The Social Risk Adjustment Debate

The previous lessons stated the design consensus: adjust payment for social risk, stratify quality reporting. That formulation is useful and it compresses a genuine disagreement worth understanding, because both positions are held by serious people for good reasons.

The case for adjusting

A provider serving patients facing housing instability, food insecurity, and transportation barriers will have worse measured outcomes for reasons that are not about the care delivered. If measurement ignores that:

• Safety-net providers score lower, earn less, and have less to invest, which the earlier lesson described as a downward spiral aimed at exactly the organizations serving those most in need.
• Providers learn to avoid complex patients, which is the adverse selection problem that appears reliably wherever measurement penalizes taking harder cases.
• The measurement misattributes a social outcome to clinical performance.

The case against adjusting

Adjusting quality scores for social risk means accepting a lower standard of outcome for disadvantaged patients as normal, and encoding it in the measurement.

If a hospital’s readmission rate is adjusted upward because its patients are poor, the report will say the hospital performed adequately, and the patients will still have been readmitted. The disparity becomes invisible in exactly the instrument built to detect it.

There is also a practical objection: some of what social risk adjustment absorbs is not immutable. Discharge planning that accounts for whether a patient has a working refrigerator is a care quality decision, not a social fact, and adjusting it away removes the pressure to do it.

The compromise, and a working example

The resolution the field has largely reached is to separate the two purposes. Adjust payment so providers serving disadvantaged populations are financially whole. Stratify reporting so the gaps remain visible and someone remains responsible for them.

Medicare’s readmissions program shows a third approach that does some of both. Since FY2019, the 21st Century Cures Act has directed CMS “to assess a hospital’s performance relative to other hospitals with a similar proportion of beneficiaries who are dually eligible for Medicare and full Medicaid benefits.”

Hospitals are sorted into peer groups by dual-eligible share and compared within them. That protects safety-net hospitals from being penalized against a baseline their patient mix makes unreachable, while preserving the underlying measured rates rather than adjusting them away. The legislation also required that estimated payments under peer grouping equal those under the prior methodology, so the change redistributed penalties without reducing them.

Worth remembering: the design question underneath this debate is what the measure is for. A measure used to set payment should be fair to the provider, which argues for adjustment or peer grouping. A measure used to identify where patients are being failed should be unadjusted, because the point is to show the gap. Trouble arises when one number is asked to do both jobs, which is most of the time. The organizations that handle this well typically maintain both views: an adjusted view for accountability and payment, and a raw stratified view for improvement, and they do not let the first quietly replace the second.

Key takeaways

• Adjusting for social risk protects safety-net providers and reduces incentives to avoid complex patients.
• Adjusting can also normalize worse outcomes for disadvantaged patients and hide disparities in the measure built to find them.
• The consensus separates purposes: adjust payment, stratify reporting.
• Medicare’s readmissions program peer-groups hospitals by dual-eligible share, protecting fairness while preserving measured rates.

Sources

• CMS, Hospital Readmissions Reduction Program (opens in a new tab)
• CMS, Contract Year 2027 Medicare Advantage and Part D Final Rule (opens in a new tab)