Interview: The importance of digital biomarkers in clinical research

By Liza Laws

- Last updated on GMT

© Getty Images
© Getty Images

Related tags Digital biomarker Clinical trials Data management Patient recruitment Research Patient centricity

Rich Christie is chief medical officer at AiCure LLC. He is a firm believer in digtal biomarkers and the audio and visual biomarkers that can be captured through a front-facing camera. OSP inteviewed him to find about more about the changing landscape and what those not entering the digital world are potentially missing out on.

What type of digital biomarkers hold the most promise in clinical research? Can you give some examples of how you’re seeing this technology used today? 

While many people think of smartwatches or other wearables when they hear the words “digital biomarkers,” the biggest opportunity to personalize care in clinical trials is through audio and visual biomarkers that can be captured through the front-facing camera on a patient’s smartphone. Many diseases can be assessed by looking at the visual and auditory elements of their symptoms. By using technology patients already have at their disposal and analyzing a patient’s behavior, we can capture these digital biomarkers such as facial characteristics, vocal patterns, movement, and language content. The combination of video and auditory biomarkers enables sites to remotely detect subtle changes in a patient’s health status and response to treatment that may otherwise be too subtle for a clinician to catch in real-time or between visits.

Detection of depression is an ideal use case for audio and video digital biomarkers. Patients suffering from depression typically exhibit reduced facial expressivity, a slow, quiet voice, and reduced overall movement. If they’re asked to use an app on their phone each day to respond to prompts or complete brief tasks, we can analyze things like whether they’re smiling, their vocal inflections, and facial expressions to get a sense of their mood and general health. Through video and audio biomarkers such as these, care providers may be able to get a better picture of whether particular medications, such as Selective Serotonin Reuptake Inhibitors (SSRIs) are improving the patient’s condition over time, and if not, can adjust their dosage.

OSP: What are three of the biggest opportunities for the adoption of these types of digital biomarker solutions in clinical trials?

Beyond providing physicians with a greater understanding of a patient’s lived experience with disease and treatment, digital biomarkers can also help pharmaceutical companies make better medications that are more targeted to patients’ needs. There are many benefits of adopting digital biomarker solutions for trial sponsors, starting with increasing a drug’s competitiveness.

In crowded drug markets, like in cardiology, it is essential to show a medication’s value over the existing standard of care. Digital biomarkers’ sensitive and objective data capture can enhance a drug’s competitiveness by identifying new endpoints that demonstrate measurable quality-of-life improvements. For example, digital biomarkers in a clinical trial could demonstrate that a heart disease drug was able to lower patients’ shortness of breath so that they were able to go up a flight of stairs stress-free. Marked, data-backed improvements in patients’ daily lives could ultimately make that medication much more attractive than other drugs on the market.

Digital Biomarkers also have great potential to enhance data objectivity in clinical trials and deepen the pool of clinical data available to interpret study findings. Rather than relying on a single moment in time when a patient has in-person site visits to gather key data, remotely collected digital biomarkers capture insights continuously in a patient’s lived environment and catch subtleties that otherwise may have been missed or aren’t visible during the time of a visit. More objective, valid data capture means faster trial timelines and improved resource allocation.

Ultimately, digital biomarkers could be a critical piece in helping deliver on the promise of precision medicine. Being able to pinpoint subtle, sensitive patient responses to therapy in a clinical trial can help inform timely, personalized care. The more we can understand the nuances of a patient’s condition and response to treatment, the more we can help match the right patient with the right drug at the right point in their disease.

OSP: What disease areas are ripe for this innovation?

Any diseases that have visual and auditory elements to their symptoms can benefit from this type of data capture. In particular, cardiovascular disease, infectious disease, central nervous system, and immunological disorders provide great opportunities for digital biomarkers to shine.

