Top 3 Takeaways:
Dr. Daniel Ford is a distinguished professor of medicine at Johns Hopkins University School of Medicine, where he also holds joint appointments in psychiatry, behavioral sciences, health policy, and epidemiology. Serving as the Director of the Institute for Clinical and Translational Research and the Senior Associate Dean for Clinical and Translational Research, Dr. Ford is widely recognized for his pioneering work on the interrelationships between mental disorders and chronic medical conditions. His groundbreaking research, which includes identifying depression as an independent risk factor for coronary heart disease and exploring the health risks of sleep disturbances, has earned him international acclaim. Dr. Ford's commitment to advancing clinical research through technology, such as his collaboration on the development of the e-IRB system, and his advocacy for treating research subjects as valued partners, underscores his leadership in the field.
Jake: Dr. Ford, it's nice to have you with us this morning. I’m thankful that you could join this Q&A session for Healthcare’s Data Innovations. I'm here in St. Louis, as always. You're in Baltimore, where you've now been for, I believe, 44 years at Hopkins University. Is that accurate?
Dr. Ford: I came as medical resident in 1982. So, you’ve got it right!
Jake: I am happy to introduce our audience to you, although many of them already likely know your name as a pioneer in translational research. For those that don't, I'll give a little bit of a background...you’re currently a Professor at Hopkins University Medical School with joint appointments in psychiatry and behavioral sciences, along with health policy and management at the Hopkins Bloomberg School of Public Health.
You're also the Director of the Institute of Clinical and Translational Research, while serving as the Senior Associate Dean for Clinical and Translational Research. That's enough to keep you busy! I believe most people do know you for the work you’ve done documenting the relationships between mental health and physical health, especially chronic medical conditions.
Dr. Ford: Yep, it is enough to stay busy. Mental and physical health relationships were the focus in the first part of my career. And I still find it very interesting today.
Jake: Given your experience, I'm looking forward to digging into your thoughts on the technical evolution of clinical research. I’m excited to have you with us today and I want to get the most out of our time, maybe I’ll kick off our first question if you’re ready?
Dr. Ford: Perfect, let’s get to it!
Jake: What has you the most excited about the future of clinical research right now?
Dr. Ford: For a long time, really for decades now, we’ve been looking at how do we stay connected to patients or research participants over the long term. We know that the more we understand about the human body, biology, diseases and coping styles that the right perspective is looking at decades more than years. I think the technology is now available to allow us to stay connected to patients in a way that, is much more affordable than ever. We can look at their real outcomes, the patient reported outcomes, connect their electronic medical records, all one place hopefully. We can be more thoughtful than ever how we incorporate clinical trials as well.
I believe there is a limit to what we can ascertain now and what we will be able to ascertain in the future about treatments and outcomes through just observational data. We have to find an easy way to move clinical trials forward in a more standardized way – the lift must get easier not harder given the technology advancements we currently have. We need to continuously push to understand not only the value of the outcomes and treatments measured in the trial but also the approaches we take to clinical trials.
Here’s an example...everybody loves a good risk predictor, but I am more than confident that most people would say, ‘yeah, it’s nice to predict my outcomes; but what I really want to do is improve my outcome.’ What are you going to do to make my outcomes better?
Jake: This is a conversation I just had with Dr. Sarah Biber, who is the Director of the National Alzheimer's Coordinating Center (NACC). A truly wonderful first step is identifying and predicting risk factors for Alzheimer's, but a much more positive long-term step is how do we delay the onset of symptoms and ultimately provide better health outcomes.
One thread I would like to pull on is the observational data you mentioned. We have digital electronic medical records that connectivity is getting better on. Interoperability is certainly still a hot button, but it's getting better. We have commercial technology that provides insights like my Apple Watch, we have remote patient monitoring – can we use that information to target clinical trials more effectively?
Dr. Ford: Certainly! A lot of scientific areas benefit from routinely collected data, but one item about the deluge of operational data is that we forget how hard it is to analyze when the timing or format of collection is not ideal. It's nice to know that we have more blood pressure measurements, right? But when you look at what's done clinically versus collected at home via a personal device or standard medical equipment, we aren’t exactly sure about the quality differences or the timing of events.
I go through this with patients all the time. They tend to take their blood pressure when they're not feeling well, or have a headache and then go check their blood pressure. That doesn't help us in the bigger picture, we must have a scientific approach to try to figure out whether a drug or a lifestyle change actually improves blood pressure.
This happens with physician visits as well, most people in this country are going to the doctor when they don’t feel well. Maybe their asthma is worse than usual or perhaps their depression is more acute than normal. Patients come to services that they need – and that makes it hard to do analysis on just clinical data. That’s why getting the right scientific approach with clinical trials is so important – in many ways, randomized clinical trials is truly the best analysis possible. The participants have a set date of treatment, current state, outcome and set timelines – their condition doesn’t impact the timing which helps us with clean data. So, to make a long story short, we do need to continue to get better at using technology to help us drive more clinical research and trials in the real world.
