Operator:
It is now my pleasure to turn today's program over to Liz Olson. Liz, the floor is yours.
Liz Olson:
Thank you, Kaylynn. On behalf of the American Heart Association and Get With The Guidelines Resuscitation, I would like to welcome you to the first in three events in our Fall Quality Improvement webinar series, Start Measuring, Start Improving. Today's webinar is Report on Risk Adjusted Survival, an Overview, presented by Dr. Paul Chan. My name is Liz Olson and I'm the national program manager for Get With The GuidelinesResuscitation. On today's webinar, we'll hear from Dr. Paul Chan, the lead in developing this groundbreaking risk-adjusted survival report available through Get With The Guidelines-Resuscitation. Dr. Chan will highlight the most recent report released earlier this month and review with us the details of the report and trends in overall survival outcomes for in-hospital cardiac arrest. This session is designed to offer an opportunity for Q&A with our speaker and we encourage your feedback and participation in this event. We invite to you submit questions throughout today’s presentation by using the “Q&A” button in the lower corner of your screen. A recording of today's webinar will be made available on the American Heart Association website, heart.org/quality.
It is my pleasure to now introduce our speaker for today. Dr. Paul Chan is professor of internal medicine at the University of Missouri-Kansas City and is immediate past chair of the Get With The GuidelinesResuscitation Clinical Working Group. He's an internationally recognized leader in cardiac arrest and cardiovascular outcomes research. Trained at both Johns Hopkins and the Harvard Schools of Medicine, Dr. Chan has a master’s degree in biostatistics and clinical trial design. He's a cardiologist by training. Dr. Chan has conducted a number of the Seminole studies examining the epidemiology process of care and outcomes of in-hospital cardiac arrest. These include three New England Journal of Medicine papers on time to defibrillation, long term outcomes and trends in survival after in-hospital cardiac arrest. He has nine JAMA papers and over 150 publications in total. Dr. Chan's work on in-hospital cardiac arrest has laid the groundwork for hospital quality improvement and national performance measure development for resuscitation care in hospitals. Most recently, he has secured NIH funding to define best practices for in-hospital cardiac arrest. It's now my pleasure to turn the webinar over to Dr. Paul Chan.
Dr. Paul Chan:
Thank you for the kind introduction, Liz, and thank you all for taking time out of your day to join us for this webinar today. Let me just lay out what we're going to do for the next 50 to 55 minutes. I'm going to spend the next 25 minutes or so describing, using some slides, what the risk standardized measure for survival for in-hospital cardiac arrest is and then we'll field questions from you all. So feel free to submit questions as we go along and I'll try to answer as many of them as possible in the remaining 20 to 25 minutes.
So let's get started. So risk standardized survival rates for in-hospital cardiac arrest. Just to start off with regards to my disclosures, I do receive support from the NIH for work on in-hospital cardiac arrest, and I don't have any conflicts to disclose with companies or industry.
So the overview of the webinar today is four fold. Number one, to first describe why we even need to risk standardize, or you could call it risk-adjust, survival outcomes for your patients at your hospital for in-hospital cardiac arrest. Why do we need to do it and not just use the proportions that you can derive off the tool online. The second thing we're going to address is how is this done in terms of the risk standardization. The third is how to interpret your specific hospital's risk standardized survival rate for in-hospital cardiac arrest. And lastly, what are the next steps from the registry standpoint and certainly from your standpoint as a hospital participant after getting these reports and how do you make use of them.
So to start off, why do we need risk standardization of survival rates for in-hospital cardiac arrest? As many of you know, the registry -- this registry was previously known as the National Registry for CPR, and it was founded in 2000, and at that time, there was really not a lot known about hospital cardiac arrests. We knew a lot more and still do to this day what goes on for cardiac arrests occurring outside the hospital. And so in collaboration with a number of entities, the American Heart Association launched this registry in 2000, and since then, there's been actually over 250,000 cases of in-hospital cardiac arrests submitted by hospitals like yourselves from over 600 sites around the country. There are about 5,000-6,000 hospitals around the country, so we have received data from about 10% of hospitals around the country in the United States over this time period. Some of you have been with us for a long time, from the inception almost, and others have recently joined. And so we have basically cases from a variety of hospitals, both academic and non-academic, small, large and medium-sized hospitals.
