Lysle E. Johnston

Aside from the common observation that functional appliances work, the literature has little to say about the means by which a correction is achieved and whether or not it differs from that produced by a single phase of fixed-appliance treatment. There seems to be relatively little interest in these basic questions. The purpose of this communication is to explore and discuss the action of functional appliances in light of their interaction with the gross mechanisms of craniofacial growth. It is hoped that a collegial discussion will ensue. Normally, the mandible outgrows the midface, even in Class II patients. This favorable pattern of growth, however, produces maxillary dentoalveolar compensations that prevent the excess mandibular growth from having any marked effect on the malocclusion. Functional appliances may avoid this mesial movement of the maxillary dentition by dissociating mandibular displacement from the condylar growth that normally accompanies it. By producing an anterior functional shift that corrects the Class II relationships all at once and by maintaining this advanced position long enough for the condyles to grow back into the fossae, functional appliances can make clever use of the normal pattern. No "extra" growth may be necessary. Relative to untreated controls, their ultimate effect is in the midface, rather than the mandible. Moreover, the final correction may not differ significantly from that produced by fixed appliances or by two-phase treatments in which the first phase is some sort of extra-oral traction to the midface. Functional appliances may utilize the normal pattern of facial growth, rather than some sort of mandibular "growth modification," to achieve a molar correction. As a result, their ultimate effect may not differ from that produced by the various one- and two-stage alternatives. The choice of which to use may well prove to be a simple practice-management decision.

I have been asked to comment on ‘how to do a retrospective study.’ A truly authoritative response to this charge is perhaps best sought in one or more reputable epidemiology texts. The present comm...

Events of the past decade or so argue that there is little support in the literature for much of the specialty's treatment portfolio. The resulting call for ‘evidence‐based’ treatment (not to mention the obvious intellectual bankruptcy of much of the clinical literature) has prompted many to argue that the randomized clinical trial (RCT) represents the future of orthodontic clinical investigation. The RCT, after all, is medicine's gold standard; what more is there to say? A popular, but ultimately divisive, corollary of this mimicry is the smug tendency to discount all other sources of data. In the face of a need for information, this attitude is also a wasteful conceit: in the end, the RCT can be applied only to a very narrow spectrum of orthodontic questions. Randomization implies equal susceptibility. Any prospective participant would have to be informed of this equality as part of the informed consent process. Unfortunately, it would be nearly impossible to enroll fully‐informed subjects into any study whose alternatives are of markedly different morbidity: extraction versus non‐extraction or orthodontics versus surgery. Thus, when measured against the most vexing clinical questions, the orthodontic RCT is almost by definition an amusing diversion ‐ expensive, but relatively trivial in scope. Like it or not, it seems reasonable to conclude that most of the specialty's comparative clinical data will have to be generated by way of non‐randomized designs in which care is taken to minimize the various known sources of bias. There probably is no other way.

Because of a perceived need to examine problems more pressing than ‘now versus later’ or ‘headgear versus functionals’, a retrospective/prospective alternative to the randomized clinical trial has been developed. In this approach, discriminant analysis is used to identify samples of ex‐patients who, prior to treatment, were equally susceptible to various contemporary treatment alternatives: extraction, non‐extraction; orthodontics, surgery; fixed versus functional. This method of ‘confounder summarization’ has made it possible for us to conduct a wide variety of long‐term comparisons and, in the process, to assess treatment efficiency, stability, esthetics, functional impact, and interaction with the pattern of facial growth. These studies are described in broad outline and serve as examples of the type of data that can be generated in a timely fashion by carefully planned, quasi‐experimental investigations.