Discussion
Early identification of children at an increased risk of developmental delay allows for more frequent monitoring of developmental milestones, earlier referral for formal evaluation, and more timely receipt of EI services for those who need them. In this study of NYC births, we reported increasing EI referral rates as birthweight and gestational age decreased. Our results indicate that both birthweight and gestational age exerted independent effects on the risk of developmental delay. Based on our findings, a birthweight threshold of <1,250 g would identify children at highest risk, as the predicted risk of diagnosed developmental delay for children below this threshold is generally more than one in two. A gestational age threshold could be used as an additional criterion in conjunction with birthweight, such as infants born <30 weeks' gestation and <1,500 g, to prevent overlooking high-risk children born <30 weeks but at higher birthweights. Our findings also suggest that certain gestational age thresholds currently used by some states, such as ≤32 weeks, may not be cost-effective given the large majority of these infants (88.1% in our study) are born with a birthweight of 1,500 g or higher and thus at lower risk of delay. These findings, combined with the fact that both are quantifiable and recorded on both the birth certificate and pediatric medical records, makes birthweight and gestational age useful indicators for identifying high-risk children.
Published data from two other studies analyzing state-level birth certificate and EI data files support these findings. The Pregnancy to Early Life Longitudinal (PELL) Data Project analyzed data from over 200,000 children born in Massachusetts from 1998 through 2000. This study, which assessed birthweight and gestational age independently, reported that birthweight <1,200 g and gestational age <32 weeks were both significant predictors of referral to EI within 12 months of birth, with birthweight the stronger predictor. The PELL Project also reported that between 1998 and 2003, 93% of infants weighing <1,200 g at birth were referred to the Massachusetts EI Program. The 71% referral rate for NYC children weighing <1,250 g in this study was lower than that observed in Massachusetts. This difference may be partly explained by varying referral patterns, policies, and practices between states, including the level of outreach conducted by the EI programs and thus the knowledge physicians, parents, and caregivers had of EI. Additionally, statewide jurisdictions may have greater ability to capture referrals due to less out-migration compared with citywide jurisdictions. Greater comparability between outcomes was found with Florida's study linking 1998 birth certificates to EI data to assess the effect of birthweight on developmental disability, which concluded that children born <1,000 g, and perhaps <1,500 g, should be automatically referred to the EI program.
The NYC EI Program handled an average of 15,852 referrals per year of birth for resident children born in NYC from 1999 through 2001. Increasing NYC's automatic eligibility birthweight to 1,250 g from its current threshold of 1,000 g would have expanded the number of automatically eligible children by 420 children per year of birth; given that 270 of these children were already referred to the NYC Early Intervention Program, this would translate to a net maximum increase in children served by NYC's EI program of approximately 150 children per year of birth (assuming all were referred). A further change in criteria to include children <1,500 g and <30 weeks' gestation would have added approximately an additional 46 children per year of birth (87 of the 133 children identified were already referred to EI). One may argue that this relatively small increase in the number of children served warrants a revision of EI criteria, especially given their heightened likelihood of developing a significant delay before the age of 3. However, each jurisdiction's determinations of eligibility criteria must occur in conjunction with other considerations. The rate of preterm delivery and low birthweight is increasing steadily in the United States, and a recent assessment of EI costs by gestational age estimated that the mean cost per infant was seven times higher for children born 24–31 weeks' gestation compared with term infants; thus decisions on birthweight and/or gestational age thresholds are likely to have escalating consequences and costs in future years.
EI services have been shown to improve childhood outcomes for both children born with low birthweight and those born preterm. In our study, the age at referral into EI among children born <1,250 g who were ultimately eligible for EI services varied by gestational age, with later referral times for children with higher gestational ages. This suggests that some children who are ultimately eligible could have been identified and received services earlier. However, it is unclear whether the lags for children born at later gestational ages are clinically significant, whether they were due to lack of knowledge of the EI Program or these children truly did not need services until later, and whether earlier services would have improved outcomes. There still remains a need for formal assessments to ascertain whether enrolling children who are automatically eligible prior to being diagnosed with a developmental delay significantly improves future outcomes. The effect of timing of program services to EI outcomes is considered one of the most important but least understood aspects of EI. However, there is evidence of better outcomes among those enrolling in intervention at a younger age.
There are several potential limitations of this analysis. Presented percentages and predicted risks are likely underestimates because we could not identify: children who were diagnosed with delays but were never referred to the EI Program; children who had significant delays and were referred but not evaluated and diagnosed; and children born in NYC who moved and participated in an EI program elsewhere. Risks may have changed during our follow-up period for various reasons, including artifacts of increasing referral patterns. As in any jurisdiction, referral patterns are influenced by current policies; in NYC, children with birthweights <1,000 g were automatically eligibility for EI services. In addition, although the validity of birthweight is expected to be high given extensive follow-back of births <1,000 g, there may be inaccuracy in the estimate of clinical gestational age due to its variable methods of estimation; this may be evident in the children classified in the more "extreme" BW–GA categories (i.e., bottom left and top right corners of Table 1). Misclassification due any inaccuracy in GA reporting may have led to an underestimation of true predicted risk among children in the extremely preterm-normal birthweight category and an overestimation in those in the extremely low birthweight-term category.