Digital Health in Clinical Trials

Posted June 9th, 2016 by Adam Rosenberg, in From The Trenches, New business models, Translational research

This blog was written by Adam Rosenberg, CEO of Rodin Therapeutics, as part of the From The Trenches feature of LifeSciVC.

As a controlled experiment, try placing a tape recorder, rotary phone or record album in front of an 8 year-old.  Then watch their futile efforts to identify technology that was commonplace even 25 years ago.

“Oh, I know this.  It’s a typewriter!  My grandfather has one of these.”

This was recently done on network television…itself at risk of becoming an anachronism in the age of Netflix and YouTube.

The rapid and now total domination of our lives by wireless devices needs no explanation, at least in this forum (where many of you are likely reading this on your mobile phone/tablet).  This mobile takeover will eventually permeate all areas of healthcare, and yes, may even help power more efficient drug development.

Silicon Valley Bank recently described Digital Health as “solutions that use digital technology to improve patients’ health outcomes and/or reduce the cost of healthcare.” Plenty of health apps, software, mobile devices and other digital technologies neither lower costs nor improve outcomes…so it is challenging to define a sector by its goals.  But broadly speaking, the definition is good enough in my book.

For “digital natives” like the 8 year-old mentioned above, it is clear that wireless technology and its progeny will reshape the future of healthcare.  The question is “when?” In many use cases, the answer is likely many years off.  But for some applications, the answer is now.

In this blog, I will focus on digital health utilization for clinical trial management, primarily in disease areas where activity and behavioral monitoring can be accomplished with reasonable accuracy and reliability, and where such data can help support whether a drug candidate produces a clinically meaningful effect.

Clinical Trial Optimization

CROs and clinical operations teams face a daunting task to accurately measure compliance, safety and efficacy across a clinical population over a defined period, balanced against a need to control costs.  Technology exists today to remotely measure some drug effects; digital health efficiencies should therefore appeal to trial sponsors and managers, at least in theory.

One needs to be realistic about the goals of utilizing digital technology, however, and to pick a thoughtful entry point.

Disease Areas

In CNS, for example, technologies that purport to measure more subjective endpoints – like mood, anxiety, or depression – will need considerable clinical data to support their utility and alignment with more traditional tracking methods.

In contrast, neurological disorders characterized by motor impairment are well-suited to the use of activity monitors.  Relatively accurate and sensitive monitoring devices can help reduce visits to the clinician, a significant cost factor and also challenge for many patients, especially in advanced stages of disease.  These devices can also capture movement at all hours, unlike clinician visits which only measure function during the appointment.

Device Accuracy

Note the above qualifier – “relatively” accurate.  Many consumer-focused activity trackers have come under attack for lack of absolute accuracy.  For the typical healthy consumer, it’s not crucial to determine whether you walked exactly 5,001 steps or 4,999.  But for clinical trial measurements and medical diagnostic devices, accuracy and reliability must be prioritized.

Peer-reviewed and data-supported examples are encouraging, one example being the multimodal work recently published by a UCSD group in Nature Communications.  See Imami, S. et al. A wearable chemical–electrophysiological hybrid biosensing system for real-time health and fitness monitoring. Nat. Commun. 7:11650 doi:1.1038/ncomms1165 (2016).

This is crucial: while time-consuming and costly, for trial sponsors, regulators, patients and physicians to trust the accuracy and utility of digital technology, device innovators must undertake these rigorous validation steps.

Clinical Adoption

As for data around the utilization of digital health in clinical trials, press releases are easy to find, but uptake is harder to measure. One encouraging sign is that many newer initiatives are being undertaken by large companies and research institutes with a leadership position.  At least anecdotally, this contrasts with prior efforts led by smaller startups without much brand recognition.

Startups often push the pace on innovation.  Buy-in from market leaders is critical, however – if at least some of these initiatives gain traction and result in demonstrable and sustainable improvement in clinical trial management, there should ultimately be a positive downstream effect on adoption.

Recent examples 


Given the ubiquity of iPhones, Apple is in pole position to play a key role in leveraging wireless technologies for clinical applications.

In 2015, Apple announced that leading academic medical centers would use its ResearchKit suite of tools to collect data for new clinical studies, and is currently recruiting for numerous trials (here).

