Medical device interfaces

7.2 General discussion 107

As so, challenges still remain in defining the common interfaces to these layers, so that these layers could be designed as interchangeable components.

The second major challenge on interfaces is to design and agree on how interfaces should be partitioned, i.e., where to draw the lines for the possible HW/SW interface and to place possible interface layers, and to specify these interfaces also in HW level. Again, the challenge in specifying these layer interfaces would be in predicting the needs of the future, so that these interfaces would not become the bottleneck in performance or limitation in expandability while still keeping them practical in the implementation sense. Were all this achieved, we could truly build interfaces from blocks, selecting the blocks required to match the performance or features needed, while still operating under a more common higher layer standard providing application level interoperability with other systems.

database. Technically, an improved solution would be to have the devices used in the practices and clinics connected directly to a desktop PC using a common digital interface. To fully automate the process, application level software support for the direct transfer of the recorded results into the patient records database would also be needed. This could be implemented by the patient record software manufactures if the medical devices would have a common interface standard, like the ISO/IEEE 11073. This would also enable automatic storage of measurement time, measurement device type, and additional context information. This kind of additional information would be useful in cases where a faulty device is detected and affected measurements would be traced.

At medical systems level, which usually means the hospital and healthcare information sys-tems and related electronic health records, the standardized digital interfaces of medical devices are tools for achieving interoperability, but they themselves do not guarantee it. Interoperability design for larger systems starts always at the business level, which describes business processes, real-world structures, functions, and conditions [Blo02]. When designing medical and healthcare systems, we have to always define the business case first, and then design the system to fulfill the real-world needs. This is why we cannot and need not to solve “universal interoperability of everything” at digital interface level using a single universal standard. For larger healthcare systems we need information systems on top of our interfaces to implement interoperability at higher level and to provide services customized to the users needs. Alternatively, middleware components can be used between the application and the interfaces, to provide independence from the interface technology for the application [Hei04]. As an example, if a home monitoring system requires video surveillance along medical measurements, a non-medical interface standard for the video-camera can be used parallel to the measurement device interface standard (such as ISO/IEEE 11073), and information from both systems can be combined in either middleware layer or in an information system to provide a interoperable application level service.

Information systems use XML and similar languages to form their messages and data de-scriptions to improve syntactical compatibility [Len07, Coy03]. While these are versatile and powerful for describing complex data structures in an unambiguous manner, they are not very effective in data representation in that a description of a very simple data entity often becomes long and complex. This is one of the reasons why medical information systems are not ex-tended to medical device level, but instead other standards are used to define medical device communication required for interoperability.

Medical device and interface interoperability through standardization can also introduce new problems. A medical device has an intended purpose for which its safety and effective-ness/performance has been conformed. An interoperable device can mistakenly or purposefully be used in combination of other devices for a purpose which is was not designed for. This is again a challenge for the usability and user interface design, topics which are not addresses by the technical interface standards like the ISO/IEEE 11073. The ISO/IEEE 11073 may be

7.2 General discussion 109

technically adequate and sound, but it can still fail because of issues not directly related to the interface specification, such as poor usability due to poor, complex, or non-harmonized user interfaces, or failure to impress the device manufactures.

There exists some overlap in the development of medical device standards and organizations defining the deployment of these standards. Different organizations focus on the area of interest to them. While this has benefits; parallel competing efforts may produce more new ideas and di-vision into smaller groups focusing on smaller application areas can yield more effective solutions due to tighter focus, this also wastes some of the limited and often voluntary human resources associated with standards development. Furthermore, it would be beneficial for the medical de-vice standardization field as a whole if some of these overlaps could be reduced. This would lead to faster development and deployment and more importantly a more uniform medical device interface standards field. This would reduce the temptation of using emerging interoperability standards from the consumer electronics field.

In home use, and especially in wellness types of applications, medical interfaces and standards may have to give way to more general interoperability standards such as UPnP if true co-operation between all home devices is desired. Otherwise, a medical device coordinator/gateway is required, but this can lead to overlapping infrastructure like separate wireless sensor networks for medical devices and other systems. This topic is addressed in the Section 7.3 describing future trends and work.

Wireless interface technologies

Critical care monitoring would benefit greatly from the wireless interfaces. Having devices connected by cables is a necessity that many would like to see go away [Pak05]. The cables limit free movement around the patient, the movement of patient to and from the operating theatre, and are just generally on the way of the doctors and nurses. However, the critical care monitoring is an area in which relative simplicity in use and stable operation is also needed.

The practitioners need to be able to identify quickly how or what devices are connected. Not having a wire between two devices can pose difficulties in understanding the device connections from the operator point-of-view, especially if and when devices are moved with the patient to another location. For example, a wireless device measuring patient A can be logically connected to patient monitoring system of patient B. These are challenges that require both development of the technology and new thinking in usability design. The battery life of a wireless device, and increased size and weight caused by the battery reduce the benefits of the cable free operation.

An additional problem is that the lifetime of complex and expensive medical systems is long, and the older systems still in use were designed at a time when wireless digital interfaces were unheard of. Some of these older systems may experience problems (EMI) if used together with wireless devices.

Reducing the amount of data exchanged is the key point in obtaining low-power

consump-tion required for battery powered wireless systems. Data transmission of a sensory device can be reduced by moving intelligence, i.e., signal processing, to the sensory device. In this way, cal-culated parameters or alert notifications need only to be sent instead of raw biosignals. The key design dilemma here is which consumes more power: the signal processing or the transmission of the raw data.

A special characteristic of medical devices used for health monitoring purposes is the twofold character of operation these devices. In normal conditions, when the patient is well, the role of the device is basically to detect any life critical changes in the subject’s condition. This kind of operation fits well to a low-power wireless technology, which optimizes power consumption over throughput and latency. In the critical monitoring state, the requirements for the interface change dramatically. Reliable real-time signal transmission becomes more important than low-power operation. This is a major design dilemma in wireless interface selection for these kinds of mobile battery-powered devices; a single interface is always a compromise. Bluetooth, for example, is tilted towards reliability and performance over low-power. As a basic of-the-self mobile phone model these days already contains multiple-radios, it would seem likely that also wireless medical devices can and will contain multiple-radios in the future. They could not only provide optimized low-power and high-throughput alternatives, but also enhance reliability by providing alternative equivalent wireless technology. The design challenge here is to group these alternative wireless technologies seamlessly under the same application level interface standard.

Reliability of interfaces

An interesting question is in which areas of health monitoring should the operation of the communication interface be most reliable and where could unreliability be tolerated. In critical care use the health of the monitored subject is in critical state, and small device errors may have serious consequences. On the other hand, the medical staff is close by and device errors caused by interfaces should be noticed quickly as they usually lead to “no-signal” types of errors. Failure of a home monitoring interface may remain undetected longer, as no medical staff is usually present to notice the malfunction and the possible deterioration of the subject’s health. Furthermore, because of the traveling required, the costs of checking and fixing a malfunction caused by an interface at home is more expensive than in the healthcare facilities. The unreliability of home healthcare systems caused by the interfaces may also be frustrating to the often nontechnical users which may lead to the devices not being used. In this respect, it is safe to say that the reliable and easy operation of interfaces is important to all forms of health monitoring devices, from home to critical care.

7.2 General discussion 111