ROLLCALL HEALTH

The Rollcall Health Index · CMS Marketplace network

Your network changes every day. Rollcall detects what changed.

Provider directories become inaccurate when the network changes faster than the data. Rollcall independently monitors the network, detects potential changes and gives health plans evidence to review and correct the published directory.

As of
30 August 2026
Published
30 August 2026
Method
v1.0
Checks
Deactivation, Exclusion
Sources
NPPES 10 Aug 2026 · HHS-OIG LEIE 10 Aug 2026

Level · as of 30 August 2026

15.58

Qualified findings per 10,000 records evaluated

Same constituents
13.22 ↓ 1.9%
Against baseline 13.69
+1.89
Panel
98 directories
Median directory
4.01
Highest directory
272.22

25,422 qualified listing records in 16,315,452 evaluated. Zero is the target and higher is worse. It is not a measure of overall provider directory accuracy, and not a compliance determination. 1 directory could not be read and carries no score; an unavailable directory is never recorded as zero. 9 directories are showing the measurement they carried forward, last measured in 2026-W34 to 2026-W35, which is outside the Index above.

Movement

The Index over time

Same-constituents seriesAll-directories series

0 5 10 15 20 W3313 Aug W3315 Aug W3421 Aug W3530 Aug 15.58 13.22

Same-constituents movement

↓ 1.9%

13.48 → 13.22

vs the prior run

Composition · as of 30 August 2026

Market distribution

The public Index reports the market. Individual directory rates and record-level evidence remain private.

98 directories scored in the current run

0 per 10,000: 11 directories00–1 per 10,000: 9 directories1–2 per 10,000: 10 directories22–3 per 10,000: 8 directories3–5 per 10,000: 16 directories55–10 per 10,000: 18 directories1010–25 per 10,000: 13 directories2525–50 per 10,000: 8 directoriesover 50 per 10,000: 5 directories50+Median 4.01

4.01

Market median

10

Elevated or priority

11

No qualified findings

Rates are qualified findings per 10,000 evaluated records; buckets widen to the right because the spread is long-tailed. Coverage: 1,325 of 1,331 files read.

Checks · method v1.0

What the Index checks today

Two checks count today, both against dated federal records, both answerable yes or no from outside the plan. The file’s own structure is next. The detail is in the method.

01 DEACTIVATION

The NPI was already cancelled

Deactivated in the CMS NPPES cumulative report before we read the file. A cancellation after our read is never a finding.

02 EXCLUSION

The provider was barred from federal programmes

An active HHS-OIG exclusion dated before the day we read the file. Matched on the federal identifier, never on a name.

03 STRUCTURE · IN DEVELOPMENT

The listing repeats or contradicts itself

Duplicate listings, provider-type conflicts and specialty overlaps inside the published file. Reported privately, and never counted in this Index’s level.

Method v1.0

How the number is made

  1. Read every provider directory a Marketplace medical issuer registers with CMS, as published.
  2. Match every NPI-bearing listing against two dated federal records: the NPPES cumulative deactivation report and the HHS-OIG exclusion list.
  3. Hold a finding only where the federal fact is dated earlier than the day the directory file was read. There is no grace period on the comparison itself.
  4. Divide qualified records by records evaluated, per 10,000, across the whole panel at once.
  5. Publish the same-constituents series beside it, so a change in membership never reads as a change in the market.

Those five clauses are the summary. The rulebook they summarise is a versioned document: objective, eligible universe, constituent selection, weighting, calculation, review and reconstitution, special cases, data sources, governance and limitations.

Read the Index constitution How we check

See what changed in your directory.

Rollcall can privately show your directory’s benchmark, current exceptions and the evidence behind each potential change.

Request private benchmark →