Thank you, Steve. I hope that you'll find many of the themes of this talk are consistent with those of the prior speakers. Regionalization is challenging. And I wanted to give you a little bit of a primer on what's required to make it successful. I think when you consider the overarching goals, being earlier access to your best care, eliminating waste and waiting, minimizing deterioration due to transport-related care delays, and increasing the number of patients served, it becomes obvious that a single administrator, no matter how gifted and talented and resourced and close to the CEO, is going to be able to really impact this. This set of goals requires not only talented administrators, but gifted clinicians and dedicated nurses. It also requires some financial folks that can set contracts that allow you the resources to effectively regionalize. Early access to the best care matters. Optimal outcomes for many presentations of critical illness, adult critical illness, and injury are time sensitive. This is just a list of the things where having early correct phenotypic recognition and implementing proper care truly matters, affects outcomes. Automation strategies are what we need. We need to identify evolving physiological instability so that we can determine which of the people outside of our emergency room need immediate care. The current approach of first come, first serve, or the triage nurse kind of peers out there and decides who's going to be next, really have to be replaced with better monitoring and better predictive analytics. To eliminate waste, we need to eliminate redundancy. Here are some examples of things that we really need to do. We not only need to do them in our own institutions, but we need to do them across the institutions that we serve. I was really excited to hear about how telemedicine's made a difference in southern New Jersey and the outreach that's been enabled. This can basically benefit both the receiving medical center and also the community hospitals. It really starts with an efficient EHR. One example is that all of the medicines are in the pharmacy database. How come they're not available in your computer? Why can't your IS folks figure that out? There needs to be national effort and resources put into making that happen. It really comes down to having situational awareness and knowing what medications the patient's on is just one example of low-lying fruit. Another example is that many patients and many families who don't have metabolic encephalopathy can actually contribute to the gathering and the organization of data and auditing it to make sure it's correct. Why aren't we providing them with electronic devices or a dictation to basically dictate into our medical record and having that information available to all the providers to double-check and make sure it's right? In terms of the crowded waiting room, we need to be applying machine learning algorithms and using our data, particularly where data is dense, such as biomedical monitor data, data using wearables to identify which patients are stable and improving and those that are getting worse, and not relying on observation from inadequately and overstretched staff. Alerts and timeline displays that enable authorization of diagnostics and monitor implementation of therapeutics are important. The folks in neurosciences and telestroke have really done a great job of this in terms of urine CT scans for those that need them. Machine learning algorithms are now increasingly available. They're dramatically better. If you take a look at the number of biomedical monitor alerts that go off in a typical critical care unit, and this is data from three independent medical centers in the United States, and you can see that the biomedical monitors generate about 120 to 160 signals a day, alerts and alarms that go off. If you then curate those based on physiological trending, you can bring that down about a log in between 10 and 12 a day for process data that's gone through a telemedicine system. If you apply machine learning algorithms, you can bring this down to one to two signals per day, and importantly, the accuracy of the signals because of the elimination of false positives is completely different. And you move from accuracy levels that are not acceptable for clinical decision making into a range of accuracy where clinical decisions can be made on those alerts. So this is an evolving advance, and it's something that to regionalize I think you need to do. To minimize transport-related deterioration, we need to have better case selection at our sending sites. And so there needs to be some curation of that to make sure that the evaluation's been accurate, the diagnostics need to be appropriate, and there needs to be pre-transport stabilization. At the receiving site, there needs to be case acceptance so that when those patients come to our receiving sites, we have a clue about what their actual needs are and that we're in a position to provide them. In addition, we need to have pre-transport communication, and we need to transport during monitored during transportation, not only with biomedical monitor data, but also with actual visualization. And the bottom line is that nobody wants to have a resuscitation while they're being transported. So there have been several technological advances that enable this kind of technology. So this particular device here in the center, it can analyze the signals, identify the device, it can extract the information, and it can put it into a variety of medical records, including Epic and Cerner and others. What's interesting about it is that in this particular example, you can see that there's an interaortic balloon pump. This is a nonstandard one. It was rented because the regular one was broken. And this system was able to