Sensors invade the medical world

Take a tour of your local hospital and there's one thing that probably won't cross your mind--sensors. It's surprising, actually, since the place is probably chock full of them. In fact, without sensors most modern medical facilities would be nearly tossed back to an era of leeches and blood letting.

"Disposable blood-pressure transducers are the single largest application of sensors in the world today," says Roger Grace of Roger Grace Associates, a San Francisco based sensor-industry analyst and consulting company. "We're talking about sixteen to twenty million devices per year."

Since Novasensor (now Lucas Novasensor, Fremont, CA) introduced disposable blood pressure transducers (DPT) to the world in 1982, this application has demonstrated that a sizable market exists for accurate, inexpensive sensing solutions. But pressure transducers aren't the only application of medical-sensing technology. Other areas of interest include:

- Stress and strain. Piezoelectric thin and thick film sensors from such folks as AMP Sensors (Valley Forge, PA) are extremely versatile. They can be turned into impact, vibration, sound, and force sensors--to name a few--used in everything from ultrasound machines to electronic stethoscopes.

- Accelerometers. Finding their way into rate-adaptive pacemakers, accelerometers can also be used for such potential applications as balance prosthesis and diagnostic equipment.

- Infrared. Silicon-based microelectromechanical systems (MEMS) technology has enabled more accurate, less expensive infrared gas analyzers used in anesthesia.

- Chemical. Non-invasive blood glucose monitors, fluorescing sensors, and other near-term technologies could be used to detect certain proteins in blood or form the basis of an "electronic nose." "Chemical sensing will be a major growth area, especially for MEMS," says Grace.

- Microfluidic. Another highly active area for MEMS research, microfluidic DNA analyzers could dramatically increase the accuracy of tests for sexually transmitted diseases while cutting the time it takes to run the tests from days to maybe 20 minutes.

Monitor on the move. Continuous monitoring of blood pressure with DPTs might be common in hospitals, but most of us are probably more familiar with the pressure cuff and stethoscope method used during doctor's office visits. Trouble is, some people are a little too used to them, and they actually show short-term elevated readings attributed to nervousness, a phenomenon called "white coat hypertension."

"It's believed that up to twenty percent of people might show this effect to some degree," says Paul Voith, signal processing engineer in the Cardiology Div. of Marquette Medical Systems. Voith and the company, though, believe they have a solution. Why not monitor a patient at discrete intervals all day and see what his real-world blood pressure is?

Still in evaluation and not yet FDA approved, Marquette's Ambulatory Blood Pressure system is intended to do just that. It consists of a small, hip-worn case containing battery-operated electronics and pneumatics used to inflate a pressure cuff on the patient's upper arm. As it deflates at a controlled rate, a piezo-film sensor sandwiched between the patient's arm and the cuff tracks the opening and closing sounds of the brachial artery--called Korotkoff sounds--similar to the way a doctor listens with a stethoscope.

From this information the usual systolic (high value) and diastolic (low value) readings can be gathered and saved to non-volatile EEPROM memory. The system is intended to be worn continuously over a 24-hour period with readings programmed to be taken every 3 to 30 minutes.

Other such systems exist, but Marquette's is patented using a technique Voith developed for extracting the signal from what can be substantial noise due to the environment and patient movement. "These types of devices are all called ambulatory, but there are not many that you can actually ambulate with," Voith says. "Most require you to stop what you are doing and hold still during the readings."

To maximize the signal and reject noise, Voith's method involves the use of two pairs of piezoelectric polyvinylidene fluoride (PVDF) film sensors from AMP Sensors laid out in a linear array 6 cm long. Each sensor measures 2 X 1 cm. The first and third sensor form one pair while the second and fourth form the other. Center-to-center spacing between two sensors in a pair is 22 mm--roughly 1/4-wave length for the Korotkoff sounds that are centered at 20 Hz.

Gathering pairs of readings serves two purposes. One, it gives a greater total signal strength. Two, each sensor will read the Korotkoff sound at a slightly different time, and this can be used to differentiate the pulse from ambient noise which all the sensors will see essentially at once.

Competing systems use a similar method, measuring two signals 1/2-wave apart and then subtracting the signals. "This gives about a 3-db gain," says Voith.

