May 14, 2018
Dr Carolyn Lam: Welcome to Circulation on the Run, your weekly podcast summary and backstage pass to the journal and its editors. I'm Dr. Carolyn Lam, associate editor from the National Heart Centre and Duke National University of Singapore.
Our featured discussion today is really a very important message, that hospitals have the capacity to influence a patient's adherence to secondary prevention and thereby potentially impacting long-term patient outcomes. Much more on this important paper coming right up.
Higher physical activity is known to be associated with lower heart failure risk. However, what is the impact of changes in physical activity on heart failure risk? The first paper in this week's journal, by first author Dr. Roberta Florido, corresponding author Dr. Ndumele from Johns Hopkins Hospital, provides us some answers. These authors evaluated more than 11,350 participants of the Atherosclerosis Risk in Communities, or ARIC, study who were followed for a median of 19 years during which there were 1,750 heart failure events.
They found that, while maintaining recommended activity levels was associated with the lowest heart failure risk, initiating and increasing physical activity even in late middle age were also linked to lower heart failure risk. Augmenting physical activity may, therefore, be an important component of strategies to prevent heart failure.
The next paper highlights the importance of bystander automated external defibrillator use. First author Dr. Pollack, corresponding author Dr. Weisfeldt from Johns Hopkins University School of Medicine sought to determine the association of bystander automated external defibrillator use with survival and functional outcomes in shockable observed public out-of-hospital cardiac arrests.
From 2011 to 2015, the Resuscitation Consortium prospectively collected detailed information on all cardiac arrests at 9 regional centers. The exposures were shock administration by a bystander applied automated external defibrillator in comparison with initial defibrillation by emergency medical services. The primary outcome measure was discharged with near or normal functional status as defined by a modified ranking score of two or less.
The authors found that among 49,555 out-of-hospital cardiac arrests, 8% were observed public out-of-hospital cardiac arrests, of which 61% were shockable. Overall bystanders shocked a remarkable 19% of shockable observed public out-of-hospital cardiac arrests. Bystander automated external defibrillation in shockable observed public out-of-hospital arrest was associated with an increased odds of survival with full or nearly full functional recovery compared to emergency medical services defibrillation.
The benefit of bystander automated external defibrillation use increased as the arrival of emergency medical service was delayed. Thus, efforts to increase the availability and use of automated external defibrillators in public locations are likely the most promising immediate ways to improve survival from out-of-hospital cardiac arrests.
The next paper suggests that the complement pathway may contain the secret to a successful cardiac regeneration. First author Dr. Natarajan, corresponding author Dr. Lee from Harvard University, and their colleagues performed a cross-species transcriptomic screen in 3 model organisms for cardiac regeneration, the axolotl, neonatal mice, and zebrafish, all of which underwent apical resection.
RNA-seq analysis showed that genes associated with inflammatory processes were found to be upregulated in a conserved manner. Complement receptors were found to be highly upregulated in all 3 species, particularly the induction of gene expression for complement 5a receptor 1. Inhibition of this particular complement receptor attenuated the cardiomyocyte proliferative response to heart injury in all 3 species.
Furthermore, following left ventricular apical resection, the cardiomyocyte proliferative response was abolished in mice with genetic deletion of complement 5a receptor 1. These data, therefore, identified the complement pathway activation as a common pathway for a successful cardiac regeneration.
The final study sheds light on the association between hyperoxia exposure after resuscitation from cardiac arrest and clinical outcomes. First author Dr. Roberts, corresponding author Dr. Trzeciak from Cooper University Hospital performed a prospective multicenter protocol directed cohort study that included 280 adult postcardiac arrest patients.
They found that early hyperoxia exposure, defined as a partial pressure of oxygen of above 300 millimeters mercury during the first 6 hours after return of spontaneous circulation, was an independent predictor of poor neurologic function at hospital discharge even after adjusting for a potential baseline and postcardiac arrest confounders.
