Table 3.
Factors associated with studies yielding significantly different failure rates between groups.
| Variable | Model 1 (Simulttaneously) | Model 2 (Backwards) |
|---|---|---|
| Fit | R2 = 0.31, p < 0.001 | R2 = 0.19, p < 0.001 |
| OR (95% CI) | OR (95% CI) | |
| Year of publication of trial | 0.78 (0.62/0.99) | 0.81 (0.68/0.96) |
| Journal impact >2 (ref.: ≤2) | 3.53 (0.72/17.3) | 3.26 (1.25/10.8) |
| Number of patients | 1.02 (0.98/1.07) | |
| Lesions per patient | 0.66 (0.26/1.71) | |
| Follow-up time (months) | 1.01 (0.98/1.04) | 1.02 (1.00/1.03) |
| Retained lesions (%) | ||
| Trials with load-bearing cavities (ref.: cervical) | 0.29 (0.12/7.32) | 0.29 (0.09/0.94) |
| Trials in permanent dentition (ref.: primary care) | 3.3 (0.04/3.00) | |
| Trials describing sample size calculation (ref.: not described) | 2.48 (0.26/23.7) | |
| Registered trials (ref.: not registered) | n/a | |
| Trials using outcome measure other than USPHS (ref.: USPHS) | 7.46 (8.69/64.0) | 4.77 (1.07/21.3) |
| Trials with sequence generation unclear/high (ref.: low) | 3.61 (0.78/16.8) | |
| Trials with allocation concealment unclear/high (ref.: low) | n/a | |
| Trial with blinding of operators unclear/high (ref.: low) | n/a | |
| Trials with blinding of examiners unclear/high (ref.: low) | 0.51 (0.14/1.87) | |
| Trials with missing data unclear/high (ref.: low) | 0.44 (0.05/3.92) | |
| Trials with selective reporting unclear/high (ref.: low) | 557 (5.6/55,442) | 12.1 (0.90/123) |
| Trials accounting for attrition in analysis/trails without attrition a (ref.: attrition not accounted for) | 1.75 (0.64/4.78) |
For regression analyses, variables were either entered simultaneously (Model 1), or entered and then removed backwards (Model 2). Model fit is indicated by R2 and p-values. Factors with significant positive (OR > 1) or negative (<1) association are highlighted in bold. For example, significant differences between groups were significantly more often found in journals with impact factor >2.