Transitioning Beyond Mercury Sphygmomanometers: Diagnostic Accuracy and Agreement of Digital Blood Pressure Monitors in a South Indian Adult Population
TL;DR
The automated digital monitor demonstrates good population-level agreement and diagnostic validity, making it a reliable tool for routine screening and epidemiological surveillance, however, given the wide limits of agreement, critical clinical decisions, especially in borderline cases, should be made with caution.
Abstract
Introduction: Hypertension is a leading cause of morbidity and mortality, and its prevalence is increasing worldwide. Accurate blood pressure (BP) measurement is fundamental to clinical diagnosis and epidemiological surveillance, and any measurement errors can lead to substantial misclassification of hypertension, resulting in inappropriate clinical management and inaccurate estimates of disease burden. While automated digital monitors offer a practical alternative to manual devices, their real-world accuracy requires robust local validation during the phase-out of traditional instruments. The objective of the current study was to assess the validity and accuracy of a digital BP monitor against the gold-standard mercury device in an adult population in South India. Materials and methods: A cross-sectional diagnostic evaluation study was conducted among 262 adults attending the outpatient department of a teaching hospital in South India. Baseline BP was measured twice using the digital BP monitor (Omron HEM-7120, OMRON Healthcare Co., Ltd., Kyoto, Japan) with a one- to two-minute interval between readings. After a five-minute interval, BP was measured using a mercury sphygmomanometer (Novaphon, NOVAFON GmbH, Weinstadt, Germany). Validity, reliability, agreement, and bias were analyzed using paired t-tests, Pearson correlation, and Bland-Altman analysis. Results: Digital and mercury measurements demonstrated exceptionally strong linear correlations for systolic BP (r = 0.945) and diastolic BP (r = 0.845). The mean difference (digital minus mercury) was non-significant for both parameters: -0.167 ± 7.46 mmHg for systolic and 0.296 ± 6.63 mmHg for diastolic. When comparing the two devices, absolute differences within 5 mmHg were observed in 135 (51.5%) participants for systolic BP and 164 (62.6%) participants for diastolic BP. Categorical assessment showed substantial diagnostic agreement (Cohen’s kappa = 0.77), with a sensitivity of 84.0% and specificity of 92.9% for hypertension detection. Bland-Altman analysis revealed a negligible mean bias of −0.167 mmHg for systolic BP (95% limits of agreement (LoA): −14.79 to 14.46 mmHg) and 0.296 mmHg for diastolic BP (95% LoA: −12.70 to 13.29 mmHg). Conclusions: The automated digital monitor demonstrates good population-level agreement and diagnostic validity, making it a reliable tool for routine screening and epidemiological surveillance. However, given the wide limits of agreement, critical clinical decisions, especially in borderline cases, should be made with caution.