A long-expected paradigm shift has arrived: health care has moved beyond the four walls of a provider’s facility and into the age of remote patient health management. Spurred by the proliferation of relatively inexpensive wearable monitoring devices, unprecedented volumes of patient data and powerful analytics, remote patient monitoring programs have now been deployed in over 66 percent of health care organizations in the U.S., according to at least one report on the topic. Another study found that close to 5 million patients are currently equipped with remote health monitoring devices, a number that is expected to grow to 36 million just within the next five years.

The most effective remote patient health management programs today leverage technology-driven precision medicine. These systems use computer algorithms that work like the human mind — that is, they “self-learn” over time by utilizing observed clinical data to predict more accurately when a patient’s condition is deteriorating. In turn, care team members have better information so that they can intervene sooner and avoid costly hospitalization. These developments are especially beneficial for individuals with chronic diseases, who now have more options for managing and monitoring care outside of the doctor’s office, and can benefit from more precise and timely interventions based on their unique health profile.

The value of precision monitoring to predict risk
Unnecessary hospitalization is a common and expensive form of treatment for people with chronic illness, and costs hundreds of billions of dollars a year. In contrast to an expensive hospital stay, remote patient health management programs that incorporate “precision algorithms” can detect and communicate early signs of health deterioration and allow a patient’s care team to intervene before hospitalization is required.

The earlier the prediction, the more moderate and affordable the intervention. In the case of an individual with congestive heart failure (CHF), for example, intervention could be as simple as an adjustment in medication. Compare these minimal expenditures to the mean cost of $13,000 per CHF readmission, as cited by the Healthcare Cost Utilization Project, and the potential value of early detection and intervention becomes quickly apparent.

An essential tool for early intervention — but not all programs are equal
So how does remote patient health management work? And what are the hallmarks of an effective program? First and foremost, such a program is based on a patient’s comprehensive health picture, including history and current symptoms, so that a reliable, patient-unique risk tracker can be created. A program without these characteristics will almost surely generate a continuous flurry of alerts — and leave caregivers with considerable alert fatigue. With predictive analytics based on machine-learning, however, precise algorithms are continuously fine-tuned based on the patient’s unique history and profile. Only then are alerts issued and caregivers notified.

Here’s how a machine-learning analytics program might work. Patient symptoms and vitals are tracked and analyzed, and assessed for their likelihood of accurately predicting risk, much in the same way that the human brain absorbs and processes certain realties. The end result is an algorithm that can predict with astonishing accuracy if a chronically ill patient is on the verge of an acute health event. Note that once this algorithm is identified and used, it minimizes the need for continuous tracking of a patient’s health status. Accurate risk predications can be made with less frequent assessments, which is especially beneficial with patients who are resistant to long-term participation. Even more importantly, it helps caregivers to intervene more quickly, should a risk be identified.

Another key characteristic of a precision remote monitoring program: clinicians have insight into individual patient risk on a population scale. Remote health monitoring programs continuously feed updated data on the individual, including vitals and symptoms and medication compliance. Advanced analytics are used to analyze a patient’s individual clinical history and correlate it with more generalized population health data. Over time, event predictability is more precise at the individual patient level thanks to machine-learning technology. Clinicians can identify thresholds that are unique to each patient and can assign a customized patient risk score. Over time, clinicians have an increasingly rapid and precise window into a patient’s current level of risk.

Tens of millions of individuals face chronic disease and managing their health has become a national crisis. We can no longer afford a reactive approach and must leverage precision medicine and machine-learning technologies to facilitate earlier and more personalized intervention. Improving the health of individuals with chronic disease not only improves the quality of their lives, it also tremendously reduces care delivery costs.

Steve Curd is CEO of Wanda, which is dedicated to helping health care providers improve outpatient care while reducing the cost of managing chronic diseases.