Dr Maddux: Including patient reported outcomes or PROs in clinical trials opens new and meaningful pathways to understanding medical care from the patient's point of view. By combining clinical and patient reported endpoints, trials featuring PROs can identify opportunities to improve patient health status and defined treatment goals. Krister Cromm is Fresenius Medical Cares Expert Director of Patient Experience for Clinical Research. He is leading the company's evolution from disease-centered to person-centered integrated trial design. Welcome, Krister.
Krister Cromm: Thank you very much, Frank. It's a pleasure to be here.
Dr Maddux: Great. Well, we're glad to have you. I'd like to start today with trying to understand how you look at the shift from disease-centered to person-centered care in research trials and certainly related to qualitative research.
Krister Cromm: That's really a very, very exciting and important topic that has evolved over the last couple of years. Few years back, most trials would have been designed really around the disease and focused on maybe understanding how a certain medication or a product would lead to physiological changes in the patient. The patient was much more the object of attention than the subject participating. Over time, of course, patients and persons said, like we all are, grew more and more interested in what was going on. Through the gap, the opportunities of electronic participation, patients also demand to be part of trials. So, this has really led to a discussion among all those stakeholders in healthcare. How to really focus on the individual and this individual's needs. That first led to some evolution towards what we would call patient-centered clinical trials, meaning that patients would be heard, given a voice. So, maybe there would be panels that involve people, for example, who have kidney disease, and they would be able to say what they would like, but often enough, that really falls short, of what is needed, because, generally speaking, all of us can become sick. That means in all sorts of circumstances, we would like to be treated as individuals, and be heard. It fell short of the other stakeholders in the care process. The nurses, the doctors in the clinics, are their healthcare, professions related to care, who have meaningful information to design a clinical trial, and also to shift the perspective and shift the actual clinical research and also the product development.
Dr Maddux: We as nephrologists, and certainly, in kidney disease care, spend an enormous amount of time looking at quantitative data. We're interested in a lot of analytical elements of this. We measure performance in many cases on numeric items that come out of blood work or physical assessment of the patient. Talk to us a little bit about the distinction between qualitative and quantitative methods in research because... I think that they're really quite different, and many people don't recognize the uniqueness of the qualitative data.
Krister Cromm: It's really a good question of whether data is that different because if we look at classical lab parameters, we would often consider them as a definite source of truth. Often enough, these parameters are also only indicators that we use as proxies, for other information about physiological processes that are going on. With psychological data there is a difference in that we draw conclusions about something that we cannot immediately measure, right? So, of course, we could do a brain scan of a patient who's perceiving pain and we could understand something about the dynamics going on there. But ultimately, it's about the perception of the individual. All of that makes this data qualitative. At the same time if individuals report this consistently and the directions of the reported data is identical, it means that you can make this quantifiable, and you can analyze it on a on a larger scale. I think it's important to realize that in that respect, what we are doing using this qualitative information as proxies for other things going on, is as valid and as important as any other data that we collect in the clinic.
Dr Maddux: Are the techniques different? When you have free text and free conversation and comments from patients, you have to use novel techniques to try to dissect the meaning from that, when comparing it to other patients that are in a trial. How do you approach that?
Krister Cromm: What happens is that, let's say, you want to understand something about anxiety. Let's say in a patient and the treatment setting, you would probably come up with all sorts of ideas, what that would mean for a patient. You would maybe discuss this with patients in a panel, and then you would conduct qualitative interviews. So, these would be concept elicitation interviews, either in group settings are with individual patients, and these would be recorded, and then they would be coded. There would be specialists, psychologists writing down every word and also understanding: okay, how often does somebody mention a certain phrase, in what context and out of this information. Specialists would phrase questions and identify different aspects. If we were to take the example of pain, there are different types of pain, it could be about the length of the pain, the intensity of the pain, where the pain happens, and all of this could be written down. That would lead to a construction of a pain survey, and that could be disease specific. It could be related, let's say in hemodialysis to, for example, needling pain are something that is related to the actual care setting. This survey would be, once again, given to patients. There would be a cognitive debriefing process to see if this is what somebody was going through, if this was meaningful, if something is missing. In the next stage, this questionnaire would then be distributed to a larger group of people in that circumstance and be asked to fill in the survey. Then there would be qualitative statistics or quantitative statistical methods used to understand if individual items contribute to understanding that particular concept. In this way, it is possible to validate a questionnaire survey for a specific subject. Then that could be used in a clinical research setting.
Dr Maddux: How do you avoid bias in qualitative assessments when you've got cultural, language, environmental conditions that are so different from patient to patient based on their own background or their own circumstance? And that can be everything from their educational level to literally what their native tongue is. Just how do you approach those elements?
