Implementation of the Hospital to Home (H2H) Heart Failure Initiative
October 30, 2012
Moderator:We are at the top of the hour so I would like to introduce our two speakers: we have Dr. Paul Heidenreichand Dr. Anju Sahayjoining us both from VA Palo Alto Health Care System. So at this time I’d like to turn it over to you Paul, are you ready?
Dr. Paul Heidenreich:Yes.
Moderator:Great. You’ll see a pop-up, go ahead and click show my screen and you should be set.
Dr. Paul Heidenreich:Okay.
Moderator:Great.
Dr. Paul Heidenreich:Well I want to thank everyone for joining. On behalf of myself and Anju Sahay, and CHF QUERI, we are pleased to discuss with you our efforts and observations in implementing the Hospital to Home system (H2H) in the VA Health Care System.
I’ll first mention that we have no disclosures related to this presentation.
I’ll begin by discussing a small amount of background, particularly for those on the call who may not be that familiar with heart failure. We’ll then discuss how the VA became interested in the Hospital to Home and how the Hospital to Home initiative came about. Anju Sahay will discuss then our heart failure network and how we decided to use this to help implement the Hospital to Home Network. We’ll discuss then a randomized trial we conducted of encouraging H2H enrollment among VA facilities, then we’ll discuss the results in terms of how successful that was and what they did. And then finally we’ll finish with our overall estimate of the impact of Hospital to Home on outcome in the VA Health Care System.
So heart failure, as many of you know, is a very common and costly condition. In the United States the prevalence is about 5.3 million. In the VA, again depending on how you define it, and we here defined it as at least two visits for heart failure within a year, that number is about 140,000 or about 2.6% of the overall population. It has a very high mortality, one million discharges per year and over $30 billion in costs per year.
In fact the costs, direct and indirect combined, are expected to increase substantially over the next twenty years, just due to the increase in the age of the population. So even if we didn’t have anything else happen to our society, we would see this marked increase just because of the older population in the United States. So it is a wake-up call, I think, that we do need to come up with ways of improving the cost of care.
Many of these costs are in the inpatient setting. And this slide shows the distribution of those costs as identified, approximately four or five years ago. And I think it’s still true today that over half of the costs are for hospitalization. Now when one looks at hospitalization, people identify that many are re-admitted. And this is data from HarlanKrumholz for Medicare population showing a wide distribution of readmission rates that were again, not only high but also spread out, with roughly about 24-25% of the population being re-admitted within thirty days after a hospitalization. And the fact that we do see this distribution suggests that some who are up near the 30% range may be able to learn something from those that are down near the 20% range. At least that was the hope.
When we look in the VA health care system, we’re seeing something similar. We see that again this shows the distribution in a slightly different way, each column is a different VA facility with at least 100 heart failure discharges over two years. Now we’re looking at their all cause 30 day readmission and we see that it varies from about 15% on the low end to over 30% on the high end with the mean in there maybe a little bit below 20 some percent, though again, a similar rate for all cause admissions following a heart failure hospitalization. And when one looks at heart failure-specific hospitalization, again did they come back specifically for heart failure? We see an even greater variation potentially from four percent up almost a four-fold or more increase from the lowest to the highest facilities. Again suggesting that the care is different across our VA facilities and it’s possible that some care best practices could be transferred to other facilities leading to better care.
So because of these observations, the high readmission rate, the wide variation in readmission rate, the Institute for Health Care Improvement along with the AmericanCollege of Cardiology developed the Hospital to Home Excellence in Transition Initiative. And again, it was based on the concern of this revolving door that many patients are leaving unprepared for this transition, the system’s not doing very much for them, and then they end back up with a preventable hospitalization.
So to counteract this, to address this, the H2H developed three core concepts: Post discharge medication management, early follow-up, and symptom management. The post discharge medication management is the idea that patients need to have access and need to know how to use their medications appropriately. They need to have early follow-up in some form within a week for the average patient. And that symptom management – for sympton management, the patient must be able to recognize the symptoms of deterioration and know what to do when those occur, know who to call, how to get a hold of someone if that were to occur.
