here, it says, bundling success versus or bundling unproven therapies, all right? So that in itself is something to think about. Now, I don't have any significant financial disclosures, but first I want to say that I'm promoting this proposition, but I'm not an evangelical zealot about this, all right? So I'm not inflexible in this thought, but rather that I'm thinking that the data that we have today leans in the direction of favoring some of these interventions that we compose in the bundles. And you know, as we've said, that science behind sepsis management is changing rapidly. I like the idea of looking for better ways to detect the best antibiotics to give patients. We're still using the old methods that we've had that are tried and true. So what is a bundle? That's the thing we need to say, what is a bundle? It's not like this asparagus that's bundled up with bacon or this here, and it's not just things tied up in a bundle, but in medicine, the terms are broader than this, and it's not like with insurance and stuff with State Farm, okay? So you're talking about using things with your best of clinical judgment. So I'm just going to bear with me as I talk about what a bundle is, and I have an issue with this because everybody's starting to bundle things together and calling them a bundle and have an evangelical sort of position on this. But the initial bundles that were started by the IHI and the VHA some years ago is that the problem is that these guidelines that have a lot of energy that people put into these to get them out, that doesn't lead to practice change, right? Because practice change is difficult. And the IOM, the Institute of Medicine, had a report some years ago that it takes up to 17 years for anything to get into practice into the hands of clinicians that take care of a patient. And a recent study I saw shows that we're just as bad as we used to be. So we're just not moving the science fast enough, and so the IHI and the VHA developed these bundled sort of terms to be able to move things quicker, to have protocolized ways to get people to care for patients using the best evidence that they have. So the goal is that they improve the quality and safety based on evidence-based practices, and that usually these bundles are usually three to five, and I'll show you my next slide that shows you something very differently, that when you perform them collectively and reliably improve patient outcomes, right? And so the idea is like with critical care, we have some for central line blood infections, things to improve quality. We use those a lot in medicine, and ventilator-associated pneumonia. And so these things are actually intended to improve the care of patients, and what has happened with patients with ventilator-associated pneumonia since we've started these bundles, that is much, much less. When I first started in critical care, when we had vent pneumonia, it was 36%. Nobody said a word about it. It was just a foregone conclusion. So they're supposed to be easy to measure, and actually should be linked with time. So when you practice, you put these together, they should be performed uniformly, and they're not equal to checklists. And they're supposed to, to your point, Dr. Remy, they're supposed to be based on randomized control trials with level one evidence, and proven in scientific tests that are accepted and well-established. And in the case of sepsis, that some of the randomized control trials that have been done in the past were used to construct the initial bundles that they put together for this initial sepsis six, six-hour and 24-hour bundles, the intervention, I mean, the resuscitation and management bundles. And that when you fail to do them, that you would not have as good of an outcome. So with these bundles, the idea is when it's an all or none sort of thing, which is kind of disturbing a little bit when you're talking about patient care, but in other words, if we use the sepsis bundle but we didn't administer enough IV fluids, you may not have the same outcome. Okay. That's what the idea is. So the bundle development in sepsis started in 2004 after the guidelines were constructed. The surviving sepsis campaign folks who partnered with the IHI to develop these bundles, and they used the evidence in the literature that I've seen about how the bundles were developed. There's not a lot of discrete information of how they came up with these randomized control trials that they used to do these, but that they did use these trials to support the guidelines for the bundles. And then in concomitant with that, I think they put the cart maybe before the horse in that they developed the bundles and then started to test to see if they made a difference. So the database that was developed, the online database was developed to evaluate the progress on these bundles to see whether or not they would have an impact on outcomes through these global collaboratives. And I don't know if you all remember the initial bundles that were up. These are unwieldy. I mean, they have everything but the kitchen sink that's in there, and those are not really inconsistent with what we would hope in a bundle that would be short and sweet and that they were proven entities. But there they are, and they had everything from glucose control to PEEP and all these other things that you would hope to achieve. But these bundles were begun back in 2005, and the first study that was done that I could find was the initial measure on people who developed the guidelines did this two-phase approach for their six-hour and 24-hour bundles and used this data from the surviving sepsis collaboratives. And they did test compliance with each element with the bundle over time. So they didn't compare things, they just looked at how well mortality improved or didn't improve over time as you implemented these bundle elements. And it was from 2005 and 2008, and they compared the initial quarter with the final quarters, and