Instability of delay classification and determination of early intervention eligibility in the first two years of life

Highlights

• We assessed development longitudinally in low- and high-risk infants using the Bayley-III from 3 to 24 months.

• It was common for Bayley-III delay classifications to switch multiple times across the first two years.

• The Bayley-III was effective at identifying typical development but poor at identifying infants with delays.

• Clinical opinion and risk factors should be valued above standardized assessments for identifying early delays.

Abstract

The purpose of this study was to determine the effectiveness of the Bayley Scales of Infant Development, Third Edition (Bayley-III) to track development and classify delays in low- and high-risk infants across the first two years of life. We assessed cognitive, language, and motor development in 24 low-risk full-term and 30 high-risk preterm infants via seven assessments performed between 3 and 24 months corrected age. The Bayley-III resulted in highly unstable delay classifications, low sensitivities, and poor positive predictive values across time. The results highlight that early intervention professionals, researchers, and policy makers should: (1) emphasize clinical opinion and prevalence of risk factors rather than standardized assessment findings when classifying delays and determining eligibility for services, and (2) develop more effective developmental assessments for infants and young children.

Introduction

The purpose of this study was to determine the effectiveness of repeated measurements using a common standardized assessment tool to track development in low- and high-risk infants in the first two years of life. This type of testing mirrors the standard of practice in the U.S. for assessing development when a potential delay has been noted or when high-risk infants, such as those born preterm or with brain injuries, receive developmental monitoring (Disabilities, 2006, Romanczyk et al., 2005). Each state has its own criteria for what defines “delay” and eligibility for early intervention services hinges on these criteria. In some cases, determination of eligibility is clear. For instance, if an infant is born with a diagnosis, such as Down syndrome, associated with future delays, the infant would qualify for services in most states. In other cases where outcomes are more heterogeneous so there is risk but not a certainty of future delay, it is more difficult to determine if an infant or child should receive services (Aarnoudse-Moens, Weisglas-Kuperus, van Goudoever, & Oosterlaan, 2009). This is the case for the growing population of infants born preterm.

Infants born preterm are more likely to have future cognitive, social, communication, and motor delays (Jongbloed-Pereboom et al., 2012, Taylor et al., 2011). Therefore, their development is typically closely monitored in the first years of life through follow-up clinics and home visits provided at regular intervals, such as every 3–6 months (Janssen et al., 2011). The standard of practice at these visits is to assess development using a standardized assessment tool, such as the Bayley Scales of Infant Development, Third Edition (Bayley-III). The Bayley-III is a commonly selected tool for classification of developmental delay because it assesses development across multiple domains, is normed on young populations (1–42 months), has a long history of use in clinical and research settings, and is often considered a “gold standard” for early developmental assessment (Bayley, 2006b, Vincer et al., 2005).

The Bayley-III is a useful tool for classification of early developmental delay because it allows professionals to standardize an individual's score to quantify degree of delay. This is important because if a child does not have a diagnosis clearly associated with future delays, early intervention eligibility is typically determined by his/her percentage of delay or the number of standard deviations his/her score lies below the mean. Therefore, the policy creates a demand for professionals to quantify developmental delay using specific measures. For all of these reasons, the Bayley-III was selected as the standardized assessment tool to track development in this study. Furthermore, the criterion for classification of delay was selected to be developmental performance more than 1.5 standard deviations below the mean because this is a common criterion in many states, such as New Mexico and Pennsylvania (New Mexico Administrative Code http://nmhealth.org/ddsd/nmfit/Documents/documents/7.30.8%20NMAC%2006-29-12%20FINAL.pdf; The Pennsylvania Code http://www.pacode.com/secure/data/055/chapter4226/s4226.22.html).

The two-year follow-up schedule, assessment tool, and delay criterion used in this study modeled the standard of practice and allowed us to test whether the Bayley-III is a valid and reliable means for the critical determination of a young child's early intervention eligibility. Although the newest version of the Bayley is one of the assessments most commonly used in practice and research, this is the first study to our knowledge assessing development longitudinally with the Bayley-III in a cohort of low- and high-risk infants (Johnson & Marlow, 2006). Previous research has identified problems when using other early norm-referenced assessment tools, including the second version of the Bayley, with young populations at risk for delays. Some of these problems include low sensitivity, poor positive predictive value, large discrepancies in scores between two proximal testing points, and instability of delay classification across time (Hess et al., 2004, Horner, 1988, McGrath et al., 2004, Vohr et al., 2012). Similar longitudinal studies have not been performed with the Bayley-III. Therefore, questions remain about the validity, reliability, and utility of standardized assessment tools, and specifically the commonly used Bayley-III, for young populations.