LAUSR

mixed effect analysis. Percentage of monitoring time in which each AR model was able to produce CPPopt recommendations were compared using Wilcoxon signed rank test. All patients were male with median age of 42 yr [interquartile range (IQR), 27e54]. The total monitoring time was 10 days with a median of 12.7 h per patient (IQR, 8.0e24.9). There was no statistically significant effect of types of AR models (PRx-DATACAR and LAx-DATACAR) on CPPopt recommendations [b-estimate1⁄40.117, 95% confidence interval (CI)1⁄4e0.303 to 0.531]. There was no statistically significant difference between the percentage of monitoring time in which CPPopt recommendations were produced by PRx-DATACAR and LAx-DATACAR (89.6% vs 87.8%, P1⁄40.463) (Fig. 2). PRx-DATACAR and LAx-DATACAR produced similar CPPopt recommendations throughout a similar percentage of monitoring time. Thus, PRx-DATACAR and LAx-DATACAR are as robust and could be used interchangeably.