As medical knowledge grows by leaps and bounds, the way in which we capture and process health-related information also continues to evolve. This is reflected in the expansion of the field of digital biomarkers, which has garnered intense interest and is poised for advances that will likely change the understanding and practice of healthcare.
What are digital biomarkers?
But what exactly are digital biomarkers? It helps to define traditional biomarkers first. “When I think of traditional biomarkers, I think of anything that measures or evaluates biological processes in order to help us understand a person’s health state,” said Bryan Cobb, PhD, FACMGG, Partner Program Lead, Roche Information Solutions. “Biomarkers capture pathogenic processes in order to identify early evidence of or diagnose disease states; they also help us evaluate responses to an exposure or a therapeutic intervention.”1
Examples of traditional biomarkers include common clinical measurements such as blood pressure readings that a nurse might measure or a lab test that measures the total amount of cholesterol from a person’s blood. 1 Total cholesterol can be broken down into different biomarkers such as low-density lipoprotein (LDL) and high-density lipoprotein (HDL) levels that can further help assess a person’s risk for developing heart disease. Other biomarkers can analyze an individual’s genetic information (DNA/RNA) to help identify inherited mutations in genetic disease like Cystic Fibrosis or susceptibility genes that predispose individuals to cancer. Molecular analysis of DNA or RNA from a virus or bacteria is used by clinicians to characterize an infection or determine if the treatment is working based on how much of the pathogen is detected.
In addition to the physiological, molecular, biochemical examples of traditional biomarkers, they can also be measured using histological (how cell changes might reflect diseases) and radiographic (x-rays) methods.1 Digital biomarkers, on the other hand, are measured using wearable, mobile, digestible, and implantable devices. They track bodily processes, but unlike traditional biomarkers they’re collected via small portable devices that capture continuous data versus traditional methods, which typically measures “snap shots” and require a clinical or laboratory setting. The use of these devices has opened the field to allow many more people to gather health-related information in new ways. From wearable tools such as watches that track steps for fitness enthusiasts to implantable tools that measure heart rhythms to digestible capsules that can record gastrointestinal data points such as gas volume2, there’s a much broader spectrum of tools to help us better understand our conditions or help predict health-related outcomes, Cobb said.
Advantages of digital biomarkers
One big difference between traditional and digital biomarkers is that the latter relies on less invasive, cheaper sensors that can use algorithms across hardware and/or software applications. 3 These applications can represent digitized versions of well-established metrics or entirely new ones. For instance, blood pressure is a known measure of cardiovascular health. Typically, a blood pressure reading is taken at a single point in time (the traditional approach) and used to assess how likely someone is to suffer a heart attack. However, if a consumer wears a 24-hour, non-invasive blood pressure monitor, which provides a continuous stream of data (the novel approach), it could be used to provide a more accurate or predictive assessment of a patient’s cardiac health.
The novelty doesn’t necessarily have to reflect a novel measurement, Cobb noted. It can also measure novel clinical insights. A well-established measurement, the blood pressure reading, could be associated with a novel insight or idea that researchers want to test. Or it could be both, for example, digitized continuous blood pressure readings (using a novel digital method) could be used to assess a patient’s response to a new drug or condition (novel insight).
Ideally, the discoveries made possible with digital biomarkers will make it easier for consumers, to use these digital devices to take a more active part in their own healthcare. Shifting towards consumer-enabled continuous types of measurements enables different types of assessments that might help us improve how we manage our own health status is what Cobb sees as the greatest opportunity for digital biomarker growth in healthcare. Two fields where there’s particular excitement about the possibilities for breakthroughs are neurology and psychiatry, in which conditions do not always present in obvious ways.3 If irregularities can be detected before problems have time to progress, experts feel, we may be able to intervene earlier which might lead to better overall outcomes. 4
The impact of digital biomarkers on neurology and psychiatry
Take the example of trying to determine whether someone has Parkinson’s Disease. “There are no specific traditional biomarker tests, only signs and symptoms and a careful evaluation by a neurologist. Lab tests can be used rule out other explanations,” Cobb said. “But instead, you could complement traditional assessments with the use of smartphone apps, where automated tests can be run over time to measure symptom fluctuations or perform motor coordination and vocal assessments. This would require software to detect the movement and to also be able to understand vocal changes. The severity, how bad it is, and it’s progression could be quantified.” These types of digital biomarkers could also be very helpful in measuring the effects of new therapies, especially for drug development.
