Medical Device Link.

Originally Published MEM Spring 2006

SEMICONDUCTORS

A simple platform built from off-the-shelf components can gather, process, and transmit patient medical data at a moderate cost.

Matt Maupin

The pressure to change is nowhere felt more acutely than in the healthcare industry. Institutions and service providers are pushed to cut costs, but expected to do so—a nice trick—while continually improving the quality of life for patients. Reducing staff, equipment costs, and power consumption and also limiting the length of hospital stays can all help control healthcare costs, but perhaps at some cost to the patient.

It is a significant challenge to deliver ever better care at a reasonable cost. Healthcare providers are counting on technology to help them meet that challenge successfully.

Semiconductor technology is having a profound effect on healthcare systems. Patient-monitoring, drug-dispensing, and other semiconductor-based information technology (IT) applications can supplement scarce manpower resources by capturing raw information relating to patients, rapidly processing the data, and then distributing information that is critical through an in-hospital network to those who need it.

Monitoring systems can send streaming data to—and receive it from—almost anywhere in the world. This not only provides pinpoint distribution without requiring additional human resources, it also allows many applications to move into any home that has an Internet connection. In some cases, a similar system can control treatment activities, such as dispensing drugs on the basis of a predetermined schedule.

Performing this kind of remote monitoring and data processing starts with a simple platform that has sensing, control, and communication functionalities, possibly including a radio-frequency (RF) transceiver for greater portability and ease of operation. Such simple networking technology is the focus of this article.

The Basic Chip Set

Figure 1. An example of a chip set for healthcare IT applications that provides sensing, control, and ZigBee radio-frequency capabilities. It is based on the MC9S08GT family of microcontrollers and MC1319x family of low-power transceivers from Freescale Semiconductor (Tempe, AZ). (click to enlarge)

An uncomplicated chip set consisting of three off-the-shelf components can provide all the functionality needed to acquire, process, and transmit physiological information at a reasonable cost (see Figure 1). The microcontroller processes the information received from the sensor and then sends it to the low-power transceivers. This could be any of a number of commercially available general-purpose 8-bit microcontrollers with an analog-to-digital convertor (ADC) and the necessary embedded memory. The memory requirement is dictated by the networking standard featured on the wireless transceiver, for example, simple proprietary wireless protocols, IEEE 802.15.4, or ZigBee.

Further, the microcontroller and transceiver can be combined into a single package to create a more compact and efficient solution. Figure 2 illustrates a system-in-package, or SiP, that features an IEEE 802.15.4– compliant wireless transceiver and a microcontroller core.

Sensor Technology

Figure 2. An SiP featuring an MC 1319x-compatible transceiver and an 8-bit HCS08 microcontroller core. (click to enlarge)

Sensors are the contact points in patient condition-monitoring applications. In a chip-set networking platform, all the other components are used to process and disseminate the information the sensors gather. Both pressure sensors and acceleration sensors can be employed in a broad range of healthcare monitoring equipment.

A typical piezoresistive transducer, a commonly used kind of pressure sensor, consists of an etched silicon diaphragm upon which a piezoresistive element is implanted. Piezoresistive materials change their resistance under physical pressure, such as increased air pressure. A silicon piezoresistive pressure sensor provides a highly accurate and linear voltage output that is directly proportional to the applied pressure.

Acceleration sensors reflect changes in motion through varying voltage output levels. An advanced acceleration sensor is significantly more complex than a pressure sensor. It consists of two surface-micromachined capacitive sensing cells (g-cells) and a signal-conditioning application-specific integrated circuit (ASIC) enclosed in a single IC package. The sensing elements are sealed hermetically at the wafer level by means of a bulk micromachined cap wafer.

A mechanical structure formed using semiconductor processes, the g-cell can be modeled as a set of beams attached to a central mass that moves between the beams. The beams form two back-to-back capacitors. As the central mass moves with acceleration, the distance between the beams changes, causing each capacitor's value to change. The ASIC uses switched capacitor techniques to determine the g-cell capacitor values and extract the acceleration measurement from the difference between them. The ASIC conditions and filters the signal also, providing a high-level output voltage that is ratiometric and proportional to acceleration.

The g-cell architecture can be used to develop various acceleration sensors. In fact, a three-axis low-g acceleration sensor that measures motion along the three spatial axes has been developed in addition to a reference board to help visualize accelerometer applications in such markets as medical monitoring. In the past few years, acceleration sensors have changed dramatically; their sensitivity has increased, power consumption has been reduced, and package size has shrunk. Tools such as the reference board that demonstrate these sensors' capabilities assist developers in making optimal use of the new technology advances.

