Impact of Low Flow Nasal Cannula Failure Criteria on High Flow Nasal Cannula Utilization: A Quality Improvement Project

Abstract

Background: High flow nasal cannula (HFNC) therapy for bronchiolitis has become increasingly popular,however early initiation did not impact patient outcomes in the two randomized trials and is reported to be15x more costly than standard low flow nasal cannula (LFNC) We hypothesized that lack of initiation criteriafor use of HFNC was resulting in overutilization Objective: We sought to increase trials of LFNC in acute viralbronchiolitis in order to decrease the use of HFNC Our primary aim is to increase the percent of LFNCutilization prior to HFNC initiation by 50% by April 2020 Design/Methods: We undertook a multidisciplinaryquality improvement project involving Hospital Medicine, Emergency Medicine and Critical Care physicians,along with respiratory therapy and nursing We revised our existing bronchiolitis protocol to include LFNCfailure defined as no improvement in oxygen saturation, heart rate or respiratory rate within 30-60 minutes ofinitiation and protocolized LFNC treatment failure to precede HFNC initiation Our pathway change took effectDecember 1st, 2019 Inclusion criteria were: < 2 years admitted to PHM with a primary diagnosis of any ICD-10 code for bronchiolitis PICU transfer rates serve as the balancing measure and statistical process controlgrouping data in two-week increments is the primary method of analysis Results: Baseline data for the seasonpreceding our intervention (2018-2019) revealed HFNC usage in floor patients with bronchiolitis to be 62%(61/98), with 39% (24/61) undergoing a trial of LFNC prior to initiation of HFNC Post-intervention datademonstrate an increase in LFNC trials to 73% (29/40) and a concomitant decrease in HFNC utilization to 48%(40/84) (Figure 1 a/b) Rules for special cause were satisfied for the LFNC measure (2 out 3 consecutivemeasures above 2 sigma) but not for the HFNC measure when the project had to be stopped due to the SARS-CoV2 epidemic which was an external pressure against the use of HFNC (an aerosolizing procedure) Ourbalancing measure appears stable with a pre-intervention PICU transfer rate of 6 1% (n=98) and a post-intervention rate of 2 9% (n=68) Conclusion: Preliminary data suggests a simple intervention consisting ofestablishing LFNC treatment failure criteria in patients with moderate bronchiolitis may reduce HFNC usewithout adverse effects