Elaboration of Screening Scales for Early Diagnosis of Developmental Delay in Four- to Five-Year-Old Children in Russia

Abstract

Background. While various screening systems are used worldwide for early detection of developmental delay (DD), Russia still does not have such a screening system in place, even though a good prognosis for the cognitive development of a child with DD depend strongly on the time of the diag nosis. Objective. The objective of this study was to create a system to rapidly monitor the mental development of four-to five-year-old Russian children, a system that allows for the use of modern information technologies to obtain reliable results. Design. This study was carried out with a sample of 1,232 children. For data collection, the multifactor study of mental development tool was used as a part of a software complex for longitudinal research. This tool included a much more extensive set of tasks than in traditional tests of abilities, allowing for a wider variation of the factor structure. For the 4-year-olds, 236 tasks were used and 349 for the 5-year-olds. Factor and discriminant analysis were carried out to construct scales for each age group (6-7 points in each), which most accurately predict the diagnosis (Norm/DD). Structural equation modeling (SEM) was used to verify the prediction model. Results. Two scales were elaborated, which coincided with the type of variables combined in each of them regardless of age (for 4-years-simpler and for 5-years-more complex): logical reasoning, motor skills, and general awareness (listed in descending order of contribution to the prediction). SEM confirmed that the selected scales are indicators of the general ability factor, which is the main predictor of the diagnosis. Conclusions. Two short scales for the rapid diagnosis of DD in preschool children were constructed, allowing the use of computer technology to timely identify the risk group among 4-and 5-year-olds with high sensitivity and specificity of the forecast (not lower than 95%).

Elaborating Screening Scales for Early Diagnosis of Developmental Delay in Five- to Six-Year-Old Children in Russia

“…Since the condition of multidimensional normality was not met (Multidimensional kurtosis ranged from 15.8 to 27.5; its C.R. from 22.7 to 30.8) in all cases, the Asymptotically Distribution Free method was used, as in our previous paper (Nasledov et al, 2018). All models confirmed the initial assumptions on the fit indices and the statistical significance of the parameters.…”

“…Data collection was carried out using the same software "Longitude" (Ivanova & Miroshnikov, 2001) as in our previous study (Nasledov, Miroshnikov, & Tkacheva, 2018). This software included a huge bank of tasks, presented in accordance with the child's calendar age, and aimed at estimating a wide range of abilities in primary domains such as motor skills, social adaptation, and cognitive abilities.…”

“…Various screening techniques have been successfully used in European countries for a long time (Frankenburg & Dodds, 1967;Griffiths, 1970). But there are no attempts for elaborate screening of child mental development in Russia today that meet modern requirements for psychometrics, except our previous work (Nasledov, Miroshnikov, & Tkacheva, 2018). The age slice being studied in scientific Russian publications begins with older preschoolers, and often DD is revealed only when they enter school, and this, unlike in most Western countries, is at the age of seven years, not five.…”

“…Statistical data analysis (as we did in our study of four-and five-year-olds (Nasledov et al, 2018)) was carried out for the following purposes: a) identification of a compact set of scales that predict the diagnosis most accurately (belonging to the group Norm or DD), and have sufficient reliability in relation to the different age ranges of five-to six-year-old children; b) interpretation of the relationships between predictors and the relative contribution of various scales in predicting the diagnosis; and c) development of statistical norms and an algorithm for rapid assessment of the probability of DD by applying the elaborated methodology. The analysis was carried out separately for the samples of five-and six-year-olds, in the same sequence, and using the same methods as in our previous study (Nasledov et al, 2018). All statistical analysis was performed using IBM SPSS software and AMOS version 25.…”

“…The main difficulty was to take into account the age, as during each period (in days), the measured indicators increase significantly. As in our previous study (Nasledov et al 2018), this problem was solved with the use of DA: for each age, a discriminative function was constructed, i.e., an axis passing through the centroids of the shared classes (Norm, DD). The linear equation of this function relates the discriminant scores for each child (DF) to the values of the previously selected scales (S1, ..., S5) and to age (in days).…”