One of the things we recently published, this was in the New England Journal of Medicine in 2015, was this paper describing trends in survival for patients who have a cardiac arrest inside the hospital setting. And we use data from Get With The GuidelinesResuscitation to describe have we gotten better over the last decade in treating patients when they have a cardiac arrest in the hospital setting. What we found when we plotted data, as you can see on the X axis from 2000 to 2009, and if you look at the blue diamonds, these are all cardiac all-comers for cardiac arrest, regardless of the initial rhythm. During the years of 2000 to 2001, the unadjusted or the raw rates of survival to discharge for all-comers within hospital cardiac arrest was about 13-15%. And by the end of the decade, by the time we hit 2008, 2009, it had climbed measurably to about 18-19%. And since this publication was done, I can tell you also that the survival rate has continued to increase to about 23-24% for all-comers within hospital cardiac arrests within the registry in the years 2014 and 2015. So we've seen a near doubling of survival rates for all-comers within hospital cardiac arrest over the last 15 years within the registry. And when we parse this out by rhythm type, if you see the orange rectangle boxes, for the shockable VF, ventricular fibrillation and pulseless ventricular tachycardia rhythms, and the green triangles for the non-shockable asystole and pulseless electrical activity cardiac arrests, we see, again, that it is increased for both the shockable and the non-shockable rhythms. For the shockable rhythms, really at the turn of the millennium, 200-2001, we were looking at survival rates where 30% unadjusted, and now they're approaching 40-45% in 2014 and 2015. For asystole and PEA, the non-shockable arrests, we've climbed from really 5-8% in the early years to the end of this study, which was closer to 12%, and since then, closer to 15%. So we've seen a concerted improvement across the board at the patient level when we look at trends for in-hospital cardiac arrest survival.
And one of the things we want to really encourage hospitals to do and one of the reasons why your hospital is participating in the registry, I'm sure, is to really focus in a more laser-like way looking at both quality and outcomes for your patients with in-hospital cardiac arrest. And the reason for developing the risk standardization reports for survival in this registry was that if you were able to go to the online tool and pull down your survival rate for your hospital, it would be a simple proportion, how many people had a cardiac arrest in the year 2015 and how many survived? But it's really hard to know what to make of that number except for the absolute proportion. How do you do relative to other hospitals, who also participate in the registry for their patients with in-hospital cardiac arrest? Unfortunately there was really no great way to really compare it because your patients at your hospital might be very different from patients at another hospital down the road in your city. They could be sicker than the other hospitals or they could be healthier. And so really, the results have been raw unadjusted comparisons of hospital outcomes for their patients with in-hospital cardiac arrest.
So let me give you an example. Take two hospitals, hospital A and hospital B, and let's say both hospitals have 100 in-hospital cardiac arrests. Hospital A, of those 100 patients, 30 of them survived at discharge. Hospital B, 40 of them survived to discharge. So on the surface, it would appear that hospital B is doing a better job for their patients with in-hospital cardiac arrest because they have a 40% survival rate compared to hospital A, which was 30%. But when we dig a little deeper, let's say for hospital A, we find of those 100 in-hospital cardiac arrests, 20 of them were shockable VT/VF and 80 were non-shockable asystole and PEA as compared to hospital B which had the opposite with 80% of them being shockable and 20 being non-shockable. And when you break it down between hospital A and hospital B, for hospital A which had the 30% overall survival rate, of their 20 shockable cardiac arrests, 14 survived at discharge. Of their 80 non-shockable, 16 survived at discharge. So their shockable survival rates were 70% and their non-shockable was 20%. Hospital B, which on the surface had a 40% raw survival rate for their 80 VF/VT patients, 38 survive or 47.5%, and for the remaining 20 cardiac arrests that were not shockable, only 2 survived or a 10% survival rate. So now if you compare hospital A and hospital B, what we're seeing is even though the overall survival rate of hospital A was 10% lower than hospital B, for both the shockable and the non-shockable rhythm survival rates, hospital A did better. And the reason why this was the case is we were able to adjust for key predictor and determinant of who survives. Because the patients in hospital A had a very different cardiac arrest rhythm makeup, their raw survival rate appeared worse when, in fact, they actually were performing better by rhythm type for each type compared to hospital B.
So what risk standardization of survival rates or risk adjustment does is it provides a fair apples to apples comparison when you take your hospital and compare it with another hospital and their patients with in-hospital cardiac arrest. And it's not only comparing apples to apples by making sure that you have a similar comparison of cardiac arrest rhythm, but it also adjusts for things that are also known to be predictive of survival, like the patient's age, how sick the patients are when they initially have their cardiac arrest, and other factors. And so what it tries to do is to provide a more level playing field for comparing similar patients from your hospitals to similar patients at other hospitals.
