Hello everyone, I'm Diana Foo from Clinical Research Centre, Sarawak General Hospital, Malaysia. Today I'm going to be talking about Heart2Miss, a decentralised echo triage in the community.

What is the importance behind the Heart2Miss study?

The Heart2Miss started off with a vision to bring the echo from the patients to the population, which is to bring the echo out of the hospital and into the community. Heart failure remains a burden globally, but the diagnostics early detection can be very challenging especially in the community due to critical understaffing and that's why Heart2Miss was created. It created a proof of concept model in decentralising the echo diagnostics for early heart failure detection in the community.

What specific triage workflow was used in the study?

So in our scalable hub-spoke approach we introduced an intermediary tier which links between the hub and the spoke. So this intermediary tier is actually an intermediary centre with an echo lab that trains novice bioscience graduates as a mobile community sonographers who actually go to the spokes diabetes primary care to perform quick echo triage — a trivial quick echo triage using the pocus with the AI guidance feature and powered with the Us2ai automated analysis software.

So the intermediary echo lab sonographers will review these results remotely and verify telehealth and trigger a second confirmation scan, which is a manual, standard echocardiography for confirmation. And the intermediary both facilitates referral to the hub for patients who need specialist care.

What was the study design and patient population?

This was a real-world implementation study in which we screened 985 patients with type 2 diabetes across a diabetes primary care over a period of seven months. These are the places where the cardiovascular imaging is not always available.

Our goal is to evaluate the efficacy in detecting the prevalence of stage B and stage C heart failure, which is previously undiagnosed, and then to assess the feasibility metrics associated with reduction in the hub loads in terms of: the diagnostic burden, patients, referral and outpatient management burden, and then most importantly, to assess the performance of the novice bioscience graduates who were trained as mobile community sonographers.

What were the key results?

Of 985 patients, 11.1% were detected with stage B heart failure and 1% with stage C heart failure, all undiagnosed. This model offloads the tertiary load significantly in terms of the diagnostic burden, outpatient management and referral burden.

In fact, along the referral cascade, only 1% out of 985 patients who require specialist care were referred to hub and most importantly, when we assess the performance of novices, we find that after just four weeks training our novice operators acquire more than 85% of the scans that were analysable by the AI. And that showed exponential uptrend in the success rate after accumulative of 400 cases with a reduction of scan time from 11 minutes to 8.3 minutes.

What are the key challenges in integrating this model within existing healthcare systems, and how could this potentially be addressed?

One of the key challenges is perhaps workforce integration. While we see a shortage of workforce especially in the healthcare systems we had on the other hand another untapped pool of bioscience graduates — many of them were underemployed. In Heart2Miss, we equipped them and upskilled them to become mobile community sonographers to address the unmet public health need.

But the key challenge here is, how do we formally integrate these unconventional diagnostic workforce, who are actually bioscience graduates, into the public health systems? Therefore, we need proper regulatory training and supervisory structure in place to ensure the quality if we were to scale this across different regions.

And besides this, another challenge I believe would be the interoperability and data sharing between the spokes and the tertiary hub referral centers, which could be potentially a practical hurdle to this. To address this in the long term, we need to have a clear task-shifting framework in place and we need to explore the policy engagement or policy support in order to integrate this in national healthcare settings.