Krister Cromm: A lot of the research takes place in limited circumstances. So, of course, the more advanced healthcare systems will be able to do more of this research. One of the things that that will happen there, is you would look for a diversity of people from different backgrounds. The age structure, the underlying cause, maybe of a disease if it's related to a particular condition, to social backgrounds, also, if somebody's a native speaker of the language or not, and all of this would then form part of the development of the tool. But it's equally important then that if a questionnaire has been developed in one place, it doesn't mean that it is valid in another. So, what then need to be done is, , you will have developed a questionnaire for patients with cancer, it does not necessarily mean that the same questionnaire is going to work in patients who have kidney care. Another round of qualitative interviews, ideally, would take place. You would assess if this business appropriate or not, if individual items need to be added, individual questions need to be added, or some have to be removed. And another aspect, of course, is that these things can change over time. So, a questionnaire would have to be revisited. Nowadays, luckily, because of the advancements in, the way we even talk, we can be in different places and still have this conversation. It is possible to not have questionnaires that are static. So 20 or 30 years ago, you would design such a questionnaire, write it down, it would have two pages, three pages, and that would be it right? And this would be distributed to an individual, they'll fill it in, maybe a question is missing or not, somebody will get this back. We'll have to interpret it, this has fundamentally changed now. You will be able to hand out a questionnaire to a patient on their own electronic device. It is also possible to get to have this read out by computer voice or by another individual, which has voice recognition. It is possible to create a toolbox with all sorts of questions that are related to the subject that you want to find something out about. As time passes by, you can replace these items if they fall out of use. So, let's say maybe 20 or 30 years ago, people would play a specific sport, or they would use a television and or a telephone and that telephone would have, a dialing space. Nowadays, nobody uses that anymore. So, you can replace that question. Then go see if another question is more appropriate. That is also something we try to do, because we want to make sure that questionnaires evolve, they are consistent, and patients, or people simply filling in that questionnaire and entrusting us with information, don't feel that it's a burden to them. So, it would just be more similar to a natural conversation.
Dr Maddux: In clinical trials, the patient's condition can be very different, where they're starting from. I'm curious whether there's the opportunity to do things like discrete choice surveys or other kinds of tools to elicit the next question from how the first question is answered, for example, and how to try to elicit what a result is by asking some corollary question about their environment, their life, how they are doing living with the disease. I'm curious whether in clinical trials, you can get that deep into the dynamic nature of the kind of questions you're asking, or is it strictly you got 10 questions, we got to finish those 10. And that's all we can ask for the clinical trial. Are the trials becoming more dynamic in their ability to capture data?
Krister Cromm: They are. It's a very dynamic process, as you mentioned, and of course, what needs to be made sure is that the data is a very high quality and that there is no bias introduced. Related to that, but it is the case. The current evolutionary stage would be to have electronic questionnaires that are adaptable. Let's say you would talk about a specific subject in that questionnaire or the questionnaire would evolve that way. There is a database behind that, that would choose a question on the same scale. You would have a scale and first you may want to narrow down and say, if it's about let's say, physical functioning, maybe the questionnaire, if it has not been filled in by a particular person before, would ask with a very general question like, can you walk up the stairs? Or how many flights of stairs can you walk up? And as the questionnaire, so to say the system behind it, gets to know the person filling it in, it would become more specific. Like how many flights of stairs can you walk up? Or what is your usual exercise routine? Has that exercise routine changed? The next stage that is probably coming, and there is development in that direction, is that the conversations become more natural, as I just mentioned, so that it would really be about, "Oh, I haven't heard from you for two or three days. Can you let me know if you went for a walk yesterday morning", to understand about the actual physical activity level of a patient. But that is something that we definitely still work on. We need to make sure at the same time that the data and the feedback we get back is really valid then.
Dr Maddux: I think certainly integration of our sensing devices, our diagnostic devices, and other sort of classical ways that that we look at physiology, combined with asking patients intermittently about their condition, how they're tolerating living with their disease, all those things look like opportunities to get a better picture as opposed to a 10, 12 or 60 question questionnaire, intermittently in the daily life of the individual having them answer one or two quick questions, every two weeks or something may give us a different way of looking at this qualitative data.
