Wearable sensors provide sophisticated insights into patients’ real-world behavior and functioning in clinical trials and healthcare settings. Drug development researchers, regulators, payors, and patients want to see meaningful, valid insights enabled by them. In this guide, we break down the following commonly used wearable sensors: including what they do, their benefits, and other important considerations for incorporating them into your clinical trial:
- Accelerometers
- Pulse-Oximeters (SpO2)
- Electrocardiogram (ECG)
- Garment-Based Sensor Platforms
- Continuous Glucose Monitors (CGM)
- Electroencephalogram (EEG)
- Multi-Sensor Integration is Essential
- Insights Enabled by Wearable Sensor Data
Accelerometers
Accelerometers are simple, tiny, conductive sandwiches measuring g-force, and are often built into other sensors precisely because they are so economical and reliable. Even when used independently, these sensors can capture useful data about functioning in terms of activity and sleep. Accelerometers have the resolution to inform us about changes in gait in neurodegenerative disorders or changes in posture associated with disease progression or therapy.
Wearing this simplest of sensors in a wrist band without a connectivity requirement, however, is still not a small ask of the patient. Failure points can occur even when the instruction is just “not to take the sensor off,” except to transfer data. Despite some limitations, however, 180 days of almost continuous accelerometer data is routinely achievable for large patient populations with very low fail rates.
Regulatory authorities accept well-collected accelerometry data as an accurate measurement of patient activity and sleep. With accelerometers being so inexpensive and yet reliably deployable, it is arguable that all clinical trials should include these robust measures of overall patient wake activity and sleep quality.
Accelerometers are valuable but can only provide a certain level of data granularity. Here are some common options for wearables that can collect more detailed physiologic data, such as cardiovascular information.
Pulse-Oximeters (SpO2)
Pulse-oximeters (SpO2’s) are some of the most common sensors in the world and represent a multi-billion-dollar business. In addition to capturing blood oxygen saturation, these ubiquitous sensors can provide PPG data from which we can extract:
• Heart rate
• Respiratory rates
• RR intervals for Heart Rate Variability (HRV)
• Complex measures of sleep, wake, performance, health, and mood
Although pulse-oximeters are inexpensive and familiar, they may be uncomfortable to deploy for extended periods. No one particularly enjoys the sensation of wearing an oximeter on a fingertip overnight or during the day. Several next-generation pulse-ox sensors hold out the possibility of alternate form-factors allowing for greater comfort.
Electrocardiogram (ECG)
One or 2-lead ECG adhesive patches provide a low-cost diagnosis of a variety of irregular heart rhythms, including atrial fibrillation (AFib). Discrete and comfortable, these sensors may be worn for up to 15 days, detecting and recording high-quality ECG signals and monitoring patient cardiac symptoms. These patches weigh just a few ounces and maximize patient compliance by requiring them to do absolutely nothing – except ignore their presence.
By recording ECG at between 250-1000Hz, often with an accelerometer included as part of the package, these sensors can give an unprecedented level of physiologic detail of a patient’s life while both awake and asleep. During a typical wear-period, ECG patches will capture hundreds of millions of data points, more than any other sensor type. This data provides us with a highly granular picture of a patient’s overall functioning for extended periods, with arguably the lowest compliance burden possible.
Due to their ease of use and the detail of the data recorded, ECG patches from various manufacturers have now been reliably deployed on millions of patients worldwide.
Garment-Based Sensor Platforms
Combining sensors into various types of clothing, often as shirts or chest straps, is an intuitive way to place sensors on-body. Garments allow the integration of sensors such as ECG, multiple respiratory sensors, and accelerometers into a single comfortable, unobtrusive package, which can be as easy to deploy on a patient as it is to remove. Several very capable, well supported, sensor-in-garment packages are commercially available today. This type of form factor has been refreshed several times. It is likely to remain in use, particularly in sports, military, hazmat, and first responder applications, where a higher user burden is less of an issue.
It is less clear that garment-sensor solutions are suitable for truly scalable deployment in clinical trial and healthcare settings. Wearing an additional garment may not be appealing to everyone, especially in warmer climates. Issues of fitting garments accurately enough to ensure exact positioning of the senor(s) can also be challenging. Humans come in an astonishing variety of shapes, sizes, and proportions, and pediatric and elderly patients may find putting on or taking off a garment just too much to comply with on any regular basis.
Continuous Glucose Monitors (CGM)
The highly granular data from pulse-ox and ECG sensors may be further enhanced using other classes of sensors, such as the new and exciting Continuous Glucose Monitors (CGM’s). They measure glucose levels and rate of change, providing insight into the impact meals, exercise, and illness on the metabolism of blood sugars. Rapid CGM adoption is improving diabetes management by minimizing the uncertainty in decisions based on a single blood glucose meter reading.
Electroencephalogram (EEG)
EEG’s are used to determine changes in electrical activity in the brain and represent an amazingly complex sub-set of wearable sensors and scientific research. They are useful in diagnosing brain disorders like seizures, stroke, and sleep dysregulation and available in a variety of electrodes with multiple configurations from one to more than 20. EEG synchronization and analysis is possible with tools such as VivoSense® Wearable Sensor Data Visualization and Analytics software. However, because the discipline of brain science and treatments is mostly a specialty domain, data analysis should be undertaken only with knowledgeable collaborators.
Multi-Sensor Integration is Essential
As more combinations of wearable sensors are deployed, the ability to integrate sensor data becomes critically important. Precise and visual alignment of data from different data streams helps researchers contextualize physiological changes and guide better patient outcomes.
However, without the right plan, tools, or processes, collecting and analyzing data from multiple sensors can be challenging to navigate. One common problem is inaccurate internal clocks in two or more sensors. The sensors collect synchronized data, but the timing is off. VivoSense® Software has a Synchronization and Merge module that resolves these types of sensor integration problems by streamlining the process of aligning files from multiple simultaneous sensor recordings. Data are synchronized using annotations, a simple point and click operation, or quick and accurate analysis to find the best match. The software hosts other specialized modules for advanced analytics used in the interpretation of ECG/HRV, complex respiratory, pulse-oximetry, continuous glucose monitoring, movement, gait, and more.
Insights Enabled by Wearable Sensor Data
The world of wearable sensors is growing rapidly as these technologies provide more sophisticated insights into the real-world behavior and functioning of patients in clinical trials and healthcare settings. Understanding the variety of wearable sensors and their capabilities is essential to operationalizing them successfully. Ultimately, researchers are tasked with making the best decisions they can about incorporating wearable sensors into their studies to enhance our understanding of treatment efficacy and the extent to which drugs and interventions improve patient’s lives.For more information about how to approach sensor selection, check out Getting Started Using Wearable Sensors for Clinical Research. Or, of course, contact us anytime to schedule a consultation with one of our experts.