The measure of treatment agreement between portable and laboratory blood gas measurements in guiding protocol-driven ventilator management.

Background: Portable blood gas analyzer and monitor devices are increasingly being used to direct ventilator therapy. The purpose of this study was to evaluate the "measure of treatment agreement" between portable and laboratory blood gas measurements used in guiding protocol-driven ventilator management.

Materials And Methods: Using National Institutes of Health Acute Respiratory Distress Syndrome network ventilator management guidelines to manage patient care, measurements taken from the Nonin 8500 M pulse oximeter (SpO2), the Novametrix-610 end-tidal CO2 (ETCO2) detector, and the i-STAT 1 (SaO2, PO2, pH, PCO2) were compared with the recommended treatment from paired laboratory ABL-725 (SaCO2, PO2, pH, PCO2) measurements.

Results: Four hundred forty-six intubated adult intensive care unit patients were studied prospectively. Except for the ETCO2 (R2 = 0.460), correlation coefficients between portable and laboratory measurements were high (R2 > or = 0.755). Testing for equivalence, the Nonin-SpO2, iSTAT-PO2, iSTAT-pH, and iSTAT-PCO2 were deemed "equivalent" surrogates to paired ABL measurements. Testing for the limits of agreement found only the iSTAT-PCO2 to be an acceptable surrogate measurement. The measure of treatment agreement between the portable and paired laboratory blood gas measurements were Nonin-SpO2 (68%), iSTAT-SaO2 (73%), iSTAT-PO2 (97%), iSTAT-pH (88%), iSTAT-PCO2 (95%), and Novametrix-ETCO2 (60%). Only the iSTAT-PO2 and the iSTAT-PCO2 achieved the > or =95% treatment agreement threshold to be considered as acceptable surrogates to laboratory measurements.

Conclusions: : The iSTAT-PO2 and -PCO2 were portable device measurements acceptable as surrogates to standard clinical laboratory blood gas measurements in guiding protocol-directed ventilator management. The "measure of treatment agreement," based on standardized decisions and measurement thresholds of a protocol, provides a simple method for assessing clinical validity of surrogate measurements.