How we score credentialing software and services
By Credentialing Scorecard editors · Rubric v1.1, published September 30, 2026 · Changelog
Who is scoring. Credentialing Scorecard is operated by Metolius Health, which is listed in our Software and Managed Services tracks. That is why this page exists: the rules below are fixed and published before any vendor is scored, and they apply to our operator exactly as they apply to everyone else. Full disclosure
The short version
- Every vendor is judged on the same 18 dimensions: 12 for capability (60% of the composite) and 6 for quality (40%).
- Software platforms and managed services are scored on separate tracks and never compared head-to-head.
- No source, no score. A dimension without cited evidence shows “Not yet verified”.
- A documented capability score appears at eight evidenced capabilities including all four core areas. Coverage is shown alongside it.
- A composite appears only when at least 12 of the track’s applicable dimensions are evidenced, including every high-weight capability and at least three quality dimensions.
- Metolius Health, our operator, goes through the same code with no override.
- Vendors can reply to or correct a listing with evidence. They can’t pay, and they can’t have a listing removed.
What does a documented capability score mean?
It measures support within the capability scope we have reviewed, on a 0–100 scale. Sources may include a vendor’s own documentation. It is not an independent performance test, service-quality rating or recommendation.
We publish it when at least 8 of 12 capability dimensions have cited evidence, including primary source verification, payer enrollment, re-credentialing, and license/expirables monitoring. This is an editorial publication threshold, not a validated reliability measure. The gate checks whether evidence exists, not whether its finding is positive.
For each evidenced capability, full support earns 1, partial support 0.5, and explicitly documented absence 0. Multiply by the relative weight below, sum, divide by the sum of the evidenced capability weights, then multiply by 100 and round once to the nearest integer. Unknown capabilities are excluded from both numerator and denominator, not scored zero.
Always read the score with its coverage. Listings show the number reviewed out of 12; profiles also show the relative weight covered out of 77 and the full/partial/unsupported counts. A score of 100 means full support within the reviewed scope, not all twelve capabilities. Different coverage makes scores not directly comparable: an unverified weakness can make a less-complete record look stronger. We keep alphabetical ordering and do not rank vendors by these scores. Open the dimension evidence to compare the scope you need.
Below the threshold, listings name up to two sourced capability areas, including partial support; “scope documented” does not mean full support. When none has a supporting source, we say “Capability sources needed.”
How is an overall rating calculated?
The overall rating (composite) remains separate. It requires at least 12 total dimensions, all four core capabilities, and at least three quality dimensions with evidence. Pricing transparency alone cannot establish service quality. No overall ratings are currently published.
Normalize the two layers separately using their evidenced dimensions: overall = round(100 × (0.60 × capability fraction + 0.40 × quality fraction)). Quality ratings 1–5 map to fractions 0, 0.25, 0.5, 0.75 and 1. Use unrounded layer fractions; round only the final result. Relative weights total 77 in capability and 40 in quality; they are not points out of 100. Separate normalization keeps the promised 60/40 split even when coverage differs. An overall rating may meet its own gate without the standalone capability score meeting the eight-capability gate.
Version 1.1 corrects the previous implementation’s pooled denominator, which did not enforce the published 60/40 layer split. No existing overall rating changed because none met the quality-evidence threshold. The code computes scores from cited evidence; no editor or operator can enter an override. Before publishing new subjective quality assessments, we must publish dimension-specific rating anchors and their evidence requirements.
Capability layer: 60% of an overall rating
Can the product or service do the work? Each capability dimension is scored explicitly unsupported (0), partial (half) or full, based on cited evidence.
