 We'll be Dr. Michael Kelly. Michael is just starting off. He's only in his seventh year as a neurosurgery chief resident at the Cleveland Clinic. Michael was a fellow with us a year or two ago in the ethics program. Michael's research interests are in surgical decision making and outcomes research. He's a graduate from Loyola University, Chicago's Strict School of Medicine. Today, Mike will talk to us on the topic who's outcome, which measure, the ethics of outcome assessment in patients with neurologic injury. Join me in welcoming Mike Kelly. Well good afternoon everyone. My name is Mike Kelly. So I'd like to thank Dr. Sigler and the McLean Center for the opportunity to talk about this topic. Who's outcome, which measure, the ethics of outcome assessment in patients with neurological injury. I'll begin with a case discussion that presents an instance of neurological injury for acute ischemic stroke. Move to kind of a discussion of what outcome measures are used to assess benefit in patients with ischemic stroke and how those outcome measures interface with decision-making at the bedside and then talk about some future directions for research and patient care. So a 75 year old female presents to your hospital with a left side at acute stroke and as you can see here on the CT scan, there's an area, a darkened area up here of hypodensity, we call it, which is where the stroke area is on the left side. Now this is where speech is in most patients. This is also where motor control over the right side of the body would be. You can also see that it's causing a substantial amount of mass effect up here. So those are the fluid field spaces, the ventricles, they're being pushed over to the side. So this is a life threatening event. This patient is likely very close to herniating and any decision-making that needs to be done needs to be done fairly quickly as far as whether or not there's going to be a life-saving procedure done. This particular patient has multiple medical problems as many of these patients do who have large strokes, heart problems, lung problems, kidney problems. Patient presents intubated hemiplegic meaning they're not moving the right side of their body and has minimal responses. And in discussion with the family in rather hurried fashion you ask them what would your mother want to do in this situation. They say that they want to do everything for mom. They're not really sure but do everything that you can. And so the question of course is what to do. You have an older patient with lots of medical problems who is going to have severe disabilities no matter what you do and if she survives that is. And so your major options are really three. It's surgery, surgery primarily being a hemicraniac to me, which is a procedure where you make a large incision over the scalp and take off a large piece of bone and open up the covering over the brain to decompress the brain. And if the patient survives at a later point in time re-implanting the skull. Another option is medical therapy so non-surgical therapy. You can give medications to help reduce pressure in the brain far less effective than surgery but another option as well. And the third option is really withdrawal of care so this patient would not want to have anything done then you could all you could proceed with withdrawing care removing the breathing tube and going from there. Of course in many of these discussions that you're having with family though the question is what's the outcome. So if we do surgery what's going to happen? If we don't do surgery what's going to happen? And as a clinician you really are the point person for providing information in this scenario. And so we do have a significant body of literature, a significant evidence body that's been growing over the past 10 to 15 years that does help in guiding decision making. In the past decade or so there's been three randomized trials one meta-analysis which suggests for patients less than 60 years of age surgery hemicraniac to me reduces mortality from 78 to 29% and increases the percentage of patients with a favorable outcome from 24 to 75% at one year. And it's important to look at how they define favorable outcomes. So they're using in these studies the modified rank and scale which you can see here illustrated in this table. It's a scoring system that's based on clinician reports of functional status measures. So you're scored from a score of zero which is no symptoms at all the six which is death. They're defining a favorable outcome is a modified rank and less than or equal to four. Meaning moderate severe disability being unable to walk without assistance and unable to attend to one's own bodily needs without assistance. The Destiny 2 trial was published just this past year in the New England Journal of Medicine and it showed for patients over the age of 60 that surgery reduces mortality from 77 to 30% and increases the percentage of patients with a favorable outcome from 18 to 38% at six months. Again favorable being defined as modified rank and less than or equal to four. And this is what the distribution of outcomes looks like. This comes from the Hamlet study but there's many studies and they all kind of show a similar distribution of outcome. So if in non-surgical therapy, conservative treatment, the mortality rate is 71% and there's patients fall into the other sub types with the modified rank and score but much smaller percentages. For patients who had surgery done, the mortality rate drops to 22%. You really don't change the number of patients who end up with a modified rank and a five but the number of patients with modified