A fresh initiative aims to make medical device connectivity "plug and play" within 3 years.  As reported in the January issue of the AAMI newsletter, the story describes efforts lead by representatives from Boston-based Partners Healthcare and Kaiser Permanente in cooperation with the FDA, other health care professionals, and the U.S. Department of Defense. Most medical devices with network connectivity (serial interfaces need not apply) already have gateways or HL7 interfaces.  But devices have very far to go indeed before they can plug and play on par with DICOM and diagnostic imaging modalities. Connectivity in health care started with the IEEE medical information bus, or MIB. How DICOM came to fruition in less time is a mystery to me.

“The concept is that individual medical devices would be networked to allow the communication of data from one device to the other and the control of one medical device by another, as well as allowing the implementation of clinical rules or guidelines,” explains Julian M. Goldman, MD, of Massachusetts General Hospital, principal investigator for the project.

The goal is not only passing data, but also interoperability between devices and between systems and devices.  Interoperability means "remote control" of a device by another device or a system.  HIT vendors have been seeking ways to do closed-loop control between medical devices and their systems for some time.  Great advances in patient safety and improved outcomes are to be gained by combining the information in an EMR and a pump or monitor. McKesson's efforts with Alaris, and Cerner's work with Hospira are examples. HIT vendor's efforts by definition, require multi vendor integration. Companies like HEI and Capsule Technologie have built businesses out of providing cross vendor connectivity. However, both these companies have built business models on connecting standalone devices with simple serial ports.  Increasingly, medical devices offer their own integrated network connectivity.

One area screaming for plug and play (and interoperability) is alarm management and alarm notification.  Nurses have many different types of devices (monitors, various pumps, vents) that they must manage.  They need one means of receiving and responding to alarms generated by any device connected to their patients.  HIT vendors are coming at this with automated worklists and alerts.  Connected device vendors are providing this control (albeit proprietary) from central stations.  Throw in currently standalone devices, cross vendor connectivity and interation with information systems and you've got one killer patient safety and nurse productivity solution.

Certainly recent technology makes plug and play connectivity technically more feasible.  Software protocols like SOAP and XML-RPC, the new low-power 802.11 a/b/g radio chip sets, and a new wave of WLAN infrastructure commoditization all contribute to making this goal more achievable.

There are barriers to achieving Dr. Goldman's connectivity goals.  Regulatory concerns about patient safety are a huge concern.  HIT vendors will go through contortions to avoid a 510k, and embedded device vendors will fear this sudden loss of end-to-end control that plug and play requires. Developing intended use statements, risk analysis, verification and validation of interoperability between devices and systems will require a monumental effort.  Strangely, vendor strategies to lock in customers by artificially erecting changing costs (usually by way of proprietary features) are not mentioned.

Every market in which I've participated in the transition from standalone devices to connected systems has been marked by a demand for open connectivity.  No buyer wants to get locked in or face artificially high changing costs.  The ability to mix and match when grandually transitioning from one vendor to another, or creating a system out of best of breed devices has always been a desire.