Developmental followup of NICU graduates has traditionally involved direct clinical assessment. However, issues such as cost containment or geographic distance may preclude such followup. This study compares telephone parental report of development to outpatient clinical assessment in a population of NICU graduates at increased risk of developmental disability. Twenty eight NICU graduates were assessed by parental telephone report, using the Infant Development Inventory (IDI) and by clinic assessment, with the Capute Scales(CAT/CLAMS; visual motor/language) and Peabody Gross Motor Scale (PGM), for 33 total visits. Infants had a mean birthweight of 1514 grams (579-3311 gms) and mean gestation age of 31.8 weeks (25-39 wks). Followup ranged from 1 month to 10 months adjusted age (mean 4.3 mos). Developmental ages (DA) and quotients(DQ) were derived for each subscale of the IDI, the CAT/CLAMS, and the PGM, with correlations performed. DQ results were ranked categorically into normal(≥85), borderline (70-84), and abnormal (<70), with chi-square analysis performed. The IDI subscale DAs consistently correlated well with the CAT DA(range.85-.92), the CLAMS DA (.80-.92), and the PGM DA (.77-.90). Modest correlations were seen between subscale DQs measuring similar skills (gross motor: IDI vs. PGM, 50; fine motor: IDI vs. CAT,.39; language: IDI vs. CLAMS,.37). Categorical performance revealed significant associations between the IDI gross motor scale and the PGM (χ2 = 9.8, df4, p =.04) and IDI fine motor and CAT (χ2 = 11.8, df4, p =.02). The IDI language scale ranking was not significantly associated with the CLAMS or the combined CAT/CLAMS. Developmental surveillance of the early development of NICU graduates by parental telephone report using the IDI appears reliable using developmental age. Categorically-ranked developmental quotients may be useful for early monitoring of motor development; early language DQ ranking using the IDI less accurately predicted CLAMS performance. Other telephone assessment tools should be compared to determine optimal sensitivity for early identification of developmental disabilities. Accuracy at later ages must also be determined.