What we learned from creating digital biomarkers for fatigue
No productive conversation about the future of personalized medicine can ignore the vast amounts of data generated and collected by digital devices. This data is raw material for creating digital biomarkers that give current, accurate information about a patient’s condition. Complex analyses of digital biomarkers can offer patients and health care professionals insights to and predictions about their health that will make truly personalized health care a reality.
What is a digital biomarker?
When you take a measurement from the human body, that measurement is a biomarker that tells you something about the body’s condition. Your mother taking your temperature with a thermometer when you were a kid, the nurse taking your blood pressure at the hospital, or a blood test result are all biomarkers. Each of these biomarkers is a bit of information and health professionals use them to understand a patient’s condition and needs.
Most biomarkers used in health care today are collected in a medical setting like a doctor’s office, clinic, or hospital. Together, a set of biomarkers give a snapshot of how a patient is doing at one moment. The patient’s history and the evaluation of their health is made up of a series of these snapshots combined with the patient’s story about what happens at home.
A digital biomarker uses information collected by sensors in mobile digital devices, like your phone or fitness tracker. The patient carries the device, which collects data all the time. There are no special trips to a clinic or hospital. Digital biomarkers can be analyzed with an app that can return results to the user or share them with health care professionals. Digital biomarkers offer information collection and analysis in real time.
Well-designed and validated apps using digital biomarkers can help reduce doctors’ visits by giving users suggestions on how they can improve their condition. They can also suggest that a user should go see the doctor if there is an early indication of a health problem. This kind of real-time medical information can change or even save a patient’s life. In fact, at least one life has already been saved by an Apple Watch that alerted its user to a heart problem.
Where do digital biomarkers come from?
The path to truly personalized medicine will be paved by organizations that develop reliable, validated digital biomarkers that can be used to measure treatment outcomes and in clinical decision making to optimize outcomes for each individual. Creating digital biomarkers depends on cooperation between patients, caregivers, health care professionals, and data scientists. In order to create digital biomarkers that can be widely used, developers need a shared a method to create and validate digital biomarkers.
Orikami is a data science company in Nijmegen, in the Netherlands. We created MS sherpa, an app that monitors and helps users manages Multiple Sclerosis (MS) symptoms. Multiple Sclerosis is a disease of the nervous system. The symptoms vary from person to person but can affect movement, vision, and energy levels. There is no cure for MS.
The MS sherpa app helps patients monitor their fatigue, mood, concentration, and stress. This tracking gives patients and their health care team insights that lead to more personalized diagnosis, prognosis, and treatment.
We developed a set of digital biomarkers for the app to measure fatigue and other effects of MS. Among them is one that measures eye movementand a two-minute walking test. In order to develop a clinically valid digital biomarker following emerging industry standards, we used the steps for novel endpoint development released by the Clinical Trials Transformation Initiative (CTTI) in June 2017.
CTTI developed a series of steps distilled from case studies. Developers can use this process to create digital biomarkers that consistently and reliable provide information that can help improve patients’ lives. The step-by-step process puts patients’ needs first and encourages corporation between patients, their caregivers, health care professionals, and developers.
With countless digital biomarkers on the horizon, many more organizations will be following the CTTI process in the future. Sharing Orikami’s experience with this process is important as we work towards a more reliable and consistent digital biomarker development process. We learned a number of useful lessons that we want to share.
Lesson 1: Digital biomarker development is a group effort
Developing a digital biomarker should take place alongside the people who will be directly involved in using it, not at a distance. The CTTI guidelines state that the search for which aspect of health a digital biomarker could address requires input from the patients, caregivers, and disease experts. Together, they can identify the aspect of health that could be improved through treatment and isn’t easy to measure or isn’t measured at all yet.
When we started developing MS sherpa, we asked MS patients which symptoms of their disease they wanted to improve. The patients answered, “fatigue.” Fatigue is a common MS symptom, which affects patients’ quality of life and can lead to depression or disability. Patients are forced to make difficult decisions about what they can do with their limited energy. Managing fatigue is a daily struggle for MS patients.
