All right, thank you so much for that introduction. My name is Bill. I've been with FRADED for about 20 years, and during that time most of my time has been spent in the Trauma Surgical ICU, helped establish clinical practice in the ED back in 2003 and 2004. And for the last three years, I've transitioned to a pain stewardship role. So this project kind of marries all of those experiences together into one bundle. I'm gonna characterize the need for harm reduction. I'm gonna share with you an example of how we leverage clinical decision support to improve prescribing of Naloxone to high-risk patients, high risk of opioid overdose. And I'm gonna share with you some of the next steps that we have in mind for our stewardship campaign. This is a canned slide that the CDC offers us, and you may have seen this before, and it essentially talks about the three phases of the opioid epidemic. I would draw your attention to the first phase, which really started with prescription opioids. And it peaked quickly and kind of plateaued, but we haven't seen a lot of movement in that variable in a number of years. Since then, there have been subsequent waves, focused on heroin and synthetic medications, but that prescription element still persists. So when you look at addressing that issue, you really need to address it across the continuum of care. And this is kind of how we've compartmentalized it at Frederick. We prevent inappropriate opioid use or unnecessary appropriate use. Sorry, those long words wouldn't fit on the slide. Really promoting multimodal strategy in an effort to spare the amount of opioid that's necessary. When opioids are appropriate and necessary, we try to prescribe them in the most appropriate manner that we can, the lowest effective dose. And in instances where they are appropriate, we try to supplement with mitigation strategies. And finally, once we inherit a patient that has an opioid use disorder, we try to offer them a wean off of medication if we can, or introduce them to medication for opioid use disorders, such as buprenorphine or methadone. And at our enterprise, we've undertaken an aggressive campaign to promote multimodal therapy. And I think that there's a relationship between that campaign and the overall reduction that we've seen in the amount of opioid that's prescribed, or the frequency at which opioids are prescribed. So as a proportion of total prescriptions that our enterprise generates, the number of opioids have actually gone down without a change in complaints of pain. When medications such as opioids are necessary, we found over the course of the last three years that we've seen about a 15% reduction in the dose of opioid that's prescribed. Finally, to mitigation strategies. Naloxone is extremely effective. We provide a very superficial overview of some of the studies that are out there right now, focusing in on primary care, the emergency department, and then population health in general. And the end result of these studies is that when you connect at-risk patients with access to naloxone, the antidote for opioids, you have improved outcomes that can be reflected in the volume or the frequency at which these patients present to the ED. If you survey the family members that have received naloxone for their loved ones, they report that they've used the naloxone to reverse an overdose. And finally, there's a correlation between the amount of naloxone that's in a given community and the volume of unintentional overdoses that you see. So based on this data, several different organizations have come up with recommendations on who the target population is to receive naloxone. We've got the CDC, FDA, Health and Human Services, Surgeon General, and the VA. And when we looked at all these and trying to decide what we're going to do at Froedtert, we essentially went along with the CDC guidelines. We found that they were clinically appropriate, there were data to support their recommendations, and we found it to be operational in the environment that we functioned. We could actually bring it to life. In addition to that, we felt that there was a need to include muscle relaxants co-prescribed with opioids as well. We've observed some respiratory depression at a higher rate in that population. So onto the project. Our design is a quasi-experimental pre-post analysis. Our enterprise has 11 hospitals and 45-plus clinics, and the intervention was leveraging clinical decision support to implement a naloxone co-prescribing best practice alert, or BPA, within our electronic health record. This happened in December of 2021. Our two groups are the pre-intervention group, six months before, our post-intervention group, six months after, and our primary outcome was the number of naloxone prescriptions written by the Trauma and Emergency Medicine Services, reported separately, compared pre- and post-implementation of the BPA. This is a screenshot of what the BPA looks like, and it fires when a prescription is placed on the patient's profile that puts the patient at high risk of opioid overdose, going back to the CDC recommendations. And in the yellow box up top, you can see that that's an alert to the prescriber saying that because of the prescriptions that have been placed, this patient is classified as being high risk for overdose. The blue text that's underlined are actually hyperlinks, and the first hyperlink is a Wisconsin DHS document that talks about risk mitigation strategies, the risks of opioid and risk mitigation strategies. The provider can print this out. It helps them facilitate a conversation with the patient real time. Should they desire to prescribe naloxone, we have a host of other hyperlinks with all sorts of educational resources, information, including patient counseling videos, so the provider doesn't even need to actually do that. Below that, you'll see the action that can be taken by the BPA. We believe that when these medications are placed, theoretically, they're appropriate. So the default is to continue the medication. The other default is to prescribe naloxone because we consider these patients at risk. So if the provider takes no other action other than to click accept, the opioid that generated the trigger will continue, but the naloxone script will be generated as well. Should the provider determine that the naloxone's not necessary, they can reject it, but they have to provide a reason, and we can track this reason behind the scenes, and the interesting thing is you'll see the lockout window, three months, six months, so on. That means that regardless of what subsequent high-risk situations are generated after the fact, the BPA will not fire. The provider's making the decision that they don't want to see this result again. When you reach that time period, it will reset, and the provider will see that again. So the beta test process was probably one of the most important elements of this project. We turned this function on within the medical record, but the providers couldn't see it. It didn't send them a trigger. We were logging it electronically behind the scenes, and on a weekly basis for eight consecutive weeks, we looked at what