Improving Health Outcomes with Biomedical Sensors

Read Transcript

Signal processing really is about extracting information from time series of data, so you can have your senses connected to a patient whether they are in hospital or whether they are making short term measurements in their homes, that gives you a time series very often of data which has got gaps in it, which has got noise and signal processing is about extracting the information in this case clinically useful information.

If we are going to be asking patients to self monitor in their home or something, I think one of the principles we have discovered in the last few years, is should be done at minimal cost, not financial cost, that cost to them to their daily routine. So we can't ask them to put on 10 different sets of electros, monitor themselves for half an hour now, we might be able to have one sensor for 30 seconds.

So we are going to have a very small, short, noisy snapshot and you need single processing rate to extract that. Does that information tell me that the patient is worse than yesterday. Every hospital patient has this finger probes where the hospital measures, your pulse rate and your oxygen levels.

We can have Bluetooth those used up in the home through signal processing. We can get other perimeters, we can surely get the breathing rate, we might be able to get some information about the blood pressure as well so signal processing is adding value to the simple probe. I work with a company that bears well which develop a patch that you stick on, you can stick on for seven days.

They use it for medication adherence but it could also record the very same vitals signs. The advantage of that it could breath on for seven days, you can keep it in the shower or even if you go swimming e.t.c. so here you have got something which is very unobtrusive. And finally the least obtrusive our wish is remote monitoring, with non contact.

When I'm skyping my children or my friends e.t.c. the webcam that is filming me, from that I can derive the pulse rate, the breathing rate and the oxygen level of the patient. No cost at all to the patient. They'll be using this tablet, or this laptop, for example, let's say to Skype or to look at pictures of their grand kids e.t.c.

and at the same time we will be getting the information for free. When we have people without a stroke, they will go to a local hospital, they will have MRI imaging, they will have CT imaging, they'll then have physiotherapy, they'll spend on average three weeks in hospital, $10, 000.

Then they go home with one measurement in the out-patient clinic, a blood pressure measurement, a single measurement usually, a $10 measurement after we spent $10, 000 on them to set their medication for hypertension, high blood pressure once they go back home. Now, elevated high blood pressure is the highest risk factor in terms of having another stroke that could kill you or maim you for life, and we do this in one measurement.

So, I went to talk to the stroke physician and said, why don't we give them one of these wireless health app? So, every patient there that leaves the hospital after having had a stroke in our local hospital is given a Bluetooth bloodpressure monitor, a cell phone and a very simple explanation of what they have to do, which is to take their blood pressure three times a day, and of course in the morning, lunchtime and evening the blood pressure varies.

So, as opposed to this one measurement in the outpatient clinic we now have three measurements a day for a month, or a maximum of three months which it might take to adjust their therapy. The frightening thing we're coming now to five years of doing this, the frightening thing is two-thirds of patients have had their medication adjusted as a result of this home monitoring and after they leave the hospital, two-thirds.

So, it was wrong in two-thirds of the cases, the good thing is we're getting far fewer patients coming back now with the major strokes.