Virtual Clinical Trials: Selecting Biometric Monitoring Technologies
Clinical trials are research studies conducted to evaluate the safety and efficacy of a medical intervention. Until 2011, clinical trials required that patients visit physicians at clinical research sites to participate. A clinical research site is a physical location, such as a private physician’s office or a research hospital, where a patient learns about the trial and can visit on a regular basis so data can be collected from the patient. The data will then be used to determine the safety and efficacy of the medical intervention.
The site model is both expensive for the Sponsor to conduct and burdensome for clinical trial participants. To operate a site-based clinical trial, Sponsors need to pay and recruit physicians to serve as principal investigators, recruit patients and also monitor the sites. Additionally, a large portion of the population is unable to participate due to transportation barriers. For example, nearly 70% of potential study participants live over two hours away from the nearest clinical trial site.
Reducing Costs and Increasing Efficiency of Clinical Trials
In an effort to reduce costs and enable the trials to be more patient-centric, clinical trials have started transitioning from a site-based, centralized model to a virtual model. A virtual clinical trial is one where participants do not have to travel to a site or interact with trial staff to participate. This shift was made possible by innovative technology, including; biometric monitoring technologies (BioMeTs) that allow for remote data collection from participants, advancements in telehealth and a suite of clinical trial management software tools.
While the first virtual study was conducted by Pfizer in 2011, not all trials across every indication can currently be done virtually. There are numerous challenges associated with transitioning clinical trials from the traditional site model to a virtual one. These challenges include but are not limited to:
- Selecting Biometric Monitoring Technologies (BioMeTs)
- Patient Recruitment
- Managing Informed Consent
- Patient Retention
- Drug Supply Chain
This review focuses on the “Selecting BioMeTs” challenge. Section 1 will overview the challenge itself. Section 2 provides solutions to alleviate this challenge. Section 3 offers investment recommendations considering the current challenges in the virtual clinical trial market.
Section 1: Defining the challenge of selecting BioMeTs
In order for a clinical intervention to be approved to market, the Sponsor needs to validate that the clinical intervention is both safe and effective. This has traditionally been done by collecting data on clinical endpoints from participants at clinical trial sites. Endpoints are direct measures of how a patient feels, functions, or survives that demonstrate a clinical intervention’s safety and efficacy.
When a clinical endpoint cannot be gathered, a surrogate endpoint is used. Surrogate endpoints are endpoints that usually utilize biomarkers (characteristics that can be objectively measured and evaluated as an indicator of normal biologic processes or pharmacologic responses to a therapeutic intervention) to act as substitutes for a clinically meaningful endpoint.
For example, in an oncology trial, increased lifespan may be the clinical endpoint that investigators would like to measure. Because increased lifespan may not be measurable during the time horizon of the clinical trial, they instead use a surrogate endpoint. The surrogate endpoint could potentially be the biomarker of decreased tumor size in trial participants.
To enable trials to move from the centralized model to a fully-virtual model, Sponsors must find ways to collect patient biomarkers and endpoint data both away from the traditional site—and without human interaction. To collect this data in a fully-virtual trial, clinical trial Sponsors are utilizing BioMeTs. BioMeTs are tools such as the Apple Watch or Proteus’ ingested “pill” that uses sensors and algorithms to measure and evaluate a person’s physiological and/or behavioral function. BioMeTs can collect digital biomarkers that can be used as replicates of biomarkers and surrogate endpoints, called digital endpoints.
While BioMeTs allow for remote collection of digital biomarkers and endpoint data, they can be risky for clinical trial Sponsors to implement
For example, a study by Bent et al., (2020) looked at the accuracy of wearable heart rate sensors and found that when participants engaged in physical activity, the data collected from BioMeT devices varied widely from device to device. Additionally, due to BioMeTs being “connected” devices, they are susceptible to security issues.
In a study by Han et al. (2020), it was found that the algorithms that create ECG readings such as those in the Apple Watch are highly susceptible to adversarial attack. To reduce the risks of implementing BioMeTs in clinical trials, Coravos et al., (2020) proposes that BioMeTs undergo a five-step evaluation before being used in a clinical trial.
We’ve gathered an explanation of each evaluation step:
- Verification: The data collected by the BioMeT sensor is compared against predetermined criteria to determine lab-based efficacy.
- Analytical validation: The BioMeT’s sensor is tested outside the lab to ensure it can measure and capture physiological or behavioral metrics.
- Clinical validation: The BioMeT is tested to ensure that it measures its intended physiological or behavioral data in a specific population./li>
- The BioMeT’s security needs to be thoroughly evaluated and a schedule for re-evaluation needs to be determined.
