From the 2021 HVPAA National Conference
Jessica Campanile (Johns Hopkins University School of Medicine), Christopher Golden, Julia Kim, Michael Crocetti, W. Reid Thompson
Background
Cardiac auscultation with a stethoscope is often used to determine need for echocardiography, however, physical examination during most telemedicine visits is limited to visual inspection of the patient. Remote auscultation is possible with an electronic stethoscope, but feasibility and effectiveness to detect heart murmurs in children have not been examined in detail.
Cardiac auscultation skills are declining among primary providers, leading to increased cardiologist referrals and echocardiography overuse, which pose financial, logistical, and emotional burdens on pediatric patients and families. The use of AI-assisted auscultation during telemedicine visits may aid in clinical decision-making when determining whether an echocardiogram is indicated.
We previously examined a new referral paradigm of AI-assisted remote auscultation by store-and-send protocol from primary care offices with later remote interpretation by cardiologist, which exhibited a net feasibility of 93%. With the onset of COVID-19 and mandatory shift to telemedicine visits, we expanded the study to examine the feasibility of in-home remote auscultation with real-time cardiologist listening and interpretation to optimize echocardiography referral while limiting exposure to high-risk individuals and environments. The value of this new paradigm, if successful, would continue post-pandemic, and is already being piloted to assist in remote areas and under-resourced healthcare settings.
Objective
Assess the feasibility and effectiveness of in-home remote cardiac auscultation during telemedicine visits.
Methods
In this pilot study, over 50 patients have been enrolled to date. Prior to a cardiology telemedicine visit, families are mailed a smartphone and electronic stethoscope. During the visit, the provider leads them through use of the equipment, having the parent or, in some cases, the patient, hold the stethoscope themselves. The provider listens to the heart sounds in real time, classifying them as pathologic or innocent, and views the information from the AI algorithm (which also provides such classification). Patient and provider usage surveys as well as clinical outcomes are being studied.
Results
To-date, findings from patient, family, and provider surveys show in-home remote auscultation to be successful in the majority of cases, with reasons for the few failures including patient factors (e.g. crying, talking, moving) or technical issues which have been iteratively addressed and resolved. In most successful remote auscultation visits the sound quality has been suitable for clinical decision-making. Additionally, most remote exams have supported the cardiologist’s clinical decision-making.
The majority of patients and families surveyed to date feel the remote stethoscope exam improved their visit and was easy to complete. All report comfort in terms of privacy, and most would agree to a future remote stethoscope exam if their physician requested.
Clinical findings of remote stethoscope exams have included tracking existing murmurs, diagnosing postural orthostatic tachycardia syndrome, and flagging new developments as indications for echocardiography. Outcome data is currently being gathered. Next steps include a large-scale multi-institutional rollout and implementation in under-resourced healthcare settings.
Conclusions
A new paradigm for in-home cardiac auscultation appears feasible, clinically valuable, and satisfactory to patients and providers alike. With iterative, robust protocols for remote auscultation, the study team has created a reliable, repeatable process for listening to heart sounds remotely within the context of a home-based telemedicine visit.
Clinical Implications
We anticipate home-based AI-assisted remote auscultation may reduce variability in clinical decision making, optimize echocardiography referrals, and reduce barriers to care for populations that are difficult to reach with traditional, in-person models of specialty care.
