I had the pleasure of meeting with a number of Biomedical Engineering and Clinical Engineers at CIMIT (Center for Integration of Medicine and Innovative Technology) in Cambridge this week. Lot’s to tell you (more to come), but perhaps nothing more timely then “time” itself. With the eventual “Meaningful Use” requirement to include the integration of Medical Devices to your EHR as a means to correlate patient vitals over time – we have big problems looming as an industry.
Under the direction of Dr. Julian Goldman at CIMIT, researcher Pratyusha Mattegunta, MS, BME and team examined 100’s of medical devices across a particular healthcare system – most of which were network attached. What the team found was an overwhelming number of medical devices with incorrect time and date. Some devices in fact were running some “very advanced firmware” that was able to predict the patient’s condition — Sometimes as much as 6+ months into the future!
So if you’re an IT guy and thinking— “yeah so what, my VCR has been flashing 12:00 for years now and I never ONCE missed taping a single episode of the Computer Chronicles – big deal”. But seriously, this is healthcare and it’s no longer the 1980’s — the problem is a much larger one when we consider the clinical ramifications and patient safety issues.
Let me put it into perspective for you. Consider for a moment a cardiac bypass patient in post-op. Typically to reduce pressure on the newly sutured arteries and to avoid a rupture and bleed-out situation, the protocol calls for the patient to receive vasodilation drug(s). These dilate the blood vessels resulting in less pressure. The larger dilated blood vessel results in lower pressure (Don’t believe me – try drinking a milkshake with a plastic coffee stirrer vs a standard straw – same basic principal). During this time, cardiac output and drug infusion rates are closely monitored along with other cardiac related factors.
When this data is electronically collected and stored in the EHR , its absolutely critical that the time stamps of the data across all of the medical devices be synchronized. If they are not, clinical decisions on weaning the patient off the drugs based on cardiac output may be performed in what I’ll call “a non-optimal manner” – and the result is never a good one. There have been a number of sentinel events where incorrect treatment decisions were made against an unsynchronized clinical dataset.
So how do we solve the problem – it would seem that a simple implementation of the NTP (Network Time Protocol) would solve the problem. The IT world has been doing this for YEARS, its not rocket science, and the service is commonly available and very well understood.
Well, sorry to inform you that the majority of medical devices have not implemented NTP. Even for those that have, there are known cases of NTP incompatibility requiring specific implementations of NTP servers for correct functionality. Furthermore, even if the time/date is manually set correctly, clock drift does occur and over time the drift can be substantial.
Faced with this issue, which by the way has been recognized as a huge issue facing the industry, I’d like to suggest a few action items for the Healthcare industry. The use of standard NTP should provide the accuracy and resolution needed and is more than sufficient. If not—there are High Performance NTP like implementations used in the financial and manufacturing industries that can provide microsecond resolution (that’s 10 to the -6 power).
1) IT and CE teams need to prepare for the delivery of NTP throughout the healthcare enterprise in a hierarchal and scalable fashion. Medical devices are and will continue to be placed on your IT network. This convergence is going to continue to expand over time.
2) Medical Device Manufacturers, we need you to implement a reliable standards based time service such as NTP. If you’ve done this, that’s great -- never mind! And in order to keep my NIST friends happy, let’s also avoid overloading the public NTP servers, so please adopt a hierarchal approach to time synchronization.
If NTP is adopted by the healthcare industry as I suggest, take a look at the NTP white paper published some time ago by Cisco. Great work by the CIMIT researchers and Dr. Goldman’s team – the industry looks forward to additional information on this topic – so it keep coming!