Abstract Rationale Prolonged oxygen therapy (O2) can delay discharge, increase healthcare costs, and raise hyperoxia risk. To promote timely oxygen discontinuation, an oxygen weaning protocol was implemented on general medicine wards in a 222-bed community hospital in mid-2025. Methods In this retrospective cohort study, patient charts from one general medicine ward were reviewed for a historical cohort (June-October 2024) to compare to a protocol cohort (June-October 2025). Inclusion Adult patients hospitalized 24 hours for hypoxia/hypoxemia, acute or acute-on-chronic respiratory failure (ARF/ACRF), or chronic obstructive pulmonary disease (COPD), who underwent oxygen weaning on nasal cannula or Venturi mask. Exclusion use of other oxygen modalities during the weaning period, transfers, hospice discharge, death, or incomplete data. Primary outcomes Time to Wean (TTW, days from weaning initiation to baseline oxygen) and Length of Stay (LOS). Secondary outcome Discharge on oxygen higher than baseline. Baseline, initial, and peak oxygen flows and comorbidities were extracted. Cohorts were compared using Welch’s t-tests, Mann-Whitney U-tests, and multivariable regression. Protocol compliance was defined as de-escalation of oxygen when SpO295%. Results Among 631 charts reviewed (protocol n = 318, non-compliant n = 36; historical n = 313), 255 were eligible (protocol n = 113, historical n = 142). Unadjusted TTW was shorter under the protocol (2. 17 ± 2. 54 days) versus historical (2. 81 ± 2. 94 days) (p = 0. 098) while LOS was longer (protocol 5. 60 ± 4. 73 days versus historical 5. 08 ± 4. 61 days; p = 0. 06). Multivariable regression showed the protocol significantly reduced TTW by ∼0. 70 days (95 %CI -1. 27 to -0. 14, p = 0. 015) but had no effect on LOS (p = 0. 58). Protocol patients had lower discharge oxygen flows (0. 79 versus 1. 09 L/min), smaller increases from baseline, and lower odds of discharge on new/higher oxygen (OR≈0. 53, 95 %CI 0. 26-1. 08, p = 0. 08). TTW and LOS were moderately correlated across all groups (Pearson r = 0. 46, Spearman ρ = 0. 43, both p 10⁻), but strongest in ARF (r = 0. 55) and COPD (r = 0. 59). Observed statistical power was low (10%). Protocol compliance was 76%. Conclusion A standardized oxygenweaning protocol modestly shortened duration of oxygen therapy but did not reduce hospital LOS. The moderate TTW-LOS correlations suggest timely weaning influences discharge timing, but patient severity and nonrespiratory factors are major drivers of LOS. Protocol patients were less likely to be discharged on increased home oxygen, which can decrease costs (from 2018-2021, Medicare fee-for-service durable medical equipment (DME) spent ∼9. 8% of 8 billion in total expenditures on oxygen products). However, larger adequately powered studies with higher protocol adherence are needed to clarify the clinical impact of protocoldriven oxygen weaning. This abstract is funded by: None
Quach et al. (Fri,) studied this question.