Jake: What concerns do you about the increase in digitization of trials though? Are there items we need to address before sprinting towards the future?
Dr. Ford: That’s a good question, one aspect I think we do have to be clear about is that most of these long-term cohorts that depend on technology are still leaving behind segments of the population. Whether it's, people that don’t have the exposure to technology, or older adults, or underserved communities – we have to work through the digital divide of society.
It’s sad to say, but people with visual or hearing impairments typically aren't accommodated in the remote technology either. It's not to say they couldn't be of course, but right now, they're not. That’s an example of how we have to improve the availability and usability of digital infrastructure within research.
Just yesterday I was part of a trial for patients over 75 that do not have any cardiovascular disease. These participants are getting randomized to take a statin or a placebo. The outcome we are evaluating is cognitive independence and for most people it is an entirely remote trial. However, there is a woman, who is very intelligent and seems comfortable with technology that we could not work through getting the consent form signed electronically. So we're ending up mailing her the consent form to sign and send back. Sometimes we don't understand as that our reliance on technology can push patients further away from us – rather it be hardware or software challenges, not everything digital is setup to accommodate all of what we do in society let alone research.
Jake: You bring up a really wonderful point, which I think there's a large population of people that have an opportunity to be hidden in what I might call a modern world. I understand older adults, but there certainly other populations of people that may not have health insurance so their access is limited, individuals that mental health challenges, and many other vulnerable populations that need to be seen both in the physical and digital realms of medicine.
Dr. Ford: You know what is scary is that technology is changing so quickly that perhaps every generation can feel that the older you get perhaps that farther away from digital inclusivity you also get. I know that’s how it’s been for the last four to six decades – we need to think differently about technical inclusion. How do we democratize access and learning and ease of use across individuals but in a manner that is uniquely comfortable to those populations of people that have to this point, been on the fringes of our digital world.
An example is a recent Alzheimer’s study, there is something like 6500 people included in this study and I have not seen the latest numbers but something like only 400 don’t have a high school education, in fact, the majority have a college or graduate degree. So, interestingly enough, the patients that have the highest risk of developing Alzheimer's disease are actually not represented well. We are at this stage where if you want to be as inclusive as possible, you must include a hands-on component in research. From my experience and what I still see, there has to be considerations like a clinical research unit where a patient can come in and just get help. Even simpler things, like participating in focus groups that are increasingly becoming online. Some people are just going to say that they can’t figure it out, don’t have access, or don’t want to be bothered with the challenges of Zoom. Early in my career, now more than 20 years ago, we developed like an extremely easy to use keyboard. This keyboard had “yes”, “no”, and “next question.” It clearly helped. You have to give people options that fit them where they are at and make the feel comfortable participating.
Jake: Not to change us to far of course, but these are system level changes you are talking about. We saw a significant burst of innovation in the pandemic, right? We saw hospital systems and academic medical centers continue to think about the value of their data in new ways, especially from an epidemiology standpoint. But technology can only take you so far, technology lives in an ecosystem of rules, processes, people, and organizations – solving system level challenges is a big effort and who is included and gets to participate in our health system in the US is certainly a system level challenge.
Dr. Ford: We tend to let the systems we have in place drive our current thoughts on innovation – perhaps we shouldn’t. But, in reality, we do the best work we can given the environment and constraints in which we work. To step back into the pandemic era, health systems and providers across the board were over worked and under resourced so using technology where we could to lower that burden made sense given that moment in time. But now, if you look at many of things we did and likely still do, almost every technology solution put more burden on the patient. The patient had to more, they had to pick up the slack in provider availability – this example may sound simple, but some of our creativity actually shifted significant responsibility to the patient. For example, we were able to do pulmonary function tests at home, we dropped off the equipment on the porch, gave them a video and had them complete the test, and put the machine back outside of us to pick up. It’s great that we are getting those tests done certainly, but how quality where they? How is that data being utilized? Should we look at or score that data differently given it’s likely somewhat flawed? These are questions that we haven’t answered yet.
Another area where I do a lot of work is smoking cessation studies. We use an apparatus that we can send to people’s homes. This device connects to their phone and help us verify smoking cessation versus having the patient come in – all of it sounds wonderful, but once the participants get started, they can get frustrated. Maybe their phone isn’t supported, maybe they need to update their phone’s operating system – they don’t want the burden to shift to them int he name of convenience. Is that at a trend that will continue? Will we prioritize convenience over perhaps lower quality data in the future in the name of technology? Is there a chance this trend reverses and we shift towards a more fully in-person or high touch health care and research ecosystem?
