As more and more patient care organizations move forward to implement the use of evidence-based order sets for electronic physician ordering, their leaders are learning a variety of lessons about what works best, and how best to plan their go-lives within the context of computerized physician order entry (CPOE) implementation within the electronic health record (EHR). One integrated delivery system that has moved forward is Signature Healthcare, a Brockton, Mass.-based system that encompasses a 225-bed hospital, a provider organization with 120 physicians and 30 mid-level practitioners, a school of nursing, and a foundation.
Marc Greenwald, M.D., chief of medicine at Signature Healthcare, has been helping to lead his fellow clinicians and healthcare IT leaders forward on evidence-based ordering. In the fall of 2009, Signature went live with evidence-based ordering in the inpatient sphere, using ProVation order sets from ProVation Medical, a Minneapolis-based division of the Philadelphia-based Wolters Kluwer Health. Greenwald spoke recently with HCI Editor-in-Chief Mark Hagland. Below are excerpts from that interview.
How many order sets are you live with right now?
We’re currently live with 42 order sets, and are continuously increasing that number.
Can you categorize the areas you’ve focused on to date?
We’ve been looking at the most common and highest-volume diagnoses, frequency, etc., which also coincided with core measure diagnoses. The ones we’ve focused on to date are congestive heart failure (CHF), COPD [chronic obstructive pulmonary disease] exacerbation, and acute pancreatitis. Marc Greenwald, M.D.
What led you to decide to implement evidence-based order sets?
When we looked at what we had to do to go live on ordering, and looked at the quality and utilization issues we were facing, in terms of optimizing utilization and quality, we felt that guiding physicians in as evidence-based a way as possible, was the way to go. We had Meditech [the Westwood, Mass.-based Meditech] for our EHR, which we kept. And we looked at a couple of different order set vendors, but in the end decided to go with ProVation. And the biggest advantage is that the order set itself is in Meditech language, and they can immediately click into links to the medical literature, and can find out, for example, why they don’t need to order a CT scan for every case of pancreatitis, and all the things they need to remember when they admit a patient for CHF.
What was the timeframe for go-live?
The plan was for a period of eight months between contract signing and go-live, and it was completed within about two or three weeks of that. We needed to go live with CPOE at the right time for meaningful use. We had used ProVation before that, because we were getting ready to translate the order sets into Meditech language, so it wasn’t really ProVation holding us up, but our CPOE go-live. We’re on the most up-to-date version of Meditech Magic.
How has acceptance been among the physicians?
The use is 100 percent. Our hospitalist team takes care of 85 percent of the hospital, and they do notes as well as orders on Meditech; it’s all computer-based now. And the order sets are actually easier for them, because they’re all on the same platform.
Was any customization required?
We actually did a lot of customization. The first thing we did was to create the pharmacy dictionary, so that the choices of medicines, as we created order sets, were limited to our formulary. I was the primary author for a lot of the initial order sets, and we modified them based on practices here and on our analysis of utilization patterns. They weren’t contrary to medical evidence, but they were refinements based on practices in our locale. We had physician reviewers, pharmacist reviewers, quality resources experts, utilization management experts, IT and compliance people, all reviewing every order set. And we created a process for rapid approval and implementation.
Were there any major obstacles, or was the rollout pretty smooth?
Actually, it was pretty smooth. The only holdup was that in the beta-test program to automatically translate ProVation language into Meditech language, they had to do the first few order sets by hand, but once they got that going, it was fine.
It seems obvious to me that using evidence-based clinical decision support in ordering should be a no-brainer, but there’s always a necessary struggle to convince physicians to adopt such practices. What are your thoughts on that subject?
The reality is that the proportion of medical practice that is actually truly evidence-based is still relatively small. For example, until recently, very rigid control of blood sugars in the ICU was considered evidence-based medicine. But what people didn’t go back and think about was that the evidence was very sketchy and never confirmed, and in fact, it has recently been disproven. In that case, it was found that rigid control of blood sugars in the ICU was actually associated with higher mortality, so people have stepped back from that earlier stance.
So in our best practices right now, I always teach the residents to be skeptical, because yes, right now, CHF has certain drugs and tests associated with it that we should use. But that may change tomorrow, so don’t take it as gospel. And the one thing I don’t want people doing is to do things mindlessly, simply to click a box. And that’s why the teaching is embedded in the order sets, and we review performance on a regular basis, and we review the order sets on a regular basis.
The teaching is embedded with regard to the way in which the CDS prompts physicians to take certain steps while ordering, right?
Right, and I ask my physicians to keep thinking as they’re ordering, so they’re not just clicking boxes.
In general, though, we should be doing as much as we can to be as evidence-based as possible, right?
Right. Just remember that the evidence is constantly changing.
What are the key conceptual, strategic, and practical things CIOs, CMIOs, and CMIOs, should be thinking about, going forward on this journey?
The first thing is to let a physician champion with the appropriate level of skepticism and leadership lead this process. And the CIO, CMIO, etc., should feel pushed from behind by the people who are using this. Have the physicians be in the lead. And this is developmental software, it’s not operational software. I’ve had physicians call up and ask me about downtime. But you don’t use this day in and day out. It’s like Microsoft Word: it may fail on you occasionally, but as long as you’ve saved everything, you haven’t lost anything. It’s a tool, it’s not a respirator. I’ve been very happy with the solution, with the vendor. But we’ve got a really good physician crew who have really been very enthusiastic about this, and I think that’s our strength.