Venous Thromboembolism (VTE) Prevention in the Hospital: Transcript

This is the transcript of Greg Maynard's web conference presentation, Venous Thromboembolism (VTE) Prevention in the Hospital. Select to access the presentation slides.

Hello everyone and welcome to this Webinar on a very important topic of designing and implementing DVT prevention programs for your medical centers. This is obviously a major, major problem that we do really need to address. Venous thromboembolism or VTE is a major source of mortality and morbidity. The estimates are that 350,000-650,000 people develop VTE per year and that up to 200,000 deaths per year occur from VTE. Many of these deaths, in fact most of them, are hospital related. It is estimated that VTE is the primary cause of fatality in half of those who develop PE.

We should keep in mind the scope of this, this is more than HIV or AIDs and motor vehicle accidents and breast cancer combined. And just as medical errors caught the attention of the Institute of Medicine when they use their analogy, VTE deaths equal about a jumbo jet crashing per day. So, a very large cost and mortality, and probably the number one preventable cause of hospital death. And in addition, very significant morbidity with high rates of DVT recurrence, disabling post thrombotic syndrome with subsequent swelling, pain and often ulceration and complications along those lines as well as chronic pulmonary artery hypertension.

There are a myriad of risk factors for venous thromboembolism. And we are not going to go through every single one of them. The important thing is to note that basically all hospitalized patients have at least one of these risk factors for venous thromboembolism. Most have more than one. These factors are more than additive, that is, if you have 3 risk factors its likely you have 5 or 6 times the risk of VTE than if you have one.

VTE prevention is a symptom of a larger problem that is if we are to do VTE prevention effectively the symptom of a larger problem that we have a hard time in health care doing a lot of simple things well. And these very simple things can make a huge difference in the outcomes of our patients. Washing hands and making sure that patients understand their medications, being sure that central lines are placed correctly using sterile technique and using a safety checklist for that making sure that patients with uncontrolled diabetes get basil insulin. All of these things are occurring with 40-60% reliability instead of with the near 95-100% reliability that our patients deserve. DVT prophylaxis is right in there. Most estimates are that that's about what the performance is in U.S. hospitals. That is, only 50 % of patients are getting the prophylaxis they should get compared to the national guidelines.

The registry data bear this out in both surgical and medical patients, you know it's really both surgical and medical patients that have this problem, it's probably even more pronounced in most studies on medical patients. For example, in a large multi-national endorse registry they looked at over 70,000 patients at 358 hospitals and appropriate prophylaxis was administered in fewer than 60% of the surgical patients, and fewer than 40% of the hospitalized medical patients. United States was somewhere in the middle of the pack of these multiple nations that were involved in this study, and we should not get too smug thinking that this is a multi-national study and that surely the U.S. does better than these other countries. The fact is we do not.

Another reason this should be on your radar and on your administration's radar, is that most certainly the stick is coming. That is, there will be measurement around DVT prophylaxis and in DVT management in your hospital in the future. The NQF or National Quality Forum already endorses these measures and public reporting and the Joint Commission measures are coming soon. In fact, the Joint Commission already has a module which people can use, which medical centers can choose as one of their quality measures. These measures include 3 factors around VTE prevention. Prophylaxis should be in place within 24 hours of admission or a risk assessment showing low risk, or some indication that pharmacologic prophylaxis is contraindicated need to be there. The same would be true for critical care unit admissions and transfers and they're also going to be asking medical centers to track their preventable venous thromboembolism events. This is starting to affect the payment world as well. CMS, Medicare/Medicaid they will they are treating DVT or PE with knee or hip replacement as a never event. That is they will reimburse knee and hip replacement that are complicated with DVT as though that complication has not occurred.

So it seems pretty simple, we have good quality guidelines in the American College of Chest Physicians (ACCP). Latest version is ACCP8, which tells you something, they've gone through 8 versions of the guidelines, and yet we're still really in many areas only at 50, 55% DVT prophylaxis, or lower in many cases. So why is that? What is so hard about this? Well, my experience in the realm of quality improvement is that this is actually doable and relatively low hanging fruit. But we have to acknowledge these potential barriers. First there is a lack of awareness or buy-in or guidelines by many physicians and nurses; second, there is an underestimation of clot-risk. Any one physician only sees DVT or PE in their patients, rarely. Whereas across the medical center, these events that are rare for any given, in the experience of any given doctor, becomes actually fairly common across multiple doctors that come to the medical center. And physicians tend to remember that one complication of bleeding risk, rather than the DVT or PE that may have occurred after the patient left their care. There've been problems because prior to last year, there was no validated risk assessment model for DVT, that is, a model that was prospectively shown to project the risk of DVT on a given in-patient was not available, and a validated protocol showing that implementation of such a protocol affected better outcomes was also not really available. Also, if you read the ACCP8 guidelines, which I do certainly urge you to do, you'll see that they're fairly long, they can be somewhat complex and it's difficult sometimes to translate those guidelines into everyday practice.

