Handheld echocardiography for rural cardiac triage in primary care: A feasibility and diagnostic agreement study of a task-shifting model with remote cardiologist validation

Abstract

Introduction Access to formal echocardiography is often delayed in geographically remote settings. This study evaluated the feasibility, image quality and diagnostic agreement of focused handheld echocardiography performed by a trained Family Medicine Specialist (FMS) in a rural Malaysian clinic with asynchronous cardiologist validation. Materials and Methods We conducted a retrospective observational study of 41 adult patients who underwent focused handheld echocardiography for dyspnoea or suspected cardiac symptoms at a rural primary care clinic in Sarawak, Malaysia, between 24 August 2025 and 5 February 2026. A trained FMS performed focused examinations during routine care. Stored images were reviewed asynchronously by a consultant cardiologist blinded to the FMS interpretation. We evaluated workflow feasibility, cardiologist-rated image quality and agreement for classification of left ventricular (LV) systolic function as normal or reduced. Results Forty-one patients were included (mean age 60.4±17.5 years; 51% male). Median scan duration was 8 minutes (range 3–14), and technical difficulties were reported in 19 examinations (46.3%). Nevertheless, all 41 stored studies contained sufficient images for cardiologist classification of LV systolic function. Endocardial border definition was rated good or fair in 87.8% of studies, whereas colour Doppler was frequently non-diagnostic. The FMS identified reduced LV systolic function in 13 patients (32%), compared with 10 patients (24%) by cardiologist review. Overall agreement was 82.9%, with Cohen's kappa 0.58. Sensitivity was 80.0%, specificity 83.9%, positive predictive value 61.5%, negative predictive value 92.9% and overall accuracy 82.9%. Conclusion In this rural primary care setting, focused handheld echocardiography performed by a trained FMS was feasible and showed moderate agreement with remote cardiologist interpretation for identifying reduced left ventricular systolic function. These findings support a hypothesis-generating service-delivery model in which handheld ultrasound may assist rural cardiac triage and referral prioritization, but larger multi-site studies are needed before practice-changing implementation can be recommended.