e-document forms, and not on the website-based forms. Documenting parental consent appeared to remain a paper-based process, even for states with web forms. Four of the thirteen states (31%) with web forms had a supplemental function to upload documents, which might have been necessary to record these handwritten signatures.

As evidence of EI agency attention to diversity, we noted that 32 of the 36 states with statewide forms (89%) solicited the family’s preferred language or need for an interpreter. Thirty-nine percent (14/36) asked for the child’s race; however, the AAP form does not.

From survey responses, we discovered that the accuracy and completeness of referral data was important to PCCs, for example as reflected in open ended-responses to the question “What about your state’s referral system doesn’t work well?”

When information about the child/family from referral source is incomplete, this delays the processing of the referral.

[PCC #16]

As paper forms and e-documents on a computer are limited in their ability to enforce data entry practices, coordinators expressed the ability or desire to improve data accuracy using web-based forms:

It would be nice to have an electronic method for submitting referrals. The electronic form would include required fields and the referral source would not be allowed to submit the form until it is complete.

[PCC #16]

Other PCCs shared challenges with data entry compliance in their statewide EI data systems, for example, affecting their agency’s ability to track referral sources accurately:

We are able to analyze all referral data, but are working on better identifying how to support common referral sources to correctly identify themselves in the dropdown selection to ensure we have accurate data.

[PCC #4]

Aim 2: Sources of EI referrals

As for the apparent intended users of states’ forms, 25% of states with statewide forms explicitly identified healthcare providers, 14% of them stated that anyone can use the form to make a referral, and 50% did not specify. The forms frequently contained requests for diagnosed conditions and medical history, suggesting that healthcare professionals would fill them out. Some states (17%) offered separate forms for different roles such as parent, professional, and EI staff member. This was most common among web forms (86%).

For parents or others who are not fluent in English, 5 out of 36 states with statewide forms (14%) offered their referral form in English and Spanish. Only an English form was found for 72% of states. Another 14% of states (5) offered a third-party auto-translator function within their web form that can render the text on that page in one of up to 108 languages by selecting from a drop-down list.

PCCs estimated that healthcare settings were the most common source of referrals overall (mean 44%, SD 17), followed by parents or family members (30%, SD 16), child protective services (11%, SD 11), childcare or early childhood school settings (5%, SD 6), and a collection of other sources (10%, SD 9). A supplemental figure (see Supplemental Digital Content 3: % EI referrals received by source) shows variability among states. For example, PCCs from six states estimated that they receive more referrals from parents than from healthcare professionals.

Aim 3: Transmission methods of EI referrals

As for forms available for use when referring children to EI, we found statewide forms at 71% of the 52 state EI agency websites, which was consistent with survey results in which 68% of responding states reported using statewide only or both statewide and regional forms. These forms were not always used; PCCs on average reported 67% of all incoming referrals used their state’s statewide form. As for types of forms, 30 of the 52 agency websites (58%) offered downloadable e-documents and 13 (25%) offered a web-based referral form. Among these two groups are six states (12%) that provided both types (Figure 1).

Sending referrals.

Survey respondents indicated that referral sources in their state can send referrals using telephone calls, facsimile, in person, email, postal mail, and websites. Thirty-four percent of states had at least one healthcare setting that can send referrals electronically. Almost all state websites contained information on how to find their local agency or otherwise initiate a referral. Two states had a mobile application available to help users find their local agency and even send referrals.

State PCCs expressed interest in outreach to families using mobile technology.

Something that could improve the process in the future is having the ability to receive referrals from texts or apps.

[PCC #21]

We do miss out on some families who only want to text or e-mail…

[PCC #10]

Ease of referral was a priority for many PCCs:

Since we started this referral system, our Child Find numbers have increased dramatically and continue to grow each month. It is a streamlined system that is easy to use and effective in moving a referral forward instantaneously.

[PCC #9]

[One of our future EI data system goals is to have] secure electronic data sharing to allow compliant referral sources to submit referrals easily.

