3D illustration Rendering of binary code pattern.Futuristic Particles digital Landscape wave Abstract background for business,Science and technology
3D illustration Rendering of binary code pattern.Futuristic Particles digital Landscape wave Abstract background for business,Science and technology

Biomarkers are indicators of biological events. They are either present, or not; they may increase or decrease; or simply change form. They relate to normal biological states, the presence of disease (or the risk of it), or in response to intrusions into the body, including medical treatment or injury. Above all else, biomarkers provide information—about a person’s current health or to give clues related to future health.

As demand for preventive and precision healthcare continues to grow, so has the need for a new class of biomarkers. Ones that are highly portable, reliable, and immediate yet suffer no lack of precision or accuracy.

Enter the domain of the digital biomarker.

Digital refers to the method of collecting information. Instead of blood tests and medical imaging, digital biomarkers use sensors and algorithms across a plethora of available connected hardware and software tools. In this space, the patient is increasingly a consumer: one who wears, ingests, or has digital devices implanted. In many cases, these patient-consumers have some autonomy over when information is collected and with whom to share it.

The role of smartphones and AI-driven data

Connection with portable, personal or home-based products defines the landscape of digital biomarkers. Given their global ubiquity, smartphones are the natural home base for collecting, analyzing, and distributing the information gained from cameras, microphones, and other sensors embedded in digital technology.

Market research backs this up. Data from Rock Health’s digital health consumer adoption report finds that in 2019:

  • Forty-four percent of respondents track some aspect of their health digitally, and those who use digital tools share health tracking information with their physician or other medical professional more frequently than those who use other tracking methods.
  • One-in-three respondents owns a wearable, and one-in-four wearable owners uses it to manage a diagnosis.

For years, smartphone and big-data rich companies like Apple and Google have been developing digital health programs of their own. Using artificial intelligence and machine learning technologies, they are forging the path of not only collecting, but also analyzing the large data sets digital biomarkers create, so much so that they have helped usher in a new class of medical device, the ‘software as a medical device’ (SaMD). The Apple Watch is perhaps the iconic symbol of a SaMD, with its ability to detect elevated heart rates and atrial fibrillation. Fitbit, Samsung Galaxy Watch, and other devices are related SaMD cousins.

Starting with the 2014 launch of its first health app, Apple has been out in front of the smartphone crowd when it comes to digital biomarkers with its suite of iPhones, Apple Watch, and companion apps. Its Apple Health app relies on machine learning to collect and process health information. The company’s HealthKit platform provides a storage site to house all of that health and fitness data while also allowing access to thousands of third-party apps if the user designates.

David Feinberg
David Feinberg, Google Health

Similarly, Google’s AI Platform and TensorFlow machine-learning strengths help app developers create new models identifying at-risk populations and in flux patient data to better predict patient outcomes.

“There’s little doubt that artificial intelligence will power the next wave of tools that can improve many facets of healthcare: delivery, access, and so much more,” said David Feinberg, head of Google Health. “We’re already making strides in organizing and making health data more useful thanks to work being done by Cloud and artificial intelligence teams.”

Expanded sensing technologies

Sensing technologies are the modes by which biometrics are perceived. Most of the first health or fitness technologies focused on heart rate metrics. Today’s techniques have expanded to new areas like sweat detectors and insole sensors to help provide information useful for doctors to help inform treatment.

The Sweatronics platform, from Eccrine Systems, is intended to be used for many drug and drug metabolite uses. The technology is not intended as a long-term wearable, but more as a spot measurement to provide rapid results that can provide immediately actionable information.

Sweat Detection. Sweat detection technologies analyze biomarkers gathered from an individual’s perspiration. Though sweat detection tools have been in development for over a decade their time for commercialization may be here. One of the leaders is Cincinnati-based Eccrine Systems and its Sweatronics Platform.

“We have done a lot of work already with sweat in athletics and dehydration, but we are predominantly focused on measuring pharmaceuticals and their metabolites for precision dosing,” explained Eccrine CEO Gavi Begtrup. “We have shown that for certain small molecule drugs we can correlate those drug levels in sweat so that one can figure out patient-specific drug responses.” By measuring drug metabolites in sweat, the idea is to keep patients on the right drug dose, over time, and to monitor for patient adherence to taking their medications.

The Sweatronics device is a wearable device that stimulates and collects sweat in an area the size of nickel. The device also analyzes on-body sweat.