rank and four, three and two all increase significantly. And so in these patients you reduce mortality and increase the number of modified rank and sub types. So the question that emerged from these studies and still kind of persist in the literature is what exactly is a favorable outcome? How are we defining this? Is it a modified rank and three perhaps or four? Most of these studies use a modified rank and a four. Modified rank and a three is moderate disability requires help but can walk without assistance. Modified rank and a four is unable to walk without assistance. Are we just trading death for disability in these patients? So you may be saving a lot of lives but you're saving lives and increasing the number of patients with modified rank and fours and threes. Is this an acceptable outcome? And really who should be making these decisions? Now, mind you, this is a clinician reported outcome measure. A clinician assesses function and the clinicians are the ones doing the research. They're assigning favorable outcomes. What do patients think about these outcomes? How do you elicit patient responses when most of these patients are aphasic? They can't talk or sometimes even minimally responsive. How do you know if this is really a benefit to patients or not? And of course, the question also is which outcome? So the modified ranking is the most well established outcome measure used for the longest time in the stroke literature, but there's certainly several other outcome measures that could be used. They basically are divided into clinician reported outcome measures, patient reported outcome measures. Clinician reported outcome measures focus largely on function. Things like the modified ranking scale, Barthol index, NH stroke scale, glass cow outcome scale, patient reported health status measures can cover several more domains, things like function or the stroke impact scale. Depression using the patient health questionnaire, the PHQ9 or quality of life using the EQ5D, the uroqual health index. And there's several more measures. Now the Destiny 2 trial, the trial that looked at the benefit of Henry Craneectomy in patients over the age of 60, they showed a benefit for patients achieving a modified rank and lesser equal to four. But they also looked at some of these other patient reported health status measures. You can see them here. The SF36, which is kind of a quality of life outcome score, the Hamilton depression rating score, and the EQ5D quality of life index. And interestingly, when they looked at patients who were able to actually complete these health instruments and follow up, they really showed no difference in benefit for these patient reported health health outcome measures when you compare the surgery to the non-surgery group. None of the groups showed a statistically significant difference. At the Cleveland Clinic, we've done work. We have a large outcomes database called the Knowledge Program, which basically follows patients in the post-operative setting and follow up clinics and has them complete health questionnaires, both patient reported health questionnaires and then the clinicians themselves report health questionnaires on the patients as they follow them up. We looked at our Henry Craneectomy patient population after surgery at three and nine months to see if there are any improvements in these scores for these patients. We found that for things like the NIH stroke scale, we were able to show a significant statistical improvement in the scores from three to nine months. But the modified rank and score in our patient population showed no significant benefit. The EQ5D, the quality of life health index, showed a significant improvement in score. But the PHQ9 depression index did not show a significant improvement. In the SIS-16, that stroke impact scale that assesses patient reported function also showed a significant benefit. So interesting discrepancies, at least in our patient population, you don't see the modified rank and benefit, but you do see a patient reported functional benefit. You see a patient reported quality of life benefit as well. So it seems to me that depending on the scale and outcome measure you're using, you may or may not see a value or a good outcome in these patients. We then looked at also functional status measures as to what patients reported if they noticed any improvement in their daily life activities and found that for mobility, self-care, pain and discomfort, anxiety, depression, really patients noticed no improvements in the post-operative course, but for usual activities, to find his family, leisure activities, social activities, things like this, they did notice an improvement. And this is consistent with a lot of published reports in the disability literature, which suggests that patients have a response shift. Maybe before surgery, before an event like this, they may have not tolerated a disability very well, but afterwards they learned to adapt to their new disability, maybe see parts of their life in different ways that they hadn't appreciated before. And so the conclusions that come from a lot of these studies is that hemicraniac benefit depends on the outcome measure used. What are you using as your primary objective for your study? What quality of life outcome measures are you using in addition? Patient-reported outcome measures, at least in our study sample and a few other reports, may show greater improvement, but you have to remember this is a highly selective subgroup. These are patients who are able to report on health status measures in the post-operative period. And oftentimes they have family support