Applications

Monitor sensors provide 3- or 5-V analog output that must be processed before it can be transmitted and stored for healthcare personnel to analyze. An inexpensive general-purpose 8-bit microcontroller with an ADC is sufficient to do the job. Essentially, a sensor generates analog signals and inputs them to the ADC, which then processes them into digital signals to be used by computers or monitors through a communication interface such as an RF transceiver or a direct universal serial bus (USB) connection. A more accurate ADC—say, a 10-bit device—will ensure cleaner input and output signals.

Eight-bit processing is generally sufficient for handling the data rates and algorithmic compression and decompression functions, but the amount of flash memory required is determined largely by the communications protocol used by the transceiver. For instance, processing ZigBee-compliant data requires more memory than simple point-to-point wireless connectivity does. The combination of sensor and 8-bit controller makes a compact, cost-effective system that consumes little power yet gathers and processes large amounts of useful data. These data can be transmitted to a remote location via an associated RF device or can be used to initiate on-the-spot medical treatment. Some sensors and microcontrollers cost so little that they can be used in disposable systems that may operate for only a few days before being discarded and replaced.

An interesting disposable application is the OmniPod insulin management system from Insulet Corp. (Bedford, MA). Designed to replace conventional insulin pumps that require users to carry around cumbersome components, the system has two fully integrated wireless components: a small, self-adhesive, disposable insulin reservoir and delivery system with a wireless receiver, and a handheld device with a microcontroller and RF transmitter that the patient uses to program the system with customized insulin-delivery instructions. It delivers insulin, monitors its own operation, and automatically stores patient records. The semiconductor and shape-memory-alloy-wire activation components in the device are so inexpensive that disposing of the entire unit when, after a few days, the insulin is gone is completely reasonable. Moreover, the wireless capabilities of a controller unit similar to that of the OmniPod could be expanded for connection to an Internet gateway to allow communication of test results and dosage information directly to a healthcare provider.

Another example involves the use of sensor technology not only to monitor patient symptoms but also to determine the conditions under which the symptoms appear. For patients suffering from incontinence, a wearable device that includes pressure sensors, a microcontroller, and an RF transceiver can detect even minute leakage and transmit time and volume data for storage and analysis at a later time. If such a system included acceleration sensors, the patient's activity—whether lying down, standing, walking, or running, and so on—could also be recorded, helping caregivers determine whether certain circumstances trigger the leaking and, possibly, devise more-effective countermeasures.

The Wireless Connection

Wireless communications technology maximizes the convenience, for both the patient and the caregiver, of a portable sensing and processing device. It also gives devices the ability to transfer data seamlessly to other networks, such as hospital Ethernet or controller-area network (CAN) systems, or to the Internet for worldwide accessibility. The family of transceivers characterized in Figure 1 features a four-wire serial peripheral interface (SPI) connection that allows direct communication with a wide variety of microcontrollers. The software and processor can be scaled to fit applications ranging from simple point-to-point systems through full ZigBee networking.

A simple proprietary wireless connection requires only a physical layer (PHY) and simple media access controller software (SMAC) for point-to-point and STAR network applications. Requiring less than 2.5 Kbyte memory, the simple wireless connection is useful for highly cost-effective applications that have extremely low power requirements, such as portable disposable medical monitoring devices.

In 2003, the Institute of Electrical and Electronics Engineers (IEEE) released its 802.15.4 standards-based simple packet data protocol for lightweight wireless networks. The robust communications and time-critical protocol consists of an 802.15.4 PHY and a media access controller (MAC). IEEE 802.15.4 works well for applications requiring long battery life and selectable latency for controllers, sensors, and remote monitoring. It provides a greater level of security than a simple wireless connection and can be used in mesh and cluster-tree network applications.

Figure 3. A comparison of wireless communication platforms suitable for medical monitoring applications that use microcontroller cores and wireless transceivers. (click to enlarge)

The capabilities of the IEEE 802.15.4 protocol standard can also serve as the basis for a fully embedded ZigBee-compliant platform. The ZigBee stack includes network and security layers, the application framework, and application profiles that can be loaded onto an 802.15.4-compliant transceiver. Figure 3 compares simple wireless connectivity, 802.15.4-compliant, and ZigBee platforms, including microcontroller memory requirements and wireless transceivers.

ZigBee has several attractive features that make it a strong alternative choice for healthcare monitoring applications. Many applications are in active operation only periodically, and then for short durations. To minimize power consumption, and thus greatly extend battery life, ZigBee and Bluetooth can sleep for long periods of time, waking up and communicating only when necessary. Wi-Fi (IEEE 802.11b) cannot operate efficiently in a sleep-wake-sleep mode, so it must be on all the time, making it unsuitable for many applications.

Both ZigBee and Bluetooth can operate at very low duty cycles (a 1% duty cycle would characterize an application that is active only 1% of the time), but latency issues work in ZigBee's favor. Although it may take several seconds for a Bluetooth device to wake, synchronize with the network, and begin communicating, ZigBee does not have to synchronize before communicating. This cuts the wake-to-communicate latency to about 27 milliseconds, which allows for more-precise scheduling and just a little more battery life for each wake cycle.