"We do a little higher-priced math to obtain just the part of the signal which is phase shifted versus that which is not," he explains. This has the advantage of rejecting common mode noise as well as all noise that is in phase with the sensors, even if it is not of the same magnitude--something simple subtraction cannot do.

The higher-math method also makes it possible to use a thin, flexible substrate, increasing patient comfort. Competing arrays often have a stiff substrate in an attempt to commonly couple the sensors for better noise rejection. "We have no need to isolate the sensors from the rest of the world," says Voith, "because we do our noise rejection electronically."

Marquette's ABP also captures readings of pressure oscillations from the inflatable air cuff. These are highly affected by motion and muscular contractions, but can provide additional data points for correlation.

Voith notes that possibly the most challenging aspect of the project involved finding the best pattern for the piezo sensors. Ultimately, he evaluated a dozen different designs before hitting upon the layout featured in the prototype. "It took awhile, but we finally found one that works quite well."

Electronic stethoscope. Piezo film also forms the basis for a versatile monitoring system developed by FlowScan (San Mateo, CA). Called the LifeFlow sensor, it can be used to track blood flow, respiratory, digestive, or other internal body sounds even under less than ideal conditions--say, inside an ambulance or rescue helicopter.

It consists of two of AMP Sensors' piezoelectric elements bonded to opposite sides of a flexible substrate. Overall, the sensor measures 31.7 X 25.4 mm with a thickness of less than 1 mm.

The package is secured to a patient's skin with hydrogel at the desired monitoring site--over the heart for blood-flow monitoring, for instance, or near the throat to track respiration. Output signals can be passed to FlowScan's portable LIfe Sound Amplifier (LISA)to drive an attached stethoscope or, for in-office situations, to a computer workstation for graphical display.

Though it might seem that the sensor is simply a microphone, in actuality it responds to minute vibrations of the skin. "It senses tiny amounts of flexure, not sound," says Jim Kassal, vice president of product development at FlowScan. Bending the center of the sensor relative to the edges by just 1 micron, he says, generates an 80-mV signal. Sounds outside the body tend to move the entire sensor as a unit, but not bend it. "And if it isn't bending, it doesn't care if it's moving," Kassal explains.

This makes LifeFlow ideal for use by emergency technicians in chaotic situations where they need to be able to clearly differentiate sounds the patient's body is making from the surrounding din.

When the sensor package is bent, each element experiences dynamic strain that is opposite in sign to the other. Those signals are sent to a difference amplifier and then the two are subtracted from each other to extract the desired patient's sounds from the background.

One challenge was to find a connector that wouldn't transmit vibration from the cord to the sensor. After looking extensively--and unsuccessfully--for a stock connector that was shielded, small, and had four spring contacts, Kassal opted to design his own. By making it much more compliant, he was able to eliminate the need to tape the cord to the patient which had been a minor annoyance.

Kassal conceived the LifeFlow sensor as a derivative of previous work designing piezoelectric sonar transmitters for the Navy. "The sensor is the crucial element," says Kassal. "Once you have a signal, you can do all sorts of neat stuff with it. The trick is getting the signal."

Anesthesia analyzer. An old medical anecdote ends with a patient saying to his anesthesiologist, "so you're the guy who makes sure I go to sleep." To which the doctor replies: "No, I'm the guy who makes sure you wake up."

It's sobering, but factual. More than 20-million Americans undergo general anesthesia during the course of a year. For the vast majority, it will be as close to death as he or she will ever come without dying.

One tool the anesthesiologist uses to keep his patient's unconsciousness to just the right level is a respiratory gas monitor. A new system just entering pilot production was developed by Ohmeda Medical Systems (Louisville, CO). It offers to improve the speed and accuracy of such units by means of a unique multi-element pyroelectric infrared detector fabricated from a copolymer of vinylidene fluoride and trifluoroethylene, more conveniently known as p(VDF-TRFE).

Infrared monitors have been used for years to measure CO(sub2), N(sub2)O, and one or two of five possible anesthetics. Usually, a small pump draws a continuous sample (100 to 200 ml/min) of the patient's respired air into the monitor. Infrared light is passed through the sample, and by measuring the amount of light absorbed at specific wavelengths, the system can determine the quantities of the various gases.