[PCC #12]

This was also reflected in the diversity of methods by which families and other sources may refer children to EI. Many states with web form referrals also made a downloadable e-document available, and only two states explicitly refused e-document or fax-based referrals.

Relationships are strong in local communities, and we accept referrals at the state and local level in a variety of formats.

[PCC #33]

Any member of the public can access this database and make a referral for a child, or they are able to call/fax/in person a referral, which goes through the Central Referral Agency- who then will enter that into the online referral system.

[PCC #4]

Receiving referrals.

Eighteen state PCCs estimated that the most common ways they receive referrals were by telephone (mean 38%, SD 26), electronic not email (23%, SD 32), facsimile (17%, SD 22), and email (17%, SD 21). Many respondents indicated that they do not collect this information and declined to estimate. Figure 3 illustrates the variability of these proportions across states.

Figure 3.

Part C Coordinator Survey: Estimated % of EI referrals received by primary methods (n=18). Note: 23 declined to estimate, saying info not available.

While the majority of electronic- or email-based referrals (58%) were electronic documents such as an online fax or PDF attached to an email, 23% of states reporting electronic referrals did receive some sort of computable referral data about the child or about the referral itself. Some of these direct computer-to-computer referrals came from healthcare settings. While few PCCs (8%) reported direct Electronic Health Record-(EHR)-to-agency data connections, other PCCs expressed interest in it when asked “How would you make [your state]’s EI referral system better?”:

Integration of the online referral system to physicians EHR system would also be beneficial.

[PCC #11]

Not linked to EHRs yet. We are in the middle of a redesign. Ask me in 6 months!

[PCC #2]

Many PCCs valued a centralized referral intake system.

(What about your state’s EI referral system works well?) Central referral contact (phone/website) which distributes referrals to regional programs

[PCC #8]

(What would you change about your state’s referral system?) Develop centralized referral lines and locations to ensure better family engagement

[PCC #17]

Other open-ended responses indicated that referral intake was often decentralized.

Most referrals are received directly by the local EI program. Local EI programs receive them through; fax, phone call and mail. Local representative would talk to the family and complete a referral form. Once a referral is received and the child is scheduled for an evaluation all the information is then stored in a statewide electronic data system.

[PCC #21]

Some state EI agencies promoting a centralized referral process may meet resistance from local agencies. One PCC described their statewide data system as a pain point in relations with local agencies.

[Our state’s local programs] are inconsistent in their approach to referrals with some not wishing to engage with a statewide system.

[PCC #7]

Statewide EI Data System Capacity.

All responding states used one or more statewide EI data systems. Most of these (≥ 73%) contained demographic, unique ID, and government-reporting data, along with EI program data such as eligibility, enrollment, service plan (IFSP), and services provided. To exchange information with other data systems, most of these systems (85%) did not or could not use the Common Education Data Standards¹⁹ supported by the US Department of Education. Messaging standards such as HL7 and FHIR²⁰ used for health information exchange were mostly not available or used (92%).

DISCUSSION

There have been repeated calls for improved communication and information system integration centered around the medical home by the American Academy of Pediatrics^23,24 and others. As an example of this need, the referral of a child to EI carries unique challenges which may promote children’s loss to follow up and exacerbate health inequities. Many of these challenges are related to communication and access to information. To begin to address them, this study presents the current state of EI referrals by describing statewide referral forms used today and the child referral process from the perspective of state Early Intervention Part C Coordinators.

We found that states offered a broad array of means to refer children to EI, with variation both within and between states. For referral sources wishing to use electronic means, statewide referral forms were available online in many states. On the other hand, EI entry points remained with local agencies in many states, and this de-centralization complicates state-level implementation of referral information technology. As we expected, medical professionals were the largest source of referrals in most states and many referral forms seem to be designed for them. Yet, we also showed the importance of parents and families as referral sources.

