IT and clinician leaders are breaking important ground at Hamilton Health Sciences, the six-facility health system in Hamilton, Ontario, Canada, an academic research teaching facility affiliated with McMaster University that serves a 2.5-million population in Ontario, and that encompasses a children’s hospital, labor and delivery, and a cancer center, with 12,000 staff.
At Hamilton Health Sciences, vice president of health information technology services and CIO Mark Farrow, who leads a team of about 140 IT professionals, has been co-leading an important initiative around the alerting of clinicians to the rapid deterioration of patient conditions, along with Alison Fox-Robichaud, M.D., a physician in the Department of Critical Care. Farrow brought together a project team of five clinical informaticists, with five clinicians led by Dr. Robichaud.
Essentially, what Farrow, Fox-Robichaud, and their colleagues have done, is this: they’ve built an automated alerting system, leveraging Android smartphones, that involves the alerting of a rapid response team to patient condition results that show the deterioration of patients who are on the regular medical/surgical floors of the inpatient hospital. The Hamilton Health Services leaders have partnered with the Toronto-based ThoughtWire and with the Armonk, N.Y.-based IBM Corporation, to develop the fully automated solution. The results have been strong, including the virtual elimination of code-blue alerts in the inpatient hospital, and a drastic reduction in ICU admissions from the med/surg floors.
For their innovative work in this area, the Hamilton Health Services leaders in January received two awards from the Intelligent Health Association, an association of information technology vendors. It received both the IHA’s 2017 Award for Improving Patient Care and Health Delivery, and its 2017 Intelligent Health Grand Award.
As ThoughtWire noted in a press release issued on Jan. 23, “Through the use of an innovative early warning score (EWS), HHS is improving hospital safety by eradicating in-hospital cardiac and respiratory arrests. With the EWS, clinicians are apt to respond to abnormal vital signs before patients progress to experience in-hospital arrests. “We believe that most cardiac arrests in an acute care setting should be considered a failure to rescue,” said Alison Fox-Robichaud, M.D., the clinical leader behind the Early Warning Score (EWS) project. Dr. Fox-Robichaud’s clinical team was complimented by members of the Health Information Technology Services group led by Mark Farrow, vice president and chief information officer at Hamilton Health Sciences.
To enhance this initiative and move toward achieving even better care outcomes, HHHS collaborated with ThoughtWire and IBM Canada to find an innovative way to address the key factors that were inhibiting the initial EWS results. Leveraging ThoughtWire’s Ambiant Health Platform, the team created a Mobile Early Warning Score Application that works in real time with Meditech. Today, nurses capture vitals at the bedside on their mobile devices. The data is then integrated into the Meditech electronic medical record system, which computes the EWS. Based on HHS research, each score prompts the Ambiant Platform to drive a standard set of notifications and responses to the appropriate members of the care team, while machine intelligence ensures that standardized best practices are consistently executed.
The early results have been powerful: HHS has seen a 17-percent decrease in the number of Critical Care Response Team consults requiring ICU admission and a 6-percent reduction in cardiopulmonary resuscitation (CPR) requiring Code Blue calls. “Before rapid response teams were in place, you would hear code blue calls on average once or twice a day in the hospital’s wards,” Dr. Fox-Robichaud, said. “Fast forward to 2016 and I can now go an entire week without hearing a code blue on the wards. While they have not been eliminated, we hear far fewer – and that means that patients are staying safe.”
Farrow spoke recently with Healthcare Informatics Editor-in-Chief Mark Hagland regarding this initiative. Below are excerpts from that interview.
Tell me about the origins of this initiative?
The Hamilton Early Warning Score (HEWS) started as a research project, and it was one of our residents’ research projects on a couple of units. We very quickly saw the merits of it and looked to see what we could do to move it forward very quickly. And one of the challenges is always, when you find evidence, how do you move it to the bedside? So, we first created it on paper to build it out, and quickly built it into the Meditech system. There wasn’t a lot of process, but it allowed us to process it and have the scores calculated automatically, and it would populate an action plan based on a score. We needed to reduce a lot of lag time. While people were documenting in Meditech, they weren’t necessarily documenting in real time. And it’s not helpful to document at the end of the shift, and have the system say, oh by the way, your patient was going to have a heart attack three hours ago, when they already did. So, it needs to be done in real time. You need to be able to create it as part of the culture of the unit. And the way to do that is to not only show that there’s benefit to this, and in a way, that doesn’t impede their workflow.
So, this had to be relatively un-intrusive to nurses. We had put in mobile carts, the computers on wheels, but they weren’t getting a good uptake; they were difficult to push around. We had the documentation in the system, but people were still putting pieces of paper in their pockets and coming back to it later. So, we needed to be able to put this data in several different places, but still have one piece of truth, with a communication link. But the second part of this is not just generating the score, but clinically, whom do you notify? And one of the challenges is, what happens if a floor nurse gets an alert at midnight, what does she do?
