Secondary Logo

Share this article on:

Wearable Devices in Clinical Trials: The Opportunities and Challenges

Owens, Sarah

doi: 10.1097/01.NT.0000521903.81399.fb
Features
Back to Top | Article Outline

ARTICLE IN BRIEF

Figure. A

Figure. A

Wearable devices replace snapshots of a patients' health with a more detailed, granular, comprehensive assessment of symptom progression. But the data lack validated measures and may contain artifacts, and issues about patient privacy remain, experts told Neurology Today.

Wearable devices that capture data in real time, already a fixture in the personal fitness, gaming, and personal technology industries, are increasingly making their way into clinical trials in neurological diseases. These devices, which can be as tiny, lightweight, and discreet as a sticker, capture a wide variety of data — from gait speed to heart flutter to electrical activity in the brain — in real time, with precision that nears or equals what is achievable in the clinic or lab through traditional measures.

But researchers who spoke with Neurology Today cautioned that while these devices have many advantages, they currently lack validated measures that ensure that therapies tested in clinical trials using wearable devices would be accepted by the US Food and Drug Administration (FDA).

Figure

Figure

Figure

Figure

In interviews with Neurology Today, researchers discussed how they are using wearable devices in clinical trials for multiple sclerosis (MS), stroke, seizure-detection, and Parkinson's disease. Here, they offer insights into the opportunities and challenges that remain from data collection using wearable devices.

Back to Top | Article Outline

WEARABLE DEVICES IN MS

One of the advantages of using wearable devices in clinical trials is they can replace isolated snapshots of patients' health with an ultra-granular, detailed account of symptom progression over time, said Jennifer Graves, MD, assistant professor of neurology at the University of California, San Francisco School of Medicine.

At the AAN Annual Meeting in April, Dr. Graves presented early results of a clinical trial to validate a wearable electromyography device that uses three sensors — an accelerometer, a gyroscope and a surface electromyogram — to provide complementary data about a patient's movement to detect upper and lower limb dysfunction in MS patients.

Surface EMG data were captured during finger taps and foot taps through use of the wearable device. Dr. Graves said the researchers observed that the metrics collected through the wearable EMG device showed clinically relevant associations with other clinical measures. They found, for example, that asymmetry in limb function detected by the device correlated with the traditional Extended Disability Status Scale score (p=0.006) and with pyramidal (p=0.018) and cerebellar (p=0.0045) function. (Analyses for self-reported patient outcomes and machine learning extracted features are in progress.)

In another trial, Jacob J. Sosnoff, PhD, associate professor in the department of kinesiology and community health at the University of Illinois at Urbana-Champaign, is studying a wireless sensor that measures gait changes in patients with MS. The sensors were compared to other validated sensors to capture metrics such as temporal gait parameters, shank angular velocity, and the number of steps in real time, and then transmit them wirelessly to a tablet.

In a validation study published in the February 8, edition of PLoS One, Dr. Sosnoff and colleagues reported that the sensors demonstrated strong accuracy and precision in measuring multiple parameters of gait across diverse walking impairment levels. The sensors were also able to detect differences in gait characteristics by level of disability (p<0.01).

Wearable devices enable researchers to capture data that is reflective of patient's daily life. “We get to look at people in their normal, everyday environment,” said Dr. Sosnoff. “Outside of a neurological exam, patients are doing what they normally do.”

Currently, participants in Dr. Sosnoff's trial are monitored with wearable sensors while they remain in the clinic. But the sensors' battery life “is about 12 hours,” Dr. Sosnoff added. “You can envision someone coming in for a morning appointment, getting the sensor on, going home and spending the day at home, and then coming back at the end of the day.”

Back to Top | Article Outline

SEIZURE DETECTION

Mark Lehmkuhle, PhD, research assistant professor of neurosurgery at the University of Utah, is currently wrapping up a feasibility trial of a small wireless, wearable electroencephalogram (EEG) patch that tracks seizure activity — even when a patient is asleep or unconscious — for up to seven continuous days. (Dr. Lehmkuhle is CEO of Epitel, Inc., which manufactures the device.)

The EEG-containing device is small – 1 inch by 1 inch – self-contained, and sticks to the scalp; it can track both convulsive and non-convulsive activity, Dr. Lehmkuhle said.

The device can capture seizures that may be missed, he told Neurology Today in a telephone interview. Patients may have a seizure while sleeping, he said, “or for temporal lobe epilepsy, a patient may have a seizure and know it at the time, but because the seizure affects short-term memory, they may have forgotten.”

Dr. Lehmkuhle noted, however, that the device has several important limitations. It uses only a single-channel EEG, compared with multi-channel EEG detection offered in traditional clinics. It may also miss some seizure types – for instance, small focal seizures that do not generalize, or deep brain seizures that do not manifest into the cortex, he said.

Dr. Lehmkuhle first presented data on the device at the 2015 American Epilepsy Society annual meeting, and he said as the feasibility study wraps up, safety and efficacy studies will follow.

Ultimately, wearables will help researchers “engage with clinical trial participants over longer periods of time or at greater distances,” said Erika Augustine, MD, assistant professor of neurology and pediatrics at the University of Rochester, who chaired the Technology and Rare Neurological Diseases Symposium at the University of Rochester Medical Center in May, which included panel discussions about the use of wearable devices in clinical trials.

