Unbiased quantitative testing of conventional orthodontic beliefs

This study used a preexisting database to test in hypothesis from the appropriateness of some common orthodontic beliefs concerning upper first molar displacement and changes in facial morphology associated with conventional full bonded/banded treatment in growing subjects. In an initial pass, the author used data from a stratified random sample of 48 subjects drawn retrospectively from the practice of a single, experienced orthodontist. This sample consisted of 4 subgroups of 12 subjects each: Class I nonextraction, Class I extraction, Class II nonextraction, and Class II extraction. The findings indicate that, relative to the facial profile, chin point did not, on average, displace anteriorly during treatment, either overall or in any subgroup. Relative to the facial profile, Point A became significantly less prominent during treatment, both overall and in each subgroup. The best estimate of the mean displacement of the upper molar cusp relative to superimposition on Anterior Cranial Base was in the mesial direction in each of the four subgroups. In only one extraction subject out of 24 did the cusp appear to be displaced distally. Mesial molar cusp displacement was significantly greater in the Class II extraction subgroup than in the Class II nonextraction subgroup. Relative to superimposition on anatomical "best fit" of maxillary structures, the findings for molar cusp displacement were similar, but even more dramatic. Mean mesial migration was highly significant in both the Class II nonextraction and Class II extraction subgroups. In no subject in the entire sample was distal displacement noted relative to this superimposition. Mean increase in anterior Total Face Height was significantly greater in the Class II extraction subgroup than in the Class II nonextraction subgroup. (This finding was contrary to the author's original expectation.) The generalizability of the findings from the initial pass to other treated growing subjects was then assessed by retesting modified hypotheses against a second database stored sample that earlier had been drawn randomly from two other orthodontic practices. The implications of the author's study strategy to the design of future shared digital databases is discussed briefly.