Accuracy, Repeatability, and Hypertension Classification of a Card-Sized, Cuff-Calibrated Cuffless Blood Pressure Monitor: Prospective Two-Stage Validation Study (Preprint)
BACKGROUND Many cuffless blood pressure (BP) monitors need calibration with a cuff device. When test readings are compared with the calibration reading, or are taken only minutes after calibration, agreement may reflect the calibration value rather than independent measurement. OBJECTIVE This study aimed to evaluate the accuracy, short-term repeatability, and hypertension classification of a card-sized, cuff-calibrated cuffless BP monitor (Card-BP) that combines finger photoplethysmography with single-lead electrocardiography. METHODS In this prospective, single-center, 2-stage study in Jinan, China, Card-BP was calibrated for each participant with an upper-arm oscillometric monitor. In stage 2 (n=88 after exclusion of 69 of 157 measured participants for reference BP variability, as specified by ISO 81060-2), 2 observers took simultaneous mercury auscultatory readings alternating with 3 Card-BP and 3 oscillometric readings, to assess accuracy, repeatability, and performance relative to a single-reading comparator (the first oscillometric reading carried forward unchanged). In stage 1 (n=726), a Card-BP reading taken immediately after calibration was compared with the calibration reading. Classification of BP ≥140/90 mm Hg was examined in both stages. RESULTS In stage 2 (264 paired readings), Card-BP minus mercury differences were mean 0.53 (SD 7.72) mm Hg for systolic BP (SBP) and mean 1.62 (SD 6.24) mm Hg for diastolic BP (DBP), meeting the ISO 81060-2 numerical criteria. The single-reading comparator also met these criteria, with differences of mean −1.32 (SD 5.88) and mean 2.72 (SD 6.15) mm Hg, as did the oscillometric monitor. The SD of Card-BP differences minus that of the comparator was 1.84 (95% CI 0.65-3.15) mm Hg for SBP and 0.09 (95% CI −0.76 to 0.92) mm Hg for DBP. Card-BP repeatability coefficients were 11.4 mm Hg for SBP and 2.7 mm Hg for DBP (mercury: 9.4 to 9.6 and 6.8 to 7.2 mm Hg). In stage 1, Card-BP readings differed from the calibration reading by mean 0.01 (SD 7.90) mm Hg for SBP and mean −0.45 (SD 2.36) mm Hg for DBP. Among 291 stage 1 participants without diagnosed or treated hypertension, Card-BP classified hypertension against the calibration reading with a sensitivity of 89% (71/80) and a specificity of 93% (196/211). In stage 2, against mercury, sensitivity and specificity were 92% (22/24) and 98% (63/64) for Card-BP and 88% (21/24) and 100% (64/64) for the comparator. CONCLUSIONS Within minutes of calibration, Card-BP met cuff-device accuracy criteria against mercury auscultation, but so did a single cuff reading carried forward unchanged, which agreed with the reference at least as closely. Card-BP diastolic readings changed little after calibration, and agreement measured against the calibration reading overstated accuracy. Validation of cuff-calibrated cuffless monitors should use reference readings independent of calibration, report a carry-forward comparator, and test the tracking of BP changes. CLINICALTRIAL Chinese Clinical Trial Registry ChiCTR2200066035; https://trialsearch.who.int/Trial2.aspx?TrialID=ChiCTR2200066035
Authors
- Shuwen Yu
- Rong Li
- Rong Wang (ORCID: https://orcid.org/0009-0002-1740-1831)
- Zhen Li
- Qing Zhao
- Huaqin Wang
- Xi Zhang
- Fengrui Yang
- Jing Gao
- Zhongyang Shi
- Jing Li
- Rui Zhang
Publication Details
- Published
- 2026-11-20
- DOI
- https://doi.org/10.2196/preprints.117285
- Primary Topic
- Non-Invasive Vital Sign Monitoring
- Type
- preprint