Preliminary Evaluation of a Mobile Application-based Tool to Aid Early Detection of Critical Congenital Heart Disease

Background This study reports the preliminary results of a mobile application (APP)-based tool to aid early detection of critical congenital heart disease (CCHD). The tool is formulated to assist neonatologists/paediatricians in bedside diagnosis by asking a series of clinical and echocardiography-based questions, leading to a conclusion. The study was done with paediatric cardiologists themselves to ascertain the validity of the algorithms. Methods Consecutive patients with CCHD seen by the participating paediatric cardiologists during the period January 2021 to June 2022 formed the subjects for the study. The routine evaluation and management of the patients were carried out as usual. Later, the treating paediatric cardiologists applied the clinical scenario and echo findings to the ‘Echo Integrated Diagnostic Tool’ of the ‘Smiling Heart APP’. The APP tool performance was recorded. Concurrence between diagnoses made by the two approaches was assessed using kappa statistics. Performance of the APP tool across the spectrum of CHDs was assessed using Fisher’s exact test. Results Out of 69 patients, correct diagnosis was reached by the APP tool in 62 (89.86%). In the rest, (7/69; 11.6%), which mostly comprised of rare diseases, an urgent paediatric cardiology opinion was suggested. The APP tool-based diagnosis was highly accurate, with a kappa value of 1, P < .001 in duct-dependent systemic and pulmonary circulations. In non duct dependent lesions requiring urgent surgery as well as non-urgent surgery, there was substantial agreement, kappa value of 0.723 (95% confidence interval [CI]: 0.549-0.897) and 0.740 (95% CI: 0.575-0.905), respectively. Performance of the APP tool did not vary significantly across the four spectrum of CCHDs (Fisher’s exact test, P value .565). Conclusions The APP tool is promising as a bedside aid in diagnosis of common CCHDs and has the potential to bridge the gap in early detection.