Pressing problem

Research being conducted at Newcastle University could lead to the development of a blood pressure device that provides physicians with considerably more accurate readings.

Consecutive blood pressure measurements from the same patient can vary significantly, whether they are taken manually or automatically.

the project's principal investigator, Prof Alan Murray, and his team plan to study the causes of blood pressure measurement variability and develop and evaluate improved measurement techniques. This, they hope, will lead to the development of a prototype 'intelligent' blood pressure measurement device.

Talking, coughing or movement can cause irregular blood pressure readings. Similarly, the team has discovered that there is a strong association between blood pressure variability and variability of pressure pulses in the arm cuff during measurements.

According to research carried out by Murray, manual blood pressure measurements often vary by more than 10mmHg (millimetre of mercury) between consecutive readings.



Misdiagnosis

This can result in serious misdiagnosis; for example, a recent US report estimated that a 5mmHg error would result in 21 million Americans being denied treatment — or 27 million being exposed to unnecessary treatment.

'The ideal of the project is that if a GP or nurse takes one blood pressure reading and this device lights up green, they won't take another one. But if it comes up with a red or orange light — not that we're planning to put lights on — they will need to take the reading again. If you have an automatic device, it might just simply not display a reading until you take another one,' said Murray.

Data for the project will come from more than 1,300 blood pressure readings taken over a three-year period. The database was compiled during an EU-funded research project to develop a simulator for evaluating the accuracy of non-invasive blood pressure (NIBP) devices by enabling real, previously-recorded oscillometric waveforms to be regenerated.

'Our initial data shows that when readings are stable, the characteristics you get, such as heart rate, give an indication that blood pressure is likely to be stable,' said Murray.

The researchers also found that in stable readings, the pulse characteristics in the cuff appeared to be smooth, for instance, due to a smooth decrease in cuff pressure. If the cuff pressure decreases unevenly, however, the pressure readings would also be irregular.

'We took readings from different groups with a variety of problems. For example, young patients are more likely to have highly variable heart rates, and those with heart disease are more likely to have ectopic beats. But the most difficult group is the elderly, with atrial defibrillation.

'With this, the heart rate is very irregular, and it would be extremely difficult to find a regular sequence. It could well be that with this group, the only answer might be to take several readings and take the average.'



Problem

While the proposed new device will be useful for manual techniques, it will be particularly useful for the more prevalent automated machines. 'There's a tendency to move over to automated techniques. Some devices are much better than others, but some are very poor, which is a big problem in the health services,' said Murray. 'There have been many warnings that medical staff should not rely on these devices because they could have the wrong value.'

So to correct these anomalies, the Newcastle team will use the analysis of its 1,300 sample-strong database to create algorithms that will be incorporated into the new device.

The researchers will receive advice and expertise from the Berlin-based Physikalisch Technische Bundesanstalt, and the Royal Infirmary of Edinburgh NHS Trust will test a prototype device when it is ready, which Murray anticipates will be by 2010.