Although studies and expert opinion support the importance of good communication and interagency collaboration for successful entry into EI,^24,25 few peer-reviewed studies to date evaluate specific EI referral transmission methods. Our results corroborate earlier findings in that while most EI agencies have statewide data systems holding demographic and program data about the children they serve, few state EI agencies report data linkages or exchanges with healthcare data systems – especially clinical electronic systems.^26,27

Part C Coordinators state that providing multiple routes of referral helps enable access to their programs, previously described as a “no wrong door” approach. Most of these methods (e.g. phone, fax, email, websites) require multiple steps of data transfer by hand, which carry opportunities for errors and inefficiencies. Although these problems may be mitigated with use of information technology, the human factor or “socio-technical” challenges of such solutions often prove more difficult than developing the software for them.²⁸ Sean Mikles writes that socio-technical challenges can be even larger for early childhood service coordination, because it requires interagency collaboration.²⁹ Challenges to successful collaboration may explain why we found that few state EI information systems use the well-established CEDS¹⁹ or FHIR²⁰ data and messaging standards, which could be important tools for offering e-referral connectivity statewide.

Our results expose specific opportunities for states to further widen their access pipeline for children entering EI, as ensuring access to their services is clearly part of their core mission. Agencies might aim to maximize parent engagement with public-facing, user-centered interfaces that take advantage of mobile technology. They might translate more referral forms to improve accessibility for parents with limited English proficiency. Automated translation—such as with 3^rd party functions on a web based form—may eventually be a viable solution; however, more study is needed to verify the quality of the translations and evaluate their acceptability, use rates, and effect on EI access inequities. Lastly, although statewide referral forms collect important data on children at a vulnerable step in the already-inequitable access pipeline to EI, only two-fifths of them request the child’s race. For this reason we suggest that race and ethnicity data be collected on every form and be used to evaluate equitable distribution of scarce program resources. State and local agencies might further facilitate “apples-to-apples” demographic comparisons by conforming with standard race and ethnicity formats such as those recommended by the US Office of Minority Health.³⁰

These results also support concrete means of addressing AAP calls for improved communication with the pediatric medical home.²³ For example, the AAP published a model EI Program Referral Form 14 years ago.²¹ Since then, we have seen much progress in internet access and functionality. An updated version of the AAP form should be developed as a template for adaptation by health organizations and state EI programs. In an electronic version of such a form template, additional fields beyond those in the three most common categories could be made optional or user-configurable to accommodate varying needs of individual states. Documenting consent within an electronic form template might include support for various methods, such as uploading a scanned document or password-based electronic authentication.

Strengths of this study include its national reach, relatively complete data collection from states, and its use of data from both surveys and referral forms. Limitations include its cross-sectional nature—which limits causal inference—and its sampling of organization representatives but not those working “in the trenches” of the EI referral process. Open-ended questions on a survey may limit the depth to which we can understand issues described in narrative responses without interviewing PCCs. In addition, three of the free-text questions were presented to respondents late in the survey, so their responses may be biased towards electronic methods by the prior questions. While not every state is represented in the survey, its high response rate and negative response bias analysis should minimize this limitation. Feasibility informed our decision to consider states as the primary unit of analysis, acknowledging that it limits the sample size while representing a national sample. Also, referral forms do not likely represent all of the information necessary to arrange a child’s eligibility evaluation. Lastly, since many of the participating state EI agencies do not regularly collect data on the way in which new referrals arrive, the referral method results in this study should be treated as estimates only.

CONCLUSION

Early Intervention can provide children the opportunity to achieve maximum academic and life success. However, too many families referred to EI in the US are still lost to follow up, denying tens of thousands of eligible children potentially life-changing services to which they are entitled. Each state has its own unique circumstances, but if information technology can improve referral completion in the United States, this study offers a place to start doing it. Additional research is needed to understand the perspectives of all participants in the EI referral process to co-create and evaluate solutions that engage families, promote communication, enable data exchange across institutions, and track children throughout the EI access pipeline. In these ways, more of our most vulnerable children can be connected to the services that can help them maximize their developmental potential and life success.