Comparison of performance of Sri Lankan and US children on cognitive and motor scales of the Bayley scales of infant development

Early intervention in children at risk for developmental delays can lead to better outcomes and improved child well-being. The prevalence of developmental disabilities in Sri Lanka has been estimated to be 12 – 29% and the majority of children with clinically identifiable developmental problems in Sri Lanka are referred to child mental health services relatively late, generally when they are over 5 years of age, due to low rates of early recognition [1]. Currently, a culturally validated Denver Development Screening Test (DDST) is being used in the Maternal and Child Health programme of the Ministry of Health, Sri Lanka to screen child development. The Denver II has been shown to have limited specificity and high over-referral rates [2] and has been considered as test of questionable value in terms of screening for developmental delay [3]. Furthermore, it is only a screening tool and is not useful for diagnostic purposes. A reliable and valid tool for diagnosing developmental delays in very young children would enable the early identification of children with impairments. The Bayley Scales of infant development is a commonly used psychometric tool for assessing the development of children between 1 to 42 months of age. It has been shown to be a valid diagnostic tool for identifying children with developmental delays at an early age and is widely used in clinical settings due to its solid theoretical background and robust psychometric properties [4]. The Bayley scales have been used in various countries such as Brazil [5], Taiwan [4] and Australia [6]; currently it is in its third edition [7]. The Bayley III scales can be used to assess infant and toddler development across five key developmental domains - cognition, language, social-emotional, motor and adaptive behaviour [8]. The interpretation of Bayley scores are based on norms established for children in the United States (US), which is based on a national standardized sample of 1700 children between 1 and 42 months of age, and divided into 17 age groups.

Despite its wide use and well-established psychometric properties, the Bayley III scale has not been validated in Sri Lanka. Socio-cultural influences can occur early in life and affect cognitive, sensorimotor and socio-emotional domains of behaviour [9] which may result in differences in the cross cultural application of the Bayley Scales for the assessment of infants. For example, Taiwanese children had lower scores on the Bayley cognitive and motor scales when compared to the US norms at six and 24 months [4] and Brazilian children aged between one and 12 months had lower motor scores [5]. These differences may be due to heredity, differences in socio-economic status and child-rearing practices [4]. Hence, the use of US norms to assess the development of children from other cultures may not be appropriate; it is important to establish normative data for Sri Lanka children for the instrument to be used as a diagnostic tool in the clinical setting.

This manuscript compares the performance of Sri Lankan children 6, 12 and 24 months of age on the cognitive and motor scales of the Bayley III with that of US children.

Male and female children performed equally well on the Bayley scales; there was also no difference in performance between children with and without siblings (p >0.05).

The test-retest interval ranged from 5 to 17 days, with a mean test-retest interval of 12.6 days. The intra class correlation (1, 1) coefficients for the cognitive, fine-motor and gross-motor scales were 0.897, 0.914 and 0.905, respectively.

The objective of this study was to assess the appropriateness of using the Bayley III among Sri Lankan children by comparing the cognitive and motor scores of Sri Lankan children aged 6, 12 and 24 months with that of US children.

There were some differences in the performance of Sri Lankan and US children on the cognitive and motor scales of the Bayley III. While there were no differences between Sri Lankan and US children at six months in either the cognitive or the motor domains, the cognitive scores of 12 month old SL children were significantly higher than their US counterparts (even after correcting for multiple hypotheses testing using the Bonferoni correction), but their gross motor scores were lower. At 24 months, Sri Lankan children had lower cognitive subtest scores than US children, but there were no differences in the motor scores. There were no differences in any of the subtests at any of the ages between male and female children.

The differences between Sri Lankan and US children in their gross motor scores at 12 months may be partially explained by differences in child rearing practices. It was observed that during the testing procedure that the mothers of 12 month olds did not encourage their children’s attempts to stand and walk, even with support, for fear of them falling. Sri Lankan parents, particularly those from rural areas, are reported to encourage clinging behavior in children as this keeps the child close to them, and hence safe, and actively discourage exploratory behavior [10]. Several items of the gross motor scale of Bayley III require children standing up with support, standing alone, etc. which may be the reason for SL children performing poorly on these items at this age. There were no differences between Sri Lankan and US children in the motor domain at 24 months when children had gained more control of their movements and mothers may have been less fearful.

A previous study among Asian children had reported that first birth order and female gender was associated with a child’s mental and motor scores on Bayley II [4]. In this study, there were no significant differences in the cognitive or motor scores between male and female children, or between children who had siblings and who did not have siblings. Socio-demographic factors such as father’s occupation and maternal education level were shown in a previous study to be associated with infant’s mental development [4]. The small sample size in this study did not permit meaningful comparisons between different groups; future studies should be designed to compare the performance of children from different socio-economic groups to identify predictors of infant development in Sri Lanka.

It has been observed that the length of the test can affect the performance of children, particularly at older ages. The recommended time for administration of Bayley III to children 13 months or older is 90 minutes. In this study, even though only three out of the seven subscales of the Bayley III were administered, in some instances, with 24 month olds, the test took about 90 minutes to administer. Some children got bored and lost attention or became restless and non-compliant. In order to maintain children’s interest and obtain their optimal performance, test administrators would need to develop strategies to give children breaks in between the administration of the different subscales without unduly lengthening the total administration time. It is important for test administrators to have a firm understanding and experience in testing young children [8]. Test familiarity is a key factor that results in a longer or shorter testing time. If this test is to be used in a clinical setting, test administrators, whether they are psychologists or physicians, need to be rigorously trained in the use of the test. Considering the length of the test, and the training required to administer and interpret the scores, it is not feasible for general physicians to administer this test routinely in their clinical practice.

Although there are no established standards for minimum acceptable levels of different statistics of reliability and agreement, the intra class correlation coefficients exceeded the recommended minimum value of 0.75 [11] for each of the subtests tested, indicating the test-retest stability of the Bayley III in the Sri Lankan setting.

While this study has provided important preliminary information about Sri Lankan children’s cognitive and motor development compared to US children, there are clear limitations in generalizing the findings to all Sri Lankan children. The sample was limited to a small geographic area in Sri Lanka; even though children from different ethnic and socio-economic groups were included, all the different subgroups in Sri Lanka may not have been represented. We also did not adjust for multiple hypothesis testing.

There were some differences between the performance of SL and US children on the motor and cognitive scales of the Bayley III, but the differences showed no consistent age-related pattern. It can be concluded that it is feasible to use the Bayley III to assess neurodevelopment in SL children. We recommend that the Bayley III be validated for different age bands, using a larger sample of children from different geographic regions and ethnic backgrounds in Sri Lanka for it to be used as a diagnostic tool in Sri Lanka.