The Geographic Risk Index (GRI) scores every US hospital's community on 18 dimensions — chronic disease, poverty, insurance gaps, provider access, isolation. It explains 60.5% of the variance in hospital mortality before a single patient is treated. What follows are two views of the same 2,490 hospitals. The y-axis — Medicare payment per case vs the national median — is identical in both. Only the x-axis changes: what we compare mortality to.
Under the current model, the system rewards hospitals that are already winning.
65% of large urban hospitals (D1, Q5) land in the "better mortality, paid more" quadrant. For small rural hospitals (D10, Q1): 1%.
The bigger the hospital, the better it looks on mortality (r = −0.61). The more rural the community, the less it gets paid (r = −0.27). These are independent forces — size and geography — but they compound against the same hospitals.
Hospitals can't choose their catchment population. The communities with the greatest burden are served by the hospitals the measurement system is most designed to punish.
Adjust for community burden and the gradient flips. D9–D10 hospitals are 3× more likely to be outperforming their peer expectation while collecting below the national median.
57% of the most rural hospitals are doing more and getting paid less. For urban hospitals, it's 18%.
The GRI was built to predict mortality. Medicare payment has no geographic input. The GRI predicts that too.
These communities can't fix a proven geographic effect. Their hospitals can't choose an easier catchment. So why does the measurement system punish them for it — and then pay them less?
The people in these communities need more resources, not fewer. A GRI adjustment doesn't change how much Medicare spends. It changes where it goes.