Krister Cromm: Absolutely. And also, It will allow us to develop a different kind of perspective on treatment routines, right. So, you might identify when physiological problems take over, create a negative spiral that you can prevent. If you if you get data, and often physiological data, changes quicker than what happens in the perception of the patient, but it can also be the other way around. It's important to understand that there can be a feedback round basically. That is definitely something that will help not only change, clinical trials become really agile, but also help to inform about treatment choice later on, and the actual clinical care setting. In the past, these two domains have remained relatively separate from each other. That is also related to product development. Classically, product development was very much driven by the technological possibilities, so you could construct a dialyzer. Then there was the opportunity to change the pore size, and that would probably lead to advancements in filtering out more, or different molecules. Now, the perspectives are changing because what does that mean? If you filter out a particular molecule, does that mean that a patient have, less of a headache once per week? Or does it mean that that patient may, for example, participate more in social roles? And even though this may at first seem to be far-fetched, it is not. Because if we develop products that, for example, do not activate the complement system, which leads to less inflammation, it could well be that a patient will feel that he is more willing and able to participate in social roles. It is, therefore, important that we start from the patient and their individual daily lives, and especially the aspect that you mentioned about, intermittently checking in with a patient every two weeks. We try to understand, at least specifically related to nephrology and clinical research with kidney patients, is to understand when do patients feel something is changing? Is that before dialysis? Is that during dialysis? Or afterwards? How do they feel during the time that they're not on dialysis? And to really understand, okay, how is the recovery time? What sort of symptoms are present at what particular moment in time? That will help us to connect these different people together. Somebody from production facility for somebody who is really in that research and product development, will be able to understand what that means. The clinician will also be able to change their strategies of care. They might want to focus more on, pruritus or fatigue, that may not have been present so much, because patients would simply accept that this is part of the condition they have. Or there may be other things that take center stage while they see a physician, for example. This is definitely something we want to turn around and identify what really bothers people and make their voices heard.
Dr Maddux: Recovery time has been probably one of the most obvious ways to connect the treatment we give, which has a lot of quantitative measures to it to measures that are really related to how patients function. Let me switch gears for a minute and ask you about two particular clinical trials and the experience of those and getting the field better acclimated with regard to qualitative data, patient reported outcomes, quality of life. So, tell me a little bit about SONG and the standardized outcomes in nephrology process and those studies and what have they done to sort of advance this part of the field.
Krister Cromm: SONG is an initiative to standardize treatment parameters in nephrology, and in the publications of nephrology. So, basically what happened was, there was a group of very ambitious people coming together, trying to figure out what is really important for patients with kidney disease. These would be academic researchers, nephrologists, but also clinicians, people in industry, and for certain, the patients themselves. And what happened was, There was something called a Delphi process, so all of these stakeholders were asked about what they saw was most important. For the first time, it came out that these perspectives were very, very different from each other. Certainly, the perspective of the physician is to improve the lifespan and reduce hospitalization. But obviously, health-related quality of life aspects would also be seen. Whereas for patients, that would be the major concern, or one of the major concerns. They have to be able to enjoy the time they have to not be prone to fatigue. So, what happened was that these were combined, and there was a process by which this was standardized. For different treatment modalities available, like hemodialysis in-center. These variables were defined. There were standard variables, like, hospitalization, but also fatigue that now have to be reported in every clinical trial that wants to be published. That has definitely made a big impact. Because in this way, we can compare the results of trials on a common metric. It would lead to real quality improvements that focus on all the stakeholders’ perspectives, and in particular, the ones who have the disease so that they can make sure we progress.
Dr Maddux: I think certainly, the expectation that quality of life patient reported outcomes in these measures will be part of clinical trials has changed the expectations in the field. Last question I have for you is, you've been involved in this but broad consortium called CONVINCE that is studying HDF hemo-diafiltration. How is that impacted the way that patient reported outcomes and quality of life measures are looked at in a large multinational clinical trial?
Krister Cromm: The CONVINCE Trial is run by a consortium that got financed by the European Union's Horizon 2020 grant. It's financed by European taxpayers, and the EU is really, really interested in understanding differences of healthcare provision and access to care among its member states. Now, the EU is extremely diverse in terms of the economic development, as well as the cultural backgrounds of its member states. This has a very big impact on the way we do research. Because the EU is not only looking at understanding basically differences between dialysis modalities, and under what circumstances they do better, but to really say, "Okay, how do patients perceive that difference?" So, within CONVINCE, we developed a new approach to measuring patient reported outcomes, and to design a toolbox of aspects related to health-related quality of life, starting from the physiological basis, and understanding, for example, symptoms that happened during dialysis as well as general symptoms via actual functionality. If patients are depressed, patients are fatigued, to the true evaluation of their situation. Can they participate in certain roles, and the real quality of life that then follows. We create a lot of transparency and also a level playing ground for everybody participating in research. It will also allow to understand the circumstances under which a patient's live and the psychological situation of the individual as well. So, that is quite important because it is truly person centered. There have been a lot of different discussions with patient organizations in different countries. We use questionnaires that were originally developed in the United States. It's promise based, but we then applied them or apply them in the European settings. So, they've got translated, and we discuss this with the patients to see if such an approach would be viable, globally. I think that will contribute a lot beyond nephrology because it will allow researchers, clinicians, patients to discuss about what matters to them irrespective of where they are. And I think that is truly important because it is not limited to the more advanced healthcare systems like the US and some countries in Europe, but it can be applied elsewhere as well. And it's open.
Dr Maddux: Today, we've been speaking with Krister Cromm about the impact of qualitative measures, patient reported outcomes on not only clinical care, but clinical research and I think it's been very informative to hear your perspective on how this evolving part of our field is developing and maturing. So, thank you very much Krister.
Krister Cromm: Thank you very much, Frank was a pleasure.