Development and internal validation of QRSense: A portable tool for manual and structured measurement of electrocardiograms in digital tracings

Abstract

Introduction

Manual electrocardiogram (ECG) measurement entails inter-observer variability that affects clinical reproducibility. QRSense was developed as an open-source, offline digital tool to standardize this process.

Objective

To describe the technical characteristics and to evaluate the measurement agreement and reproducibility of QRSense compared with conventional manual measurement.

Methods

A client-side web application based on HTML5 Canvas was developed for vector-based measurements without automation. Subsequently, a cross-sectional concordance study was conducted using 50 consecutively selected ECG tracings. Eight independent observers measured QRS duration and voltage, divided into two groups of four observers each (four using QRSense and four using manual methods). Agreement between methods was assessed with Bland–Altman analysis (bias and 95% limits of agreement) and Lin's Concordance Correlation Coefficient (CCC) as the primary metric of exact agreement.

Results

Excellent exact agreement was observed for both parameters: Lin's CCC was 0.9527 for QRS voltage and 0.9221 for QRS duration. Bland–Altman analysis showed a negligible mean bias of +0.0482 mV for voltage (95% limits of agreement: −0.0958 to +0.1922 mV) and a small bias of −1.32 ms for duration (95% limits of agreement: −16.24 to +13.59 ms), without proportional trends across the measurement range.

Conclusion

QRSense is a valid and reliable tool. It offers measurement agreement comparable to expert human measurement and supports reproducible quantification of ECG parameters, facilitating the standardization and auditability of electrocardiographic data for research and clinical practice.