A New Tool for Assessing the Level of Evidence

Objectives: To develop a new level of evidence (LOE) classification system that recognizes the value of certain types of observational studies, and pilot its use in an assessment of the LOE of clinical studies published in leading orthodontic journals.
Methods: A new four-category LOE tool was developed based on the Oxford 2011 LOE and GRADE (Grading of Recommendations Assessment, Development and Evaluation). Then, a hand-search of the official online archives of five leading orthodontic journals was carried out to identify clinical studies published during 2015 to 2017. Studies meeting our pre-determined eligibility criteria were included and assessed with the new LOE tool. Generalized estimation equation (GEE) analyses were performed to identify the association between the LOE of each article and 6 potentially related factors.
Results: A total of 637 studies were included and assessed. According to the new LOE tool, 8 of these (1.3%) were level-1 studies, 160 (25.1%) were level-2, 326 (51.2%) were level-3, and 143 (22.4%) were level-4. The mean LOE of all included studies was 2.95 (95% CI: 2.89 to 3.00; median: 3.00). A significant difference in LOE was found among different journals (P<0.001). According to the multivariable GEE analysis, the geographic origin of the first author (P<0.001) and existence of financial support (P=0.002) were significantly associated with LOE.
Conclusions: A new LOE assessment tool has been proposed and it could be helpful in future EBD (evidence-based dentistry) related research, communication and practice. In addition, there is still need for further improvement in the overall LOE of clinical studies in orthodontics.