Reducing Treatment Burden Among People With Chronic Conditions Using Machine Learning: Viewpoint

Harpreet Nagra,Aradhana Goel,Dan Goldner

doi:10.2196/29499

Abstract

The COVID-19 pandemic has illuminated multiple challenges within the health care system and is unique to those living with chronic conditions. Recent advances in digital health technologies (eHealth) present opportunities to improve quality of care, self-management, and decision-making support to reduce treatment burden and the risk of chronic condition management burnout. There are limited available eHealth models that can adequately describe how this can be carried out. In this paper, we define treatment burden and the related risk of affective burnout; assess how an eHealth enhanced Chronic Care Model can help prioritize digital health solutions; and describe an emerging machine learning model as one example aimed to alleviate treatment burden and burnout risk. We propose that eHealth-driven machine learning models can be a disruptive change to optimally support persons living with chronic conditions.

Highlights

The COVID-19 pandemic has surfaced multiple concerns present within our health care systems, including the high infection risk prevalent among people with chronic conditions, and the fact that practitioners can only provide specialized responses to acute illnesses [1]
These, in turn, leave people with chronic conditions to experience fragmented, poorly coordinated, and limited support in their treatment [2], which exacerbates the treatment burden patients experience as they encounter decreased support for their ongoing medical care [3]
Increased treatment burden can heighten the risk for illness-related burnout—a chronic affective state comprising symptoms of emotional exhaustion, physical fatigue, and cognitive weariness, often the outcome of depletion of energetic resources resulting from prolonged exposure to medical distress [4]

Summary

Introduction

The COVID-19 pandemic has surfaced multiple concerns present within our health care systems, including the high infection risk prevalent among people with chronic conditions, and the fact that practitioners can only provide specialized responses to acute illnesses [1]. No eHealth offerings focused on people with chronic conditions integrate a combination of multiple ML models to (1) forecast outcomes up to 6 months into the future; (2) provide insight into which lifestyle behavior modification may most significantly impact a person’s desired clinical outcome; and (3) use various data inputs (eg, biometrics, behavior, outcomes, and engagement) to support behavior change and reduce treatment burden for people with chronic conditions. 2 (page number not for citation purposes) any reason, the most effective mode could be selected In this way, the system can support people with chronic conditions in making progress toward better health outcomes using tailored interventions that adapt to each individual, evolve with time, and are informed by their predicted effect on the individual’s health. These coaches are typically more readily available for appointments between clinic visits, and the use of their services may reduce the burden on the health care system by increasing patient engagement in self-care and supporting health care providers with additional patient health insight

Findings

Discussion

Conclusion