Well what's been shown to be effective in DVT prevention and boosting the rates of DVT prevention and DVT prophylaxis? Well, certainly we know that E-Alerts can increase prophylaxis. In this New England Journal of Medicine study, in over 2500 hospitalized patients, with the VTE risk score, that was elevated, that is they had risk factors for DVT. They were randomized to intervention or control groups and as you can see the E-Alert patients received higher prophylaxis. And because of this, they also had a decrease in subsequent development of hospital acquired VTE, over time. Unfortunately the effectiveness of the E-Alert or electronic alert can wane over time and here you see that overall the effectiveness of the E-alert went down as time passed. More recently, Dr. Piazza reported that human alerts can also increase prophylaxis. There's nothing magical about the electronic interface and perhaps sometimes in some settings the human alert can be better. Again in almost 2500 hospitalized patients they pretty much replicated the study that was done with E-alerts and they randomized patients who had high risk who were not on prophylaxis to human alert and once again they demonstrated that this alerting system resulted in higher rates in prophylaxis and they also decreased venous thromboembolism in those who received the human alerts. Note that these results are relatively modest, but then again these are patients who were at risk who were not on prophylaxis to begin with. So the bottom line about alerts, is, it is a useful strategy, no doubt, E-alerts and hume-alerts can work but they are not a panacea and you can get alert fatigue. You have to be cautious about how you use these alerts and in reality you probably need a multifaceted approach where you're tracking the problem of low prophylaxis rates in more than one way.

Medical admission order sets have been used and certainly, I'm a big proponent of using medical admission order sets and integrating protocol guidance and order sets. I think that is probably the number one strategy to boost DVT prophylaxis rates across your institution. You just have to be careful about how that's done. This study was done by O'Connor et al in Canada and at baseline, only 11% of their in-patients were on any VTE prophylaxis. They put a simple prompt for unfractionated Heparin or mechanical prophylaxis into voluntary admission order sets. That is they just made these order sets available; they didn't mandate that people use them. Post intervention they did secure a good amount of improvement. 44% were on some prophylaxis, only 26% were on pharmacologic prophylaxis, but nonetheless this was a definite improvement. The problem is I don't think you or your patients should be satisfied with that level of DVT Prophylaxis. That is not nearly the kind of reliability that our patients deserve.

So there some things that they could've done differently, in my opinion that would've boosted their DVT prevention rates to a much higher degree without a whole lot of extra effort and those are some of the things that we'll be talking about in the rest of the Webinar. First of all, in their order set best practice for prophylaxis was not defined, so they simply received a prompt that had a menu of options and one choice was deemed as good as the other. There was no guidance about which choice was better, no prophylaxis versus mechanical, versus Heparin, versus low molecular weight Heparin. So this prompt is not really the same as having protocol guidance at the point of care. The implementation and reliability of that use of the order set was poor, at 15 months only about half their in-patient admissions even used the order set. So if you've got an order set that's perfect, but it's only being used 50% of the time, that is not going to get the job done. Has to be used much more reliably, and you need other methods to enhance the performance of the order set.

Education is always going to be necessary. You're going to have to have some sort of an educational effort go on for your providers and quite possibly your nurses and pharmacists and it's essential to optimize these other strategies that are effective. Standardized order sets, computerized decision support, E-Alerts, human alerts, raising the situational awareness that we've got a patient at risk in a bed, they're DVT and their not on prophylaxis. That's going to be a key strategy as well as audit and feedback.

