Chronic diseases are lifelong ailments that can’t be cured but only managed. Treatments can be complex and inconvenient — and may have significant side effects. Furthermore, poor self-management, particularly among new patients, can result in faster disease progression and preventable acute episodes. With continuous monitoring enabled by connected devices and an engaged and informed clinician, chronic diseases can be managed more effectively, often adding years of functional life.
Continuous biometric monitoring provides data that supports two key interventions. The first is to proactively identify patients who are likely to have trouble with adherence. The second is to customize effective interventions early on, before poor adherence becomes a long-term and intractable problem.
Predictive modeling based on known patterns of behavior, as indicated by both provider interactions and the data from connected devices, identifies patients who might require directed coaching from a specialist in disease-specific education.
Patient adherence to prescribed therapy may be low because of the lack of usable feedback on the effectiveness of health interventions for patients and clinicians. Integrating information from biometric sensors with physician analysis may provide patients with clearer, fresher insight on how treatment is affecting their condition. Patients can visualize what is improving — and what behaviors might make it worse. Self-management is easier when results can be tied to actions.
The ideal form of monitoring of elderly patients is constant and unobtrusive, not requiring deliberate self-testing on the part of the patient, and instantly detecting the first signs of developing conditions. The most effective signs of incipient health events are often not from implants or other biometric monitors, but from changes in activities of daily living (ADL), such as movement, toileting and sleep.
Passive infrared motion detectors, pressure sensors in beds and chairs, sensors for CO2 concentration, sound (vibration) and video — anonymized as necessary for privacy — can all be used to first establish a baseline of normal variability, and then be applied to detect significant deviations from that baseline. This continuous and nearly invisible sensing can be surprisingly effective in assisting in care.
For example, the onset of congestive heart failure can be signaled by reduced use of the bed, as patients having trouble breathing at night switch to sleeping semi-upright in a recliner, while changes in toilet flushes can detect a urinary tract infection or incipient dehydration.
One in four Americans over 65 falls each year, but only half tell their doctor. Motion sensors can not only detect and alert caregivers immediately of a fall, but with proper algorithms they can also detect unsteadiness, changes in gait and balance and other issues early, enabling early intervention to prevent falls. Twenty percent of falls in the elderly result in a disruptive injury, often the first step to significant health decline.
Proactively sensing, diagnosing health problems
While sensors are becoming more pervasive, sensitive and precise, data analytics are extracting much more information from these signals, combining them, and enabling early detection of health problems.
Handheld, wearable, or patch ECGs detect arrhythmia, myocardial ischemia and some negative drug interactions. Increasingly, they can non-invasively and continuously determine blood pressure as well. Connected biometric devices in concert with deep learning are particularly useful in analyzing ECG data, which is complex and time-consuming to analyze properly, particularly if the intent is to detect signs of an approaching acute episode. Companies like Biotricity, with its beat and arrhythmia-detecting cellular IoT sensors, and Cardiolog, with its device-agnostic deep-learning ECG algorithms, are among the innovators in this area.
For orthopedics, the problem has been how to design smart joint implants containing sensors for pressure, force, strain, displacement and other parameters without compromising joint functionality. Smaller and more robust sensors are finally providing the ability to see into an implanted joint over time, providing invaluable data for early intervention before problems get serious and to individualize physical therapy regimens. Sensors and the data they collect are also assisting in developing future generations of improved joints and improved surgical techniques. Hip, knee and spine interventions will see significant evolution as a result.
According to the VA, the prevalence of Type 2 diabetes is more than twice as common among veterans than the national average, leading to a large rate of diabetic foot ulcers that can result in sepsis and limb loss. A veteran can step for 20 seconds on a foot mat from Podimetrics that measures foot temperature. The specific and longitudinal data is evaluated using an algorithm that can detect growing inflammation an average of five weeks before it evolves into an ulcer, enabling a proactive intervention.
Other examples include soft contact lenses containing strain gauges that measure intraocular pressure to manage glaucoma (SENSIMED Triggerfish), seizure-preventing electrical pulse generators (Neuropace), and closed-loop systems for diabetics that detect blood sugar levels to aid in self-management and support regulated doses of insulin (Medtronic SmartGuard).
Home health in an institutional context
Evolving systems that optimally share risk and reward among the complex network of stakeholders, comprising the modern healthcare system, is as important to ensure the best therapeutic outcomes of home health as any blood pressure detector or insulin pump.
And the word “evolve” is deliberate. Such a system cannot be designed and imposed. But it can certainly be guided and nurtured. This evolution is further aided by the connectivity of biometric sensors generating continuous data that combines with computer intelligence and clinician insight to improve patient outcome on an ongoing basis.
Part 3 of this series examines ways to make sense of all the information that flows into the clinical encounter, including patient physiological data as well as the patient’s health history, social context, genetics and individual treatment responses. Part 4 will assess how specific treatments will be created, starting from patient data, clinical trials and real-world evidence and progressing through the control systems of the pharmaceutical plant of the near future. Part 1 covers the human foundation of connected wellness.
This article was written by Rodan Zadeh, Head of Connected Care Strategy within Cognizant’s Digital Business, Brian Williams, Chief Digital Officer within Cognizant’s Life Sciences Practice. Also contributing to this article were: David Staunton, Sashi Padarthy and Naveen Nayar from Cognizant’s Life Sciences, Healthcare Consulting and IoT practices.
To learn more, visit the Healthcare, Life Sciences and IoT sections of our website, or contact us.