For example, CO(sub2) absorbs IR radiation at the 4.26 micron wavelength. To detect this gas the system has an optical filter with a pass band at 4.26 microns matched with a detector sensitive at that wavelength. Past respiratory gas monitors used a discrete number of filters centered at each absorption wavelength desired to be measured. But as the number of anesthetic agents has increased, it has become more difficult to monitor them.

Ohmeda's solution was to increase the number of IR filters from about 6 to 74, covering the spectrum from 7 to 10 microns in wavelength. The broad spectrum makes it easy to simultaneously identify and quantify not only CO(sub2) and N(sub2)O, but all five common anesthetic gases.

The heart of the new system is a Linear Variable Filter (LVF) Detector. It consists of 74 separate elements, custom ASICs for signal amplification and sampling, custom optical filters, and a ceramic PGA package. "It's the only detector out there that has been made from p(VDF-TRFE) with this many elements," claims engineer Chuck Logan.

The detector arrays are created by silicon micromachining. The active structure is built up on top of the wafer, and the wafer is back etched to release the copolymer membrane with the detector pixels already in place. The membrane measures 3.5 X 6.5 mm, but is less than 2 microns thick.

In operation, an IR light source is alternately directed via a rotating chopper drum to one of two optical paths consisting of spherical and flat mirrors. One path passes the IR beam through a cell containing the continuously flowing sample of the patient's respiratory gases. The other path passes the beam through a reference cell containing no absorbing gases. In both cases, the beam strikes a 74-pixel detector array. By comparing the detector output of the sample to the reference, the quantity of the various gases can be determined.

Not limited to medical applications, LVF Detector could be designed to measure semiconductor processing gases, says Logan.

Accelerating the pace. Designing and manufacturing pacemakers would probably give an ordinary person a heart attack. They must be supremely reliable, since liability and competition are so stiff that many manufacturers will discuss their devices with the press in only the most guarded fashion. However, one thing is certain--sensors can make pacemakers better.

Many pacemakers now employ accelerometers or other motion sensors which allow the unit to adjust the heart rate to the patient's perceived activity level. "The original fixed-pace pacemakers maintained people's lives but limited the quality of life," says Stuart Ferguson, applications development manager for VTI Hamlin (Helsinki, Finland).

Engineers at the company have adapted their 2g automotive airbag accelerometer to use with pacemakers. A micromachined capacitive design, it works by suspending a small silicon mass, plated with electrodes, between two sheets of silicon. Under acceleration, the mass deflects slightly, changing the relative capacitance of the sensor in proportion to the acceleration event.

VTI Hamlin's design offers several advantages, Ferguson says, over competing designs. Primarily, their sensor, featuring hermetically sealed elements, can be mounted directly onto the pacemaker circuit board, and can be handled with conventional pick-and-place manufacturing equipment. It also has the advantage of low power consumption and the ability to output dc information--static position--as well as ac acceleration values. This allows it to provide an accurate measurement of a patient's position through a range of ±90 degrees, thus enabling the user's heart rate to be boosted when changing from a prone to a standing position.

Protective of his clients, Ferguson mentions that the sensor is already in use in Europe and is under evaluation with companies in the U.S. "The ability to integrate the sensor with the pacemaker circuitry is a nice advantage," he says.

Cybercontacts

You can reach the following companies mentioned in this feature on the Internet. Please tell them that you were referred by Design News.

AMP Sensors: http://www.amp.com/sensors/sensors.html

Cepheid: http://www.cepheid.com

Marquette Medical Systems: http://www.mei.com

Ohmeda: http://www.ohmeda.com

VTI Hamlin: http://www.vti.fi

For more information

To learn more about the technologies described in this article call 1-800-828-6344, x 011 and key in the specific Product Code below

Piezoelectric film from AMP Sensors:Product Code 4286

DNA analysis systems from Cepheid: Product Code 4287

Accelerometers from VTI Hamlin: Product Code 4288

IR gas sensors from Ohmeda: Product Code 4289

Ambulatory Blood Pressure systems from Marquette Medical: Product Code 4290

LifeFlow sensors from FlowScan: Product Code 4291