Generalizability of trial results based on randomized versus nonrandomized allocation of OME infants to ventilation tubes or watchful waiting.

The objective was to study the generalizability of trial results by comparing randomized patients to eligible but nonrandomized patients who received the same management. Implementation of trial results is only justifiable when the results can be generalized to the total domain population. The design was a multicentre randomized controlled trial on the effect of early screening and treatment with ventilation tubes on infants with otitis media with effusion. Randomized (n = 187) and nonrandomized eligible patients (n = 133) were followed up. The study population comprised children who were detected by auditory screening at the age of 9-12 months and who were subsequently diagnosed with persistent bilateral otitis media with effusion for 4-6 months. A significant difference was found in the distribution of some prognostic factors: more randomized children had older siblings, did not attend day care and had mothers with a lower educational level than the nonrandomized children. These factors, however, did not modify the outcome. No differences were found in mean hearing levels between the randomized and nonrandomized children: in both the randomized and nonrandomized children ventilation tubes improved the hearing level, especially after 6 months. However, in the long term (12 months), the hearing levels were equal again. The results of the randomized and nonrandomized patients were comparable. The results of this trial appear to be generalizable to the total domain population. The procedure of following up both randomized and nonrandomized patients is recommended when there is concern about selective participation and reduced generalizability.