Ways Wearable Data Can Become Incorporated Into Your EHR

Every year at national medical conferences, there are a few presentations on wearable devices. The typical presentation goes like this: there are so many new wearable devices, new apps, new security concerns, yet nobody is monitoring or regulating these so there is a massive void of guidance for patients. Each year more wearables and more apps are developed, there are more urgent calls for standardization of data privacy and evaluation strategies. But even if the most robust evaluation database existed, most patients probably use a simple heuristic when deciding whether to engage: did someone recommend it? Is the brand well-known and trusted? Is it easy to use? A simple mental shortcut is what most rely on because the number of variables and features make it nearly impossible to compare.  

There are a few main categories for wearables currently on the market, which collect and analyze data pertaining to: physical activity, sleep, physiologic measurements, and subjective parameters.
 
Physical activity is notoriously diminished in patients who experience depression. Exercise can feel impossibly difficult during an episode of depression, and psychotherapeutic techniques can employ motivational interviewing strategies to encourage movement. Utilizing the actual data in the session could be more helpful to understand why certain goals were met or not met on any particular day. It could provide additional context about how inactive a patient is throughout the weeks since the last appointment, since the appointment only provides a very small snapshot of the patient’s well-being. Measuring physical activity is very accurate with wrist watches now and monitors movement even when the individual is not exercising.
 
Sleep is affected in many mental illnesses, including depression, anxiety, and bipolar disorder, and is therefore an important parameter to measure when tracking symptom changes. The data is mixed in support of actigraphic measurement as a proxy for sleep quality when compared to polysomnography. Furthermore, subjective measures of sleep quality are not always consistent with actigraphic measurements. For example, a user may experience sound sleep, but an app may alert to a poor night sleep, or vice versa. In a clinical setting, there may be more time spent trying to interpret the accuracy of the data, rather than using the data for actionable recommendations to enact change.

Physiologic parameters that are accurately measured by wearables include skin temperature, skin conductance, and heart rate variability. Skin temperature fluctuates on a circadian cycle and can be easily impacted by ambient temperatures. Skin conductance and heart rate variability have been found by multiple studies to be reflective of autonomic system functioning and can correlate to levels of depression. These data can then be compared to other measures such as rating scales of depression to help identify episodes of depression early and in real-time. Most users would welcome this alternative, rather than complete a more time-consuming rating scale. However, if the user experiences too many “false positives” where the app alerts to symptoms that don’t register with them, they may start to reduce engagement with the app.  
 
Subjective parameters would allow patient-reported outcome measures to be gathered and sent to the clinician prior to the appointment. Most evidence-based rating scales consist of 10 or more questions which can become a repetitive experience in an acute situation and the user is prompted too frequently. There are many apps that have a simple mood question on a Likert scale or even an emoji. The advantage is that real-time changes can be tracked and monitored more accurately, reduce recall bias, and give the clinician a better understanding of mood fluctuations in between appointments.

There is no shortage of potential new applications of wearables: speech analysis, gait variability analysis, and motor abnormalities. Because there is no limit to the amount of data that can be collected, the challenge in utilizing wearables is that data must be meaningful to both the user and the clinician. For data to be meaningful, it needs to measure key variables, be analyzed accurately, and be presented in a way that a human can easily interpret. All these components need to be met adequately for the user to engage with the app consistently. Once the app is seen as valuable and useful, then there will be continued long-term engagement by the user, and it will become the recommendation of choice by the clinician to other patients.