Practice nurse-led use of portable echo to detect early forms of heart failure in primary care: insights from a prospective, multicenter, surveillance study
S Stewart, B Worsley, C Kosch, S Barry, J Elliott, L Mullen, E Johns, A Wilson, L Collison, R Perry, D Playford, Y K ChanAbstract
Background
Despite our best efforts, the initial detection of heart failure (HF) typically occurs when a person is hospitalized with acute HF. Concurrently, even among high-risk cases, HF remains chronically undetected in the primary care setting.
Purpose
To determine if Practice Nurses (PNs) can learn to reliably apply artificial intelligence-mediated portable cardiac ultrasound (AI-PoCUS) to acquire sufficiently detailed cardiac images to identify previously undetected HF in primary care patients aged ≥60 years with at least one major antecedent for the syndrome.
Methods
We are conducting a prospective, pragmatic, multicentre, HF surveillance study across metropolitan and rural-remote primary care clinics. After an initial training program of >6months, we examined the pattern of success in generating clinically informative AI-PoCUS reports from the first 300 cases being formally screened for undetected HF by novice PNs. AI-assisted 2-dimensional imaging included parasternal long-axis and apical 4- and 2-chamber views. Manually acquired imaging included colour Doppler imaging of left heart valves, mitral inflow velocities and both septal and lateral mitral annular tissue Doppler velocities.
Results
Overall, 4 of 8 PNs from diverse clinical backgrounds who underwent initial training took responsibility for HF screening. The screened cohort comprised 150 men and 150 women of a similar age (70.7±9.2 versus 71.7±10.7 years) and anthropometric (BMI 31.0±6.4 versus 31.1±6.2 kg/m2) profile. Overall, cardiac images were not generated in 27 cases (9%). However, this decreased from 8 (16%) to zero in the first 50 to last 50 cases screened, whilst the number of complete AI-PoCUS reports (generating >30 cardiac indices) rose from 13 (26%) to 34 (68%), respectively. On an adjusted basis, there was no difference in those cases with clinically meaningful AI-PoCUS reports generated (>20 cardiac indices) in respect to their age (aOR 0.99, 95% CI 0.96 – 1.02 per year; p=0.322), biological sex (aOR 1.03, 95% CI 0.61-1.76 men versus women; p=0.910) or BMI (aOR 0.98, 95% CI 0.91-1.06 per unit; p=0.595). Alternatively, a full AI-PoCUS report was significantly more likely to be generated in the last versus first 50 cases screened (aOR 3.95, 95% CI 1.38-11.3; p=0.011). A clear ‘threshold of competence' was evident once a PN had conducted 30-40 HF screenings – with the generation of full AI-PoCUS reports being 2.43-fold (95% CI 1.03-5.72; p=0.042), 5.25-fold (95% CI 2.17-12.7; p<0.001) and 6.30-fold (95% CI 2.58-15.4; p<0.001) more likely in those screened in the 151-200, 201-250 and 251-300 positions, respectively.
Conclusions
We provide compelling evidence that, with appropriate training, time and support, PNs can apply AI-PoCUS as part of an innovative program to find undetected HF in the primary care setting. As such, there is proven capacity to expand the scope of practice of PNs to apply this new technology to improve the earlier detection and treatment of HF.