"CardioVal: A Clinician-Verified Multimodal Dataset of Single-Lead IoMT ECG and Clinical Metadata for Cardiovascular "

"This work directly addresses a critical bottleneck within the scope of JBHI: the ongoing tension between large-scale automated data repositories and real-world ecological validity in healthcare informatics. While artificial intelligence in cardiovascular medicine relies heavily on benchmark datasets, many available resources are either collected under highly controlled conditions or depend on automated diagnostic labels that strip away real-world signal noise and missing clinical variables. To bridge this gap, we introduce CardioVal, a pilot-scale, clinician-verified multimodal dataset linking 60-second single-lead Lead II ECG recordings (acquired via a custom AD8232-based Internet of Medical Things hardware prototype from 71 adult patients at Bir Al-Ater Public Hospital) with 29 structured clinical metadata attributes. By preserving real-world signal artifacts, acquisition noise, and high clinical missingness (such as laboratory variable missingness reaching up to 74%) rather than artificially imputing them, CardioVal provides a vital, clinically grounded resource for benchmarking data quality, missing-data handling, and model robustness under practical IoMT constraints."