That brings us to the end of our summaries. Now, for our featured discussion.
Medication nonadherence is a common problem worldwide and, indeed, the very topic of our featured discussion today. Our featured paper is so interesting because it tells us that hospitals may have the capacity to influence a patient's adherence to secondary preventive cardiac medications, thereby, potentially impacting long-term patient outcomes, and there are a lot of implications of that.
I'm so pleased to have with us the first and corresponding author, Dr. Robin Mathews, from Duke Clinical Research Institute, as well as the editorialist for this paper, Dr. Jeptha Curtis from Yale University School of Medicine, and our associate editor, Dr. Sandeep Das from UT Southwestern. Lots to talk about.
Robin, could you perhaps start by telling us what made you look at this issue of nonadherence and what did you find?
Dr Robin Mathews: The issue of medication adherence has been something that I think we've been dealing with in healthcare for some time now and, traditionally, we looked at factors that, on a patient level, you sort of also have an idea that maybe they might provider level factors that contribute to nonadherence, so we started thinking about this, what's the health system's role in adherence and is there a role? Do hospital and do providers have more of a role in promoting adherence than we acknowledged in the past?
We are fortunate to have a lot of great clinical data sources available, and the one that we used for this study is the ACTION-Get With The Guidelines Registry, and this is a quality improvement registry that's been around for some time. It's a great source of research and observational studies that has produced a lot of data over the years.
ACTION is a voluntary registry; there are several hundred hospitals that participate, and it gives us very good data, detailed data on the patient experience in the hospital for patients who come in with acute coronary syndrome, so we looked at patients who were enrolled in ACTION over the course of 3 years, from 2007 to about 2010, and looked at the typical patient level factors, medications that were given on admission, how they were treated and what medications they went home on.
What ACTION doesn't give us is longitudinal data, which is really what we were trying to get at here, so we were able to link this clinical data set using CMS data, which is administrative data, claims data, in order to ascertain longitudinal adherence, so we ended up, after exclusions of about 19,500 patients or so, and this spanned about 347 hospitals, of patients that we followed up to 2 years out, and our objectives of the study were 2-fold, one to assess adherence at 90 days for cardio vascular medication, secondary prevention medications that are typically used, so, in this case, we looked at beta blockers, ACE inhibitors, ARB, phenoperidine, and statins.
We looked at 90-day adherence, and then the question we had specifically was does adherence vary among hospitals? The second thing we wanted to knowledge was, if adherence does vary among hospitals, is there a relationship between hospital adherence and cardiovascular outcomes at 2 years, so we looked at MACE, which is MI, revascularization, readmission, stroke. We also looked at death and all-cause readmission, and also mortality.
What we found is that the adherence actually did markedly vary within the medication classes, but also among hospitals, and once we divided these groups into essentially high adherence hospitals, low adherence hospitals, and moderate adherence hospitals, there were these typical differences in terms of patient characteristics that one would expect in terms of comorbidity, socioeconomic status. Patients who were in the high adherence hospitals were more likely to be from ... to have a less comorbidity burden. They had higher income based on ZIP code, and they were more often represented from non-southern hospitals in the United States.
When we then correlated these two outcomes, what we found is pretty interesting. Patients who were in the low adherence hospitals were more likely to have the outcomes that I mentioned earlier. That's not too surprising, yeah, because I had mentioned that the patient mix in terms of the ... their case mix varied among these hospitals, so the logical question as well, maybe the hospitals that are ... have low adherence have low adherence because the patients are generally just sicker.
We know that there are certain high-risk groups and we know that the patients who are treated at some hospitals might be sicker than others, so we did our best to adjust to these, so we did a multivariable model. We adjusted for various patient differences, and we also looked at hospital-level differences, the best that we can ascertain based on the ACTION Registry. That was sort of where the interesting finding was the rates of major adverse cardiac events and death at readmission were mitigated somewhat closer to the null, but they remained significant.
