The association between the Ages and Stages Questionnaire 3 assessment at age 2 and the Early Years Foundation Stage at age 5: A longitudinal observational study using routine data

This longitudinal study of 47,046 children in England reveals that achieving a Good Level of Development on the Ages and Stages Questionnaire-3 at age two is a strong predictor of similar success on the Early Years Foundation Stage Profile at age five, highlighting the critical need for early intervention to address developmental inequities among boys, children in deprived areas, and specific ethnic groups.

The Gist

Imagine a child's development as a long journey toward a big finish line called "School Readiness." In England, when children turn five and finish their first year of school, they take a big test called the EYFSP (Early Years Foundation Stage Profile). If they pass this test with flying colors, they get a "Good Level of Development" (GLD) badge. This badge is a strong hint that they will do well in school later on.

For a long time, experts knew that some kids struggled to get this badge. But they didn't have a good way to spot who was going to struggle before they even started school.

This study is like a time-traveling detective story. The researchers wanted to see if a smaller, earlier check-up at age two (called the ASQ-3) could predict who would get the big "GLD" badge at age five.

The Setup: Two Different Check-Ups

Think of the two tests as different checkpoints on a hiking trail:

1. The Age 2 Checkpoint (ASQ-3): This is a "parent report." It's like a parent filling out a checklist about their toddler's skills (can they stack blocks? do they say words?). It happens when the child is about two years old.
2. The Age 5 Checkpoint (EYFSP): This is a "teacher report." It happens when the child starts school. The teacher watches the child and grades them on reading, math, social skills, and more.

The Big Discovery: The Crystal Ball Effect

The researchers looked at data from nearly 50,000 children in Bradford, a city in the north of England. They linked the records of the kids who took the age-2 test with the records of the kids who took the age-5 test.

Here is the magic finding:
If a child got a "Good Level" on the age-2 test, they were more than three times more likely to get the "Good Level" badge at age five.

Think of the age-2 test as a weather forecast. If the forecast says "sunny" (good development) at age two, there's a very high chance the weather will still be sunny at age five. If the forecast says "stormy" (developmental concerns) at age two, there's a high chance the storm will still be there when they start school.

Who Struggles the Most?

The study also looked at who was having a harder time on the trail. Just like in real life, some groups faced more headwinds than others:

• Boys: They were less likely to get the "Good Level" badge at both ages compared to girls.
• Poverty: Children living in the poorest neighborhoods (the bottom 20% of the map) were much less likely to pass the tests. It's like trying to run a race with heavy backpacks while others run with light ones.
• Ethnicity: The results were a bit mixed here. At age two, children from South Asian backgrounds struggled more. But at age five, children from "White Other" backgrounds (like Eastern Europeans) struggled more. The researchers suggest this might be because the age-2 test is filled out by parents (who might speak different languages or have different cultural views on development), while the age-5 test is done by teachers who focus heavily on English skills.

Why Does This Matter? (The "Aha!" Moment)

The most important part of this paper is the call to action.

For a long time, the age-2 test was just used to get a general "vibe" of how kids in a city were doing. It wasn't used to help individual kids. But this study says: "Wait a minute! This test is actually a crystal ball!"

If we use the age-2 test to spot kids who are "stormy" (struggling), we can send them a rescue team before they even start school.

• Current situation: We wait until age five, see the child is struggling, and then try to catch up.
• Proposed solution: Spot the struggle at age two, give them extra help immediately, and help them arrive at school ready to learn.

The Catch

There are two big hurdles to making this happen:

1. Missing the Bus: About 1 in 4 families in England never even get the age-2 check-up. If we don't get these kids into the system, we can't help them.
2. The Gap: We need to make sure the "rescue teams" (doctors, teachers, and support workers) can talk to each other. Right now, the health system and the school system often don't share information. This study says we need to build a bridge between them so no child falls through the cracks.

The Bottom Line

This study tells us that many children are already falling behind by the time they are two years old. We can't wait until they are five to fix it. By using the age-2 check-up as an early warning system, we can give every child, especially those from poor backgrounds or boys, a fairer shot at a successful start in life. It's about fixing the foundation before we build the house.

Technical Summary

1. Problem Statement and Context

• The Policy Challenge: In England, a national priority is to increase the proportion of children achieving a "Good Level of Development" (GLD) on the Early Years Foundation Stage Profile (EYFSP) by age five and to reduce equity gaps in this outcome. Currently, approximately 30% of children do not reach a GLD.
• The Knowledge Gap: While the EYFSP (assessed at age 5) is a well-validated predictor of later educational attainment, the predictive utility of earlier developmental assessments is less understood. Specifically, no prior study had explored the longitudinal association between the Ages and Stages Questionnaire-3 (ASQ-3) (administered at age 2) and the EYFSP using individual-level linked data.
• Data Barrier: Historically, a key barrier has been the lack of linkage between health data (0-5 Healthy Child Programme) and education data due to different identifiers (NHS numbers vs. Unique Pupil Numbers).
• Objective: To determine if achieving a GLD on the ASQ-3 at age 2 predicts achieving a GLD on the EYFSP at age 5, and to identify sociodemographic predictors (sex, ethnicity, deprivation) for both assessments.