This is from the University of California San Diego experience. We published this in the Journal of Hospital Medicine. And getting to very high levels of performance took us almost 3 years. It's taken many others who have been using our road map often 6-12 months and sometimes shorter. So we expect that you can reach very high rates of DVT prophylaxis in a fairly short period of time if you follow the guidelines and use the lessons that we've learned over the course of our experience and the collaborative that we've been involved in. What you see here is that at baseline we're at about 50% by the time we built consensus, good education, and worked with key med exec leaders we got up to about 70% adequate prophylaxis with the fairly rigorous definition of adequate prophylaxis which we'll get into in a bit. We put that protocol guidance into an order set and implemented and we got up to 80% and then we tweaked the order set further and got up to 90% adequate prophylaxis. And then we used another strategy where we identified patients who were not on the right prophylaxis in real time while they were still in the bed and we could do something about it. And that has boosted our rates of prophylaxis to over 98% and this graph doesn't go out that far but we're now at sustained 98% plus adequate prophylaxis for over 2 years. So we have basically validated the UCSD VTE prevention and protocol, it's easy to use it only takes a few seconds, it has excellent inter-observer agreement, that is if you think a patients at low risk, I will think so also, if you think that they're on the right prophylaxis I would think so also, if we're using the same tool.

The risk assessment model I'll be showing you is predictive of VTE, the implementation rate is high and because the implementation was good, we went to this very high rate of prophylaxis, it was safe; there was no discernible increase in HIT or bleeding, HIT being Heparin Induced Thrombocytopenia and it was effective. We had big reductions of hospital acquired VTE and an even bigger reduction in preventable hospital acquired VTE. Realize that you won't be able to prevent all cases of hospital acquired VTE. All the randomized control trials show about a 50%-60% efficacy or so of pharmacologic prophylaxis and not everyone in your hospital is going to be able to be on pharmacologic prophylaxis because of some very real contraindications.

So, this slide depicts the rate of preventable hospital acquired VTE per quarter and you can see that we had at the onset across all services around 13 preventable hospital acquired VTE per quarter, and we've driven that down now to the point where it's about 0 or 1 or 2 per quarter. And this is just a table depicting what's happened to hospital acquired VTE here, cases with VTE, cases with PE, cases with DVT, and cases with preventable VTE. They've all come down substantially and statistically significantly and basically has done that without any increase in bleeding or HIT. So in our smallish Hillcrest Medical Center at UCSD which is only 350 beds, the bottom line is that you're probably going to have a reduction of about 50 hospital acquired events in a 350 bed hospital.

These publications, these tools, are implementation guides that have been crafted through our experience and through the experience of others like Jason Stein at Emory, and many, many others who've contributed to building this sort of implementation guide for how to build a good order set and to put that order set into place and how to monitor it, how to measure its effectiveness, and how to tweak it until you get very high levels. So I encourage you to have a look at these, they are sister publications that are quite similar, the Society for Hospital Medicine (SMH) one is surrounded by other materials like slide sets and a bibliography and things like that. Whereas I think the AHRQ publication that you see on the other side of the page may be a bit more concise and may be a bit more updated.

Now, on this presentation today I'll be updating you even further than what these publications have presented and we've evolved our way to measure I think DVT prophylaxis event that makes it even more doable and practical across your many institutions, so be attuned to that but nonetheless I think these are very worthwhile to pull down and try to look through if you're part of any leadership of an improvement team.

This is just a screenshot of the Society of Hospital of Medicine quality improvement resource room. VTE is only one of the resource rooms available. And basically these efforts have spurred collaborative efforts using those toolkits as sort of the center piece. The concepts and tricks and methods I'll be discussing with you have been tried out at many different sites across many different kinds of settings; small and large institutions; academic and community centers, paper and computerized order sets. So, really I think these concepts have been tested on the front lines across a variety of different settings and should be applicable to your setting.

So we're going to get into some specifics, but let's hit the big picture first. The framework for quality improvement in general, it's important to sort of make sure you've got these things in place and sort of you know if you're running into problems go back to this basic check list and make sure nothing's really missing. So first of all, you need to have some good institutional support. By institutional support I mean the institution has to have high VTE prophylaxis rates, as part of its priority list and that means they have to give you as the improvement leaders, for this effort, some time, rooms to meet, and more importantly they have to give you support to standardize the process, standardize the order sets, etc. They have to back you up in that regard and they have to back you up when you tell them you need some help and some expertise in helping monitor the progress of the project. This multi-disciplinary team would ideally have physician leadership and buy-in and physician leaders who we've commonly recruited have been hospitalists but also critical care physicians, anesthesiologists, surgeons, I mean many people are interested in this. DVT Prophylaxis cuts across a wide variety of patient care populations in a medical center. Specific goals and metrics are important; we'll be talking about these metrics and these ways to measure prophylaxis soon.

