TABLE 3.

Simple and multiple linear regression of phenotypic characteristics of PB/RB isolates and duration of bacteremia

PhenotypeSimple linear regressionMultiple linear regressiona
Regression coefficient (%)b (95% CIc)PRegression coefficient (%)b (95% CI)P
hNP-1 (% survival)158 (111-229)0.01148 (112-197)0.007
tPMPs (% survival)103 (102-105)<0.001103 (102-105)<0.001
Fibronectin binding (A570)11 (0-285)0.17
Endothelial cell binding (% binding)102 (92-113)0.69
Biofilm formation (A490)185 (88-389)0.10191 (111-329)0.021
  • a Model P value < 0.0001; adjusted R2 = 0.492.

  • b A log transformation was applied to make the outcome more normally distributed. Once the final model was identified, the regression coefficients and their confidence intervals were transformed back. Thus, the coefficients of the models correspond to a multiplicative factor (percentage change) in days of bacteremia associated with a 1-unit change in each predictor variable.

  • c CI, confidence interval.