Back to Top | Article Outline

HOW TO VALIDATE THE DATA

Identifying meaning in what may be an overwhelming amount of data is the primary challenge facing the field of wearable disease-monitoring devices, experts agreed. Currently, there are no validated measures for the data captured by most wearable devices, and establishing validation may be a long and rigorous process, researchers interviewed by Neurology Today said.

“How do we separate the signal from the noise?” asked Dr. Augustine.

“I think we're going to be able to measure a tremendous amount of data, but what that data means will mean could be difficult to discern,” Dr. Sosnoff said. “I think we're going to need a lot of different types of data to be able to attach meaning to it. We're going to have to triangulate with other sources of data to gain real insight.”

“Establishing the utility of a wearable device requires rigorous data collection in diseased and healthy individuals,” Dr. Graves said. “Will need to determine the test/re-test reliability of the measurements, and assess the association with meaningful disease outcomes in cross-sectional and longitudinal studies over time.”

In the eyes of the FDA, the data must be high-quality and meaningful to patients and families, said Dr. Augustine. While researchers may interpret data very quantitatively — for instance, they may consider a modestly lower time on a six-minute walk test an improvement — the FDA places great weight on “how a patient feels, functions, or survives.”

“We're collecting massive amounts of data; in the end, it comes down to what's important for an individual patient,” Dr. Augustine said.

That “feels, functions, or survives” standard may be achievable if the granular data captured using wearables are used to create a more holistic picture of a patient's life, Dr. Augustine noted.

The rapid advances in technology may also be an issue from a regulatory perspective, Dr. Augustine noted. “The rating scales and measures tend to develop one at a time for one kind of symptom or one specific disease. It can take quite a lot of time to develop and validate. In technology, there's a much more rapid pace of development and discovery. The regulatory process is still grappling with what that process for approval should look like.”

Validation is crucial for finding meaning in data, and having confidence in those conclusions. “Much like the tools that we use in clinic in a standardized way, that same kind of rigor is needed in clinical trials using wearable devices so we know how valuable, reliable, reproducible the findings are,” Dr. Augustine said.

“We need to demonstrate that the data we are collecting from wearables reflect what we want the device to measure,” said Tanya Simuni, MD, chief of movement disorders and the Arthur C. Nielsen, Jr., Research Professor of Parkinson's Disease and Movement Disorders at Northwestern University Feinberg School of Medicine.

Dr. Simuni is currently recruiting for two large clinical trials funded by the Michael J. Fox Foundation for Parkinson's Disease Research, which aim to use smartwatches to detect tremors, dyskinesias, and other movement changes, and to examine how data captured by wearable devices can impact the routine care of patients with Parkinson's.

For example, she said, when someone is wearing a watch, the algorithm that analyzes the collected data needs to accurately distinguish if someone had a change in their movement pattern because they picked up a cup or whether that was because of their tremor,” Dr. Simuni noted.

“We need to demonstrate in controlled environments that what was recorded by the device and interpreted by its algorithm actually represents tremor,” she explained.

“We may be able to do this by ‘coaching’ the device and the algorithm. So the device is recording the data while an experienced investigator is rating the symptoms. If the device is interpreting the data and saying, ‘This is grade 4 tremor,’ researchers can reference the clinical assessment and see if they correspond. If the tech has misinterpreted the data, then the data analyst can go to the algorithm and revise it,” Dr. Simuni explained.

Dr. Lehmkuhle noted that in his study of a seizure-detection device, “We do see all kinds of artifacts, especially when someone's talking or chewing, but those are very stereotypic artifacts that are easy to pick up and distinguish from seizure activity.” As opposed to spikes, he said, “a seizure is a progressive electrographic chain.”

Data blips shouldn't necessarily be dismissed, however, because they lead to new insights, Dr. Sosnoff noted. “What we think are blips and aberrations can reflect important data. A change in the norm may be indicative of an oncoming relapse” in a patient with MS, for example,” he said.

The surplus of data that can be captured by wearable devices also raises ethical questions and privacy concerns. Dr. Sosnoff said trials involving wearable devices will have to grapple with issues around privacy protections: Will patients be comfortable with researchers livestreaming their health?

Figure

Figure

“We need to find out what level of information patients deem acceptable,” Dr. Augustine said. “Technology allows many kinds of integration, potentially including, for example, between smartphones and the electronic health record. That may not be something patients are comfortable with.”

“Development of these devices will require close follow-up and input from patients to weigh in on that level of integration,” she said. Many of the same ethical questions raised in debates about precision medicine may also apply to wearable devices, experts agreed.

“We may learn new things about you or your family that aren't what we set out to learn, or make conclusions that aren't what the device intended,” Dr. Augustine said. Thoroughly understanding the potential of wearable devices, and defining the limits of data capture, will be crucial, she said.

Back to Top | Article Outline

LINK UP FOR MORE INFORMATION:

•. AES Annual Meeting, 2015 Abstract 2.158: Lehmkuhle M, Elwood M, Wheeler J, et al. Development of a discrete, wearable, EEG device for counting seizures. http://bit.ly/2NT-AES-EEGpatch
    •. AAN Annual Meeting Abstract P1.377: Graves J, Arjona J, Gourraud PA. http://bit.ly/NT-AAN-EMG
      •. The Fox Insight Wearables study: http://www.michaeljfox.org/fox-insight-form.html
        •. Moon Y, McGinnis RS, Seagers K, et al Monitoring gait in multiple sclerosis with novel wearable motion sensors http://journals.plos.org/plosone/article?id=10.1371/journal.pone.0171346. PLoS One 2017; Epub 2017 Feb 28.
          © 2017 American Academy of Neurology