Value-Based Care and Chronic Conditions: Leveraging Artificial Intelligence and Machine Learning to Address Social Determinants of Health

How can AI and ML data analytics reduce cost and improve health outcomes for individuals living with chronic conditions?

Chronic health conditions require ongoing medical care and significantly impact the quality of life. Costs of care are substantial, and are influenced by various factors, including Social Determinants of Health (SDoH), lifestyle, and co-morbidities. Value Based Care and the future health of populations rely on innovative and widespread use of Machine Learning (ML) and Artificial Intelligence (AI) tools to accomplish goals of high-quality care, exacerbation of disease, and the well-being of millions.

Chronic Conditions Are Our Norm, Not Our Exception

Chronic conditions account for nearly 75% of aggregate healthcare spending. . Many challenges exist in managing these conditions and healthcare payors and providers must deepen strategies to address aging populations with multiple risk factors such as: tobacco use, poor nutrition, lack of physical activity. Conditions include: cancer, diabetes, hypertension, stroke, heart disease, respiratory diseases, arthritis, obesity, and more.

70% of all deaths are caused by one or more these 5 conditions: heart disease, cancer, stroke, chronic obstructive pulmonary disease, and diabetes

The prevalence and severity of chronic conditions are rising:  

• Approximately 45% (133 million) of all Americans report at least one chronic condition and the number is growing

• 25% of U.S. adults have two or more chronic conditions

• 50% of older adults have three or more chronic conditions

• Persistent chronic conditions lead to hospitalization, long-term disability risk, reduced quality of life, and are the leading cause of death and disability in the US

• Chronic diseases kill 1.7 million Americans annually (7/10 of all deaths)

• Nearly 70% of all deaths are caused by one or more these 5 conditions: heart disease, cancer, stroke, chronic obstructive pulmonary disease, and diabetes

Compounding the cost and complexities of these high-risk variables, medical advances extend longevity yet fail to improve overall health and quality of life. In addition, health plans and most organizations have inadequate coding and rely on claims data that fail to offer predictive insights regarding aggravating factors to alleviate chronic conditions.

Strategically driving Value Based Care requires a deeper understanding of enrollees and members by leveraging ML and AI to forecast care needs (e.g., costs, conditions, SDoH) before claims are ever submitted.  

ML Changes Our Relationships to Chronic Conditions

Social Determinants of Health (SDoH) critically influence chronic conditions. Advanced ML data insights for social and environmental factors enable strategies reducing health burdens and costs of care by addressing social disparities impacting health outcomes.

Our relationship with chronic conditions is now more deeply informed as data and clinical science improve. ML incorporates insights from behavioral health, mental health, social and financial demographics to inform the complex nature of chronic illness progression and applied therapies. Importantly, the power of ML insights can be harnessed to address our leading causes of mortality and cost - cardiovascular disease, diabetes, cancer, pulmonary disease, mental health, obesity, and chronic kidney disease (CKD). SDoH factors correlate with behavioral choices related to food, fitness, finance, employment, and a wide range of specific consumer activities playing an important role in risk and management of these conditions.

 Beyond claims data, SDoH factors correlate with behavioral and consumptive choices related to food, fitness, finance, employment, and a wide range of specific consumer activities playing an important role in risk, understanding, and management of chronic conditions.

Cardiovascular Disease

• Heart disease and stroke negatively impact quality of life significantly and drastically, not to mention cost. Effective ML protocols evaluate the behaviors and choices of members with specific data reflecting food consumption, sedentary vs active lifestyles, cigarette and alcohol purchases, and more.

  Diabetes

• $237 billion dollars of spending is in direct relation to diabetes care in the US. The personal cost of this disease is 2.3 times greater for individuals with diabetes -- an absolute pain point for quality of life and ever growing concern. The aging population and their propensity to develop type 2 diabetes accelerates expenditures for individual members, payors, governments, and providers,  placing increasing strain on the Medicare system. Lifestyle data insights produced by ML  tools reflect choices impacting diabetes such as: consumption of sugary drinks, processed and fast foods, likelihood for obesity due to sedentary lifestyles, drug purchases and treatment adherence (such as regular purchase of insulin or medical monitoring devices).

 Cancer

• Early detection, prevention and treatments are closely linked to SDoH and lifestyle variables. ML insights point toward influences such as expenditures and income, education, and geography (i.e.,access to services). Consumption patterns relevant to tobacco and poor nutrition offer predictive insights regarding cancer risk management. Cancer patients face financial pressures leading to medical debt accrual, and even bankruptcy. ML insights can predict, mitigate, or all out prevent many negative consequences related to cancer care. 

