Implementation study reporting diagnostic accuracy, outcomes and costs in a multicentre randomised controlled trial of non‐expert lung ultrasound to detect pulmonary oedema

BACKGROUND: Lung ultrasound experts claim that 'B line' artefacts herald pulmonary oedema, but links between early recognition and improved outcome are unconfirmed, particularly for non-expert clinicians. OBJECTIVE: Assess individual and system impacts of early, non-expert lung ultrasound (LUS) in breathless older patients. METHODS: Prospective single-blinded randomised controlled trial of point of care LUS by non-expert clinicians, augmenting ED assessment of patients over 60 years, presenting with dyspnoea. After brief training, clinicians at three hospitals used computer-generated block randomisation envelopes to allocate a convenience sample of consenting presenters to conventional or LUS-augmented management. ED provisional diagnosis was compared to a post-discharge chart audit diagnosis, blinded to ultrasound findings. Secondary outcomes were the length of stay in ED and hospital, costs and discharge destinations. RESULTS: From three sites, 224 controls were compared with 218 interventions. LUS improved diagnostic accuracy, with a small effect size in favour of LUS (risk difference: 6.5%, 95% CI 0.9-12) and a number needed to scan of 16 (95% CI 8-107). With LUS, ED and hospital stay increased non-significantly, while discharge destination trends improved. No harm ensued. CONCLUSIONS: Non-expert LUS augmenting dyspnoea workup may improve diagnostic accuracy, but did not significantly alter costs or outcomes in the ED or the hospital. TRIAL REGISTRATION: Australia and New Zealand Clinical Trials Registry (ACTRN12613001023741).