2. Methodology

• Study Design: Longitudinal observational study using routine pseudonymized data.
• Data Source: The Connected Bradford dataset, which links health, education, social care, and environmental data at the individual level for over 600,000 individuals in Bradford, UK.
• Study Population:
  • Eligibility: Children born between Sept 2013 and May 2025 registered in Bradford.
  • ASQ-3 Cohort: 47,046 children with ASQ-3 records (completed between 24–30 months).
  • EYFSP Cohort: 9,469 children with EYFSP records (end of Reception year, age 5).
  • Linked Cohort: 6,021 children (13% of ASQ-3 cohort) with both ASQ-3 and EYFSP records. Note: The low linkage rate is attributed to education data only being available up to the 2018/19 academic year.
• Variables:
  • Outcomes: Binary GLD status (Achieved/Not Achieved) and continuous scores for both ASQ-3 and EYFSP.
    • ASQ-3 GLD: Defined as all five domains (communication, gross/fine motor, personal-social, problem solving) being "No Risk" or "Monitor."
    • EYFSP GLD: Defined as meeting expected levels in five core areas (communication, physical, personal/social, literacy, math).
  • Covariates: Sex, Ethnicity (2011 Census categories), Index of Multiple Deprivation (IMD) quintiles, and age at assessment.
• Statistical Analysis:
  • Multiple logistic regression models to estimate odds ratios (OR) for achieving GLD.
  • Linear regression for sensitivity analysis using continuous scores.
  • Interaction terms tested to see if the ASQ-3 to EYFSP association varies by sex, ethnicity, or deprivation.
  • Complete-case analysis was used (no imputation).

3. Key Results

• Descriptive Statistics:
  • ASQ-3: 83.3% achieved GLD. Boys, children in the lowest IMD quintiles, and South Asian ethnic groups were less likely to achieve GLD.
  • EYFSP: 67.7% achieved GLD. Boys, children in the lowest IMD quintiles, and "White Other" ethnic groups were less likely to achieve GLD.
• Predictive Association (Primary Finding):
  • Achieving a GLD on the ASQ-3 at age 2 is a strong predictor of achieving a GLD on the EYFSP at age 5.
  • Odds Ratio: Children who achieved GLD on ASQ-3 had 3.18 times higher odds (95% CI: 2.70–3.75) of achieving GLD on EYFSP compared to those who did not, after adjusting for sex, ethnicity, age, and deprivation.
  • Continuous Scores: A one-unit increase in ASQ-3 score was associated with a 1.7-point increase in EYFSP total score.
• Interaction Effects:
  • Ethnicity: The strength of the association varied significantly. It was strongest for Black/African/Caribbean/Black British children and weakest for "White Other" children.
  • Deprivation: The association was stronger for children in the lowest IMD quintile compared to the highest.
  • Sex: The predictive relationship did not significantly differ by sex.

4. Key Contributions

• First Longitudinal Link: This is the first study to use linked, individual-level longitudinal data to quantify the predictive validity of the ASQ-3 (age 2) for the EYFSP (age 5).
• Evidence for Early Screening: It provides empirical evidence that the ASQ-3, currently used primarily for population monitoring in England, has significant utility as an individual screening tool to identify children at risk of poor school readiness.
• Equity Insights: The study highlights that developmental disparities are already evident by age 2, with specific ethnic and deprivation groups showing distinct patterns of risk and predictive strength.
• Methodological Innovation: Successfully bridged the data gap between the NHS (health) and Department for Education (school) datasets using the Connected Bradford infrastructure.

5. Significance and Implications

• Policy and Practice: The findings suggest that relying solely on age-5 assessments is too late for intervention. The ASQ-3 results should be utilized to trigger early interventions (before age 5) to improve school readiness.
• Targeted Support: Since many children have already fallen behind by age 2, interventions must focus on the "first 1001 days" (pregnancy to age 3) to address inequities effectively.
• Systemic Changes:
  • Screening: Re-evaluating the ASQ-3 as a clinical screening tool rather than just a population metric.
  • Information Sharing: The study underscores the need for standardized information sharing between health visitors, family hubs, early education settings, and schools to ensure at-risk children are identified and supported continuously.
• Limitations & Future Research:
  • The study is observational; causality cannot be definitively proven.
  • Linkage rates were low (13%) due to data availability constraints, though the authors argue the association is likely consistent across unlinked years.
  • Differences in assessment methods (parent-reported ASQ-3 vs. teacher-assessed EYFSP) and potential cultural/language biases in the tools require further investigation.

Conclusion: The study confirms a robust link between developmental status at age 2 and school readiness at age 5. It argues for a paradigm shift toward earlier, data-driven interventions to close the equity gap in child development before children enter formal schooling.