The use of health apps has increased considerably in recent years. In 2018, there were 62.7 million monthly health app users in the United States. Just three years later, in 2021, that number jumped to 86.3 million monthly users. Likewise, the number of consumers willing to wear smart fitness technology has more than tripled in the last four years, from 9 percent to 33 percent of U.S. consumers.
Lately, there has been more talk about how these technologies are expected to play a greater role in the healthcare industry as a method of screening for potential diseases. Health practitioners have access to a constant stream of data about their patient’s health, including blood pressure, heart rate, daily physical activity, and even blood oxygen levels. In theory, we’d expect that level of information to help practitioners treat and even prevent illness and disease in their patients.
Yet, there are some naysayers who suggest this kind of tech isn’t all it promises to be.
Why some think data isn’t enough
The primary concern is that this influx of data isn’t helping health practitioners diagnose their patients correctly. Many doctors still have questions about the reliability and the accuracy of the information that health apps and wearables provide.
One recent Forrester report claims that these health wearables and tech don’t actually help clinicians due to the design of these devices. Because these technologies are typically created with consumers in mind, their primary use cases are to be user-friendly and marketable, not industrial-standard technology that clinicians can rely on. “These products neither promise nor deliver health care,” the report says. “They offer data, not answers to consumer questions, diagnoses, or treatment suggestions to help restore one’s health.”
Another common concern is that all the data in the world isn’t enough to make a reliable diagnosis. Clinicians need to speak to the patient, not just look at a collection of statistics to make a potentially life-changing decision on their patients’ health. “I didn’t get into health care to look at data,” one physician told Forrester researchers. “My passion is helping people.”
And some current scientific consensus is not helpful. One meta-study that looked at the effectiveness of wearables suggests that while wearables can help motivate physical activity, current data doesn’t demonstrate any other health benefits.
But there is a light at the end of the tunnel, and it is NOT a train.
So how can health apps and wearables help?
Data by itself cannot diagnose a patient. Nor can a doctor look at the patient and provide a diagnosis without the need for testing to confirm their anticipated findings.
But when that data is processed using scientifically proven machine learning algorithms (just like a confirmatory test) the resulting information can be incredibly valuable and accurate.
In a study of 14,011 participants using 57,675 person-weeks of data, it was determined that an AI model could be trained to accurately pre-diagnose four conditions: high cholesterol, hypertension, sleep apnea, and diabetes. The results were between 77-83% with very few false-positives and have been confirmed with data from a Top 5 Health Insurer.
Additionally, specific, one-use wearables with high acuity are very accurate but are not widely available to consumers. And most people do not want to wear 5 devices to accomplish 5 results. Hence, the adoption of multi-use devices that accomplish many health monitoring functions and even aid in condition pre-diagnosis when combine with software.
Most wearables and smartwatches use optical heart rate monitors (OHRMs) or sensors to provide the user with their heart rate can also be used to provide sleep information to users. Photoplethysmography (PPG) is the process by which light shone onto the skin is reflected and absorbed by blood cells, these lights measure volumetric changes of blood below the epidermis and estimate heart rate based on the fluctuations. While OHRMs may not seem like a reliable measurement, in certain devices they are surprisingly accurate.
A study done by the National Cheng Kung University in Taiwan found that the Garmin Vivosmart HR+ “exhibited acceptable overall accuracy” as compared to the “baseline” HR metrics obtained from a chest strap. The study was done using both a young (20-26) and senior (65+) population and found that the Garmin OHRM was more accurate.
To be effective, health tech must be accessible, accurate, and specific. Let’s break each of those criteria down.
To be useful, health apps have to be at least somewhat accessible. For instance, at ilumivu, we develop a health app that works with any standard smartwatch. Smartwatches aren’t free, of course, but it’s possible to buy a decent-quality one for $100. Devices like this are growing in popularity, so we know our app could reach a wider range of patients.
Health tech also needs to be accurate in order to engender trust with clinicians. We know that their heart rate capture approaches 100% accurate in Apple and Samsung watches, and 94% accurate in Fitbit watches. But sleep data, by contrast, is significantly less reliable. Health apps and devices should only use data they can count on.
And finally, health tech needs to be specific. For example, there are many different health conditions that wearable technology could focus on, but our app focuses specifically on conditions such as diabetes, hypertension, atrial fibrillation, and sleep apnea. What do all of these conditions have in common? They all affect the cardiovascular system, so there’s a very discrete focus to what our app is measuring.
Most apps try to be all things to all people and attempt to track too many different variables. This contributes to physicians’ overall distrust and reluctance to rely on wearables and apps to help make medical diagnoses. But the more focused these technologies can be, the greater help they can be for clinicians and patients alike.
Work together to confirm diagnosis
Consumer health apps and wearable devices have the potential to contribute significantly to medical diagnoses. And clinicians can increase the confirmatory tests in their bag by leveraging proven technologies to help educate them and their patients and point them to potential problems well in advance.
But the data needs to be consolidated and presented in a way that helps the clinician review and use the data correctly.
Clinicians and health professionals know that they hold their patients’ lives in their hands. That’s what makes health tech so appealing, and yet so cumbersome. Doctors and nurses are used to using their intuition and experience to detect health conditions, and they may be reluctant to outsource any decisions to unproven technologies. But being able to rely on longitudinal data in patients to help detect conditions like diabetes and hypertension early can literally be a lifesaver.
While many apps and wearables are nothing more than glorified habit trackers, clinicians can benefit from additional data to directionally aid and reduce the time to a confirmed diagnosis.
David Smith, President, and CRO at ilumivu, the company that provides healthcare decision support applications using the psychology of behavior change, combined with real-time data from smartphones and wearables.