Jake: That’s very interesting, we do spend a significant amount of time having conversations, at what I’ll call, the top-level of technology. The strategy of getting outcomes, architectures, integrating workflows, etc....but it only matters if the rubber hits the road in the intended manner. I’m curious to get your thoughts on research being driven by academic medical centers which typically are in large population areas. Do you think technology is helping democratize research for populations outside those locations?
Dr. Ford: You really do need a small army of personnel to conduct research and care at scale it seems. That being said, the ability for research team to work together without geographical restrictions has been critically important, so maybe the patient’s data benefits a significant amount in the name of research advancement, but it’s another story for the patient themselves. I do not believe that that technology will magically bring in everybody no matter of their location, some care requires extensive resources that just aren’t found in every geographic location in our country – to think otherwise is likely a little naive.
That’s not to say that I don’t think things are getting better in terms of spreading care advancements, just that technology alone won’t solve this problem. From a quality research perspective, I still think, that the best approach hybrid type studies which certainly leaves more room for inclusion, but does make it impossible for everyone or anyone to join a study.
Jake: There’s still a bit of I want to see and hear and feel a doctor who will be treating or following me right? We have this love/hate relationship with technology when it comes to health.
Dr. Ford: Yes, we love that advancements have gotten us to a place where my medications come straight to my home. their home. But, how much do we let technology drive our health encounters is largely driven by how much we trust technology generally. For example, we have to reach out to individuals for trial engagement and that process is always changing. For individuals that are not active patients within our providers, I would venture to guess most people's first thought on our outreach is ‘this is a scam’. We have a research participant experience survey that we've been doing for a long time, and we get a response rate of about 28 percent. But what we find is that the younger individuals had the lower response rate.
People over 65, they're very dependable and they may respond at a rate of 40%. People less than 45 were more like 10%. So, we decided to start texting them. Texting them the link to the survey. It didn't improve the response rates at all. But this last time they got the young adult rates up into the high 20s by sending a text that told them to go check their emails and complete their survey.
Making a connection through technology isn’t a given and it’s a moving target in our digital world.
Jake: That makes so much sense, I likely get 300 emails a day and I typically start by checking all of them and maybe pick out the 20 or so I need to respond to but then mark the rest as unread. I just don’t believe there’s value in those messages enough to check them. I wonder if our changing attitudes towards recognizing the negatives in technology, like screen time in children, or social media use impacts are making critical technical communications even harder.
Dr. Ford: We are lucky here to have the Johns Hopkins history and community trust because it’s so easy to dispell anything you receive in a digital communication.
Jake: I wouldn’t be doing a very good job if I didn’t ask you about your views and perspectives on artificial intelligence currently.
Dr. Ford: My general thoughts are that the workflows, processes, and intelligent integration is likely more important than the exact technology used. Providing care and doing research at scale are complex things that have real impact on our patient’s lives. Getting the right result is critical, we can’t negatively impact the patient or participant.
For example, we look into the health system EMR and say, we believe these patients are eligible for a specific study. That’s wonderful, perhaps you’ve used AI to make that patient matching more automated and accurate – but how are you building in appropriate delays for contacting those patients? You can’t just have a patient you are monitoring, get a positive on a certain test and you hit them with trial messaging before their physician has event talked to them.
Leveraging AI certainly may help with one burden of patient matching, but create an even larger problem if deployed without incredibly thoughtful workflows and guardrails.
Another example would be some things I see around dementia and cognitive impairment. This is a very sensitive area for patients, telling someone that you believe they are eligible for a trial because they are experiencing memory or cognitive declines is not easy. We know that there there is a high number of objections from these patients. I would also include studies that have sought out participants that would be considered obese adolescents – further automating those workflows with technology, no matter how advanced, is likely not a fruitful endeavor.
Jake: The human element is critically important in those circumstances, we’ve seen some success in using AI to A|B test outreach methods to find what does work. What are your thoughts there?
Dr. Ford: That’s an area of interest, can AI help us understand how effective our current is? I do think so, if you can understand what works well in terms of trial recruitment then you gain the ability to do clinical trials more efficiently and ultimately save money, time, and resources. I'll give you some examples. We have a pretty good clinical trial management system that would allow us to know whether that patient has been in a study before. And we think, a tailored message around whether it's their first experience with research is, maybe the best approach. Our experience and intelligent automation has taught us that prior study participation is probably more important than if that participant hits all ten of the eligibility requirements. That's at least, our experience. The other piece we’ve learned is that we need to develop digital recruiting messages that that makes it clear that we are seeking a broad and diverse population, we want participants with a disability, varying education, and experiences.
Jake: Dr. Ford, we could do this all day but I know we are at time. I want to again thank you for time your time today. I know our audience will appreciate your thoughts and perspectives. This has been a wonderful conversation and I’ll be looking forward to more!