As I've alluded to what I think really helps is to build a VTE prevention protocol and build that protocol guidance into the front lines of care via order sets that are widely distributed and widely used if not mandated. Education is needed as well as consensus of course and then we'll spend a bit of time talking about how you can actually measure prophylaxis and spur further improvement and intervention on people who are not on prophylaxis in real time. So, basically, find the high volume providers. If you're worried about some niche group, perhaps a stubborn orthopedic group or some other group that you think doesn't want to play ball initially, I wouldn't sweat that too much. I would focus on your high volume providers; realize that the medical patient population is at very high risk by virtue of their prevalence throughout your hospital and by virtue of the fact that they're likely not prophylaxes at a rate as high as your surgical patients.

So be sure you incorporate the medical population into this effort. You should become familiar with this key reference which is the ACCP8 American College of Trust Physicians, 8^th version of the guidelines. And some of the key things are that you don't want to use aspirin alone for DVT prophylaxis, it's not nearly as effective as Heparin or low molecular weight Heparin. Another key concept is that mechanical prophylaxis alone is not to be a first line prophylaxis agent in the absence of contraindications to pharmacologic prophylaxis. This is true for a number of reasons, one, it's not as effective in some populations as pharmacologic prophylaxis and two, it's very hard, to keep mechanical prophylaxis in place reliably. In our experience, we've never reached a month where more than 60% of our in-patients who have mechanical prophylaxis ordered have it actually on when we go and audit the patient.

I would encourage the champions of the effort to be alert to identify the powerful anecdote as well as the objective data. You can quote statistics; you can quote ACCP8, etc. But likely it will resonate just as much or more with your physician population and your nurse population if you can identify that case that occurred in your hospital where a patient developed a bad DVT or a pulmonary embolism and subsequent mortality or morbidity, needlessly. That is they should've been on prophylaxis and they weren't. If you can identify that case and put a face on it, it can really be a powerful tool. Certainly you should look for data about how often VTE occurs in your medical center, and I can guarantee you it likely occurs more often than the doctors think that it does.

Alright, well let's get down to brass tacks, so I say you should have a protocol, I say you should have an order set, what do I mean by that? And how should that be done? Well, first question is, what's the best VTE risk assessment model? And our take on this that a simple text based model with only 2 or 3 layers of VTE risk is the way to go as opposed to a more complicated strategy, a more complicated risk assessment model. And I'll show examples of this in just a moment, of both types.

Second question is who should do the VTE risk assessment? In some centers they're trying to have the nurse do the risk assessment and then they have the nurse put the risk assessment on the chart for the doctor to complete and take action on. And again, in our experience, and the experience of many, many, sites cross country, we think that it's probably best that the doctor or the admitting health care professional, the one who does the admission and transfer orders do the primary risk assessment. They already know the patient, they're in the best position to do it quickly and efficiently and there's no reason why this should not be on the responsibility of the front line physician who's admitting the patient. Now that doesn't excuse the pharmacists, the nurse, from being part of this effort or part of the team. In fact, we think a very key strategy that we'll be lying out for you is to have front line nurses or pharmacists focus on those who are not receiving prophylaxis as a backup mechanism. And we'll go into details of that in a bit.

It may be best to sort of put this in context of a construct we call hierarchy of reliability. If you want to get to very high, high levels of reliability of providing anything, in this case DVT prophylaxis, you should sort of see where your current efforts fall in the hierarchy of reliability and try to move your way up from level one up to level 5. So, not only can you sort of gauge where you're at in terms of process, you can probably get a pretty good estimate of what percentage of your patients are on mechanical prophylaxis just by figuring out where you're at in the process. So, basically, Level 1 in the hierarchy is that you have no protocol, you have great guidelines but nobody's taken the trouble to translate those guidelines into a local protocol that says who should do what kind of prophylaxis for what kind of patient. In level 2 you've got some sort of decisions for it, somebody has done that, but that guidance is on a shelf somewhere, it's not integrated into the flow of work. Or, you have what we talked about earlier where you just have a simple prompt in the order sets. It doesn't give any guidance; it's just a prompt to do something about DVT prophylaxis. With that you'll get a little bang for your buck, but not much.

