Value of clinical assessment in the diagnostic evaluation of Global Developmental Delay (GDD) using a Likelihood Ratio Model

Abstract

Objective: A selective approach is recommended for investigating children with GDD. Our objective is to identify clinical markers to improve the diagnostic yield of evaluation of children with GDD. Method: Children with GDD (delay > 2 S.D. in > 1 domain) followed up in our centre were reviewed retrospectively. We selected nine clinical markers (sex, severity of GDD, parental consanguinity, family history, behavioral problems, head size, facial dysmorphism, non-facial anomalies and neurological deficits) and looked into the likelihood of finding an underlying etiology during follow-up. Results: There were 577 children with 63%, 33% and 4% having mild, moderate and severe grade GDD. An identifiable etiology is detected in 53%. Genetic disease (25%) was the commonest cause identified. We have found that severity of GDD (severe and moderate versus mild grade [LR+ = 1.92 (95% C.I. = 1.49-2.48); LR- = 0.72(0.64-0.81)], behavioral problems [LR+ = 0.24 (95% C.I. = 0.17-0.34); LR- = 1.67 (1.48-1.88)], facial dysmorphism [LR+ = 2.66 (95% C.I. = 1.10-3.54); LR- = 0.65 (0.58-0.73)] and neurological deficits [LR+ = 2.85 (95% C.I. = 2.32-3.50); LR- = 0.31(0.25-0.39)] were clinical markers associated with increased chance of identifying an underlying etiology by multivariate analysis. Conclusion: These four clinical markers are useful in selecting patients with GDD for further diagnostic tests. Using the LR model, clinical markers in the first clinical evaluation of any child with GDD can potentially improve the etiological yield using targeted investigations. © 2010 The Japanese Society of Child Neurology.