Effects of implementing a protocol for arterial blood gas use on ordering practices and diagnostic yield

Background: The arterial blood gas (ABG) is a valuable and commonly used laboratory test. This prospective cohort study examined the variability of ABG ordering through the implementation of an evidence-based protocol.

Methods: The study consisted of two 6-week periods. The protocol consisted of evidence-based and consensus opinion based indications for ABGs. In the first phase (initial 6 weeks), respiratory therapists recorded the indications for ABGs ordered by clinicians. In the second phase, all medical and surgical physicians were trained on the clinical rationale behind the protocol and were instructed to write the indication for each ABG with the order. Rates of ABGs/patient/day were measured in aggregate and per indication. Multivariate regression was used for adjusted comparisons between indications within the protocol.

Results: After protocol implementation, there was a significant decrease in ABGs from 2158 to 1674 (p=0.001), and after adjusting for daily census, there was a significant decrease from 35.3 ABGs/100 patients/day to 26.5 ABGs/100 patients/day (p<0.001), with no change in mortality or demographic characteristics between the populations. The percent of ABGs with normal range values for pH, PaCO2, and PaO2 decreased from 13.3% to 9.6% after implementation (p<0.001). Multivariate analysis revealed a 14% decrease in daily ABGs (p=0.001), a 15% decrease in weaning trial ABGs (p=0.039), and a 15% increase in ABGs ordered following a change in minute ventilation (p=0.004). Cost minimization analysis estimated annual institutional savings to be $87,565.

Conclusions: Implementation of an evidence based protocol for ABG use resulted in fewer ABGs/patient/day largely from reduction of routine, daily ABGs. Ordering patterns for ABGs appeared to shift towards more clinically appropriate/relevant indications.

Level of evidence: 2b.