A Study on Variances in Multivariate Analyses of Oral Implant Outcome

BACKGROUND: Elaborate studies have shown that interdependency exists between implants being placed in the same patient/jaw. Therefore, interdependency ought to be an important aspect to address, whenever performing statistical analyses of oral implant outcomes. A Jackknife method could be an option when conducting statistical evaluations of oral implant failure prognoses. PURPOSE: The aim of this study was to evaluate whether a statistical difference can be detected by using the Jackknife method in conjunction with life table analyses and/or a log rank test of four different combinations of jaw density and quantity. MATERIALS AND METHODS: Four multicenter studies were pooled and adjusted in order to create a research database consisting of 486 patients and 1,737 implants in preparation for the Jackknife resampling method. Combinations of jaw shapes and bone qualities were constructed to select at-risk patients. STATISTICAL METHODS: Life tables with confidence intervals were calculated and a log rank test was used to determine whether a statistical difference between the combinations could be established. RESULTS: Both statistical analyses, after the Jackknife resampling method, showed that patients with poor bone quality and resorbed jaws (combination IV) had a statistically higher risk of implant failure. CONCLUSION: By rearranging data using the Jackknife method, standardized statistical tests seem to work well even when the study population tested was affected by interdependency.