in reporting on these measures. So this isn't strictly just a patient telling you how they're thinking, how they're feeling, how they're doing. And it is important to recognize that at least in clinician-reported measures, that disability is often emphasized over ability. So if the modified rank and score is clearly a set of instances or scoring where you have this deficit, you have this deficit, you have this deficit, instead of you're able to do this or you're able to do this or able to do this. So there's clearly a framing effect, a bias that goes into even the outcome measures that are used. So again you're brought back to this patient bill. This patient presents their comatose, they need an urgent operation, and what do you do? And certainly listening to their preferences is helpful and helps to direct care. But you as a clinician also, knowing all this data and knowing the outcome measures, certainly have your own opinion as well, in addition to your own clinical experience. And this is where of course our decision-making paradigm comes into play, and certainly our most basic ethical principles. In instances where patients can't make decisions for themselves, it's been well worked out that you start with an advanced directive, see if they've declared their wishes in some way they may help guide your therapy. And that's not available, move to a substituted judgment standard, where we attempt to elicit preferences and then try to use them to make the decision. And if that's not available and there's really nobody has any idea what this patient would have wanted, move to a best interest standard. But of course in patients with neurologic injury you have several limitations. They all lack decisional capacity in the severe instances. Advanced directives are of limited utility, so according to the support study about 25 percent of the population has an advanced directive. Those of the advanced directives, those that specify stroke care is really unknown, but it's probably extremely low. And those that specify which outcome measure that's being used currently matters to them is probably zero. And so really you don't have the specificity to really use advanced directives maybe as effectively as you'd like. Decisions are time sensitive, surrogates are often overwhelmed in these situations, and duties of beneficence are unclear. Not just to the surrogates, but to the doctors making the decisions themselves. Which outcome measure matters to you as a clinician? Are you okay with a modifier ranking of 4 or 3? Now it's certainly you don't want to impose your decision making on the patient, but you've seen these patients before, you certainly have an opinion. Are you valuing the EQ 5D quality of life metrics you read about, or maybe the depression scales? And how are you coming in these decisions? Substitute of judgment itself has fallen under some critique over the past decade or so. Shallowits in 2006 reported basically he did a meta analysis showing that in studies where patients, surrogates are put in a room and patients are put in a room, and they're asked to make decisions about what their healthcare would be. That the accuracy rates are poor, but they're particularly poor for patients who have stroke. An accuracy rate of about 58%, the lowest of all the disease types. And they're particularly low for intervention type, like surgery, where it's 62%. So patients, surrogates are oftentimes very inaccurate in even figuring out what preferences would be in a situation of a stroke requiring a potential surgery. Dr. Stilmase in 2007 published an article that showed that, and he looked at 147 terminally ill patients, and to try to figure out which model of decision making they would prefer if they were to become incapacitated in their decision making. And the results were somewhat helpful at least for those of us who deal with neurological injury because they overwhelmingly said they preferred a model of best interests and substituted judgments. That you try to elicit their preferences as best you can, but as a clinician, try to equally weigh in on your opinion about what the best course of treatment would be for these patients. And then that kind of brings us back to this classic paternalism versus autonomy debate. And we see this play out, I think, in a lot of the neurological outcome metrics that are used. The movement historically traditionally has been to focus on clinician-reported outcome measures. But there's also this emerging patient reported outcome measure paradigm that's important, I think, but also very limited in patients who have neurologic injury. And perhaps looking at them as either clinician or patient reported outcome measures may not be as helpful in figuring out what to do at the bedside. It may be that there's a mixed dynamic that's going on where clinicians have an opinion, and patients, of course, have their preferences and these things need to be brought together to best figure out what outcome measure we should be looking at in this population. And so the question then is if we have these two competing outcome measures, you have a clinician-reported, patient reported, how do you best bring them together? And I think it's important to acknowledge, first of all, where we're at. So outcomes depend on the perspective and the type of instrument used. It hasn't been acknowledged in these randomized trials, really, how much of a role did the modified rank in place in that it's a clinician reported outcome measure in that it emphasizes deficits and disability over abilities. And that certainly influences clinicians' perspective on the surgery to begin