Objective
To explore new methods of remote congenital heart disease (CHD) screening, our study investigated the feasibility of applying the “Iron Man” remote CHD screening system to CHD screening.
Methods
135 CHD patients and 57 non-CHD patients undergoing transthoracic echocardiography (TTE) at two different hospitals (from 2018 to 2019) were evaluated. We examined competency in diagnosing simple CHD cases with the "Iron Man" system by comparing cardiac sounds collected and analyzed in the following three ways-remote auscultation by a physician, automatic analysis by computer, and the online analysis system combining physician auscultation and computer-based automatic analysis.
Results
The physician’s diagnosis of CHD (based on remote auscultation of cardiac sounds collected by the "Iron Man"), had a sensitivity, specificity, overall diagnostic consistency rate and area under the ROC curve (AUC) of 84.44%, 71.93%, 80.73%, and 0.782, respectively. In comparison, the automatic computer analysis had rates of 89.63%, 73.68%, 84.90% and 0.817, respectively, and the remote CHD screening system had rates of 92.59%, 61.40%, 83.33% and 0.770. The remote CHD screening system had a higher sensitivity in diagnosing CHD than physician auscultation (92.59% vs.84.44%, P=0.001<0.017), and automatic computer analysis had a higher specificity than the remote CHD screening system (73.68% vs.61.40%, P=0.016<0.017). The automatic computer analysis demonstrated a higher rate of positive cases than physician auscultation (74.48% vs.67.71%, P=0.041<0.05).
Conclusion
The remote CHD screening system integrating physician auscultation and automatic computer analysis were more sensitive than physician auscultation in diagnosing CHD.