Can guidelines rein in oxygen use? A retrospective cross-sectional study using routinely collected data

Abstract

Oxygen is one of the most commonly used emergency therapies. Like other therapies, oxygen can cause harm if used inappropriately. During the COVID-19 pandemic, guidelines were released to optimize oxygen and medication use. In the current study, we examine whether oxygen and medication use during the first wave of the COVID-19 pandemic was in concordance with new guidelines.

A retrospective cross-sectional study was conducted using routinely collected data from University of Birmingham NHS Foundation Trust in England. Patients were admitted between April 2020 and September 2020, were over the age of 18 years, and had a confirmed diagnosis of COVID-19. To assess adherence to the oxygen guidelines (i.e. SpO₂ adherence), the percentage of times oxygen therapy was administered within, over, and under guideline specifications were calculated for patients overall, and then for patients with and without chronic obstructive pulmonary disease (COPD)/pulmonary disease separately. Next, two multinomial regression analyses were conducted to assess whether clinical processes, pre-admission diagnoses, and other demographic factors were related to oxygen use. Analysis 1 included patients not diagnosed with COPD/pulmonary disease. Analysis 2 included patients diagnosed with COPD/pulmonary disease. Results are reported as tallies, percentages, and odds ratios with 95% confidence intervals. To assess adherence to a new medication guideline, the percentage of patients administered oxygen and dexamethasone was calculated for those admitted after 25 June 2020.

The overall number of patients included in our SpO₂ adherence analyses was 8751 (female = 4168). Oxygen was used within guideline specifications less than half the time, i.e. 41.6% (n = 3638/8751); non-adherence involving under-administering (3.5%, n = 304/8751) was markedly lower than over-administering (55.0%, n = 4809/8751). Adherence was higher for patients without COPD (43.7%, n = 3383/7741) than with COPD (25.2%, n = 255/1010). Under-administering was low across groups (non-COPD 3.5%, n = 274/7741 and COPD 2.9%, n = 30/1010). Over-administering was markedly lower for non-COPD (52.3%, n = 4084/7741) than COPD (71.8%, n = 725/1010) patients. Diagnoses associated with over-administering varied across the groups. Regarding the dexamethasone guidelines, of the 6397 patients admitted after the 24th of June, only 12.6% (n = 805) received dexamethasone.

Suboptimal use of oxygen and medication were common during the first wave of the COVID-19 pandemic. As found in previous studies, over-administering was more common than under-administering. The new guidelines issued during the COVID-19 pandemic were not by themselves sufficient to optimize oxygen use. Behavioural strategies are explored which may help policymakers optimize oxygen use.