The first real big bang for your buck, which is why we highlighted this level in yellow, is that you crystallize your protocol succinctly and you integrate that protocol guidance into orders at the point of care so that the provider who's ordering DVT prophylaxis as well as a whole range of other admission and transfer orders has that guidance available to them when they need it and where they need it. It has to be simple, it has to be easy to use, but it has to be virtually ubiquitous and hard to get around whenever they order whenever they admit or transfer someone they should have to sort of go through that order set process. In level 4, you work hard to make sure that that's the case and you close off the loops that allow people to go around your protocol, and in level 5 is the next big bang for your buck we'll talk about a bit more where you identify people who have somehow gotten around the protocol and are not on the right prophylaxis and you catch those and address them in real time, which boosts you up to that very high level of 95%+.

When I say VTE protocol, what I mean is a standardize VTE risk assessment that's linked to a menu of appropriate prophylaxis options as well as an out for when you have contraindications of pharmacologic VTE prophylaxis in which case you would be promoting mechanical prophylaxis as a first choice. The challenges I was saying is to make this VTE protocol easy to use and automatic to use, to make sure it captures almost all patients, and you have to walk that guideline between ease of use and efficiency. So what I'm asking improvement teams to do is to basically do an audit or a survey of all your currently existing admission and transfer order sets, that is or can be a fairly big job. But I think that's a good exercise to walk through and what you're doing is you're looking for all reference to VTE prophylaxis. You're looking for if they have any orders or prompts to use DVT prophylaxis in each one of those order sets.

The idea is that you will design a protocol driven DVT prophylaxis order set that integrates the Risk Assessment Model or R.A.M, RAM. And that you integrate that protocol driven risk assessment into your order set, you vet it, you pilot it, and then you replace all those other order sets that were already in place, and replace them with your standardized order set module. Of course you will have to monitor and tweak and do some plan, do, study, act around whether your order set is even being used and if it's being used correctly and that's part of the process.

This is from a 250 bed hospital that I administered a process, this is courtesy of Christine LunLung from Cooter Valley in Colorado and she looked for all their order sets, they're on paper at the time and all these order sets had reference, had some allusion to DVT prophylaxis in them. So that, was her task is to take the non standardized items out and replace them with the standardized DVT order sets. This would be an example of too little guidance that I was referring to, where you just have a prompt. All these choices look the same; there is no real guidance about what's better. In fact no prophylaxis looks just as attractive or appropriate as low molecular weight Heparin.

Many places that I've worked with use a point based system when we start this process. I'll show you an example of that here. So, it's hard to get it all on the screen at the same time, but this was an order set that was in actual use at an institution, and they, like many other centers, when they actually assessed how widely this was being used or how properly it was being used, they found out that it was only being used a very small minority of the time. It was an order set that is detailed, that probably has validity in terms of predicting VTE risk, but it ends up just being hard to use, and bulky to use. So if you see what they're asking their doctors to do, they're asking them to look through all these contraindications, and then come down here and score, look at each one of these lines, and determine how many points they have for these risk factors associated with the clinical setting and so they're adding up points for each one of these things and putting it in this box and then you go down here and look at the risk factors associated with the patient, and they add up the points and put it in this box, then they add up this box and this box and this box, and then they finally come down to "okay if it's this score, then I look at these choices." So, it could work, it's just that in practice, it doesn't work very well. People stop reading the different risk factors line by line and they stop doing the math and end up sort of gestalting it anyway. In our view, this is a much less effective way to do things. And in fact, if you look at the differences or the resolution between say a score of 2 and a score of 3 and 4, it hardly seems worth the effort, plus, note that they offered sequential compression devices as a first line choice for both moderate and high risk patients, and that is really not the approach the ACCP8 guidelines would take which would relegate mechanical prophylaxis to an ancillary or adjunct choice instead of being a first line choice. That's an example of one with too much guidance. Now again, I really think these can be problematic. I admire Dr. Capprini and his group who first came up with this model, and they still promote this model, but I really think that it's less practical to use.