But this wasn’t the only stage at which input was helpful. Disease experts pointed out some clear physical symptoms of fatigue and patients’ partners helped us discover a new way to measure fatigue. Working as a team, we were able to create a tool that is uniquely suited to help MS patients in ways they need help most.
Lesson 2: New digital biomarkers offer new opportunities
Fatigue is a subjective symptom, which means you measure it by asking patients how they feel. Then, you look for changes in how patients report their experiences. But this self-reporting is unreliable and could reflect a patient’s mood or some other variable instead of isolating fatigue. Digital biomarkers from data collected by sensors built into a mobile phone or fitness device offer an opportunity to find new ways to track and measure fatigue.
It’s not easy to find a way to measure fatigue. However, both the scientific literature and experts on MS agreed that MS fatigue impacts reaction time. When MS patients are fatigued, they react slower. In our discussions with patients and their partners, the partners mentioned that they could see whether or not the patients were fatigued by their facial expressions. This led us to investigate the possibility of using some kind of facial change as a way to measure fatigue.
The MS sherpa app measures eye movement. A delayed reaction can be measured by using a mobile phone to measure how far an eye moves as it’s following something and to time how long that movement takes. Together, these create a digital biomarker for fatigue. Working together with patients and disease experts, we are working towards an objective measure for fatigue where is none.
Lesson 3: Technology follows patients’ needs
With all the technology available, it is tempting to start this process by looking first at what the technology can do. However, the patient and measurement needs come first when you’re developing a digital biomarker. That means technology follows, it does not lead. The patient’s interest drives the process, not possibilities offered by technology.
Once the technology has been selected, the developer’s goal is to produce exact, precise results. The technology must be able to perform the measurement conveniently and accurately. MS patients can use the MS sherpa app on a mobile phone to measure eye movement. The mobile phones have all the sensors necessary and patients tend to keep them close by. That means that if an MS patient is feeling fatigued, feels they might become fatigued, or is just curious, they can grab their phone and perform the test easily.
This is perfect for the patient but introduces challenges for the developer. The developer must figure out how to make sure the mobile phone used for the measurement has enough light and distance from the eye to measure well. They also have to deal with things like hand movement and the phone hardware. Sensors in the phone have to be sensitive enough to complete a reliable measurement. For MS sherpa, we created a set of guidelines that measured lighting and the distance between the phone and the eyes, guaranteeing an accurate measurement.
Lesson 4: Developers need a clear shared process
The path to creating a widely digital biomarker is demanding, and it should be. CTTI provides useful guidelines to make this process more transparent and the digital biomarkers themselves more reliable.
Orikami followed the CTTI steps to develop a set of digital biomarkers that use technology to measure a specific aspect of health, fatigue, and create a result that’s reliable and useful for MS patients. They can measure and track their fatigue using objective measures instead of relying on self-reporting. This information helps the patients, their caregivers, and the health care professionals who work with them.
Where are we now?
The next stage for any group that creates digital biomarkers is testing and validation. A small group of healthy people participated in the first round of MS sherpa testing. The next phase is with a larger group of MS patients.
This gives us an opportunity to test the MS sherpa app in different conditions to make sure the app is still accurate and precise. Orikami is now validating the MS sherpa biomarker as a patient-centered measurement for fatigue and disease activity. The question is whether the app works to give patients a chance to measure their own fatigue and how active their disease is. Patients will tell us about their experiences with the app. Clinicians will give report on the usefulness of the results. This study at the VU University Medical Center in Amsterdam (VUMC) includes 100 patients who will use the app and have several MRIs throughout the year.
We hope to have conclusions by the end of 2020. In the meantime, we continue to use the CTTI process for developing novel endpoints to develop digital biomarkers that will be useful in other areas of health care.
