Echocardiographic Screening for Diagnosis of Rheumatic Heart Disease by Nurses: A Diagnostic Accuracy Study in Nepal

Abstract

Background and Aims: Growing evidence suggests that training nurses in echocardiographic screening for latent rheumatic heart disease (RHD) detection could be a promising strategy, especially in settings where cardiologists are often inaccessible. This diagnostic accuracy study evaluated the feasibility and accuracy of nurse-led RHD focused echocardiographic screening in Nepal.

Methods: The prospective diagnostic accuracy study consisted of three phases: training, screening and testing. Four nurses were trained in RHD-focused echocardiography for six weeks using a structured training module. Nurses were tested on an enriched, conveniently selected sample of 100 children aged 6–16 years (36 with RHD, 64 without) to ensure adequate statistical power for sensitivity estimates. Findings were compared with standard echocardiographic assessments performed by an expert pediatric cardiologist. Diagnostic accuracy was assessed using sensitivity, specificity, area under the ROC curve (AUC) and Cohen's kappa.

Results: Among all nurses, mean sensitivity and specificity for any RHD were 87.5% (95% CI: 84.3–90.7%) and 95.3% (95% CI: 93.2–97.4%), respectively, with almost perfect inter-rater agreement (κ = 0.84, p<0.001). AUC ranged from 0.90 to 0.95. Sensitivity for borderline RHD was markedly lower at 47.9%. The mean time difference between nurses and the cardiologist was modest but statistically significant (00:32 min:sec for RHD-negative, 01:31 for RHD-positive cases; p<0.05).

Conclusions: Nurse-performed echocardiographic screening following brief structured training demonstrated acceptable diagnostic accuracy for RHD in school-age children. Within a two-step model (nurse screener followed by cardiologist confirmation), task-shifting represents a scalable and equitable strategy for early RHD detection in resource-limited settings such as Nepal.