So they focused on these three areas and primarily because I think a lot of it is non-randomized data, but there is some data to suggest that these things will be effective. Here are some again, non-randomized data on early follow-up in readmissions published two years ago where they looked at Medicare patients and first saw very large variation in follow-up within seven days in some form, within the Medicare population. So again, a range from under 20% to 60% and then when one broke those down into quartiles, there was a small but statistically significant increase in readmission for that lowest quartile group where patients were being seen less than 32% of the time compared to the other facilities.
So again, it wasn’t a … we were not looking at a huge impact here, but at least there is some data to suggest that the seven day follow-up that is recommended by H2H will be effective.
Their goal was to reduce thirty day all cause risk standardized readmission rates for patients discharged with cardiac condition by 20% by December 2012, so two months from now. So this was launched at the end of October of 2010 and so the goal was, within two years to have this reduced, the readmission rate by 20%.
Now H2H asked for strategic partners and many people signed on and the VA Central Office agreed to this, thought it was a good idea and wanted to be a strategic partner. But beyond that, there were no requirements, no mandates that any facility do anything regarding this. In fact, sort of the extent of it at the time was that H2H was mentioned on calls to VISN and facility leadership but it was up to them at that point to go forward and initiate projects. And I’ll let you know what enrollment means is representative of the facility, go the H2H website, and basically confirm their dedication to those three areas, to agree to work on at least one of those areas if they are not doing it already, to start new projects to improve care in one of those three areas. And then to periodically respond to surveys that the H2H group would send to them.
Now our group at CHF QUERI recognized that there was an opportunity here to do more than just go the top down method that I think is common within the VA, and that we could use our existing VA Heart Failure Network to combine that top down with more of a bottom up approach to help implement H2H within the VA health care system. So at this point, I’m going to turn it over to Anju Sahay who is then going to discuss the VA Heart Failure Network and some of our initial findings with the network. So Anju, please go ahead.
Dr. Anju Sahay:Thanks Paul. So I believe everybody can see my slides. Correct. Okay. So I’ll be talking about the heart failure network which has been formed by the CHF QUERI and we have facilitated the implementation of the VA H2H Initiative to the Heart Failure Network at all the facilities.
So in July 2006 we formed this Heart Failure Network, or HF Network, which is a social, informal network of VA providers who are interested in improving the quality of care for heart failure patients. We currently have over 800 providers – specifically 862 providers at 150 facilities. And at each facility, our number of members or providers range from one to two providers at each facility. The purpose… we use this Heart Failure Network in a variety of ways and in our presentation, in this context we have been using this Heart Failure Network to facilitate the implementation of the VA H2H Initiative.
This network operates through web-based teleconferences or live meetings. We have them every two months, which the format typically is we share information, any updates, and then there are two presentations typically one focusing on a quality improvement initiative and the other one on research findings on this project. And these presentations are made by members of the Heart Failure Network at different facilities. So we also have emails and surveys which we do through the Heart Failure Network.
Now we view the Heart Failure Network as a community of practice. These communities of practice are a type of informal learning organization or organizations – they can be more than one community to practice within a formal organization. According to Wenger, communities of practice consist of groups of people who share a concern, a set of problems, or a passion about a topic, and who deepen their knowledge and expertise in this area by interacting on an ongoing basis. These communities of practice are gaining popularity in the health sector and there is growing evidence of these communities of practice. They are developing and functioning within social networks, such an example is our Heart Failure Network which has this community of practice focusing specifically on improving the quality of heart failure care through the involvement of providers or community practice members. And through this involvement, the members are learning from each other through explicit and tactic knowledge. They are interacting and the members are actively involved in heart failure practice or quality improvement projects.
This slide shows the composition of the current heart failure network members. The members are multi-level within the VA system and multi-disciplinary. We have physicians, chiefs of cardiology, chiefs of medicine, nurses and nurse practitioners, pharmacists. We have facility level members and VISN level members who include administrators.
We have looked at the participation by the members at the facility level and as you can see, about 20% of the facilities, though we have members there, the members are not participating. About 40% of facilities do have members, but they are participating at a low level, and the remaining 40% have members are participating at the high level.
Let me specify what really we mean by participation. We define participation as being actively involved in at least one activity of the Heart Failure Network. So by just being a member, that means they are receiving all communications and information, but they may not be actively participating in the activity. And again, participation is, what we are talking here is active participation. That means they have attended at least one session or responded to an email or participated in other activity. We do know some of the members are participating passively because we know they go to our website and access resources that we’ve learned through formative evaluation. But in this context we are talking about active participation by our members.