Dr Carolyn Lam: What a detailed summary. Thanks so much.
Jeptha, I love your editorial that accompanied it. Could you put the study into context a bit for all of us? Why are these finding so impactful?
Dr Jeptha Curtis: It's rare that you get to review and editorialize a paper that has so many implications both from a clinical practice and policy standpoint, so I think they really hit on a understudied area, and really this paper should cause people to reflect on what's going on in their practice and at the institutions that they practice in.
I would say that adherence is just such a challenging problem that, as Robin articulated, has been refractory to change over 15 years. We've been studying this for a long time, and we know that the numbers had not improved over time.
What's different about this paper is that it really suggests a completely different approach to addressing nonadherence among patients, and if this is ... if their findings are true, if nonadherence is really actionable at the hospital level or attributable to the hospital level, it really opens up new avenues both for research as well as for quality measurements.
As I read this paper for the first time, I was really struck by thinking about how invisible adherence is to frontline clinicians. We just don't have the information to tell us are our patients taking their medications on a day-to-day basis, and we know that most of them are not because the research has consistently shown that a large proportion failed to take their medication, and Robin's paper showed that yet again, but I can't say that there's any steps that our hospitals are really doing to address that in a systematic fashion.
All of our efforts for quality improvement have really been towards making sure that patients are prescribed the medication on discharge, and in the setting of readmission and trying to prevent readmission to our hospitals, we are now having follow-up phone calls with patients to assess failures to taking medications and follow-up, but it's really ... That's it. There's really no systematic way that we're trying to ... if an individual patient or a group of patients are adherent to their medications, so this is really a whole new avenue.
What we don't know is how to improve it, right? I think that the first implication of this paper is that there are differences at the hospital level. Some hospitals seem to be doing this better than others. That could be driven by differences in case mix, but it could also be driven by differences in hospital practices, and I think this is a wonderful opportunity for future direction of research perhaps using positive deviance methodologies to go to those hospitals that have high adherence rates in really trying to understand what differentiates their practices from those of other hospitals.
Dr Carolyn Lam: Indeed, Sandeep, I remember some of the conversations we had as editors about this paper. We, too, were struck by the novelty, and you've mentioned before, Sandeep, that the novelty of perhaps nonadherence or adherence as a new performance measurement. Would you like to comment on that?
Dr Sandeep Das: Yeah, first thing, what was kind of interesting about the discussion surrounding this paper, there were some people who read it and just sort of read it as the message being nonadherence associated with worst outcomes, and I thought like that was pretty established, known, but then there were some people like Jeptha and Erica who really got it, who really understood what was novel and interesting about this, and I also congratulate Robin on a fantastic paper.
One thing I think that's really interesting, in my day job, I wear a couple of quality hats. I am the cardiology division quality officer, and health system quality officer for UT Southwestern, so I spend a lot of time thinking about quality, and I'll tell you there's quite a bit of metrics that he ... there's just a lot of things that now you feel they're not particularly substantive and they're very difficult to change, you have, you know, if aspirin on discharge, as Robin mentioned discharge adherence, aspirin on discharge is 99% and getting people to document the last 1% rather than fail to document it, there's not really a fulfilling challenge where you think, "I'm really impacting patient endpoints."
I was really struck by the opportunity here. We know that from studies like MI FREEE that adherence to medications even at a year is probably about a third of patients are not adherence, so it's really kind of interesting to take that as an opportunity. We should fixate on what are these therapeutic option or not therapeutic option can move the needle by a fraction of a percent, but these are medications that are proven to prevent MI and change lives, and there's a massive delta here that we can address. The concept that this is addressable on the hospital level is fascinating, and I'm a big fan of coming up with sort of systems level approaches to addressing problems.
Dr Carolyn Lam: Congratulations once again on this great paper. Just tell us what do you think of the next steps and what would your message be to those of us who practice outside of the US?
