Reasonable risk adjustment helps ensure fair payment

Millions of lives will enter the insurance market through new state exchanges with unmet health needs and rather unpredictable futures. Even the most sophisticated predictive model isn't perfect, and health plans will be doing a lot of correcting as they gain experience.

In the next few years, arriving at reliable risk scores for lives within the exchanges will be critical for fair payment. Plans that fail to analyze their populations accurately will be giving money away to their competition.

Medicare Advantage adjustments compare a plan's population against fee-for-service benchmarks then figure in an intensity reduction and county adjustment. In the exchanges, plans won't be compared to a benchmark, but to each other. Certainly, plans with less risk will be paid less. That unspent money, however, will be reallocated to pay other plans more.

So the equalization is more a case of subsidization.

I recently spoke to John Steele, managing partner for HealthScape Advisors, and he cautions that plan revenue increasingly will be driven by risk adjustment, which will depend on the ability to obtain current, accurate and complete diagnostic information as early as possible. Plans will need to act quickly.

"This is not a 2013 or 2014 issue," Steele says "It's something to focus on sooner rather than later."

He says the experience period for encounter and health status information for the initial risk adjustments could start in the middle of 2012. Plans will have 18 months to model for risk in the insurance exchanges that launch in 2014.

"The lead time you need to change your medical management and provider processes is not something that happens in six months or a year," Steele says. "It's a very long lead time."

DIAGNOSIS DATA NEEDED

Steve Young, managing director for HealthScape Advisors, also tells me reliable diagnoses, which will be key to prospective risk adjustment, usually only come through inpatient codes. Other sources that feed into the risk adjustment model can have data gaps, such as pharmacy data that might only code a single health issue associated with a particular prescription drug.

While pharmacy data might be the only fallback in some states, because it's coded for payment, it lacks a more telling indicator of risk, such as a chronic condition, for example.

"The big risk adjustment comes with comorbidities, when they're all added up," Young says. "It's very additive, particularly when you get into behavioral health issues."

Medicare Advantage plans that are doing fairly well with their risk adjustment still need to rethink their approaches for the exchange market, Steele says.

It's unlikely that one risk system will fit across all exchange populations. Anticipate the need for painstaking recalibration. Comprehensive, prospective data and adequate analysis of the modeling results will be key to accurate risk scores.

Julie Miller is editor-in-chief of MANAGED HEALTHCARE EXECUTIVE. She can be reached at julie.miller@advanstar.com