Data Rights and Governance
- Patients and Sponsors need to know who owns the data collected from the BioMeT and how it is shared overtime.
Utility and Usability
- Before BioMeT implementation, it must be clear that engineers know how to analyze the data the BioMeT produces, physicians feel comfortable recommending it to trial participants, and trial members know how and are comfortable using the BioMeT as the protocol indicates.
Sponsors must think about and plan for the upfront hardware and recurring software costs of a BioMeT to ensure data can be collected properly for the duration of the trial.
Currently, the information that would allow Sponsors to run through these five evaluation steps to select BioMeTs for their trials is either unavailable or located in disparate sources.
Section 2: Solutions
Below are the companies working to facilitate implementation of BioMeTs in clinical trials
Scripps: Scripps has created a database of devices that can be used in clinical trials that lets users filter by what the device measures, which users it is ideal for, and the manufacturer of the device.
Clinical Trials Transformation Initiative: CTTI has created a database of feasibility studies showing the trials that BioMeTs have been used in and what they were measuring in order to reduce duplication of costly feasibility studies.
DiMe: The Digital Medicine Society (DiMe) has created a crowdsourced database of digital endpoints in Industry-Sponsored studies. In their database, viewers can see the indications and endpoints collected from specific devices.
Elektra Labs: Elektra Labs is working to implement the five-step evaluation process outlined earlier that would allow BioMeT devices to be compared to one another and properly evaluated before implementation in a clinical trial.
Litmus Health: Litmus Health allows for data collected from wearables to be standardized in ways that make it easy for analysis and regulatory submission. Additionally, they have written a white paper comparing BioMeTs for clinical trial use.
Section 3: Future State
Clinical research lacks the infrastructure to properly evaluate BioMeT solutions for virtual clinical trials. A similar problem is occurring with digital therapeutics and traditional pharmaceuticals in clinical care.
For digital therapeutics, there is currently no infrastructure in place to properly evaluate the digital therapeutics (DTxs) being created. Currently, PBMs Express Scripts and CVS Caremark have created digital formularies, yet provide no insight on the DTx such as who it works best for, how it works (mechanism of action) or how the DTx compares to similar DTxs. As the number of DTxs grow, confusion will remain around which DTx is best for a specific patient unless the proper evaluation framework is built and the right people all have access to that framework.
Additionally, there is currently no infrastructure that allows for traditional pharmaceutical effectiveness to be tracked overtime and compared to other drugs. In order to move to a more value-based drug system, infrastructure is needed to enable real world data and evidence to be used to improve outcomes and more cost-effectively prescribe medications that are best suited for each individual.
Investment opportunity lies within companies that provide the infrastructure to allow all appropriate stakeholders to easily understand and compare BioMeT, DTx and pharmaceuticals.
- National Institutes of Health. https://grants.nih.gov/policy/clinical-trials/definition.htm
- Adams, B. Sanofi launches new virtual trials offering with Science 37. Fierce Biotech. https://www.fiercebiotech.com/cro/sanofi-launches-new-virtual-trials-offering-science-37
- Coravos, A., Goldsack, J. C., Karlin, D. R., Nebeker, C., Perakslis, E., Zimmerman, N., & Erb, M. K. (2019). Digital medicine: a primer on measurement. Digital Biomarkers, 3(2), 31-71.
- FDA-NIH Biomarker Working Group. (2016). BEST (Biomarkers, EndpointS, and other Tools) resource. https://www.biostatsolutions.com/wp-content/uploads/2016/11/Bookshelf_NBK326791.pdf
- Goldsack J, Coravos A, Bakker J, Bent B, Dowling AV, Fitzer-Attas C, Godfrey A, Godino JG, Gujar N, Izmailova E, Manta C, Peterson B, Vandendressche BV, Wood WA, Wang KW, Dunn J Verification, Analytical Validation, and Clinical Validation (V3): The Foundation of Determining Fit-for-Purpose for Biometric Monitoring Technologies (BioMeTs) JMIR Preprints. 29/11/2019:17264
- Bent, B., Goldstein, B. A., Kibbe, W. A., & Dunn, J. P. (2020). Investigating sources of inaccuracy in wearable optical heart rate sensors. npj Digital Medicine, 3(1), 1-9.
- Han, X., Hu, Y., Foschini, L., Chinitz, L., Jankelson, L., & Ranganath, R. (2020). Deep learning models for electrocardiograms are susceptible to adversarial attack. Nature Medicine, 1-4.
- Coravos, A., Doerr, M., Goldsack, J., Manta, C., Shervey, M., Woods, B., & Wood, W. A. (2020). Modernizing and designing evaluation frameworks for connected sensor technologies in medicine. npj Digital Medicine, 3(1), 1-10.