In March 2016, Apple also launched CareKit, a healthcare tracking system that allows individuals to track drug adherence and motion, and to share that information with doctors or family members.  Apple also recently launched a number of CareKit apps, one of which – the OneDrop app – tracks glucose, food, insulin and activity in diabetics.

Pfizer / IBM collaboration

Pfizer and IBM recently announced a remote monitoring collaboration focused on Parkinson’s disease.  As mentioned above, motor symptoms may be highly tractable for digital health measurement. Pfizer/IBM are seeking to obtain better, more complete movement data to better inform trial design and, ultimately, medical management decisions (here).


A Cedars-Sinai team is also exploring a new tracker to monitor Parkinson’s disease patients and quantify drug compliance and movement over a sustained period.  Notably, Cedars-Sinai is evaluating the wristwatch-sized technology through a clinical trial format, starting with a 60-patient study initiated in 2014 (here).

ALS-TDI / Denali

As noted above, engineering efforts by market leaders such as Apple, IBM and Pfizer, along with rigorous peer-reviewed clinical data-sets, will help lead digital health into greater mainstream acceptance and adoption.

If realized, digital health-driven efficiencies will help everyone.  But for smaller organizations in particular, access to validated translational tools could be a game-changer by enabling early clinical proof of concept in a timeframe that would be prohibitive under costly, traditional clinical trial protocols.  This may be particularly true in neuroscience.

One example is the ALS Therapy Development Institute (ALS-TDI), which has directed donations from the Ice Bucket Challenge towards validating accelerometer data to track disease changes.  As with Parkinson’s, ALS is characterized by diminished motor function; movement sensors could thus provide valuable data related to disease progression and drug effectiveness.

ALS-TDI’s clinical study has already enrolled almost 300 patients.  This is impressive given the relatively rare incidence of ALS, but also a strong indicator of the need for new tools for clinical trial measurement.

ALS-TDI is also collaborating with venture-backed Denali Therapeutics to establish the utilization of wearable technologies to measure drug effectiveness. Patients, providers, start-ups like Denali, and foundations like ALS-TDI all benefit if acceptance of these devices can lead to smaller, faster trials with more complete and accurate datasets.

Conclusions and Next Steps

There are numerous other examples.  In combination, these clinician-focused and patient-focused initiatives should help chronic disease management and spur adoption in clinical trials.  This will hopefully contrast with earlier attempts at leveraging digital health, which often focused on software built around single modality hardware, requiring a change in patient behavior and multiple, unfamiliar devices (each with its own design issues and utilization hurdles).

Yet exciting press releases do not necessarily drive broader adoption and buy-in from sponsors, physicians, patients, and regulators.  And to be clear, these technologies should not initially be viewed as a substitute for sound clinical judgment, or for validated disease biomarkers.  Nor are they likely to accepted immediately by regulators as a primary endpoint data source for new drug approval.

Rather, these new technologies should be viewed as incremental tools to measure clinical progression and drug responses, especially in areas such as CNS where placebo effects and often subjective endpoints complicate clinical development.  As validation and comfort with these tools increase in particular settings, they may very well become more central to managing patients and evaluating new therapies.

As noted above, it is preferable to start with reasonably objective endpoints.   Remotely measuring movement in neurological disorders or validated biomarkers such as glucose levels and weight in metabolic disorders are most likely to generate reliable data and demonstrate robust cross-correlation of device accuracy.

Yes, caution is warranted, healthy skepticism is valid, and baby steps are required.  But in 20-30 years, today’s 8 year-olds who open a tape recorder and wonder if it can toast a waffle will have grown up and will be developing drugs and running clinical trials.  And just as they struggle to figure out our prehistoric, pre-digital technologies, they will struggle to justify spending billions of dollars on drug development without the quantification and 24/7 monitoring tools available even today.

So consider digital health approaches for clinical trial design, along with a sensible biomarker strategy, and other technologies such as EEG (see Vanessa King’s recent blog).

Even in its infancy, digital health is already enabling drug developers to better manage clinical monitoring costs, and ideally to make more informed decisions regarding clinical prioritization.  Hopefully it will also soon lead to more and better approved therapies.


This entry was posted in From The Trenches, New business models, Translational research. Bookmark the permalink.