The change in endurance time induced by PLB was significantly correlated with the variationin IC (∆IC) in the isotime (=0.755; p<0.05) in IMPROVER subgroup. Moreover, the IC inducedby PLB in this subgroup correlated well with the variation in dyspnoea intensity in the isotime (Figure1; = 0.791; p<0.05). No other association between variables in the two subgroups wereobserved. We performed a ROC curve analysis of the peak expiratory flow (PEF) as a percentageof the predicted values (% pred) to identify cutoff values that had greater sensitivity and specificityin differentiating between IMPROVER and NONIMPROVER. At the point showing the greatestsensitivity and specificity, we observed 61% sensitivity and 89% specificity with a 47.7% pred PEF(Figure 2). Based on this analysis, we can conclude that the sensitivity value was very low, indicatinga large percentage of falsepositive results. However, for PEF values above 55.9% pred, thespecificity values were 100% and the sensitivity values were 38.7%, indicating an excellentthreshold, without falsenegative results in identifying patients who did not benefit from PLB.
การแปล กรุณารอสักครู่..