So how is this done in terms of the risk standardization? The way it was done was we created a predictive multivariable model using biostatistical calculations to basically determine what the predictors were for survival to discharge overall for patients with in-hospital cardiac arrest. Then we ultimately use statistics to determine which were the most important variables to keep in the model to adjust for it. And so in the adult setting, there are nine key variables that we are risk-adjusting for, risk-standardizing for when we compare one hospital versus another. The nine variables are age, obviously the older your patients are when they have an in-hospital cardiac arrest, the more likely they're to die, regardless of whatever else is going on, so we want to make sure that your age -- your hospital's patients' age is similar to other hospitals' ages when we make these comparisons of survival rates at the hospital level. The initial cardiac arrest rhythm, hospitals that have a higher rate of shockable cardiac arrest rhythms are going to have a higher crude survival rate, because we know that VF and VT have a higher survival rate to begin with. The location of the arrest, whether or not the patient is hypotensive at the time of arrest or have a diagnosis of sepsis, whether or not they have a malignancy, whether or not they have hepatic insufficiency or whether or not they were in a mechanical ventilator at the time of the arrest, or required vasopressors. These are the pivotal factors that determine whether or not a patient, when they have an arrest, is likely to live or die, all things being equal across hospitals.
In pediatrics, I'm sorry this is a little fuzzy and we can certainly forward to you the actual paper document, there were 13 variables. So somewhat different from the adults, but again the idea is the same, to try to create a similar comparison from your hospital to another hospital when we look at how your patients do when they have a cardiac arrest inside the hospital setting.
So you may or may not have seen your hospital's risk standardized survival rates. But how do you interpret it once you've seen it and obtain the result? This is a snapshot of the online sites at which you can download your hospital's risk standardized survival rate report if you haven't seen it already. If you can look at the second column where it says “My Reports” towards the bottom, what you can do is click on the -- this is a blowup of the “My Reports” section. There is a link at the bottom of the “My Reports” section that says “Get With The GuidelinesResuscitation Risk Adjusted Survival to Discharge Report.” When you click on that link, you get this pop-up which, again, you'll click on and when you click on this, you would actually get a downloadable PDF for your hospital's risk standardized survival rate. This is what the document will share with you. The first page of the document will, again, reiterate, if you're in the adult setting, the nine key variables that are used for risk adjustment that we validate every year when we re-do the reports for risk standardization and we calculate your hospital and other hospital's rates for risk adjusted survival. And in this report, you're going to see something that appears like this. This is for a hospital where we reported their cardiopulmonary arrest events occurring in the year 2012, for instance. The site number is given, the year is given, and then you're going to receive the risk standardized survival or the risk adjusted survival rates for your hospital, so it's about 20%, which means that after adjustment, your hospital's survival rate for patients with in-hospital cardiac arrest is 1 in 5. And it also provides the risk standardized survivor quintile for your hospital. It ranges from 1 to 5, with 5 being the top 20% of hospitals in survival rates for in-hospital cardiac arrests, and so forth. And so what you will also see graphically is this diagram, and I'm sorry it's a little busy, but just to orient you, the X axis, the bottom row, provides the risk standardized survival rates. The out of boundaries on the left and the right tells you what the minimum and the maximum values for risk standardized survival was for year 2012. So all hospitals have risk standardized survival rates somewhere between 8-9%, to 34% in year 2012. The solid very dark black line in the middle is the median. So this is the midpoint of the risk standardized survival rate. And then you'll see some dotted or dashed lines, and that separates the quintiles. And then the red arrow describes where this specific hospital fell for year 2012 in the risk standardized survival rate. Remember that they were 19th percentile, which means they're in the boundary between the first and the second quintile, and that's why they're near the dotted line here. If your hospital had a risk standardized survival rate of 30% instead in year 2012, you would be located further to the right and you would be in the top quintile or quintile 5. If you were closer to the middle, you would be in 2, 3 or 4. And in this report, it also summarizes what the quintiles mean. So if your hospital's risk standardized survival rate for year 2012 fell on quintile 5, that means essentially your percentile ranking for your hospital for that year, for survival was 81 to 99th percentile, kind of like the reading and math reports that you see for your kidswhen they get their assessments on an annual basis. That means that if you're in quintile 5, it was higher than 80% of hospitals for that year. If you're in quintile 4, the percentile ranking will be somewhere between 61 percentile to 80th percentile. Again, that means that it would be higher than 60% of hospitals that year. The same goes for the middle quintile, the third quintile falls between 41 to 60th percentile. And again, it means that that percentile -- or that survival rating that year would be higher than 40% of hospitals that year, and so forth.