Customer Solutions

Cardiovascular diseases, such as hypertension, are very common pathologies in developed countries. Automated blood pressure measuring instruments are widely used not only in the hospital setting but even more frequently in the home-care environment and stand-alone, self-measuring facilities. These instruments are microprocessor-controlled and do not need expert personnel to perform the measurement. A microprocessor-driven air pump inflates the occlusive cuff at a fixed-pressure value. Then, it records the pressure oscillation pattern during a stepwise deflation. The measuring principle relies on the brachial arterial wall nonlinear properties, which induce nonconstant oscillations of the cuff pressure during deflation. Oscillometric waveform identification algorithms determine the mean, systolic, and diastolic arterial pressure values.  

Despite its diffusion, the oscillometric method accuracy is still controversial. Identification algorithms are mainly empirical and appear to estimate better mean and systolic pressure than diastolic pressure. Subject movement strongly influences the law of deflation and must be accurately controlled to avoid artifact presence in the oscillometric waveform. For these and other reasons, scientists still are conducting theoretical and experimental research to improve instrument reliability and establish the oscillometric-technique theoretical basis. Our team decided to develop a flexible, transparent, and easily customizable oscillometric blood pressure apparatus for research and didactical purposes.

Developing a Customizable, PC-Based Automated Blood Pressure Measuring Device      

We created the device mechanical unit by assembling an occlusive cuff, two pressure sensors, an air pump, and an on/off pneumatic valve to control the law of deflation. Moreover, we added a finger pletismograph and photopletismograph to the experimental setup for a redundant systolic pressure and arterial pulse wave velocity measurement method. We assembled the sensor amplifiers, filters, and valve driver circuitry on a breadboard for clear visualization, easy signal conditioning path customization, and the opportunity to use an oscilloscope to measure the signals at different stages.

Because we needed to acquire signals from four sensors and generate an output signal to command the valve, we chose an NI E Series DAQ device and output generation and implemented NI LabVIEW control and postprocessing software using a PC running Windows 2000. We chose NI data acquisition and LabVIEW because they work together to create a transparent and easily customizable control, processing, and storage system. Moreover, we needed to implement several postprocessing blocks to reproduce the various commonly used identification algorithms. Because we needed a comprehensible oscillometric method illustration and the capability to simply change the control law, NI software was the best choice for its flexibility, cost-effectiveness, and simple programming. Using LabVIEW, it was easy to design a friendly graphical user interface displaying the interest parameters, graphics, and waveforms in real time.

Creating Virtual Instruments for Flexible Control and Postprocessing

To impose a specific law of deflation, our application continuously measured cuff pressure and compared it to an operator-defined, software-generated law of deflation value. Assuming the application control was not time critical, we completely implemented it via software by including the control algorithm in a for loop, which every time it executed, acquired one sample from all sensors and generated one sample output to open or close the valve if the cuff pressure value was higher than or equal to the one desired. The algorithm assured the cuff pressure never exceeded an operator-defined value on the front panel (for example, 150 mmHg for normotensive patients), and, once it started deflation, the cuff pressure reduced according to the operator-defined equation and time slope. With nondeterministic control, the system recorded the true sample period continuously, and if it exceeded a critical time of 50 ms, the execution stopped and the operator must restart the measurement procedure.

We performed several trials in vitro and in vivov, finding an average sample period equal to 20 ms and never exceeding 50 ms. This sample period resulted in optimal application performance, as we can set a slope from -0.4 up to -5.0 mmHg/s with an actual resolution of 0.1 mmHg/s, which is a quite high resolution for our purposes if we consider that automated commercial blood pressure devices usually deflate the cuff at 2 or 3 mmHg/s. If the pressure falls below the 40 mmHg threshold, the control algorithm generates a signal to open the valve and quickly deflates the cuff at the end of the task. Moreover, sensor data displays on the front panel in real time, and can be saved in a file with relevant patient information for off-line postprocessing.

We implemented flexible VIs for postprocessing. One VI executes all oscillometric method calculation (heart frequency values, positive and negative peaks, and peak-to-peak amplitudes) and feeds the data to other VIs with which the operator can choose specific algorithms for calculating mean, systolic, and diastolic pressure. Using the described application, we comparatively evaluated the automated blood pressure measuring device identification algorithm efficiency.

For more information, contact:

Sergio Silvestri

Engineer

Faculty of Biomedical Engineering

University Campus Bio-Medico

Tel: +39-06-22541217

E-mail: s.silvestri@unicampus.it