A dataset of paired auscultatory and oscillometric blood pressure measurements from routine clinical practice

OSC-AUS-BPM is a dataset of paired auscultatory (AUS) and oscillometric (OSC) blood pressure measurements collected during routine outpatient cardiology visits. The dataset was acquired using a CE-certified oscillometric module integrated into a custom data acquisition system, enabling AUS and OSC measurements to be obtained during the same cuff inflation–deflation cycle. For each measurement, an experienced cardiologist determined systolic and diastolic blood pressure by auscultation during device-controlled cuff deflation. The auscultatory values were entered before the oscillometric results were displayed, preserving independence between the observer-derived AUS readings and the device-derived OSC outputs. The oscillometric module recorded systolic, diastolic and mean blood pressure, heart rate, and a device-output cuff-pressure trace sampled at 5 Hz. Version 2 of the dataset contains 1,319 paired measurement records from 247 pseudonymised subjects. Each record includes contextual metadata supporting stratified analyses, including age, sex/gender, height, weight, BMI, arm circumference, cuff type, rhythm classification, hypertension classification, device measurement method and preprocessing indicators. Missing values are standardised as `null`, and paired AUS and OSC blood pressure values are complete for all retained records. The dataset was processed using a semi-automated validation pipeline that removed test-phase records, screened physiologically implausible or logically inconsistent values, corrected only unambiguous typographical errors, standardised missing values and flagged records affected by missingness or preprocessing. A public correction log is provided with the release. Directly identifying personal data are not included in the released dataset. Patient identifiers were replaced by study-specific pseudonyms, dates of birth were removed after age derivation, and timestamps were shifted before publication. The dataset is intended for method comparison, device benchmarking, signal-processing research and algorithm development using paired AUS and OSC measurements from routine clinical practice. It should not be used for population-prevalence estimation, causal inference or clinical outcome inference.