Thoughts on the Future of Medical Devices at the Point of Care
Nathan left a very meaty comment tying together the post on Trends in Nursing Units with the one on Secondary vs. Primary Alarm Notification. There was so much there that I couldn’t bear to leave it buried in the comments. Let’s take it point by point.
The biggest trend in central station patient monitoring now is to create a ‘war room’, such as VISICU, where many patients are monitored at once. While you do have the cost of new equipment and a full-time tech, your nursing staff can be freed to be much more productive.
The ROI on these two remote monitoring applications is very different, but in both cases if there was a better solution (like enough intensivists on staff to eliminate the need for VISICU, or better alarm delivery to eliminate monitor techs) they would take it. Monitor techs are expensive and inefficient; most of the time they watch normal waveforms. War rooms have been around for years – possibly the biggest (covering a half dozen hospitals) is at Parkview in Fort Wayne, Indiana.
Central stations are a mature undifferentiated market, especially regarding clinical features. With demands for remote access and integration with VISICU (they’ve done that at Parkview too) and other information systems, central stations are becoming more like health care IT software that’s still regulated like a medical device. The cost and complexity this represents for device vendors is significant. When you do your central station market study, track how many vendors support features like RADIUS/LDAP integration, single sign on, full featured role based access control (including consulting and covering physicians), and full HL7 support (bi-directional ADT, data export, orders). Add these IT features with features they should have as a medical device – high availability, hand held alarm management, wide area network support for remote war rooms – this is way beyond what medical device vendors have been able to bring to market on their own.
On top of this, efficient secondary alarming systems can be implemented as you’ve discussed in a recent post. In the end, I think that the costs for these systems are acceptable when you consider that this technology solution solves the communication dilemma created by decentralized nursing.
I don’t know that secondary alarm systems can be efficient when customers can’t rely on those capabilities because they’re “secondary” – that’s one of the conundrums faced by the market. The greatest challenge to improving patient safety and lowering operating costs is the high cost of new technology. We need a wave of commoditization so hospitals can afford to adopt current technologies, or it will be 10+ years before most hospitals have EMRs, let alone all this other stuff. Secondary alarm systems are expensive, several thousand dollars per caregiver – more if you include infrastructure. If you’re going to continue with monitor techs, where’s the ROI for secondary alarms? You are right, Nathan, it may not be prudent for a hospital to depend on a secondary alarm system for primary alarm notification – at least without taking the proper steps. And you’re right, central stations won’t do much for decentralized nursing.
These alarms can be carried on top of Vocera badges or smart phones that also provide rapid communication responses. It’s better for the patient and better for the nurses.
Vocera makes a great product, and like a lot of wireless communications systems, delivers solid benefits at the point of care. However, in its current incarnation the Vocera badge would make a poor alarm delivery device, at least for some medical devices. You really can’t replace a person with a machine, and so medical devices tend to produce false positive alarms (some times a lot) in an effort to never miss a real alarm condition. Alarm fatigue is a common outcome of mostly false positive alarms that are broadcast to every caregiver on the unit – this is the reason there are war rooms. An effective alarm notification device needs to be able to represent a meaningful set of data that caused the alarm so caregivers (like monitor techs) can screen out the false positives.
If this is the way of the future, then maybe this will reach the threshold of FDA involvement as you discussed.
Here is my fantasy: medical devices on a nursing unit (all types, all vendors) will interoperate with a third party server that replaces multiple proprietary servers/central stations. Caregivers will carry alarm management devices that also provide wireless voice communications, with alerts and alarms in text and graphics (waveforms). Anyone wishing to see retrospective or “full disclosure” data will go to any PC in the hospital with a browser and pull it up. High availability medical device servers will serve up waveforms – retrospective and near real time – to EMRs. Devices will interoperate with clinical information systems setting device parameters based on physician orders – without human intervention – and generate clinical alerts (and maybe even life saving interventions at the medical device) by combining the many things the CIS knows about the patient (the patient’s diagnosis, latest lab values, etc.) with what the medical devices connected to that patient know. All of this will be tied together with an enterprise service bus. Proprietary WMTS systems will be replaced with 802.11. In my fantasy it’s a patient centric world with no artificial barriers resulting from the product portfolios or installed base of the device vendors.
The FDA will not be doing anything different from what they’re doing now – investing their limited enforcement budget in areas that pose the greatest patient safety risk, and servicing their “customers” by regulating products. I don’t think it likely that the FDA’s going to ride in and force a structural change on the industry (like making secondary alarm vendors get 510(k)s) – but they could.
