Table 6.

Multiple linear regression analysis with change of CAL as the dependent variable in the all subjects (N = 45).

Variables	B (95% CI)	β	P value
Intercept	1.068 (0.470, 1.666)	—	0.001
PPD (mm) at baseline	−0.534 (−0.853, −0.216)	−0.451	0.002
PCR (%) at baseline	0.008 (0.0003, 0.016)	0.262	0.042

We used backward elimination method with a P > 0.1 threshold for exclusion to construct the model. The candidate variables for the model were ΔTeeth present, PPD at baseline, and PCR at baseline. The F-statistic was 7.374 (P value was 0.002), the R-square was 0.260, and the adjusted R-square of the model was 0.225. The variance inflation factor was <5 and condition index was <15.

Δ represents differences (value at follow-up minus value at baseline).

N: number; CI: confidence interval; B: unstandardized regression coefficient; β: standardized coefficient; PPD: probing pocket depth; CAL: clinical attachment level; PCR: plaque control record.