ZigBee can also be scaled from simple point-to-point applications to large mesh network environments through the addition of memory and the appropriate software.

Extending the Communications

Hospitals and other healthcare facilities increasingly are relying on network communications to move massive amounts of data and make information more accessible for those who need it. A CAN, a type of network originally developed for automotive applications, can connect a cluster of instruments and monitors that rely on shared information. The CAN, in turn, can port to an Ethernet backbone for facilitywide distribution. But most of the equipment used is portable in one way or another; usually, it is stationed on a cart or rolling stand. Short-range wireless communication allows a permanent connection between the instrument and the facility backbone, even while a cart is being rolled from one room to another.

In a similar respect, home-healthcare monitoring devices can connect to the Internet through cable or digital subscriber line via a Wi-Fi access point or home networking gateway. Once placed on the Internet, critical patient-monitoring information can be accessed through dedicated Web sites that can also provide healthcare personnel with quick access to the patient's medical records. It is all a matter of managing the flow of data from multiple sources so that it can be analyzed efficiently for a faster and more accurate doctor-to-patient response.

Medical Monitoring Gateways

Figure 4. Block diagram of a medical monitoring gateway that uses an embedded processor. (click to enlarge)

Healthcare personnel today have a wider variety of sources of information about patients than ever before, yet the multitude of facts can be an impediment to clear medical decision making if the data are not interconnected, analyzed, and recorded for useful reference in the event of an emergency. An important emerging application is the patient-monitoring gateway that enables healthcare personnel to monitor and interface with equipment both locally and from a central base station (see Figure 4). Equipment such as respirators, heart monitors, and medicine dosage controllers located in different hospital rooms can be securely and remotely monitored from a central base station. The local bedside medical equipment may be connected to a nearby microprocessor-based medical monitoring gateway via Ethernet or, in the case of legacy medical devices, through a serial RS-232 port. The gateway in each room is normally wired to a central router that is in turn connected to the central base station.

Microprocessors used in medical gateways must offer connectivity, security, and product-lifetime longevity. The latest generation of microprocessors designed specifically for this application offer such connectivity options as multiple 10/100 Fast Ethernet controllers, USB connection, and queued serial peripheral interfaces. They help to provide secure point-to-point communication across an insecure Ethernet network without any sacrifice of system performance. The microprocessor encrypts the data received from the bedside equipment before they are transmitted by Ethernet to the central base station. At the station, the data are decrypted and interpreted.

In addition to being able to monitor bedside medical equipment from the remote central base station, healthcare personnel may want to access the same data while they are in the patient's room, by means of a USB interface. For instance, the system could be accessed through the gateway's USB-to-Ethernet adapter port using a personal digital assistant with a USB On-The-Go (OTG) port. USB-OTG is a new supplement to the USB 2.0 specification that augments the capabilities of existing mobile devices and USB peripherals by adding host functionality for connection to USB peripherals. With this additional USB feature, a healthcare provider can access the record of the patient's condition upon every visit to the bedside.

VoIP Applications

Often there is no substitute for vocal communication. An interesting technology spreading through the consumer and business markets that has great potential application for healthcare delivery is voice over Internet protocol (VoIP), which provides the additional service of voice communications. VoIP can be built right into a hospital bed, for instance, to provide wireless voice communication with a nurse station or points further removed. Because information is traveling as packets over the Internet, it can be routed almost anywhere—to a central handling area that may serve a number of hospitals or to an individual's cell phone, accommodating an extended doctor-patient relationship. Also, with VoIP, text and voice data can be stored directly in a database, allowing rapid retrieval and avoiding bulky tape storage systems.

Conclusion

To record, process, transmit, and store massive amounts of medical data more efficiently than ever before while at the same time cutting healthcare costs is a tall order. But developments in semiconductor technology are helping the healthcare industry to do just that in hospitals and clinics around the world.

Medical monitoring, networking, and other IT tools that allow patients to spend more time at home and less time in hospitals and other healthcare facilities helps patients get on with their lives. Moreover, through advanced network communication, doctors can reach inside the home, as it were, to make virtual house calls. The benefit to healthcare providers of semiconductor-based networking technology is improved access to timely, accurate information that can be used to administer treatment more efficiently. Extending the doctor-patient relationship both into the home and, through the Internet, to any part of the world, networking technology offers advantages to patients and caregivers alike.

Matt Maupin joined Freescale Semiconductor (Tempe, AZ) in 2001. He has spent more than 14 years focusing on wireless connectivity, including Wi-Fi, Bluetooth, and ZigBee.

Copyright ©2006 Medical Electronics Manufacturing