those reports looked at. What was the volume that it was firing at? What was the environment? What was the reason why it was firing? And we made lots of tweaks along the way. We were hypersensitive to sensitivity and specificity, to alert fatigue. We didn't want more static in the medical record. As a result of that process, we made several changes. We retired the concurrent muscle relaxant prescribing. It accounted for about 50% of the alerts. It was just too much. We didn't want that much noise firing at providers. It would certainly hurt buy-in. We retired the ICD-10 codes associated with dependence and tolerance, not because it was clinically inappropriate per se, but because we do a poor job of utilizing these diagnosis codes within our medical record, and it was actually not helping us find our risk population that we wanted to find. We excluded Scripps less than five days. This is largely the short course perioperative patient population. Again, more noise, alert fatigue. And we had initially excluded hospice and palliative care patients, and we went to those groups and said, we're turning it off, that's our plan, but do you want these patients included? And a little bit surprising to me, unanimously they all wanted it turned on. They found that their patients were in fact at risk. We didn't initially find data in the literature to say that they were at risk. That population wasn't teased out in any clinical trials, but they felt like they wanted their patients to at least have that question asked about their care. So the final criteria that we had were MMEs greater than 50 for at least five days, or an opioid at any dose, co-prescribed with a benzodiazepine, or an opioid, again, at any dose in the setting of a patient that has a history of opioid toxicity or overdose. Again, if the provider had previously suppressed the alert, the BPA will not fire. If there's already an active naloxone Scripp or a naloxone Scripp pended, the BPA will not fire. And if it's a short course Scripp less than five days, it will not fire. And I think the messaging was really important here. We leveraged existing pathways for communication, an email from the CMIO went to providers. We did active learning sessions with the pharmacy staff. And again, the practice briefs, which is a normal standard operating procedure for nursing communication, went out in anticipation of this. The result was that we had a significant increase in naloxone prescribing across the enterprise, almost instantaneously. We have a lot of demographic variables, and frankly, we're still sorting through some of them. We did find some interesting trends, and I'm not reporting all of them. But we do see a high rate of comorbidities that put patients at risk, such as COPD and obesity. We included insurance, such as Medicare Part A and B, because we know this medication, naloxone, can be expensive in the ambulatory setting if you don't have proper coverage. We do know that Medicare A and B covers it to a pretty high degree, often zero to $3 copay. So that's important for us to consider as we look at fill rates. We also really try to promote all of our scripts going to one of our internal pharmacies. Not only does that generate revenue, but it also increases compliance of patients for their prescriptions at our retail pharmacy before they're discharged from the hospital. Compliance goes up as a result. We can also track outcome data a little bit better. And what we found was that when we turned this BPA on, the highest risk population, that at-risk of overdose population that we're most concerned about, the rate of naloxone prescribing went up significantly in both the trauma surgery group and the emergency department. We went from about 1% to 50% in trauma, and we went up about fourfold from 20 to 80% in the emergency department. Now the question, why did that happen? What was the trigger that caused these BPAs to fire? And interestingly, you see very different stories. In the trauma patient population, prior to this intervention, we had implemented prescribing guidelines, and we demonstrated about a 20% reduction in MMEs with no additional requests for refills. We brought our average MME threshold from above 50 to below 50, right? So in this cohort, the average MME prescribed for trauma was about 40 MMEs, morphine milligram equivalents per day. And the average duration was about 3.1 or 3.2 days. So, you know, a lot of the patients were not actually eligible for the BPA, but when you look at the standard deviations of those numbers, when you include the standard deviation, it puts us north of 50 MMEs, it puts us north of five days. And given the trauma population, some of those patients are in quite a bit of pain, and that's why the bulk of this was driven by the MME criteria being satisfied. Transitioning over to emergency medicine, the average MME is 33 per day, quite a bit lower. Duration of therapy is about two days of therapy, again, quite a bit lower. So when the emergency department generates Scripps, they're low dose, short course, but a lot of these patients already have preexisting prescriptions for benzodiazepines. So the addition of an opioid at any dose in that cohort will trigger the BPA to fire. Now we need to put this in context, right? I kind of cherry pick the highest risk population to report. When you take all comers, all opioid prescriptions that were generated by the service, you see that the rate of prescribing comes down quite a bit, right? We're looking at about 4% for trauma surgery and just under 2% for emergency medicine. And I think that speaks to the fact that the bulk of our prescribing is low risk prescribing. And we've done a good, efficient job of identifying those highest risk patients and generating naloxone prescriptions for them. So in conclusion, BPAs can be effective. It was a very robust multidisciplinary team with stakeholders from different areas and environments in the hospital that helped drive this and really contributed to the messaging. We learned a lot from the beta testing process. And anytime you leverage clinical decision support, you want to observe the behavior before it's forward facing to the group that it's going to affect. But gaps remain in care. We did identify that. We're currently evaluating disparities in care and there are certainly some that are out there. One of the biggest gaps was in the ED. This BPA is exclusively designed to fire when a prescription is placed that puts the patient at high risk. We also know that there's a tremendous number of patients that come in through the ED with acute opioid overdoses that will never get a prescription from our enterprise. So we need to address those patients. We've actually implemented a pharmacist driven naloxone prescribing protocol in the ED to target that population. We're a couple months into that project. The results are extremely exciting. I look forward to reporting those next year. And I'd like to acknowledge all my co-authors, our multidisciplinary team for contributing to this project. Thank you.

clinical decision support system

prescribing naloxone

high-risk patients

Froedtert hospital

CDC guidelines