The Science Journal of the American Association for Respiratory Care

2010 OPEN FORUM Abstracts


Carla Wollens, Robert L. Chatburn; Respiratory Institute, Cleveland Clinic, Cleveland, OH

BACKGROUND: In our tertiary care academic ICU we have a standing order for arterial blood gas (ABG) sampling per protocol. The protocol lists acceptable indications for ABGs. A review of the literature showed that protocols reduce ABG sampling frequency but there are no data about adherence to protocols or use of results. The purpose of the study was to assess the degree of respiratory therapist adherence to the protocol. A secondary goal was to determine how many times the ABG result was associated with a change in the ventilator management. METHODS: ABG results were collected in a 35 bed MICU over on 8 random dates during a 10-day period. On each date, all ABGs on all patients were recorded resulting in 200 sets of data. An individual ABG sample was deemed indicated if it met one of the 12 protocol criteria, otherwise it was non-indicated. We also noted when there was a ventilator change associated with the ABG result. RESULTS: Review of the data showed 80% (160) of the ABG’s drawn were indicated by our protocol. For all ABG samples drawn, 41% (82) resulted in some ventilator adjustment. The distribution of indications and actions taken are shown in the table. Interestingly, 33% of the non-indicated ABGs (13) resulted in a ventilator change. CONCLUSION: 20% of ABGs were either not indicated or no indication was documented. We speculate the most of these were issues of clinician’s emotional assurance of patient condition rather than poor documentation of prescribed indications. Given that 33% of these actually resulted in ventilator adjustments, perhaps our list of indications should be expanded. On the other hand, 59% of indicated ABGs resulted in no action. Given that the majority of ABGs were driven by prior “out of range” results, perhaps our ranges should be modified. In particular, many patients had metabolic acidosis which would not be expected to result in ventilator changes. These data should serve as benchmarks for future studies. Sponsored Research - None