This is more an example of what we have used at UCSD and what we've used in many, many hospitals across the country. It's basically a 3 bucket model. And by 3 buckets I mean that with simple text cues you clue people into what bucket of risk they're in, high, medium, or low. On what end of the spectrum is the very high risk patients and these are the hip fracture patients, the acute spinal cord injury with paresis the multiple major trauma patient, the abdominal pelvic cancer surgery patient, and those patients who generally get low molecular weight Heparin or Arixtra, Fondaparinux, also known as, in the generic terms, or Warfarin. And, in addition there is some evidence that suggests that they should be on not only low molecular weight Heparin for example, but also mechanical prophylaxis at the same time. In fact that was our definition of adequate prophylaxis for these patients they had to have both to be considered adequate prophylaxis at UCSD. On the other end of the spectrum, is the low risk patient, and these patients are rare in the hospital. These patients basically have, they're fully ambulatory or they're only expected to be in the hospital for less than 24-48 hours. Examples would be the ambulatory patient with no other risk factors, the same day surgery patients, and those patients you don't really need anything. Most the time if you find these patients you should discharge them; they don't need to be in the hospital, at least not for very long. Usually in our patient population, this is far fewer than 10% of the in-patient population and usually less than 5%. In the middle, in this medium risk group is almost everybody else. Medical surgical patients in general fit in this middle group. They have some risk factors the best studied have been CHF, CHOPD, Pneumonia, in the medical side, and basically in this group we don't care if you use unfractionated Heparin or low medical weight Heparin, but you should use something, you should use one of those, you should use pharmacologic prophylaxis.

We basically relegated unfractionated 5,000 Q12 to a secondary role because in some populations the evidence is not as good for those, for that dose is as is for unfractionated 5,000 Q8. And because all the trials that compared low molecular weight Heparin to unfractionated Heparin, at unfractionated 5,000 Q8 in the comparator arm. So basically you've got high on one end of the extreme, you've got low on the other end of the extreme, and you've got everybody else.

So see what this looks like in an order set, it looks something more like this. This is a paper example and we've also done this in many places including UCSD that have computerized physician order entry, but the concept is basically the same. You're stating the risk in text, and then linked directly to that risk, are the choices that are most appropriate, for that level of risk. You can look at this, and complete it in a matter of seconds and you and I will agree almost always about whether or not the patient is at low, medium, or high risk. Note that there is an opt out, if there is a risk for pharmacologic prophylaxis in which case we're trying to get them to use mechanical prophylaxis. So you can see the contrast between this kind of an order set and this one.

If you're going to do this on paper ideally you would try to make this order set portable enough that you can actually insert it as a module and integrate it into order sets that are already in wide spread use. Instead of having a stand-alone order set. If you do use a stand-alone order set technique then I suggest that you clip it to all the admission and transfer orders and the history and physical forms, for example, you know you have to make readily available to the point where it's everywhere and it's hard to get around it.

The most common mistakes, I think, in constructing VTE prevention orders are using the point based risk assessment model, getting too complicated in your guidance or else providing no guidance, failing to revise old order sets, and you know, I really think you have to do some homework and look and see what's out there, trying to get to find what the definitions of risk can be a problem. If you get to six levels of risk, then you're getting way too busy. Allowing non pharmacologic prophylaxis is something that I see commonly in order sets the first time they're handed to me and that's clearly not generally the way to go so they should be relegated to a secondary role or an adjunct role. And finally I've seen some people, some centers, who have the method to do the risk assessment in one area and then several steps later and sometimes several minutes later they'll expect people to remember what that risk level was and choose the prophylaxis based on that level of risk. In other words, the integration or the link between the risk level and the prophylaxis choices is separated in time or space. And that just doesn't work. You need to have these linked as closely as possible.

Using the same framework of hierarchy of reliability, we're going to talk about getting to Level 5 now which is really the nirvana; it's the highest level of reliability where you're going to have really high levels of reliability reliably delivering DVT prophylaxis where it's needed. The concept is known as measure-vention and that's putting together two words, measurement and intervention. It's really daily regular measurement of patients who are on DVT prophylaxis and then acting or intervening on those who are not adherent to the protocol. So, it's suitable for on-going assessment, and reporting to a governing body. You can take that kind of data and report it to a ward by ward basis and at the same time you can roll it up and use it as a report to your governing body. So it's archive-able data and yet it can be used for real time intervention on the front line. Now this can be done on paper very easily, but in this electronic age we encourage you to go after electronic means to do this, to do the measurement piece.

So the map to reach this level 5, 95% prophylaxis is you know, you can start the measurement any time, but the intervention piece should probably wait until the protocol is in place and the way to start the measurement is that you use the medication administration record or automated reports to classify all patients on the unit, and being in 1 of 3 zones; green, yellow, or red. Fairly self explanatory, if you're on the green zone you're on anticoagulation on either a therapeutic or prophylactic dose, if you're in the yellow zone you're on mechanical prophylaxis only, and if you're in the red zone you're on nothing.

