Study Designs and Efficacy Levels in Dentomaxillofacial Radiology

Objectives: Radiologists and others use published results to help make important clinical decisions that affect patients' well-being. The strength of evidence is dependent on the study design, and the impact on patient care is dependent on the level of diagnostic efficacy. The purpose of this study was to evaluate the study designs used and the level of efficacy investigated in the oral and maxillofacial literature as represented by Dentomaxillofacial Radiology (DMFR).

Methods: Two investigators used 2 hierarchical classification schemes to independently assess 325 articles published in 1996-2005 in DMFR. One scheme classified articles by study design using the traditional epidemiological hierarchy; the second by diagnostic efficacy using the criteria presented by Fryback and Thornbury (Med Decis Making, 1991). The country of origin and number of authors were noted.

Results: Of the 325 articles, 167 were clinical research. Of these, 47% were case reports/series, 49% were cross-sectional, and 4% were case-control or experimental studies. Ninety-six articles addressed diagnostic efficacy, of which 33% were technical efficacy, 64% were diagnostic accuracy, and 3% were either diagnostic thinking or therapeutic efficacy. Thirty-three countries were represented: Japan (23%), USA (17%), UK (9%), and Sweden (7%) were the four major contributors. The average number of authors/article was 2.8-3.8 for USA, UK, Sweden, and others. Japan had a statistically significant greater average of 5.4 authors/article (ANOVA, F=18.8, P<0.01).

Conclusions: The study designs used in the majority of clinical research articles published in DMFR do not provide strong evidence for clinical decision-making. Similarly, the majority of efficacy papers did not address the impact or relevance of diagnostic imaging to patient care. In order to make wise choices regarding the allocation of resources and sound clinical decisions, diagnostic imaging research needs to be shifted to study designs and efficacy levels at the higher end of their respective hierarchies.