recognize it and to appropriately display the signals in the ICU and to the receiving CV surgery team to allow safer and more stable transport. Importantly, and not shown on this slide, is the output that goes into databases that generate curated reports that allow leaders to accurately measure what the workforce is doing and to have situational awareness about patients in transport who may be becoming unstable. When we regionalize, there are really four C's that we need to focus on. The first one is communication. It's actually the very most important. It has to happen at the right level. And so it does mean that the responsible clinician at ascending side and the receiving clinician need to be talking to one another. There needs to be a commitment to teamwork. There needs to be coordination and effective logistics. I'm going to say a little bit more about that. And there needs to be true collaboration. I'll give a couple examples of where this matters. A logistical primer really starts with situational awareness. So many medical record systems now provide flow sheets or data dashboards that can go across institutions. So you need to know what's out there. You need to have an automated system to identify patients with evolving acuity. And clinician tools include privileges and fiscal support to order indicated diagnostics and therapeutics for patients who are at risk for deterioration. And sometimes that can happen automatically without the community hospital clinicians getting in the way of it. And lastly and probably the most important thing for critical care professionals is there needs to be expert care plan review early in the first couple hours after the incidents of physiological instability so that patients with time sensitive conditions can be identified and remediated. We need logistical centers, more than just organizations, but an actual center that's staffed, that can monitor across the region to identify patients where there are opportunities for rescue and using automatic alerts and a leverage team approach. So in this example, the woman who is above the man who's sitting there, when the patients have low acuity and aren't predicted to deteriorate, they can be immediately triaged. When they do have high acuity, they need to be evaluated and there needs to be more drill down. A couple of things about team assignments in your logistic center is that the folks that do that need to be authorized to directly admit the patients to hospitalist teams. And one of the things that we found is that if the discussion about whether the patient's coming occurs after a full history and physical are present in the chart, there's a lot less patients that don't get accepted. And that patients need to have appropriate support devices. So another rule that we have implemented is the 100, 100 a liter rule. So if the heart rate is greater than 100, the blood pressure is less than 100, and they've received a liter of a fluid, then they need to have a stable central line before they come. This is an example of how we can, this logistical center can help respond to a surge. So in this example, there've been an unfortunate series of auto accidents during a snowstorm outside of Worcester. Our emergency department was overwhelmed. We had 15 patients that needed to go to the operating room and they were managed in our ORs and PACUs. Several other patients who didn't require the operating room were able to directly go into the trauma and neuro ICUs. To get patients out of those ICUs, we moved most of the patients to our medical ICU, and we took many of the patients who were in the medical ICU who had stabilized, and we moved them to the medical wards and to our community hospital ICUs. So this forms a sort of two-way system to get more capacity. So in our former system, it was all one way, up or out, and that was our only way. And now we have a two-way system where sometimes the patients can go directly to the community hospital ICUs, sometimes to their ER, sometimes not, and patients at the medical center can find their way into the community. In order to do this, it requires a restructuring of the financial agreements between the institutions and the third party payers in order so that it can be a financial win for everyone. Expert case review is really one of the keystones of what we offer. The example here is a neuro intensivist making a decision about TPA. This has been very well worked out. To really make this work well, though, across the region, there needs to be standards for how the neurological exam is done by the folks that are helping. So you have to have standards and competencies across your organization to achieve efficiency, more efficient operations. We do know from our outcome studies and from a validated instrument that we looked about what folks were doing that the most important thing is early intensivist case involvement, that phenotypic recognition is important. It's one of the only studies that's shown that interprofessional rounding actually affects outcomes, so we're very excited about that. And early warning and recognition of at-risk phenotypes is also one of the predictors. The more things you change, the better you get, and you can see that the people that changed more did get better. The key points are that regional coordination increases access to critical care by eliminating waste and redundancy, that logistical support, including ICU bed assignment and ICU care plan review by an expert clinician improves outcomes, and that early access to critical care professionals improves outcomes and therapeutic efficiency when it's time sensitive. Thank you very much. I appreciate the opportunity to be here. Thank you.

regionalizing healthcare

early access to care

automation strategies

effective communication

improving patient outcomes