A biomarker is a piece of anatomical, physiological, or biochemical data that is used to diagnose or develop a treatment plan for a patient. A digital biomarker is simply a biomarker that is developed from data that are analyzed using advanced analytics and AI to extract previously invisible insights. The key defining elements of a digital biomarker include:
- Continuous and noninvasive collection of data via a digital device like a smartphone, voice recorder, camera, or any wearable device
- Using advanced analytics and AI algorithms to analyze the data – such analysis would not be possible via humans alone
- Extracting novel or previously invisible insights
It is worth noting that "digital biomarkers get defined" not via the act of how the data is collected (which does not intrinsically add more or less value to a biomarker) but rather through the key differentiator of the act of analyzing the data streams to extract novel insights.
While digital biomarkers are slowly finding application in multiple health conditions, such as high blood pressure and sleep apnea, these elements are especially conducive (although not exclusive) to tracking behavioral changes. For some time, experts have recognized that behaviors and social interactions speak volumes about disease states, but researches were previously unable to compile and analyze these data. However, it has only been with the near-ubiquity of wearables, smartphones, and social media and the availability of advanced communication and computing technologies that the medical research community now has access to the data sets and analytics that allow for a direct diagnosis to be made based on nontraditional data sets. Traditional diagnostic tools for depression require tedious questionnaires that are easy for the patient to intentionally bias and only take a snapshot in time. Researchers have shown that the simple monitoring of an activity tracker was able to detect depression more quickly and accurately with step counts alone with fewer time and money resources. Digital biomarkers derived from text analysis are allowing for the earlier diagnosis of dementia – again, a diagnosis that has historically taken months to make and careful monitoring by caregivers. Even gait analysis, using a camera or shoe-worn sensors, is detecting minute changes in hearing ability that can predict cognitive decline early enough to treat the condition before it becomes permanent.
With new data sets and new diagnostic capabilities based on that data, researchers are uncovering new ways to diagnose conditions. Some of these conditions, such as dementia and depression, can take months or years to diagnose using traditional diagnostic tools – but digital biomarkers can shorten that time to a range of hours or days, or even minutes. A faster, more precise diagnostic can do a few things:
- It can get help to people who need it faster – meaning that patients can receive care in earlier stages of a disease, when it is typically less expensive to treat
- If can foster a faster understanding of population demographics and disease states within a population in order to more efficiently allocate resources
- It personalizes the care to the specific need level the patient has
- It makes the diagnosis and care more patient-centric rather than physician-focused, as these biomarkers can be measured and tracked remotely
Because of these diverse capabilities, emerging tech-based modes of care, such as digital therapeutics, are being increasingly informed by these markers. Especially in the case of cognitive and mental health conditions, we expect future diagnostic and treatment protocols will be underpinned by "digital" biomarkers the way that current treatment protocols for conditions like diabetes and myocardial infarction are underpinned by biochemical biomarkers. Additionally, digital biomarkers can have a degree of separation from traditional healthcare facilities – shifting the balance of who is determining the need for care from doctors and hospitals to tech companies and patients.
Digital biomarkers are emerging in the research space, most of them being identified in papers but not widely used for diagnostic purposes. Companies should be monitoring this research and begin to strategize how these biomarkers can be used to identify new business opportunities. However, rather than getting caught up in the idea that "digital biomarkers" will bring dramatic changes to healthcare, they should be viewed as just another data set. It is true that they are data sets derived from new analytics tools, but they are still simply data sets. Digital biomarkers, like any other data set or combination of data sets, have the potential to allow companies to personalize products to their clientele, streamline operations, and develop alternative revenue streams.