then they looked at all kinds of hospital characteristics. And these kind of what the findings show, they had 15,000 patients. Now keep in mind, these were people who voluntarily joined the collaboratives. So you may have a little bit of bias towards really being pro-bundle if you're joining these collaboratives, but they had 165 sites that joined this with a median sample size, 57, and they were in for basically 14, 15 months. And what they found in this study is that they had an absolute drop in mortality of 0.8% per quarter, and then over a two-year period they would find a 5.4% drop over that period of time. So their findings were basically they supported this bundle because the adjusted odds ratio improved the longer that someone stayed in this collaborative, that their mortality tended to decrease. And they saw as more you compliant with it, you had more of a decrease in hospital mortality. Now it's interesting, most of these studies looked at compliance versus mortality, and we all know in this room there's sometimes things that are worse than dying. And so that's an easy-to-measure binary outcome. So that in itself might be problematic. But when you look at this, if you're looking at compliance, you can see in this graph is as you increase your compliance using all these crazy things in this bundle they started out with, that mortality did favor the intervention in going down over time. Now whether the bundle had everything to do with it or not, I can't say. But those are the results. So later they got more folks in this collaborative, and after 77 and a half years of patients in this, the database grew. So they studied some of the same patients from 2005 to 2012, and they again studied compliance and mortality, which is a little different when they looked at people who did the bundles versus the people who didn't do the bundles. Now at least we know if compliance with the bundles, we know that they did all the elements in there as opposed to something else. And most patients would have received some of the elements that are in the bundle anyway. So that's kind of a mixture of you don't know what those patients are getting. But it included 219 hospitals in the United States, South America, Europe, and included med surg, ED, and ICU patients. And so their results with the resuscitation bundle, they found that there was lower mortality in the people who had high bundle compliance over the lower. Now remember that in this particular study, low compliance was less than 15%. That's really low. And higher compliance was greater than 15%, so it could have been 16%. So that's one of the things that might take into consideration when you're looking at these results. But despite that, in those two groups, 29% of mortality versus 38% was significantly different. Now in that management bundle of 24 hours, there was the same sort of thing that they saw a little bit less in compliance, I mean mortality protection, but 32 to 38, 33.8% still showed a statistically significant difference. So their results showed that they had 25% relative reduction in mortality, and for every 10% increase in compliance, an additional quarter was associated with significant odds ratio. And here's a graph from that particular study, and you can see here that as the resuscitation bundle increased, it looked like there was a bit of a plateau here, right? So about after two years or something involved, then they didn't get much better with that over time. Whereas with both of these, actually there seemed to be a little dip, but then regardless, the mortality tended to fall. So, you know, that suggests a relationship with that, that every quarter was associated with like a 0.7% mortality reduction. Now there's some other folks that have done studies, I'll go through these really quickly, just to show that, again, these are systematic reviews and meta-analysis of observational studies because that's all they can find, again, that we don't have these randomized controlled trials on this. But they looked at the impact of these PIs that people were having for sepsis in their organizations and did a study on those. And most of them were before and after, and observational trials, again, you know, to support what the con side of this is that there's no randomized controlled trials out there. And they looked at changes in compliance with bundle mortality, and they found 50 studies that included, and they found that with these performance improvements that they increased six-hour bundle compliance, and it reduced mortality and odds ratio of 0.66. And that's with like a half a million patients that they put in this meta-analysis. And here's the forest plot showing all of these studies, and if you don't have to be a rocket scientist to see most of this stuff, it's definitely less than one and favors the intervention of using the PI. So the IMPRESS study that was also done was a global international study. It was, again, observational. Compliance was really the intervention versus non-compliance in about 1800 patients in 62 countries. And their compliance, again, was very low. Again, that's kind of an issue when you're trying to compare things, but if you're seeing whether the bundle works, we're saying if the bundle is compliance with the bundle versus non-compliance with the bundle, that's what the studies are about. But again, they saw three-hour bundle compliance was lower mortality by almost 11%. And then the six-hour bundle, again, this is a three and six-hour bundle as science changed and saw very, very same sort of results in the six-hour bundle as well. And so it does show that there's improved hospital mortality for both the three and six-hour bundles with lower odds ratios. But they said that in this study that there was a 40% reductions in odds of death for those two bundles. Now let's talk about what happened in New York. We were talking about mandated things and the Rory, of course, what happened to that poor child. It happens