1997 OPEN FORUM Abstracts
THE IMPACT OF AN ARTERIAL BLOOD GAS ALGORITHM ON ORDERING PRACTICES IN A MEDICAL ICU.
Douglas K. Orens, MBA,RRT. Pat Perez-Trepechio, BS,RRT. James K. Stoller, MD, Cleveland Clinic Foundation, Cleveland, OH
BACKGROUND: Previous studies have suggested that adherence to an arterial blood gas algorithm can lessen the frequency of inappropriately drawn arterial blood gases in a Surgical Intensive Care Unit. As part of our ongoing assessment of the impact of respiratory care protocols, we conducted a controlled study to examine whether an algorithm for arterial blood gas sampling lessened the volume and associated costs of arterial blood gases. Methods: The arterial blood gas algorithm was developed by members of the Section of Respiratory Therapy in accordance with relevant AARC Clinical Practice Guideline's. Overall, the algorithm suggested sampling an arterial blood gas under specific conditions,: e.g. hemodynamic instability, pre-and post-extubation, during active weaning from mechanical ventilation, increasing FI02 requirements. Using a management information system to tally the number of arterial blood gas samples and a cost accounting system (Transition Systems, Boston, MA) that allowed for variable and fixed costs associated with arterial sampling to be calculated, we compared the volume and cost of arterial blood gas sampling in two similar periods, both before (8/95 - 3/96) and then again after (8/96 - 3/97) implementing the algorithm. Prior to implementing the algorithm, arterial blood gases could be obtained on request by physicians, nurses or respiratory therapists. In implementing the algorithm, respiratory care practitioners in the ICU were specifically trained and the algorithm was made available to other health care providers. Results: During the 8 month control period (before algorithm implementation), 531 patients accounted for 2,334 ventilator-days in the 12-bed Medical ICU. During this period, 11,435 arterial blood gas samples were drawn (mean 4.9 arterial blood gases per ventilator-day). For the 8 month period after implementing the arterial blood gas algorithm (8/96 - 3/97), 462 patients accounted for 2665 ventilator-days and the number of arterial blood gas samples decreased to 6,620 (mean 2.5 arterial blood gases per ventilator-day, for a 49% reduction). The observed decrease in the number of arterial blood gases was associated with a cost savings of $62,210. Furthermore, comparison between the study and control periods showed no significant difference in Medical ICU mortality rates (24%[study period] vs. 23.3%) or ICU re-admission rates (13.6% vs. 11.3%). Conclusions: We conclude that the availability of an arterial blood gas sampling algorithm can be associated with significant decreases in the volume of arterial blood gas sampling and substantial cost savings without recognizable adverse events. This reduction occurred despite a 14% increase in the number of ventilator-days during the compared time periods. Implementation of an arterial blood gas sampling algorithm appears to be a cost-effective tool in a Medical ICU, extending previous observations from the Surgical ICU experience.