One of the applications Eccrine is pursuing first is in the area of opioid drug monitoring. “This means one could do rapid adherence monitoring, instead of the traditional urine-based monitoring which can take time,” said Begtrup. With the company’s technology, opioids are measured in the sweat with results immediately available in the clinic to the treating professional.

As a platform technology, the Sweatronics platform is intended to be used for many drug and drug metabolite uses. The technology is not intended as a long-term wearable, but more as a spot measurement to provide rapid results that can provide immediately actionable information. Results are electronically delivered to a tablet-like device and into the patient’s electronic health record.

Gavi Begtrup
Gavi Begtrup, Eccrine

“The big picture with our technology is to get away from one size fits all dosing and get everyone on the right drug dose faster,” said Begtrup. “That improves adherence and minimizes side effects for each patient individually.” Immediate applications could include medications with a narrow therapeutic range, like some blood thinners, and anti-rejection medicines following transplantation.

Other companies involved in sweat sensing technology include Xsensio with its Lab-on-Skin chip, Epicore Biosystems and its Discovery patch with electrochemical sensors for measuring sweat analytes and inflammation biomarkers for skin health, and Sweati’s patch for non-invasive glucose monitoring.

Several academic laboratories, including those at Stanford University, UC Berkeley, and the John Rogers laboratory at Northwestern University, have created wearable sweat detectors intended to provide real-time information on the user’s pH, sweat rate, and levels of chloride, glucose and lactate—which could be useful in diseases like cystic fibrosis and diabetes.

The Rogers laboratory partnered with Gatorade for its wearable sweat analysis technology intended to provide athletes real-time performance data. The Stanford team, led by biomaterials engineer Alberto Salleo, has developed a wearable patch that can measure a person’s cortisol levels from sweat. Cortisol levels rise during stress and are crucial in certain diseases like Addison’s disease and Cushing’s disease. Similarly, Ali Javey and colleagues at the University of California, Berkeley, designed a patch that can measure sodium in sweat and determine sweat rate directly from the skin.

Insole Sensing. For other digital biomarkers, companies have developed digital pedometers, or step counters. But some are taking a different perspective on tracking foot mobility.

Feet Me insole
An example of a Feet Me insole able to collect data on step pressure and gait.

Shoe insole sensors are now in use that capture data of a person’s gait to help with diagnosis and interventional treatment for conditions like multiple sclerosis, Parkinson’s disease, and even diabetes.

Feet Me, based in Paris, France, has created a wearable foot insole to assess and improve mobility. Now commercially available in Europe, the first technology was developed when the company’s founder was working with diabetic patients suffering from diabetic neuropathy. The FeetMe insole is a solution for ambulatory gait assessment involving a connected insole and mobile software to allow measurement of real-time gait parameters. It measures pressure distribution across the bottom of the foot, as well as movement patterns including walking speed, cadence, and load.

“We quickly realized the sensors we embedded in the foot soles could be useful for not only diabetic patients but also others with mobility disorders,” said Alexis Mathieu, co-founder and CEO. The team realized that current methods of assessing gait and walking patterns are insufficiently accurate. “The way people walk under observation in a short amount of time is not how they walk in real life,” he noted. “We wanted to find a way to grasp real-world relevant data to make clinical conclusions.”

Alexis Mathieu
Alexis Mathieu, co-founder and CEO, Feet Me

The company provides the insoles to physicians for home-based assessments of patients to help them understand a patient’s true mobility and to measure the efficacy of rehabilitation programs and drug treatment in a more sensitive way. “It’s important for everyone—patients, clinicians, and payors—to understand what works, and what does not, to improve the way patients are cared for,” Mathieu said.

Patients use the insoles which store the collected data. The insole is then given to the clinician who uses the FeetMe app which synchronizes the data and makes it available to physicians in a usable format.

Currently, the FeetMe insole is in a clinical study for multiple sclerosis in partnership with Novartis. Trials for other diseases like osteoporosis, diabetes, stroke, and Parkinson’s disease, are on the horizon.  “We want to replace the assessments people have used for years to monitor patients,” said Mathieu. “By making it real-world, digital, and continuous monitoring, we can find the small changes early that highlight treatment or physical rehabilitation efficacy or inefficacy.”

Other companies working in this area include 3L Labs (FootLogger), Moticon, RunScribe, and Digitsole.

Apps predicting disease onset

With Apple, Google, and others providing sensor and data receiving hardware and software, other companies use this information for their own companion medical apps. There are countless companies working in this area. The following represent just a glimpse into the new apps that either measure behavior or provide testing methods to collect data that can indicate either the risk, or pending onset, of disease.