For example, stroke can be a frustrating thing for patients to deal with and for pharmaceutical companies to develop effective medications for, because it is difficult to reliably keep track of symptom changes between doctor visits to understand if the medication is accurately treating their disease. With digital biomarkers, care providers can track a stroke patient’s symptoms over time and create a trajectory of their health so they can understand how a patient is responding to medication sooner.

AI-powered digital biomarkers can have a powerful impact on immune diseases like asthma, especially when that data is layered on top of environmental data that could impact symptoms. By tracking digital biomarkers such as shortness of breath and factoring in patient dosing behavior and environmental elements like weather and pollen count, care providers can get a much more comprehensive picture of a patient’s lived experience with their disease.

OSP: There are many diseases that have a wide range of symptom expression, which can make it especially hard for trial sponsors to decipher whether a medication is working for a particular patient or not. Why is it that digital biomarkers are so good at cutting through these symptoms to identify the true effectiveness of a drug?

Diseases with an array of symptom expressions, particularly lend themselves to the kind of predictive modeling machine learning and AI can provide, which can help clinicians untangle the dimensions of diversity in how symptoms present in different patients. For example in immunological disease, these tools can monitor and analyze digital biomarkers across several coexisting symptoms in a patient to identify how multiple comorbidities may be influencing each other and help care providers optimize interventions accordingly The precise, personalized insights delivered from digital biomarker data help sponsors determine a drug’s effectiveness on a personal patient level, as well as for a broad patient population.

Immunology clinical trials in particular can also struggle with apparent “non-responders,” or patients who don’t react to treatment with little explanation. By exploring the predictive capabilities of digital biomarkers and gaining an intimate understanding of a patient’s lived experience with their disease, we may be able to catch these “non-responders” much earlier to understand what adjustments in their care plans may be necessary.

OSP: Bias in AI has been a big topic of conversation recently. How can tech companies ensure these AI-powered digital biomarkers can provide quality data for all patient populations, and that they aren’t perpetuating any human biases?

When developing an app that will analyze digital biomarkers using AI, it is essential that the algorithm is developed using diverse data sets reflecting the broad diversity of the underlying patient population. Algorithms are only as inclusive and intuitive as the data they are fed, so it’s essential that computer vision-based algorithms are trained with an array of skin colors, tongue pigmentations, room lighting, and other factors so that derived biomarkers are able to capture critical patient health indicators across all racial, gender and age demographics.

Until recently, the enforcement of scientifically-sound, ethical practices when developing AI has largely relied on good faith in developers to do the right thing. Fortunately, the FDA recently issued​ draft guidance in an effort to promote transparency and address some of the root causes of biases in AI. Their latest approach would ensure factors such as race, ethnicity, gender, age, and beyond are more vigilantly addressed in the ongoing development, validation, implementation, and monitoring of AI-powered devices. Regulations and federal oversight of this growing industry continue to evolve, but this is a promising step toward advancing equitable technology in clinical research.

OSP: Where is digital biomarker technology in clinical research headed in the next 3-5 years?

In the next few years, we’re really going to see a shift in how digital biomarkers are applied in clinical trials and patient care. Today, digital biomarkers can give us insights about a patient’s general health and response to treatment, but as they advance, they will be able to transform that knowledge into predictive power, including predicting a patient’s outcome and the trajectory of their disease.

For example, there is a lot of promise for digital biomarkers’ influence on immune disorders that have a high risk of disease flare-ups, such as Crohn’s disease, dermatitis, rheumatoid arthritis, asthma, and multiple sclerosis. With enough high-quality data from visual and auditory digital biomarkers, we can visualize the trajectory of the patient’s health and predict these breakthrough events to proactively mitigate them before they occur. This predictive power could be strengthened by combining digital biomarker data with lifestyle data such as diet, exercise, and environmental data, including weather, or pollen count, allowing for further personalization based on an individual’s lifestyle, environment, and predicted response.

Related topics Contract Manufacturing & Logistics

Related news

Show more

Follow us