Digital Biomarkers & the Future of Mental Health

 In Articles

By Simona Mellino

 

With many industries currently undergoing a digital transformation, it should come as no surprise that digital biomarkers have piqued the interest of the pharmaceutical and healthcare industries. But what are digital biomarkers? What promise do they hold? And what issues arise with their implementation?

Background

Digital biomarkers are consumer-generated physiological and behavioral measures obtained using digital tools used to inform healthcare decisions. According to the National Institutes of Health (NIH), a biomarker is a characteristic (e.g., physiologic, pathologic) that is objectively measured and evaluated as an indicator of normal biologic processes, or biological responses to a therapeutic intervention.

The term “digital” thus refers to the means of collection, which includes wearables, implantables, and even ingestible devices. The increased enthusiasm stems from the opportunity of having access to vast amounts of new longitudinal data sources (the collection of a same type of data over time). These can be transformed into actionable insights.

Digital biomarkers have the advantage of extending beyond clinical settings, giving the possibility to include individuals not formally enrolled in clinical research and thus excluded from the data collection process. According to Markets and Markets, the global market for medical wearable devices is projected to reach $14.4 billion by 2022, with the US being the largest market.

Opportunities and challenges

The opportunities behind digital biomarkers are immense. Digital health technologies can enable more convenient, cost-effective means to collect discrete health measures, such as blood pressure and glucose, that clinicians can use for the decision-making process.

Examples of digital biomarkers include risk, diagnostic, monitoring, prognostic biomarkers as highlighted in a recent publication in Npj Digital Medicine. Some digital biomarkers are directly targeting clinical management.

For example, Empatica launched a wrist-device called Embrace2 (cleared by the FDA), which measures sympathetic nervous impulses at the skin. Their algorithm is then able to detect seizures and alert care providers.

It is not surprising to see that a wide variety of applications are in the field of neurology. A strong body of scientific literature indicates that cognitive, behavioral, sensory, and motor changes can help detect neurological or neurodegenerative diseases.

One example of a risk biomarker is a computerized cognitive test that helps classify adults at high-risk late onset of Alzheimer’s.

Other applications, which would simply require a microphone as a sensor, are in the domain of speech and language (including voice features and cognition, such as voice tremor, vocabulary, semantic qualities). The usage of vocal biomarkers of depression, dementia and Parkinson’s disease is for example under development. Evidation used 7000+ recordings from the Framingham study over a time span of 10 years (2005-2016), selecting a subset of patients diagnosed with dementia. Companies such as Sonde Health have already performed pilot studies and patented their technologies. A small study of 34 young patients, undertook an analysis of many speech features (e.g., confusion, word choices) to predict whether patients at risk of schizophrenia would transition to psychosis and outperformed clinical ratings.

Digital biomarkers could significantly help shifting healthcare from reactive to a more preventive approach, as more and more data will be available to analyze what healthy, normal states look like, and predict future health outcomes.

However, digital biomarkers come with challenges. Discovering and validating insightful digital biomarkers surely can require a long validation process. As brilliantly pointed out by Rock Health, the complexity will depend on the novelty of the insight being measured, as well as the novelty of the measurement itself.

For example, discrete blood pressure is a previously validated insight for predicting the risk of a heart attack. However, it would potentially be novel for predicting the risk of other conditions. Thus, in certain cases, both the novel insight as well as the novel data acquisition method will undergo a rigorous validation process.

A variety of organizations have been trying to establish solid frameworks to “certify” digital health applications and have issued guidelines which cover clinical validation, product development best practices, safety and privacy topics (e.g., NODE.Health, Digital Therapeutic Alliance).

Some of the key challenges are presented below:

  1. Test and evaluation: After design and creation, the product would need to undergo verification (evidence that specified requirements have been fulfilled) and validation (evidence that the requirements for a specific intended use have been fulfilled) before being ultimately commercialized. Solutions that claim to perform the functions of established medical devices should be able to demonstrate equivalence vs. technical gold standards. However, who can confirm that the gold standard of a traditional medical device is really the right comparison? One issue that might arise, for example, is that the biomarker is established across a large and heterogeneous population.
  2. Addressing biases: It is key to be aware of any potential biases that may exist in how datasets – including those generated by biomarkers – are developed (the same applies to mHealth design and use more broadly). In addition, there is currently human bias in neurology and psychiatry research, for example when using established scales to assess depression, psychosis, and cognition – but this technology offers the possibility to remove or at least reduce this bias.
  3. Validation Testing: A clinical assessment is necessary to show that the data collected can be mapped to a health-related outcome. All validation testing must be performed not only in healthy individuals, but in the target population and relevant environment. A recent study revealed that some existing or emerging technologies might not be clinically ready. For example, measurement which could be captured by various digital health tools (e.g., Fitbit wearables, Apple Watch), are affected by the activity state and showed low accuracy for energy expenditure. Another element to consider is that outside of the clinical setting, biases could be generated when the consumer (or patient) is tracking a specific metric. This could lead to changes in his/her behavior and consequently outcomes.
  4. Integration: Other elements to be considered are security and interoperability. There are different levels of interoperability, but ultimately semantic interoperability, the ability of two or more systems or elements to exchange information and to use the information that has been exchanged, is essential. When designing digital health services, it is important to ensure that data can be shared across clinicians, labs, hospitals, pharmacies, and patients regardless of the application or application manufacturer/vendor. Many digital health services are designed to be integrated into clinical services with already a workflow in place, and thus should enhance rather than burden the existing workflow. Design guidelines in this space are provided by companies such as PCHA (Personal Connected Health Alliance).

 

Other aspects include cost and usability. The entire cost including the technology, lifecycle, and integration in the clinical setting needs to be estimated to make a comprehensive assessment of the benefits. Usability is an important aspect, as these technologies will be used by patients (or consumers) and it is critical to ensure they are designed with a user-centric approach and they are ultimately liked and used by the patient.

 

The Women’s Brain Project considers digital biomarkers as an area with significant potential for mental health – and expect this to be the source of a revolution in terms of research and treatment. That’s why, with the support of Roche, and the participation of expert Petranka Krumova, there will be a panel dedicated to biomarkers at the upcoming International Forum on Women’s Brain and Mental Health, taking place this weekend, on June 8-9 in Zurich, Switzerland.

 

If you are interested, do not miss this opportunity and join us! www.forum-wbp.com

 

[Featured image courtesy of Pixabay.]

Facebooktwitterlinkedin
Recent Posts
Showing 3 comments

Leave a Comment

Contact Us

We're not around right now. But you can send us an email and we'll get back to you, asap.

Not readable? Change text. captcha txt