In a prospective study of premature neonates of birthweight <1700gm, a) we validated a 12 point neurodevelopmental assessment score (NDAS) assigned at 3, 6, 9 and 12 months corrected age, against a Griffiths developmental quotient (DQ) at 12 months corrected age; b) we assessed the value of early physiotherapy intervention during the first year. According to the NDAS at 3 mths, infants were classified as normal (score 0-4), at risk (5-9) or probably impaired (10-12). Normal (N) and at risk (R) infants were randomised into intervention (+) and control (−) groups. All infants with scores of 10-12 were treated. N+ and R+ infants received physical therapy as outpatients and at home. A physical therapist assigned the NDAS at 3, 6, 9 and 12 mths; the Griffiths test was performed at 1 year by a psychologist blinded for infant group. Fifty one infants have been tested (28N, 23R). Infants classified as R according to the NDAS had a mean DQ of 95.9±9.9 at 1 yr vs 102.6±8.7 for N infants (p<.02). Intervention had no effect in N or R infants: N+ and N− had similar NDAS's of 2; 2; 1; 0 at 3, 6, 9 and 12 mths resp. Similarly R+ and R− had similar scores (6; 3; 2; 1 vs 6; 4; 2; 1 at 3, 6, 9, 12 mths). Our NDAS appears to predict poorer outcome as assessed by the Griffiths DQ at 1 year, while early intervention has no apparent effect on development over the same time period.