How to deal with the drop-out in clinical follow-up studies: Results of a long-term follow-up study of orthodontically treated patients

The drop-out in a long-term follow-up study of former orthodontically treated patients was analyzed to avoid biased results. Since structural conformity of the responders (n = 299) and the potential study population (n = 1464) cannot be presumed on the basis of the clinical results, roentgenologic results, and questionnaires, analysis of the participating patients, the follow-up method was used. By means of questionnaires sent to all nonresponders, a comparison of responding and nonresponding patient characteristics was completed to find a presumed selectivity bias. The comparison of the participating patients (n = 299) and the total sample revealed that the responders live closer to the health center, had a longer treatment time, and a shorter posttreatment interval. Compared with nonresponders (n = 266), the responders were satisfied with the treatment result and had a higher dental Intelligence Quotient (IQ, concerning tooth and jaw position). Although a definite answer to the question of validity of the examined subsample was difficult to give, the description and evaluation of the parameters that might be reasons for missing at random respectively, not missing at random was recommended. (A M J O RTHOD D ENTOFAC O RTHOP 1995;108:415-20.)