OBJECTIVE This study set out to determine the accuracy with which tone pip ABR and click ABR, carried out in babies referred from universal newborn hearing screening, is able to predict the hearing outcome as determined by follow-up hearing tests. METHODS The cohort of babies studied were all babies referred for hearing assessment from the universal newborn hearing screen in Sheffield, UK for the period January 2002 to September 2007, who were found to have a significant hearing impairment. METHODS The results of hearing assessment following referral from the newborn hearing screen were collected together with those of follow-up tests carried out up to an age when behavioural testing had established ear- and frequency-specific thresholds at 0.5, 1, 2 and 4 kHz. RESULTS The standard deviation of the difference between the follow up and the tone pip ABR thresholds was 10.5 dB for the 4-kHz tone pip, 16.8 dB for the 1-kHz tone pip, and ranged between 21.7 and 24.7 dB for click ABR. CONCLUSIONS The results of the study show that tone pip ABR following referral from newborn hearing screening has a similar accuracy to that reported in older subjects, and is a much better predictor compared to click ABR.