4 Objections to HHS' Two-Stage Weighted Risk Adjustment Model for the Marketplace

Analysts say HHS favors "models that make accurate predictions at the enrollee level within a sample population... that tend to miss the ranges of severity and spending within specific conditions like diabetes."

In its annual marketplace rule, HHS proposes a new "two-stage weighted approach to … [risk] model recalibrations." While some support the change announced in the Notice of Benefit and Payment Parameters (NBPP), most payers, their affiliation groups, and key policy analysts do not. The following are four key oppositions to HHS' proposed change.

1. The proposed two-weighted model is not appropriate for risk adjustment.

In its comment letter to the proposed rule, the Blue Cross Blue Shield Association (BCBSA) noted that "a two-stage weighted approach is not a standard procedure for risk adjustment and actually worsens the fit [predictive capability], although the effect is small." In making this claim, BCBSA further notes that the model's interaction with other factors that influence risk payment transfers (several noted below) should be considered.

2. Current risk adjustment predictions are already accurate.

The opposition of multiple payers—Anthem and Health Care Services Corporation in addition to BCBSA—suggests that there isn't anything substantially flawed with HHS' current risk-adjustment model given the relative health and competitiveness of the exchange market. Even so, payers support the continued discussion of possible risk-adjustment model changes that HHS introduced in an October 2021 Risk Adjustment (RA) Technical Paper that predated the 2022 NBPP.

3. Current administrative adjustments are also sufficient to achieve HHS's goal

BCBSA suggests that an existing administrative adjustment designed to "reduce … transfers from low-risk enrollees to higher-risk enrollees … already addresses some of the under-prediction" that HHS seeks to correct.

4. It will have the opposite effect and increase adverse selection.

Echoing other payers, BCBSA states: "The first priority for the risk adjustment model should be that issuers have no incentives to avoid higher-risk enrollees." HHS believes that its new two-weighted model will correct under-estimates of low-risk enrollees, BCBSA—in addition to multiple payers and the Brookings Institution believe that HHS' new model will do the opposite.

The result could be increased adverse selection.

Brookings Institution analysts Matthew Fiedler and Timothy Layton note: "Economic theory implies that these changes would have three types of effects:

• Shift enrollment toward skimpier plans …
   
• Shift premium burdens toward higher-risk enrollees …
   
• Reduce choice and competition."

In an interview with HealthLeaders, Fiedler notes that the HHS approach "is to prefer models that make accurate predictions at the enrollee level within a sample population." He adds that while this result may result in estimates that are "perfectly accurate on average," they tend to miss the ranges of severity and spending within specific conditions like diabetes.

In summary, Fiedler notes: "The core issue is that the agency is misreading what's happening under the current risk adjustment system so in thinking about changes, they're moving in the wrong direction."