Recently, a team of researchers in Milan published a study in which they used virtual reality devices and artificial intelligence to detect early gait alterations in older adults. These changes potentially signaled neurological or cognitive problems that, without intervention, could advance to dementia.5 And scientists at Dartmouth College and Harvard University described a trial that included an app designed to track a person’s movements, along with their pattern of phone and text usage, to predict subjects’ levels of social anxiety.6 Such digital biomarkers could be collected continuously, providing points of comparison that could inform in a much broader sense than data captured at a single point in time.
Effectively managing biomarker data
But whether it’s the brain, heart, or other bodily systems being monitored, the challenge now is to harness the vast trove of data generated by the $30 billion wearables industry in a way that demonstrates what’s meaningful and what’s noise, according to Cobb: “You have this explosion of measuring devices, and there's this question of how much of the data generated by the things that we as consumers go and buy will be meaningful and how long will it take for us to be able to fully leverage it within the healthcare space. Vast amounts of digital data have already been generated and clinicians and researchers are studying how to integrate data into healthcare IT solutions, present the data to clinicians in a meaningful way, and make sure clinicians can easily use the data. Even if we know a given solution can improve patient care, human factors need to be considered in how digital tools get designed and implemented to make sure that clinicians will use them in routine practice”
While the field of digital biomarkers is exploding, Cobb is realistic about the obstacles in taking advantage of this data, at least in the short term. “Currently, we don't have the infrastructure to be able to store, analyze and really understand how these digital biomarkers can be integrated, validated, and used most effectively,” he said.
"Imagine the insights we will gain into our health status with digital biomarkers that measure and regulate processes inside our bodies at the cellular level--truly personalized medicine--and I think that's one really exciting area that we can look forward to in our lifetime.”
As these hurdles are overcome, one goal is to use digital biomarkers to help solve todays problems, like separating patients who need to be hospitalized from those who don’t. During the COVID-19 pandemic, hospitals nationwide are overrun, struggling to be able to meet the demand for those that need care with unnecessary hospital stays a concern. Cobb envisions a multitude of digital tools that remotely monitor people will be used more and more to help better assess symptoms, the need to go to the hospital, or for clinicians to determine the trajectory once admitted. “I think one day digital monitoring devices integrated along with other patient data could help clinicians distinguish sooner those that are likely to deteriorate and need further care from those that can be sent home” he said. 7
The inclusion of digital biomarkers in healthcare could benefit not only patients but hospitals themselves, given the current focus on value and outcome-based healthcare and the need to improve overall hospital metrics. “We have to have ways to properly assess the quality of digital tools, be able to integrate the data, while carefully producing clinical evidence to really be able to take advantage of their utility. Imagine the insights we could gain into our health status with digital biomarkers that measure and regulate processes inside our bodies at the cellular level. Truly personalized medicine.” Cobb said. “And I think that that's probably one really exciting area that we can look forward to in our lifetime.”
- FDA-NIH Biomarker Working Group BEST (Biomarkers, EndpointS, and other Tools) Resource (Food and Drug Administration (US) Silver Spring (MD) 2016 http://www.ncbi.nlm.nih.gov/books/NBK326791/
- Dickson I. “Gas-sensing gut capsules.” Nature Reviews Gastroenterology & Hepatology. 17 Jan 2018. 15:131. 131(2018) https://www.nature.com/articles/nrgastro.2018.3
- Babrak LM, Menetski J, Rebhan M, Nisato G, Zinggeler M, et al. “Traditional and digital biomarkers: two worlds apart?” Digital Biomarkers. 2019;3:92-102. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7015353/pdf/dib-0003-0092.pdf.
- Taylor, K.I., Staunton, H., Lipsmeier, F. et al. Outcome measures based on digital health technology sensor data: data- and patient-centric approaches. npj Digit. Med. 3, 97 (2020). https://doi.org/10.1038/s41746-020-0305-8.
- Cavedoni S, Chirico A, Pedroli E, Cipresso P, Riva G. “Digital biomarkers for the early detection of mild cognitive impairment: artificial intelligence meets virtual reality.” Frontiers in Human Neuroscience. 2020. 14:245. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7396670/.
- Jacobson NC, Summers B, Wilhelm S. “Digital biomarkers of social anxiety severity: digital phenotyping using passive smartphone sensors.” Journal of Medical Internet Research. 2020;22(5):e16875). https://www.jmir.org/2020/5/e16875/.
- NODE Health Website. The Changing Regulatory Landscape during and after COVID-19 featuring Dr. Eric Topol, Bakul Patel and moderated by Dr. Aenor Sawyer [August 19th, 2020]. Available from: https://nodehealth.org/2020/09/16/the-changing-regulatory-landscape-during-and-after-covid-19/