Wearable gadgets have provided a seismic shift in healthcare delivery. These devices have found a base across eyes, ears, wrists, skin, and clothing — to offer therapeutic benefits, or measure progress for better health insights.
In 2020, around 30% of American adults used wearable devices for healthcare, with nearly 50% of these users using their gadgets daily. As our constant companions, wearable devices know us best — they track our habits, heart rates, activity levels, and other key metrics, using self-reporting or via integrated sensors. This data can potentially revolutionize how we study human behavior, and power interventions to promote better health. Recognizing this, around 80% of U.S. adults have reported being open to sharing wearable-derived health data with their providers.
In this guide, we’ll be examining the requirements for mining this data from large populations, and the possibilities this holds for healthcare.
The information aggregated from fitness trackers, smart watches, biosensors, glucose trackers, and other devices is known as big data. In connection to wearables, big data is sourced from wide datasets generated in real-time. The following are different ways wearables and big data can transform healthcare and patient lives.
Big data can play the ultimate trend forecaster in healthcare. Using individual health information aggregated from structures like Electronic Health Records (EHRs), Health Information Exchanges (HIEs), disease registries, health surveys, and wearables — researchers can pinpoint a demographic population’s health needs, or predict the spread of disease.
At the pandemic's peak, Korea was a noted leader in flattening the curve of COVID-19 by adopting smart contact tracing and express testing. In particular, the use of remote monitoring and data collected from IoT and wearable devices proved instrumental in identifying individuals needing medical assistance.
Highlighting the role of wearables in disease detection, specific devices like smartwatches and rings are fitted with audio sensors for cough detection and lung health assessment. Likewise, novel developments like eccrine sweat sensors can measure cytokine markers to determine an infected case. This positions wearables as an effective alarm and detection systems should disease outbreaks occur.
Wearables widen the net of individuals otherwise unavailable for clinical trials and research. These devices are a valuable trove of health information that can be adopted for health research. Wearables tick many features that simplify the information-gathering process. These gadgets are non-invasive, comfortable to the user, and considerably less expensive than standard research instruments.
As wearables continue to improve in quality, so does their recognition across clinical settings. Wearable devices like Apple and Google watches are approved by the Food and Drug Administration (FDA) for their AFib History feature, while Fitbit has FDA clearance for its ECG application.
When researching links between societal influences and sleep — researchers carried out a study of 8000 users using a sleep-tracking app. They looked to uncover social factors, light exposure, the circadian rhythm, and other potential influences on sleep. Using information derived from the app, researchers discovered connections between social pressures and the power to delay the biological need for sleep.
Wearable devices hold value for generating data in large study populations, with the added benefit of observing patients in their natural environments and interactions.
Wearables could be the key to redefining patient interactions in healthcare. This technology promotes remote patient monitoring, minimizing the need for in-person observations in traditional healthcare settings. Through gamified features like leaderboards, personal goals, and fitness rings, users are encouraged to incorporate healthy lifestyles like daily walks, regular sleep, and and regular sleep into their daily lives.
In a survey of 500 users, 86% reported that wearable devices improved their quality of life and health outcomes. By using wearables with EMR/EHR integrations, patients can also self-report their health metrics, or share them with relevant healthcare providers.
Fitness trackers, Virtual Reality (VR) headsets, continuous glucose monitors (CGM), and others generate a continuous stream of valuable patient data per user. Across a large population, this high data volume requires specialized infrastructure and data processing techniques.
The following standard measures are used to analyze big data across care settings:
Not to be confused with the Food and Drug Administration, FDA — Functional Data Analysis — refers to a collection of methods for analyzing data over a continuous curve or functions. In simpler terms, FDA analyzes data that changes over time like temperature readings, or glucose levels throughout the day. These units are termed ‘functional data’. FDA steers clear of considering every reading, instead taking the efficient route of representing hundreds, thousands, or millions of wearable and other health information on a graph or curve.
This approach accurately identifies trends and patterns that exist throughout the data sequence, allowing for some flexibility in comparing different sections of a demographic. For example, using the FDA-derived curve, researchers can identify running trends and compare distances covered between middle-aged men and women of a specific region to track community lifestyles. FDA can also identify the percentage of users that prefer morning to evening runs from this dataset.
By examining the curves of the graph, FDA identifies patterns and links that signify trends or differences in an individual or population's health behaviors. Researchers can predict trends and identify otherwise undetectable changes across a large dataset.
GFPCA tracks data changes over time. Like FDA, this system employs graphs and curves in aggregating health data to identify patterns and trends. However, while FDA focuses on function and time alone, GFPCA goes further by considering other variables.
For example, while FDA is likely to focus on the patterns and trends observed in computing sleep data from user wearable devices, GFPCA would consider other variables such as age, occupation, and level of education to determine patterns and trends and potential impact on sleep quality.
Because of its focus on variables, researchers will likely turn to the GFPCA model to fill in the gap to determine patient behaviors when they aren’t wearing the device, and track health metrics. This approach provides accurate reconstructions by observing external factors that can influence health readings.
PCA finds common meaning in large datasets by breaking them into smaller units called principal components. The unique thing about the smaller units, however, is their capacity to retain most of the information from the larger dataset.
PCA is not used to track direct health data variables like heart rate or blood pressure. Instead, this tool captures how these individual variables are connected to understand potential trends. Each principal component summarizes the most important measurements in the dataset. These components function as Russian dolls, with the largest doll containing the most information about a dataset, and smaller dolls containing smaller, yet vital information about the measurement in view.
A principal component could contain information about the patient's height, weight, or age. Put together, these principal components provide a full view of the demographic data, yet apart, they contain key moments of the data sequence.
Wearable technology is growing into an increasingly influential player in patient and population welfare. On a large scale, these devices provide key insights into community behaviors and disease trends, and require specialized statistical structures to collect information.
On the individual level, users are now encouraged to proactively approach health goals, without relying on healthcare professionals to take the lead. Healthcare professionals can stay ahead of the curve by getting up-to-date information on patient needs, through tools such as Metriport’s Device API, which provides seamless integrations with popular wearable devices, to give a complete picture of a patient’s health profile.
Get the latest updates and blog posts from the Metriport team.
