Reliability of video taped interviews with children suspected of being sexually abused.

OBJECTIVE--To determine the reliability of judgments about the likelihood of child sexual abuse based only on video recorded interviews. DESIGN--Blinded rating of likelihood of abuse by seven professional groups and comparison with consensus rating. SETTING--Child and adolescent psychiatry centre. SUBJECTS--Four people from each of seven professional disciplines: specialist psychiatrists, general psychiatrists, experimental psychologists, trainee social workers, trainee clinical psychologists, lawyers, and police. MAIN OUTCOME MEASURE--Rating of 12 recorded interviews. RESULTS--Agreement between the consensus panel and professional groups was 83% (151/183) for high likelihood cases (seven cases) and 89% (118/132) for low likelihood cases (five). Specialist psychiatrists and the police were better able to identify high likelihood cases than were other groups with less experience of interviewing sexually abused children (91% (48/53) v 79% (102/129); p = 0.05). CONCLUSIONS--Raters could accurately distinguish children with low likelihood of abuse on interview evidence alone, but those with more experience of dealing with sexual abuse were better at identifying high likelihood cases.