Quality improvement of functional diagnostics in dentistry through computer-aided diagnosis: a randomized controlled trial
Becker, K.; Jakstat, H.A.; Ahlers, M.O.
International Journal of Computerized Dentistry 21(4): 281-294
2018
ISSN/ISBN: 1463-4201 PMID: 30539170 Document Number: 695639
Functional diagnostic examinations such as clinical functional analysis and manual structural analysis ('orthopedic tests') allow the dentist to establish a structured diagnosis. Previously, the process of correlating findings with the appropriate diagnoses was guided by human thought processes alone. The experimental diagnostic randomized controlled trial (RCT) in this study investigated whether computer-aided diagnosis (CADx) of temporomandibular disorders (TMD) offers quality advantages over traditional diagnosis (TRAD). Thirty-nine 5th-year dental students (examiners) at a university in Hamburg, Germany, received joint training in the diagnosis of TMD by clinical functional analysis and manual structural analysis ('orthopedic tests'). This study is based on anonymized data from 10 patients who were consecutively recruited at a specialized TMJ treatment center. The examiners were randomly allocated to two groups. Each examiner established a structured diagnosis through a traditional diagnostic method and by computer-aided diagnosis (CMDfact 4 functional diagnostics software) of five cases, each using the AB/BA crossover design. The diagnoses established by each individual examiner were then compared with the corresponding reference diagnoses (gold standard) and with those of the other examiners. Cohen's kappa coefficient analysis showed that median agreement with the reference diagnoses was significantly higher (P < 0.001) with computer assistance (median 0.692) than without it (0.553). Fleiss' kappa showed that the median interexaminer consistency of diagnoses was significantly higher (P < 0.001) with computer assistance (0.497) than with traditional diagnostic methods alone (0.271). Likewise, the number of false-positive and false-negative diagnoses was significantly lower with computer assistance (P < 0.001). This study determined that dentists who are less experienced and not specialized in dental functional diagnostics achieve a significantly better and more consistent diagnostic quality with computer assistance by means of the system used in this study. Therefore, it seems advisable to extend computer-aided diagnostics to further functional examination techniques (condylar position analysis and jaw motion analysis).