latest content

featured content

filter by categories

how risk adjustment and predictive models drive effective value-based care

by
November 10, 2016

Healthcare technology is constantly evolving and improving. Although risk adjustment and predictive models have long been established in the industry, their power and accuracy cement their status as a cornerstone for healthcare strategy and operations. Risk adjustment and predictive models provide critical insight into risk, utilization, and cost, as well as ensure fair comparisons between populations. Developments in modeling advance these fundamentals and provide new insights into healthcare trends that propel value-based care strategies.

history of risk adjustment and predictive modeling

In the 1980s, actuaries used basic methods to create payment models, only factoring in age, gender, and Medicaid eligibility. After the passage of the Tax Equity and Fiscal Responsibility Act (TEFRA) in 1982, however, the Health Care Financing Administration (now the Centers for Medicare & Medicaid Services or CMS) reached out to some of the nation’s top analysts to create a better way to assess medical risk to reflect the future chance of illness.

 

A trio of researchers in Boston then developed diagnostic cost groups (DCGs), adding principal diagnoses found in inpatient hospitalization claims to help predict costs. This led to the development of hierarchical condition categories (HCCs), which use inpatient, outpatient, and professional claims to characterize a patient’s complete disease status and predict his or her future cost. In 1996, the first commercial version of HCCs were adopted by CMS based on the work of this team, whose models are now known as DxCG. Additionally, CMS has adopted other versions of the HCC methodology for use in the healthcare exchanges, which are part of the Affordable Care Act.

importance of risk adjustment and predictive modeling in value-based care

In recent years, there has been tremendous growth in the number of lives covered under risk-bearing contracts. There are currently more than 800 accountable care organizations (ACOs) in the United States, and by 2020, the number of lives covered under risk-based contracts is expected to account for roughly 40 percent of the total insured population. This sea change in the healthcare industry has led many provider organizations to start their own insurance plans, requiring them to understand risk in a way they didn’t have to before.

 

Success in value-based care depends upon an organization’s ability to use healthcare data effectively. Risk adjustment and predictive models can enable organizations to accomplish both financial and clinical goals across three key areas: budgeting and underwriting, medical management, and performance assessment.

 

In addition to effective payment and contracting, risk adjustment and predictive models allow organizations to control healthcare costs through targeted interventions directed toward high-risk population members. Once those members have been identified through analytics, one can then apply additional filters and analysis to focus on the subset of people that case managers can impact through a defined intervention. A well-performing predictive model will be able detect not only who will be high-cost in year one, but who is most likely to remain high cost and continue to benefit from targeted intervention.

emerging insights on challenges in value-based care contracting

The shift to value-based contracting fundamentally changes the business model for providers, who are encountering new practical challenges in making this shift: mitigating unexpected volatility in risk scores and managing mixed risk and fee-for-service (FFS) populations during the transition. Organizations across the value-based continuum can find value in applying risk adjustment and predictive models to their data.

 

Organizations with complete claims data sets can utilize three advanced applications of risk adjustment and predictive models: using control charts for relative risk score to track changes over time in the context of past performance, monitoring changes in the population stability index for condition categories to identify key population drivers of risk and disease burden, and monitoring cost trends by condition categories to identify cost drivers. These approaches unlock the power of models and inform effective care management.

 

When managing a mixed population that includes both at-risk and FFS members, many providers only have administrative claims for a subset of their overall population. To overcome this challenge, risk scores can be calculated using billing or electronic health record (EHR) data for the purpose of supporting risk contract negotiations when claims aren’t available, obtaining history on patients moving from FFS to risk contracts, determining quality bonuses for FFS contracts, and simplifying operations for care management programs across payer types.

conclusion

Risk adjustment and predictive modeling have become foundational to modern healthcare and are critical to achieving the goals of value-based care: aligning payment with disease burden, identifying high-risk patients, and evaluating performance. Predictive models provide insights that help address practical challenges in value-based care contracting, mitigating unexpected volatility in risk scores and managing mixed FFS and risk populations.

 

For a closer look at advances made in predictive analytics and how they could potentially benefit your organization, watch our on-demand webinar: “But Wait... There's More: Risk Adjustment and Predictive Modeling Advances.”

watch the webinar

 

don’t miss an update

leave a comment below: