Use case 6 · Shift-share decomposition

Finding the cause behind a posted number

The pattern

Plans now post prior authorization metrics every year. The posted rate shows where a plan stands, not why it moves. A decomposition separates a change in mix from a change in each service.

Question: Why did our posted overturn rate move?
Lane: All lanes
Basis: Federal reporting rule and KFF data

What METRIS does

Rebuild the posted metrics monthly from the plan's own records
Decompose every change into mix and within-service effects
Trace the largest effect to the tool, vendor or criteria change behind it

Monthly overturn rate on appeal

Simulation study · synthetic data

The overturn rate rose from about 50.4% to 56.0% after month 12. The posted annual number would show only the rise.

Method

ΔR = Σ_s (w2_s − w1_s)·r1_s + Σ_s w2_s·(r2_s − r1_s) mix effect within-service effects

Services: 5
Appeals: 900 a month
Change: Imaging overturn 45% → 62%; post-acute share 15% → 24%
Interval: Wilson, 95%

A single posted percentage cannot separate a change in what was appealed from a change in how each service was decided. The decomposition does, exactly and additively.

What moved the rate

Exact calculation · assumed parameters

Of the 5.7-point rise, 4.8 comes from imaging and 1.0 from a shift in the mix of services appealed. The other services add up to -0.1 points.

Reproducibility

Monthly appeals are multinomial by service and binomial within service. Figure B decomposes the change using the true simulation parameters. Random seed 20261013; Python with NumPy 2.5.3 and SciPy.

Sources

KFF, Prior Authorization Metrics Provide New Insights into Insurer Practices, but Gaps Remain, August 13, 2026

Study design and analysis: Suneeta Modekurty, SANJEEVANI AI

A statistical study of a pattern, not of any event. It identifies no organization or person. Simulated figures use synthetic data; published figures are industry-wide.