The issue with FDA involvement that I see is the lack of regulation around wireless IT in hospitals. What exactly would be regulated in a system that is created by a patient monitoring vendor working with a middleware developer working with an IT provider? In my mind, the most important issues are to build redundancy into a WiFi network, document everything the alarms and the responses, and always have the primary alarm remain visible at the central station and the patient’s room.
There is no need to regulate wireless IT in hospitals – if the medical device includes the wireless network, the network is regulated as part of the medical device. I’m doing some business development work with a vendor who is grappling with this very issue, and it really comes down to this question: how do you develop software systems with parts that the FDA classifies as a medical device without turning the entire system into a “medical device,” increasing complexity, development costs, and time to market? Fortunately it can be done safely, with existing regulations. You just have to be smart about it.
What’s really at issue here is – and I love this term – “piercing the veil of commerce.” The FDA’s current regulatory oversight ends with the sale of a regulated device – they don’t go into customer’s sites to make sure they’re conforming to the intended use or maintaining a validated configuration. But they might in the future. They’d need a lot bigger budget.
Ultimately, decentralized nursing may be exactly the kind of operational trend on the provider side that is necessary to push demand for these patient monitoring alarm notification systems to the next level.
Demand for improved point of care workflow automation, including nurse carried alarm notification, exists. They sold well in the ’90s before
people figured out you couldn’t count on paging systems for alarm delivery. In a previous life I sat across the table from customers who demanded (rather forcefully) nurse carried alarm notification. All the PDA prototypes flashed at shows or rumored in the field are a response to that unmet demand.
Ultimately what will break the market lose will be small innovative vendors who can visualize and create solutions completely from the customer’s perspective. The path from today’s market to my fantasy of a patient centric world is as yet unknown – blazing that trail will be an adventure.
Another RFID Market Forecast
Frost & Sullivan pegs the RFID market at $201.3 million in 2004 and forecasts a market of $1.26 billion in 2011.
also suggests that channel partners in the RTLS industry are not very
active and need to be trained more effectively. The successful
implementation of this technology requires a deep understanding of
business operations, quality functions, system support issues,
standards, IT infrastructure issues, and ERP systems. Ineffective
installations lead to customer dissatisfaction and increased customer
support issues later on.
“The apparent inefficiency of channel members adversely affects the
perception of brands among consumers,” warned Durga. “Though technology
and pricing factors remain favourable, ineffective deployments reduce
consumer confidence and ruin brand images.” To overcome these issues,
solution providers must train their channel members. Effective and
customised training would reflect on the level of customer satisfaction
and help in brand building, the report says.
Pushing a disruptive new technology through indirect channels is never easy.
PDAs Called Most Overrated Technology
Mr HIStalk unveiled the HISSIES yesterday – this is his tongue in cheek send up of the HIT market subtitled, the Brutally Honest Healthcare Information System Awards, Nominated and Voted On by the Readers of HIStalk. In among the nose tweaking and insider jokes, this item caught my eye – in the category of Most Overrated Technology, PDAs won big. When looking at workflow automation at the point of care, I must agree.
What’s wrong with PDAs? Well, they’re expensive, most you can’t drop because they’re not ruggedized, and most you can’t wipe with disinfectant because they’re not moisture resistant. The PDAs that are ruggedized and moisture resistant are about the size of a brick – or at least that’s what it feels like after a 12 hour shift. The typical PDA would last less than a week on a nursing unit.
Battery life is also a problem – few PDA batteries can support an 8 hour shift, let alone go the full distance of 12 hours. That means you’ve got to have battery chargers and extra batteries for all your PDAs.
Screen real estate is also a big limitation. Given that tiny screen combined with the eyesight of the average nurse (who was 45.2 years old in 2001) there’s not much you can do with a PDA. Forget data entry applications, or accessing patient records – you get either too much tiny text or too many screen flips.
An often overlooked weakness of PDAs is their poor suitability as a general purpose computing platform. The operating systems on PDAs are tweaked to a much greater extent than laptops, tablets or desk top PCs. This means that application code must be written for specific PDAs – that’s the specific model made by a specific vendor. The processing horsepower in a PDA is also more limited than larger computers. Running an app while connecting to a device via Bluetooth while connected to an 802.11 WLAN will bring many PDAs to their knees. In practical use this means that PDAs are single use devices – you can have one for meds administration, and one for nursing worklists and results reporting, and another one to go with your “smart” pump. Caregivers will have so many holsters they’ll look like scrub-clad commandos.
At HIMSS I’ll be checking out all the caregiver carried computing devices and reporting back on the lay of the land.