Dr Robin Mathews: Jeptha talked about where our focus should be in terms of what we can do on a hospital level. I think the ultimate answer is there's a lot of heterogeneity in terms of what is done, and I think that, expanding on his point about better investigating practices that currently exist, and whether that's surveying things, and we have a lot of great professional societies and registries that we can sort of reach out to these hospitals, find out what they're doing, what makes them different from the hospitals that are not doing those things and then really doing some rigorous testing to figure out if in fact these specific interventions that these hospitals have put in place are with the high likelihood leading to the effects that we've seen, so I think that surveying sort of what's out there, understanding what works in a rigorous way and then being able to systematically apply this or distribute this to other hospitals to share the knowledge and say, "Hey, this is what we think. We've actually done it."
Like Sandeep said, with the inpatient management of patients who come in with acute coronary syndrome, we've done it well. I think it sort of contributed. Our guidelines and adherence to these guidelines and the metrics that we've used have really demonstrated that we've sort of achieved high levels, but we sort of reached I think the ceiling for a lot of that, and you always have to be open to novel metrics and then the idea of focusing in on the transition from hospital to home and what we can do once they leave their door, once they leave the door of the hospital, I think would be useful.
In terms of the rest of the world, I mean, the US has very unique problems based on our payment models and access to care and whatnot, but I think a lot of the themes that we sort of have seen with medication nonadherence when it comes to patient-level factors and provider-level factors are sort of universal.
At the end of the day, patients need to be empowered, and they also need to have the tools to allow them to be successful in my opinion. I think we've for a long time in this space often said, "Well, this is sort of a patient that there's only so much that we can do as providers," but I think that papers like this highlight the possibility that there's probably more that we can do to make these impacts.
Dr Sandeep Das: One of the comments or a question that I had was the controversial thing is to what extent hospitals should be accountable for things that happen well after discharge? I think readmission is one that always comes up. There's factors that are outside our control, so one question is kind of to what extent should we be responsible for stuff that happens forward of 6, 9 months down the road?
The second question that I had or a comment that I had was I do think that there's going to be a generalizability to non-US settings because there's elements of this ... For example, this now would incentivize hospitals and discharging physicians to make sure that patient education is substantive, right? If the metric is, "Did you provide discharge instructions, yes or no?" then that's sort of trivially accomplished by handing them a piece of paper and checking a box, but, now, if we follow a metric like this, we're really going to be accountable for making sure people understand what they're supposed to be taking and have a path to get it and things like that, so it makes some of the transitions of care stuff, and that's a great point, some of the transitions of care stuff much more substantive.
Dr Robin Mathews: Sandeep's point is a very good point, and it's very difficult to come up with a clear answer for that and, like you said, the issue with readmissions and all sort of the factors that are involved from a social level and research level cloud that, so ... and, hence, I think something like readmission is controversial, and I think this sort of question will generate a lot of further questions about whether using medication adherence and holding hospitals responsible.
I will say that when we looked at adherence sort of in the short term at 90 days and we looked at it in the long term at a year, we saw there was sort of a drop off, but it wasn't as substantial it was earlier, so I think a lot of adherence in the short term after hospital discharge continues to decline over time, but it doesn't drop down as precipitously downstream as it does early on, and I think that, just like with readmission, there's been some data to suggest that near term readmission are more likely things that "could be preventable" as opposed to maybe a readmission toward the end of the month.
At the end of the day, it's a very difficult thing and there's a lot more discussion that needs to be had about this topic, but I think that with this, it gives me some hopefulness and I think everybody else on this call that at least we wouldn't then be able to prevent every adverse outcome that happens 2 years down the road, but we might be able to at least affect a substantial portion of them.
Dr Carolyn Lam: Listeners, you heard it. There's lots that we can do. This paper says a lot. Please do pick it up. Read the editorial as well.
Thank you so much for listening today, and don't forget to tune in again next week.