Chronic Respiratory Disease

• Asthma and Chronic obstructive pulmonary disease (COPD) are frequent and chronic conditions which are commonly linked to cigarette smoke, pollution, and inflammatory factors like stress. Air quality is linked to social determinants such as geography and housing conditions and can exacerbate these respiratory conditions. Importantly, ML models offer detailed data insights regarding smoking (e.g., purchase of cigarettes, smoking accessories, vaping), occupational hazards (e.g., job related exposure to poor air quality), and environmental factors by geography which inform treatment and prevention.

 Obesity

• Not a chronic condition unto itself, obesity represents major risk factors for other chronic disease including diabetes, heart disease, and cancers. While genetics and epigenetics impact likelihood of obesity risk, lifestyle, behavior, and economic systems represent the primary drivers of obesity. Lifestyle data within ML models provides remarkable tools for understanding behaviors and trends related to consumption, sedentary activity, geography and housing, and more. Pinpointing at risk groups and scenarios can influence likelihoods of obesity and mitigate severity of the condition.

 Chronic Kidney Disease (CKD)

• With 37 million adults in the US with CKD, and Medicare bearing the brunt of CKD cost, early detections can drastically impact costs representing 7% of the Medicare FFS budget. Predicting and preventing advanced stage of CKD through social and behavioral factors leads to significant cost implications and reduction of potential CKD hospitalizations. ML promises to make identification more precise, including evaluating SDoH variables like demographics, economic factors, and even political affiliations as correlates.

 Advanced and robust ML is a crucial tool in reducing costs, improving informed decision for care, identifying high risk individuals, improving the management of chronic conditions and potential exacerbations, and forecasting patients at risk for chronic diseases.

Outcomes and Ultimate Successes

Advanced and robust ML is a crucial tool in reducing costs, improving informed decision for care, identifying high risk individuals, improving the management of chronic conditions and potential exacerbations, and forecasting patients at risk for chronic diseases. Payors, providers, and caregivers can develop strategies based on ML to derive meaningful information to build workable tools in treatment and guidance of longitudinal care:

• Capture of nonstandard data in support of clinical suspicions, socioeconomic data, patient preferences, lifestyle factors and indicators, and structure and related information hinting at related diagnoses and care

• Leveraging rich streams of information enhancing the experience of the patient and buttresses better management and ultimately, lower cost and higher quality outcomes

Social and Behavior detail have an incredible impact on chronic disease development, treatment, and outcomes; competitive strategies for payors, providers, and care organizations include the integration of ML platforms capable of identification, prediction ,intervention, and engagement long before a claim for chronic condition care is even submitted.

Ultimately, ML algorithms with the right data sets offer advanced and innovative solutions needed to collect and assess non-standard data to produce insights leading to action. Improving quality of life, care, and costs are a real outcome for the future of ML and Value Based Care

ML Tools Applied to Chronic Conditions:

• Develop insights for lifestyle data including behaviors and trends linked to consumption habits, housing, and other factors impacting likelihoods of chronic disease including obesity

• Provide nuanced and predictive support for social behavioral factors associated with specific chronic conditions

• Produce insights for essential behavioral habits including food choice, exercise, cigarette use, and alcohol use

• Calculate probabilities of developing and managing conditions

• Assess adherence to medication and pharmaceutical purchasing habits

• Predict occupational hazards for chronic conditions within groups, including exacerbation of various respiratory conditions

• Manage Medicaid populations more completely with insights for mental health and individuated care; this includes occupation, media consumption, dietary habits, exercise, and financial pressures

#ValueBasedCare #vbc #MachineLearning #ML #AI #PersonalizedCare #MedicareAdvantage #Medicaid #MCO #PatientExperience #SDoH #MemberExperience #Medicare #OpenEnrollment #PREMs #PROMs #Lifestyle #VBCenablement #MemberEngagement

A note on Verikai:

Verikai data assesses more than 700 unique social and behavioral pattern fields to provide extremely specific qualitative insights; firms adapt their strategies to better serve current and future members by understanding: 

Personality tendencies such as: Impulsive, Value Seeker, Risk-taking…

Consumption patterns such as: Groceries, Stores, Online Shopping…

Spending trends such as: Alcohol, Tobacco, Dining, Education, Entertainment, Furnishings…

Technology usage such as: Type of cell phone, Social networking, Email use…

Wellness indicators such as: Sleep quality, stress levels, Lifestyle diet, Risky health behaviors…

Financial patterns such as: Spending, Investments, Taxable assets, Real Estate, Credit card type and use, Bank account types, Debt management…

Societal behaviors such as: Charitable donations, Political affiliation...

 More on Verikai