every day in this country. It was a retrospective cohort study in New York State when mandating the programs. And they looked at all payer discharge in New York State and control groups. I believe these particular states, Florida, Maryland, Massachusetts, New Jersey, had somewhat robust sepsis programs in place between January to March of 13 and then April 13 to September 15. And they compared, it was interrupted time series approach. They looked at a million patients basically in 509 hospitals and before was their mortality in New York State was 26 versus 22% in those controls. So New York State was worse than their control. And then after New York State went up, I mean, they went down in the mortality and also the control went down in the mortality. So, you know, basically they did show that there were positive outcomes with this. The other study that was done on New York mandated hospitals was done again by some of the originators of the surviving sepsis campaign. Also looking at three and six hour bundles for mortality outcomes. And this was done in 14 and 15 with 91,000 patients. And at least 81% of patients had sepsis protocols initiated and their compliance rate went up. That's an incredible amount of compliance for the sepsis bundle actually. Went up from 53 to 64%. And for the six hour bundle went up about 7%. And the risk adjusted mortality was from 29 to 24% in those who had the sepsis control protocol initiated. And then, you know, not to belabor this point, but a recent study was done with Medicare beneficiaries in sepsis compliance with the SEP1 bundle. A lot of patients in this study, there were 3,252 hospitals using the SEP1 bundle. And they did some complex matching with propensity scoring with standard and stringent matching techniques, which is Barbayama understanding of statistics, but they compared the bundle compliant cases with the same bundle non-bundle compliant cases. And you can see in the standard group of 122,000, 123,000 for both groups demonstrated a decrease in 30 day mortality, 21.8 versus 27.5. And the absolute reduction was 5.67%. And in the stringent, the most stringent criteria came in 107,000 patients in both groups also demonstrated a decreased mortality as compared to the ones without the bundle compliance. Now, what about pediatrics? There's very little in the literature about pediatric sepsis. There was one that was done in Thailand in children, and they did 188 assigned to an intervention group and 331 in historical case control, which can get kind of messy. But they found in those cases that they tended to use more IV fluids, of course, because you're using the bundle, and more basal pressors, again, compared to control. And they did see a significantly reduced mortality. And this is what their graphs look like with that. So, and again, Dr. Workman did an interesting study with, it was again a retrospective study in 321 kids, and they looked at surviving sepsis compliant care in the first hour versus those that were a little slower. That was not really defined that well, but their outcomes measure was not, for a primary, was not mortality, but progressive multi-organ dysfunction syndrome, and then secondary was mortality and need for mechanical ventilation or vasoactive medications. But they didn't find any differences in multiple organs dysfunction syndrome, nor the secondary outcomes, but they did say that most all patients received their interventions within three hours. So, they claim that this, you know, at their shop, they were able to deliver timely care in these patients. So, what's the rub? I think we're going to find out what the rub is in just a minute. There's significant rubs, and I totally understand those. But, you know, some of the problems about these bundles is that there's been changes, six changes in the guidelines since they started 20 years ago. And the science is continuously changing. There's a lot of ambiguity about the amount of IV fluids, what types of IV fluids, steroids, no steroids, this or that. And, of course, when we went into CMS core measures as a mandate, really, for most hospitals to be measured against, that changed the scene a bit when people felt like they were compelled to use these bundles. And there's been changes in the bundle elements, timing. The bundles have gone from this to this now. So, you know, to Dr. Remy's point, why can't we get some randomized control trials done since we have such a smaller group of elements? And then one of the problems with the one hour, hour one bundle is that this has not really been tested. I think the whole point behind that was to get people moving faster, to take care of these patients quicker, and that they would presume that this would result in better outcomes. And then one of the problems is that the non-compliant patients probably got some elements of the bundle anyway. We just don't know what they are, and they are not reported. So, and then again, the weak, where there are a lot of weak recommendations in these that are retrospectives and comparing compliance with non-compliance. This is the one caveat that I do want to say that is in every guideline that's talked about bundles, is these recommendations are intended to provide guidance for the clinician caring for a patient with severe sepsis or septic shock, but they are not applicable to all patients. Recommendations from these guidelines cannot replace a clinician's decision-making capability when he or she is provided with a patient's unique set of clinical variables. So, in summary, given the outcomes, I'm leaning towards the balance towards favoring the bundles, given that we have seen a concomitant change in mortality over time. Not that all the elements are equal, neither are the patients, but I believe it's a reasonable approach to start sepsis care, and then I will turn this over to my friend.

bundling

sepsis interventions

standardized care

evidence-based methods

bundle compliance

clinical judgment