Page 170 - Peri-implant health: the effect of implant design and surgical procedure on bone and soft tissue stability

CHAPTER 7
Statistical analysis
Data analysis was performed in SPSS statistics 26 (SPSS Inc., Chicago, IL) and R Core Team 2019 (R: a language and environment for statistical computing; R Foundation for Statistical Computing, Vienna, Austria). Outcomes are reported with descriptive statistics (mean, standard deviation [SD], median, and range) and visualized through boxplot representation. The analysis on bone level is carried out after 6(t1) and 36(t2) months using a linear mixed-model analysis. The mean bone level per implant type at t1 and t2 were the dependent variables and the patient as random intercept. The independent variables for this analysis were the bone level at baseline (t0), implant type (I-MT, I-NMT, E-MT, and E-NMT), and implant position (anterior and posterior).
An analysis of the measurement error for the continuous variable bone level between the different observers (S.S., M.G., and R.D.) was performed by the use of a scatterplot representation and a paired t test. The random error or duplicate measurement error (DME) was calculated with the formula: (sd of differences)/√2. Also, the paired samples correlation between the observers was calculated.
The analysis on probing pocket depth, bleeding on probing and plaque were, respectively, carried out after 36 months using a linear mixed-model analysis for the probing pocket depth and a mixed-model logistic regression analysis for bleeding on probing and plaque. The probing pocket depth, bleeding on probing and plaque at site level was the dependent variable and both to the implant and patient a random intercept was assigned. The independent variables for this analysis were implant type (I-MT, I-NMT, E-MT, and E-NMT) and implant position (anterior and posterior). To show the effect of the fixed factor, the fixed effects estimates (linear mixed-model analysis) and the odds ratios (mixed-model logistic regression analysis), all with their confidence limits and p-value, are given. The level of significance was set at 0.05.
■ RESULTS
Study population
Initially, 25 patients were included in this study (Figure 2). The study population consisted of 15 men and 10 women with a mean age of 62 years (SD 9.98, range
168