PREDICTIVE VALUE OF THE GRIFFITHS ASSESSMENT IN EXTREMELY LOW-BIRTH-WEIGHT INFANTS

Objective: To assess the relationship between the Griffiths Mental Dev elopment Scales at 1 and 3 years and the Stanford-Binet Intelligence S cale (S-B) and Beery Test of Visual-Motor Integration (VMI) at 5 years in extremely low birthweight (ELBW) children. Methodology: Prospectiv e study of 45 ELBW infants, without severe neurosensory impairment, ca red for in a single Level III neonatal intensive care unit. Results: A t 5 years, 36 (80%) children were of average intelligence, 8 (18%) had borderline intelligence and one was mentally retarded. The Griffiths general quotient (GQ) at 1 year had a weak correlation with the 5 year IQ (corr. coeff. = 0.47), with only 17% of children with a GQ < -1 s. d. at 1 year receiving an IQ < -1 s.d. at 5 years. In contrast, the Gr iffiths GQ at 3 years correlated strongly with 5 year IQ (corr. coeff. = 0.78). Among those children with a 3 year GQ < -1 s.d., 67% had a 5 year IQ < -1 s.d. and all had a 5 year IQ < 89. The 3 year hearing an d speech subscale correlated strongly with the 5 year S-B verbal compr ehension factor (corr. coeff = 0.753) and the 3 year combined eye/hand co-ordination/performance quotient had a moderate correlation with th e S-B non-verbal reasoning factor (corr coeff = 0.597) and with the Be ery VMI (corr coeff = 0.49). Conclusions: The 3 year Griffiths GQ is a good predictor of 5 year S-B IQ in ELBW children and can be used to i dentify children who may benefit from intervention prior to school ent ry.