Metrological Characterization of a Method for Blood Pressure Estimation Based on Arterial Lumen Area Model
Autor: | Shervin Shirmohammadi, Izmail Batkin, Hilmi R. Dajani, Iraj Koohi, Voicu Groza, Saif Ahmad |
---|---|
Rok vydání: | 2017 |
Předmět: |
Mean squared error
020208 electrical & electronic engineering Diastole 02 engineering and technology 030204 cardiovascular system & hematology Standard deviation law.invention 03 medical and health sciences 0302 clinical medicine Blood pressure Pressure measurement law 0202 electrical engineering electronic engineering information engineering Electronic engineering Oscillometry Measurement uncertainty Waveform Electrical and Electronic Engineering Instrumentation Mathematics Biomedical engineering |
Zdroj: | IEEE Transactions on Instrumentation and Measurement. 66:734-745 |
ISSN: | 1557-9662 0018-9456 |
Popis: | Accuracy of blood pressure (BP) measurement is a challenging issue in oscillometry. Most of the automated noninvasive BP monitors estimate BP from envelope of the measured oscillometric pulses. The peak and the trough of the oscillometric pulses are very sensitive to noise caused by breathing, heart-rate variability, motion artifacts, muscle contraction, and environmental noise. Therefore, accuracy of the estimated BP based on the oscillometric waveform envelopes is not reliable in some cases. Recently, employing a modeling approach to estimate BP, we obtained the accurate results for a set of healthy subjects. The method is based on the lumen area oscillations model and estimates BP by comparing the actual and corresponding simulated waveforms. The method’s accuracy worsened when we tested it on a broader range of healthy subjects, while a significant drop was observed when the method was used for patients with chronic cardiovascular diseases. The work presented in this paper represents an improved version of our previous approach. We tested the proposed method on both healthy subjects and patients with chronic cardiovascular diseases, and compared the results to two popular BP estimation algorithms: maximum amplitude algorithm and maximum/minimum slope algorithm. We observed up to 56.7% and 57.3% improvements in mean absolute error, 98.9% and 64.4% improvements in mean error, 50% and 59% improvements in standard deviation of errors, and up to 57.6% and 55.8% in measurement uncertainty for the estimated systolic and diastolic pressures, respectively. |
Databáze: | OpenAIRE |
Externí odkaz: |