So, what does happen now, since the deployment of this solution and process?
So, with the handheld, the nurses are now entering seven elements into the data. Those seven are heart rate/pulse; systemic blood pressure; respiratory rate; temperature; the O2 stat (saturation level of oxygen in the body); and O2—oxygen therapy (are they getting by room air or mask, and volume they’re getting); and a visual observation by the nurse. Is the patient alert? In pain? Is the patient’s voice fading? Is the patient unresponsive? So, there are six items that can be automated, and there is still a visual assessment by the nurse that has to be entered into the system.
And essentially, what happens is that everything turns on the score that is created from the seven elements. A normal score would be 0. With a slightly elevated heart rate, you’d get a 1; if the patient’s heart rate is up by 10 percent, you’d get a score of 2. The same applies with blood pressure. If any particular measure goes outside the normal parameters, the score would end up with a 3. The measures are combined into a HEWS score, which leads to an action plan. If the HEWS score reaches 5, the nurse is required to call the rapid response team to intervene on the patient’s behalf before the patient goes into a code blue situation.
When did this program go live?
The original version of the process went live with the Meditech side of the system in 2013, and then we rolled it out across our organization, and began discussions. We were able to show a 17-percent decrease in the number of critical care response team consults requiring a patient to be transferred into the ICU; and we were able to document a 6-percent reduction in cardiopulmonary resuscitation requiring a code blue. And that was in the very early stages of this, before we became fully automated. The paper version was live four years. The fully automated version involving ThoughtWire and IBM, we took live in November 2016.
So the metrics were documented over four years?
That was over the initial data collection on testing the HEWS score, to see if the score worked. That was the review we did of the first 400 patients who required the HEWS data, in 2014.
And the new system basically allows the nurse to go in; she just fills in the data, the HEWS score pops up, based on entering the seven data elements—it’s very easy to do. And as soon as she or he does that, it creates a notification. So, based on the score, it will notify the charge nurse and the rapid response team. The next step will be to add a notification to the MRP, the most responsible physician. Right now, the system tells them who the MRP—most responsible person—is. Later this year, we’ll add that notification. And we want the notification to also go straight to the device of the responsible physician.
With the automation, the nurse doesn’t have to leave the bedside, because she can see who’s been notified, and at that point, either the charge nurse, or if it goes up to the rapid response team, they can all see the vitals. And if bedside nurse records new vitals, those will be documented. And the rapid response team is in charge at that point.
Have there been any technological or process challenges in this initiative?
We did this with ThoughtWire and IBM, using their design-think principles, using their people to design the ergonomics, the device, the screen layout for it; we have had some challenges in that it is an Android application, and we’ve found that certain versions of Android are not working as well as others, so we’re ensuring that all devices are working up to the proper level of Android. We also use this device to do bedside meds verification in Meditech; the object there was to have a multi-use device. And this was for front-line staff, so the buy-in was very high. They were very supportive of it, so it’s been very successful. It’s an Android smartphone device.
When did nurses start using the Android smartphone for this?
In November 2016. And it used to be very common to hear code-blues, but you don’t hear those anymore. The last two I’ve heard have been in an ambulatory clinic area that this doesn’t cover.
And essentially, Dr. Fox’s take is that if a patient ends up in a code-blue, that’s a failure to treat. We should have could intervene and see what’s going on. There will always be that odd case that wasn’t predictable, but in most cases, we should be able to determine this in advance, based on trending.
What have been the biggest lessons learned in rolling out initiative so far?
I think the big thing is that when you create that actionable knowledge, technology can provide a shortcut to assist with all the manual calculations. We’ve found that you need supportive leadership to be able to do this. It certainly works best with those rapid response teams to do this. But from a technology standpoint, when we finally provide nursing with a clear outcome of what can happen, they’re very receptive to do the work. We had been documenting for years, but all that documentation just went into the charts, and nobody knew whether it was being used. But automating this brought out the real power in it.
And we’re looking at future projects, including a pediatric version of this that we’re working on right now. And we’re looking at potentially creating a version of this for sepsis. And certainly, having the frontline staff involved from the beginning was essential. And ThoughtWire’s ability to duplicate this without heavy HL7 interfaces, was a major leap forward in our creating something very quickly, and still have something that wasn’t duplicative.
What would you like to say to our audience of CIOs, CMIOs, and other healthcare IT leaders, about this?
This shows that we are now at the age where we can harness this information and start to become useful, rather than just being large storage lockers of data. And the new technologies available to us are ones that we need to embrace and harness and get out to the bedside, so that we can change how we’re doing our business. Because if we can avoid a code blue call or an ICU admission, that’s huge for the organization—and for the patient. So IT is coming together with the clinical side of the business, to really transform how we deliver care going forward.