Alzheimer’s Disease. The first thing one notices on the Altoida website is the banner: “Reinventing Digital Biomarkers for Better Brain Health.” The Swiss-based company has developed a phone- or tablet-based app that tests visual acuity and other measures for the early detection of Alzheimer’s onset.

“We are using active biomarker technology, and artificial intelligence machine learning to provide the Altoida neuro-motor index (NMI), a medical device to detect mild cognitive impairment (MCI) due to Alzheimer’s disease six to 10 years prior to the onset of symptoms with 94% accuracy,” said Oliver Krause-Huckleberry, Altoida’s VP of marketing and business development. The NMI is comprised of an iPad or tablet accelerometer, gyroscope, and touch screen sensor that detects user ‘micro-errors.’ Users engage with the table for approximately ten minutes generating data for the app. Using machine learning algorithms, user-specific information is compared with pre-loaded data from a healthy population to assess MCI risk. Users are instructed to perform the tasks every six months from a baseline assessment.

The Altoida tool is based on earlier research that shows as brain function changes, motor functions follow suit. “These motor function can be detected at a very early stage,” he said. “Our tool detects these early indicators and classifies the patient with an MCI risk due to Alzheimer’s Disease.” The app can be downloaded on any iOS- or Android-based smartphone or tablet. The company is developing a professional version for clinicians and a home-based app for consumers interested in assessing their brain health.

Examples of others developing similar tools for assessing brain disorders and brain injury include NeuraMetrix, Sage Bionetworks, Cerora, and Sway Medical.

Mental Health Disorders. CompanionMx has developed a mobile app-based tool intended to help clinicians monitor and treat their patients with mood disorders like depression and bipolar disorder. The app measures a patient’s speech patterns and cell phone metadata looking for quantitative symptoms of mental health disorders. Cell phone data include patterns on the frequency, diversity, and timing of cellphone interactions.

“We use subjective data from cell phones, specifically around voice and cell phone metadata, to measure behavioral health,” said CEO Subhrangshu Datta.

CompanionMx applies cloud-based AI algorithms to quantify symptoms from the collected data to provide measures of four symptoms of behavioral health: interest, social isolation, mood, and energy level. The information is relayed to health providers to act on if necessary.

“Clinicians will use the information to monitor patients, to risk-stratify, and detect changes in behavioral health early,” said Datta. Just like a continuous glucometer measuring changes in diabetic patients, Datta sees CompanionMx as a similar measurement tool. “We don’t have anything like a glucometer yet for mental health disorders and we should have a system in place to manage these early symptoms and provide better outcomes for these patients.”

In September 2019, the company entered a clinical study with the US Department of Defense using its tool for suicide prevention in active duty naval personnel.  In the future, it plans to modify the tool to assess for anxiety and stress.

MindStrong, Ellipsis Health, and BrainCheck, among others are also developing apps for measuring human-computer interactions to proactively identify mental health issues and provide users and clinicians with alerts of worrisome changes.

In addition to app-based companies, many established pharmaceutical and diagnostics companies are running pilot studies to test the feasibility of using digital biomarkers.  A few examples:

  • Roche Digital Biomarkers, a unit within the company’s informatics group, is building apps for multiple sclerosis, Parkinson’s disease, and others neurologic disorders
  • In August 2019, Evidation Health, a big data firm working with Apple and Eli Lilly & Co showed that digital biomarkers could be useful in detecting symptoms of Alzheimer’s disease and dementia earlier
  • Pfizer is developing a wrist-worn device to track tremor
  • Sage Bionetworks has partnered with Novartis to study multiple sclerosis, Lilly to study Parkinson’s disease, and Celgene to study asthma

As digital biomarkers continue to gain in popularity and sophistication, success hinges on the long-term view. In a May 2019 paper published in npj Digital Medicine, researchers working on digital biomarkers research at Boston Children’s Hospital write:

“To accrue maximum benefit to the patient, a safe and effective digital biomarker ecosystem requires transparency of the algorithms, interoperable components with open interfaces to accelerate the development of new multicomponent systems, high integrity measurement systems. The time is now to give forethought to strong incentive structures to promote the safe and effective use of digital biomarkers. Generally, the verification and validation of a digital biomarker should be not construed as a one-time process, but rather, a learning digital health system should continuously collect data and handle modifications and updates overtime. Industry, researchers, regulators, clinicians, and patients have a joint responsibility to design such a learning system that can improve digital biomarker products, empower patients, and improve health and healthcare delivery for everyone.”

This site uses Akismet to reduce spam. Learn how your comment data is processed.