Ideally you would empower the nurses to move people out of the red very quickly. You're trying to raise situational awareness. I've got a patient in my hospital bed who is at risk and is not on prophylaxis, you're trying to raise that awareness so that you can do something about it. You empower the nurses to place them on sequential devices at least as a standing order and to contact the doctor if there is no anticoagulant in place and no obvious contraindication. So, it says, most places have used nurses, some places have used pharmacists, so it is important of course that since you're adding something to their plate that you get their buy-in that this is an important thing to do. They can become the most effective champions of the whole process when they realize that it's making a big difference.

The interventions simple, it's generally a templated note that says “Dear Doc, we note that this patient seems to be at risk, they're not on prophylaxis, we think that they should be, please order DVT Prophylaxis, or pharmacologic prophylaxis, if not, say why here. And many places either do a text page or some kind of contact to reinforce the templated note. You really need administration to back this up, you don't want doctors yelling at nurses or pharmacists for doing their job or taking action in the best interest of the patient.

This is what it looks like, what you can expect to see. This is Jason Stein's work. This is 3 different wards, retros from 3 different wards and 2 different hospitals, and what you see is in all 3 wards they had a protocol in place already, they had an order set that was well done, and they already raised levels from the 50% range of prophylaxis to the 70% range or so. The arrows show where they started the measure-vention technique, where they explicitly for every patient in every bed, classified them as red, yellow, green, and took action on the yellow and red ones. And they got them to prophylaxis virtually within a couple days, they're in the high 90's, and that effect is sustained.

So, if you're going to measure DVT prophylaxis there are many ways to do it, there are complicated ways and there are easy ways. The most generalize-able and easy way albeit, perhaps not as granular and detailed a way is to just use the simple red, yellow green method, and you can enhance this with ruling out or including people who are in the grey, who have an INR of 2 or something, but basically we're recommending you do this simple method because we're sure you can do that. The most common mistakes of measurement are not doing it at all, the second most common mistake is to try to do measurement across time and report it once a month, in which case it's not really helping the patient. You want to do your measurement at the same time on a nearly daily basis, so that you can intervene on patients who are not on the right prophylaxis. So in other words, you want to make any poor performance or non adherence an actionable event.

A few things in summary then, key points and recommendations. Quality improvement building blocks need to be in place. The basics: a team, a physician leader, institutional support, a protocol. Multi-faceted approached are needed, we know some things from the literature that help boost VTE prophylaxis rates. Standardized order sets, E-alerts or human alerts, and so on. There are good and bad ways to implement those order sets and to design those order sets, so I'm asking you to learn from the lessons of all these collaboratives that have taken place and try to use something that is simple, and gives simple text cues about what's appropriate to use and what's not appropriate to use. 3 levels of risk really should do it. If you're just dealing with medical patients only, even just two levels of risk should do it.

Ideally you would have an institution- wide protocol that you can use across all or virtually all med-surg populations. You may run into situations like with the psychiatric population, OB-GYN population, or sometimes ortho, where you went to carve them out from the rest of the medical center, and they might have special niche orders, but by in large we are telling you that it is feasible, and probably desirable to have as uniform of a DVT prevention protocol across the medical center as is feasible for you to pull off. It's okay if you don't follow exactly same choices that we use at UCSD as long as you follow the big picture guidelines and try to make sure you're aligned with the ACCP8 guidelines. The details in some of those grey areas are not that important. The most important thing is that you are getting most people on pharmacologic prophylaxis when it's feasible to do so, with those few exceptions of those who are at very low risk, and with those few exceptions of those who are at such high risk that they should perhaps be on low molecular weight Heparin plus mechanical prophylaxis. So this concept called measure-vention I think is a very powerful one, we're finding that it's useful in not only DVT prevention, but also in many other sorts of quality improvement projects and you can tuck that away and think about it and think how you might use it for other projects. But for sure we found that it's useful in DVT prevention.

I've provided you with some of the references that I've been involved in, certainly the ACCP8 guidelines, and many others, references are available on the SHM Website, and in the bibliography of some of these other articles. I wish you the best of luck; I will be involved to some degree with the rest of this project and look forward to working with some of you in the future and looking at some examples of order sets and giving feedback about some sample order sets and how that might go. With that I'm going to close and thank you very much for your attention tonight.