1439 Dutch children were included in a randomised trial to evaluate the efficacy of preschool screening for otitis media with effusion (OME) by 3-monthly tympanometry. Children with bilateral OME on two consecutive occasions were referred for further investigation and then, if parents gave their consent, allocated at random to treatment or non-treatment groups. The effect of childhood screening for OME and subsequent treatment was evaluated by assessment of language performance; no benefit was found, mainly because of the large degree of spontaneous recovery.