We’ve also looked at the academic characteristics of the facilities, VA facilities in the Heart Failure Network, and participation in this network by the members. Specifically we’ve, here we’re looking at the ACGME which starts for the Accreditation Council for Graduate Medical Education. This body is responsible for accrediting the majority of graduate medical training programs for physicians in the U.S. The other dimension is COTH which stands for Council for Teaching Hospitals and the board certifications from over 75%. And as the slide shows, consistently, facilities which have higher percentage of ACGME, COTH, and board certification are participating more actively in the Heart Failure Network. One reason could be the smaller facilities have fewer providers and they have less time to participate actively in the Heart Failure Network.
Similarly, we’ve also looked at the volume, the bed size, and the heart failure discharges at these facilities and linked that to participation in the Heart Failure Network. Like before, our facilities which are larger and have more discharges, the members, or providers, at these facilities participate much more actively compared to the smaller facilities.
This slide looks at the mission critical measures for heart failure inpatients by CMS and Joint Commission. And these four measures are LVEF documentation - that’s left ventricular ejection fraction, documentation offered. And ACEi if LVEF is less than 40. Discharge education and smoking counseling while they’re still in the hospital and inpatients. We see across board that all across and all types of facilities that they are doing very well. But if you wanted to look very closely then we see definitely that facilities that are participating more have, are doing better but again, across board, all facilities are doing very well in that.
I’ll hand it back to Dr. Heidenreich to continue the presentation.
Dr. Paul Heidenreich:Okay, thank you Anju, so our thought was that given this Heart Failure Network which included again many front line clinician providers and interested in heart failure and quality managers at their facilities, that we would want them involved as they would most likely be those who would be implementing any particular project related to H2H.
So the purpose of our first randomized trial was to see if we could use the Heart Failure Network to help implement the Hospital to Home initiative. We wanted to see if we can encourage facilities to enroll using the Heart Failure Network and then we wanted to see what types of programs they implemented and give an estimate of the resources used to do this.
So we identified 124 VA facilities with at least 100 heart failure discharges over the last two years. Again this was in 2010 and we had launched this in … in the start of 2011. Again a few months after the main launch of the H2H initiative. And hospitals are randomized 1:1 to usual care for Heart Failure Network activation. Our usual care of Heart Failure Network activation and we at first paired the facilities by number of beds to make sure we had balanced on size of the hospital.
Now the usual care group received what I had mentioned previously in that H2H was presented byCHF QUERI on two national calls, one was the Chief of Staff Director call, and one the VISN CMO Chief Medical Officer call. And where it was mentioned that the VA was a strategic partner and facilities were encouraged to enroll in this.
Now the intervention again, we first emailed Heart Failure Network members in the randomized facilities, we set up web based teleconferences announcing H2H, we did follow-up surveys asking what were planned and then several months into it, we asked individual facilities to present their progress on the web based teleconferences so that facilities could then learn what other facilities had done. So we sort of had the early adopters give a discussion.
And the second survey follow up was asking about the status of the projects. And for this then we asked both intervention and control sites if they had started any projects related to H2H. Now the main outcome would be hospital enrollment on the ACC/IHI website. So that was out of our control, we do have affiliations with that group so they would periodically let us know all the different VA facilities around the country that had enrolled.
Secondary outcomes were projects that were initiated in response to H2H and this is through our surveys of Heart Failure Network members at each site including both initiated plan and planned projects related to the H2H goals.
So looking at some baseline characteristics, this slide just shows the outcome in 2008, sort of leading up to it. And we see that the facilities are reasonably balanced, the mortality, the 30 day mortality, and 30 day readmission for heart failure, not all cause readmission, was very similar between the two groups as we would have expected.
Our results – we’re going to jump to the main results here and that is, we were at six months successful for the primary outcome is at least getting facilities to enroll. So we had pretty much half of the facilities were in the intervention group enrolled compared to about ten percent in the control group. So if we look at that over time, and that’s shown on this graph, we see enrollment in blue for those randomized to network activation and in red for the control group – that there’s early and consistent increase in enrollment up to about four months and then it plateaued. But it was a clear demonstration that at least for this type of intervention, we could get facilities agreeing to be involved by activating the Heart Failure Network members, members of the disease-related network.