with. Isolated patient and clinician-reported outcome measures may not be helpful because they don't probably reflect actually how the decisions are being made, even how the follow-up is being done. Patients and clinicians and oftentimes their family members work together on making these decisions. Clinicians have opinions, patients have opinions. We need to better measure that to see where this is leading us. And shared outcome measures between the doctor and patient, kind of a hybrid measure, are needed. Not just process outcome measures, and these are growing today, there are shared decision-making measures that say, how well are we using decision-lays? How well are we emphasizing patient preferences and care? But this is slightly different. This is asking us, when you follow up in clinic, what does the doctor think about your outcome? What does the patient and their family think about their outcome? And what does that say about the surgery that we're doing and the care we're giving in the post-operative setting? These outcome measures have not been described really at all in the neurologic literature, perhaps maybe in other domains, but certainly not neurologic literature. And I think a shifting focus towards ability rather than disability. And not just singularly defining ability over disability, but giving both sides of the coin. We give a more complete picture of maybe what we're dealing with when we decide to operate or not operate on patients who have a large ischemic stroke. I'd like to thank some of my colleagues at the Cleveland Clinic for all the help and effort they've given, as well as Dr. Salmeza here at the McLean Center and open the floor to any questions. Thank you. Questions for Dr. Kelly. I noticed in the destiny two study that two-thirds of the patients or their surrogates elected to have surgery. And one-third did not. That was that based, and now those negotiations sort of taken place between the doctors and the surrogates presumably. You mentioned that that decision was time-sensitive. Can you just comment on how much time sensitivity and why you think two-thirds elected surgery? So there's a well-described literature which looks at variability in surgical decision-making. It's better probably described for interstitial hemorrhage, the kind of stroke where there's bleeding instead of just ischemia. But there's wide variability across institutions and even countries and how aggressive they offer surgical intervention. So certainly the participating trials that probably agreed to this study to begin with were probably the more aggressive institutions. And so the two-thirds agreement rates to have the surgery done reflects more aggressive care institutions probably than all institutions in its entirety. And oftentimes it's clinician-dependent as well. So some clinicians are very negative on this. Some clinicians will emphasize dominant low, dominant hemisphere, like the left side, if there's going to be speech affected, they may be less willing to proceed with an operation versus non-dominant side where speech might not be affected. So there's a lot of institutional variability. Thank you. Please. Hi. I have a question from a slightly different perspective. I'm a staff chaplain advocate at Good Shepherd Hospital in Barrington. I study bioethics through the Nice Wenger Institute at Loyola. So I'm there in the emergency room as the chaplain. The woman comes in. She's had a massive stroke. There is an advanced directive. Just last week she and her husband discussed we don't want excessive interventions. The EDDoc does a beautiful job of taking the husband, showing him the monitor, showing him the extent of the stroke, and says, your wife is not going to survive this. You're going to have to prepare yourself. Husband starts to cry. Turns to me, I make sure their sacramental needs are met. The priest comes. I'm comforting them. All of the children are in agreement. This is what mom wanted. The neurosurgeon comes in. Well, we have a surgery we can offer. It's probably not going to do a lot of good, but it's the only chance we have. So we have to go ahead with it. Family under duress. You know, they're frantic. They agree to the neurosurgery. And then they go on a five-day roller coaster ride that involves staying in the ICU and a lot of ups and downs. And finally, the outcome, of course, is the same. The poor lady dies. So how do we better address these situations so that people get consistent messages that allow them to really make the decisions in accordance with these advanced directives and so forth? I think that's a great question. And I think part of the problem we have, at least in neurosurgery and perhaps this is true in other surgical specialties, is protocolized decision-making. Not that it tells you what to do, but just that there are certain factors you need to look for in making your decision. Maybe you bring thresholds to match that you need to say, we've looked at these parameters, and this is why I'm offering this operation. We do a poor job of that neurosurgery, and that probably accounts for a lot of the variability that you see from institution to institution or even surgeon to surgeon. And so it's very difficult, I think, for some patients. It's difficult even for residents, those of us who rotate at different hospitals and then rotate with different attendings, seeing the different decision-making processes and trying to come up with your own way of figuring this out. The criteria that people use shift sometimes, and that lack of consistency can be frustrating for families, I think even for trainees as well. One mic. Thanks so much.