3 Tips For Designing At-Home Monitoring Devices That Deliver Actionable Information
By Anna (Ronning) Cohen and Hilde Viroux, PA Consulting
We’re all seeing it in real time: Healthcare is increasingly shifting from clinical settings to our homes. The COVID-19 pandemic has accelerated the development of at-home monitoring products and many are collecting data in the background without requiring specific user action. This acceleration is not likely to slow down post-pandemic. In this article, we define “passive at-home monitoring tools” as non-invasive products that continuously monitor elements of health and collect the user’s data without the direct supervision of a healthcare professional. Take, for example, smart devices such as the Apple Watch or a biosensor patch — neither requires operation from the user to collect data, so we consider them to be passive at-home monitoring tools.
Generally, diagnostic tools can provide insights that are either informative or actionable to the user. A good example of an actionable product is a blood sugar monitor. It notifies a patient when their blood sugar level is either high or low so that the patient can add insulin or eat a glucose-rich snack to compensate. Compare this with an informative diagnostic tool such as an at-home blood pressure cuff. The user may know their blood pressure is running higher than normal ranges but cannot take immediate action to lower it. Which of these product types is more valuable to the patient?
The large amount of data collected from passive at-home monitoring products could be leveraged in a proactive capacity to assess and manage more serious health conditions. For example, the Apple Watch collects ECG data in order to recognize atrial fibrillation, a serious heart arrhythmia. Most at-home passive monitoring products, such as wearables, provide bits of informative data with little context rather than actionable insights. Currently, tools that measure discrete elements are low-risk devices for patients because they do not provide actionable information for the user to make improvements to their health.
From a regulatory perspective, there is a very different approach between the U.S. and the EU. The FDA does not consider software used “for maintaining or encouraging a healthy lifestyle and [that] is unrelated to the diagnosis, cure, mitigation, prevention, or treatment of a disease or condition” a medical device. In addition, for apps that fall within the definition of a medical device, FDA states in its Policy for Device Software Functions and Mobile Medical Applications that it plans to apply oversight authority only to those software applications whose functionality could pose a risk to a patient’s safety if the software applications were to not function as intended.
In the EU, all software apps meeting the definition of a medical device must follow the approval route of regular medical devices. The Apple Watch mentioned above will be classified as a medium-low risk device (class lla) as it provides information used to make decisions with diagnosis or therapeutic purposes.
This difference in regulatory pathway should not stop manufacturers from designing and developing at-home monitoring devices because the data that they generate can give a wealth of information on patient behavior, disease evolution, and disease prevention.
1. Target Biomarkers That Underlie Disease
First, the design should focus on understanding causes of disease and how these progress in various patient archetypes. Patient archetypes can be categorized in a several different ways depending on the product, including by medical complexity, underlying conditions, health status, social or behavioral patterns, and/or diagnosis to allow for personalized or precision medicine, just to name a few. For example, the Apple Watch scenario could focus on patients with underlying heart conditions.
The National Institutes of Health (NIH) defines precision medicine as the “approach for disease treatment and prevention that takes into account individual variability in genes, environment, and lifestyle for each person.” The NIH places emphasis on individualized prevention, diagnosis, and treatment for each person based on their health history, genetics, lifestyle, etc. There is an underlying assumption that personalized medicine will provide healthcare to patients that yield a decreased rate of adverse outcomes.
Digital biomarkers are user-generated physiological and/or behavioral measures collected from digital health products and are used to explain or influence health outcomes. Longitudinal data gathered via digital biomarkers offer invaluable information to understand and manage disease prevention or disease management. As healthcare and medicine continue to shift to the digital space, digital biomarkers are an opportunity to collect physiological measures such as sleep-related activity, heart rate, blood pressure, or electrodermal activity. Behavioral measures are intimately related to physiological measures, which include changes in physical activity or behavior. For example, body mass index is a physiological biomarker used to assess the risk for metabolic diseases such as diabetes and hypertension.
2. Translate Biomarkers Into Actionable Insights
Secondly, you should consider leveraging these digital biomarkers to understand what metrics your device should be capturing in order to create actionable insights for patients. Because these tools gather months’ and even years’ worth of continuous data from the user, understanding the user’s baseline biomarkers is vital to creating actionable insights and developing preventative health measures or interventions. Further, passive at-home monitoring tools continuously monitor the patient and collect longitudinal data, which offers invaluable information to better understand, detect, and manage diseases. For example, blood pressure measurements can be indicative of cardiovascular risk. This biomarker could also be used to gather additional insights by linking blood pressure to depression.
As consumer-facing health products become more accessible and supported by physicians, the products collect and analyze even larger amounts of data from diverse patient populations. Scientifically proven correlations between data and health outcomes can be leveraged to alert users to take precautionary steps or to contact a healthcare professional for high-risk conditions. Our earlier Apple Watch example is an example: It has an irregular rhythm notification feature identifying an irregular rhythm suggestive of atrial fibrillation and the feature confirms the data with multiple readings.
Biomarkers and various data inputs enable organizations to demonstrate real-world value and present the opportunity for more accurate and timely interventions at the onset of symptoms. With such a diverse patient population, generic recommendations will likely not work for all patients.
3. Use AI For Ongoing Monitoring And Patient-Specific Recommendations
Lastly, by combining various types of data elements from physiological and behavioral factors, recommendations can be patient-specific by understanding how their behavioral inputs effect their physiological response. To do this, you may leverage artificial intelligence (AI) to monitor longer-term trends in wellness monitoring, allowing for highly personalized recommendations for improved health outcomes.
For example, when a patient is diagnosed with congestive heart failure (CHF), minimal weight gain (~2–3 pounds) can indicate critical changes in fluid retention, which may be an initial sign of heart failure. A remote monitoring scale that leverages AI and Internet of Things can log and assess changes in a patient’s weight. This data can detect the small variance in weight and automatically notify the on-call physician, therefore preventing hospitalization or worse.
In conclusion, if you are creating at-home monitoring products and tools, you’ll need to find novel ways to utilize data and create actionable insights that enable patients to change their behaviors. In turn, this will lead to better health outcomes. With the right technology and processes in place, these products can provide information to patients that enable more effective programs and optimized care. Leveraging patient data is no small feat, but it’s the future of healthcare.
About The Authors
Anna (Ronning) Cohen is a life sciences expert at PA Consulting with a background in data analytics, business development, and medical devices. Her experience ranges from supporting biopharma clients with end-to-end product development to using analytics to solve challenges related to market access, pricing, promotional effectiveness, and commercialization of pipeline products. Combining her strong analytical and project management background, Anna implements solutions to challenges that are effective and valuable to all stakeholders and end users. Connect with her on LinkedIn.
Hilde Viroux is a medtech expert at PA Consulting and a leading expert on the European Medical Devices Regulation. She is a senior leader with a broad experience in regulations, quality, manufacturing, supply chain, and project management in the pharmaceutical and medical device industry. She has an outstanding track record on successful implementation of major projects and building up new capabilities within an organization. Hilde has an MSc in medical technology regulatory affairs from Cranfield University in the UK, and a BS in biochemistry engineering. Connect with her on LinkedIn.