Large random measurement errors (differences between in vivo and in vitro SvOJ resulting in wide 95 percent confidence intervals cannot be corrected by simple software manipulation. We have presented both 95 percent confidence limits for both total variability and intrasubject variability. The 95 percent confidence interval based on total variability defines the limits of agreement present in circumstances when oximetry PA catheters are used to guide therapeutic interventions when the goal is to identify the patients actual Sv02.
For instance, if the therapeutic goal is to ensure that a patients actual Sv02 is 70 percent, it would be necessary for the Oximetrix 3 catheter to read 76 percent to ensure, with 95 percent confidence, that the patients actual SvOa was at least 70 percent. In the case of the HEM0PR02, it would be necessary for the catheter to read 80 percent to ensure the same level of certainty regarding the target Sv02 of 70 percent. Similar analogies apply in cases where the decision to treat may be triggered if certain low Sv02 values are reached. Conversely, if the oximetry PA catheters are used as trending devices, a smaller change can be used to predict an actual trend in the patients Sv02. For instance, all three systems will allow the detection, with 95 percent confidence, of a directional trend with a change in readings of approximately ±7 percent.