Despite widespread use of oxygen (O2) therapy, there is relatively little available information on routine O2 administration and monitoring; this is an issue particularly when considering the potential risks associated with inappropriate O2 utilization. A rapid health technology assessment (HTA) was conducted to inform the Respiratory Health Strategic Clinical Network Oxygen Summit in Alberta on aspects related to current practice in the use of O2 therapy in acute care, including administration, safety and quality, and inappropriate practice. Clinical audit is a tool used to determine deviations in practice and to identify opportunities for improvement. The objective of this presentation is to describe the experience and lessons learned from including clinical audit studies in the rapid HTA.

A standardized rapid review approach was used to identify, select, and synthesize evidence from studies published in English from 2005 to 2016. A supplementary literature search conducted in 2018 provided additional background information on the value, applicability, and limitation of using results from clinical audit studies to inform questions of good practice.

Twenty-four clinical audit studies on O2 therapy were identified; the majority were conducted in the United Kingdom. The studies varied in design, methodology, and data and outcomes reporting. Ten studies investigated the appropriateness of O2 therapy prescription pre- and post-implementation of local initiatives and interventions, which helped pinpoint major gaps in current practice, and identified general recommendations for improvement of practice. A list of reporting criteria is proposed for improving the reporting of clinical audit studies results.

Conducting clinical audit studies is resource-intensive. In the absence of other research evidence and local practice data, translating results from clinical audit studies conducted in other jurisdictions, while challenging, can help address appropriateness questions. However, inferences from these studies may be suitable only for certain topics or an operating context.