| # | Dimension | Relative weight within layer | What we look for |
|---|---|---|---|
| 01 | Primary source verification | 12 | Verifies licenses, education, board certification and other credentials directly with the issuing source |
| 02 | Payer enrollment automation | 10 | Prepares, submits and tracks payer enrollment applications |
| 03 | Re-credentialing | 8 | Runs the re-credentialing cycle, not just reminders |
| 04 | License and expirables monitoring | 8 | Tracks expirations and alerts before they lapse |
| 05 | OIG / SAM exclusion and sanctions monitoring | 7 | Screens exclusion and sanctions lists on a schedule |
| 06 | CAQH / DataSpring integration | 5 | Reads or maintains the provider’s DataSpring (formerly CAQH) profile |
| 07 | NPDB query | 4 | Queries the National Practitioner Data Bank and handles ongoing enrollment |
| 08 | Privileging | 6 | Privilege delineation, approvals and board routing |
| 09 | Provider data management | 5 | Keeps a single, current record of each provider |
| 10 | Delegated credentialing support | 4 | Supports delegation to or from health plans, with oversight reporting |
| 11 | Reporting and analytics | 4 | Status, turnaround and audit reporting a team can use |
| 12 | API and integrations | 4 | Documented integrations with the systems a buyer already runs |
Quality layer: 40% of an overall rating
How well does it do the work? Each quality dimension is scored from cited evidence.
| # | Dimension | Relative weight within layer | Evidence we accept |
|---|---|---|---|
| 13 | Ease of use | 8 | Moderated user reviews and documented product walk-throughs |
| 14 | Support quality | 8 | Moderated user reviews and published support terms |
| 15 | Implementation time | 6 | User reviews, and vendor-stated timelines we could check |
| 16 | Security and compliance | 7 | Verifiable credentials only: SOC 2 reports, HIPAA attestations, HITRUST, NCQA status from NCQA’s directory, KLAS reports |
| 17 | Pricing transparency | 5 | A published price list on the vendor’s own site ranks above published ranges, which rank above quote-only |
| 18 | Value for money | 6 | User reviews set against published pricing, where it exists |
Weights are fixed. Weights were published before any vendor was scored. They change only in a new rubric version, logged in the changelog with the reason. A buyer-type toggle that re-weights scores for health systems, groups, payers or solo practices is planned, and when it ships the fixed weights above remain the default.
What counts as evidence?
Evidence is a source a reader can open: the vendor’s own documentation or pricing page, a regulator or accreditor’s record (such as NCQA’s directory), a filing, a named press release, or a moderated user review. Each dimension score stores the source’s URL, its publisher and the date we read it, and the profile shows them next to the score.
What we don’t count as evidence:
- A vendor’s own claim about its NCQA status. We use NCQA’s directory.
- Third-party price estimates presented as a vendor’s price.
- Performance figures a vendor publishes about itself (for example, speed or accuracy claims) unless an outside source confirms them.
- Ratings we could see only in a search snippet.
- Anything we could not date.
Why do some profiles say “Not yet verified”?
Because we only score what we can source. If we have no cited evidence for a dimension, the profile says “Not yet verified” for it instead of guessing, and the dimension is left out of the composite rather than counted as zero.
When does a composite appear?
A composite appears once a vendor’s evidence covers at least 12 of the dimensions that apply to its track, and that must include every capability dimension worth 8 points or more: primary source verification, payer enrollment automation, re-credentialing, and license and expirables monitoring. It must also include at least three quality dimensions, because capability evidence usually comes from a vendor’s own documentation and a score should not rest on that alone. Until then the profile shows documented capability scope, coverage and “Overall rating: Not published.”
We set the threshold so that no composite rests on a handful of easy dimensions. We will revisit it after the first quarter of publishing, and any change will appear in the changelog.
Do software and services use the same rubric?
They use the same 18 dimensions and the same weights, but they are scored and listed separately.
A software score and a services score sit on separate scales. We never rank one against the other.
How is NCQA status recorded?
From NCQA’s own certification directory, re-pulled every quarter. We record exactly one of: certified (with the number of verification elements), accredited, scheduled, expired, denied or not listed. Accreditation and certification are different NCQA programs, and we label them separately. What each status means
How is Metolius Health scored?
Exactly like everyone else. Metolius Health, which operates this site, is a vendor record in the same data files, scored by the same code from the same kind of cited evidence. There is no override, manual adjustment or separate rule for it. Where it lacks evidence, its profile says “Not yet verified” in public, like anyone’s. Changes to its scores are logged in the public changelog like everyone else’s.
We also keep two further rules for our operator: it does not submit user reviews about itself, and any Briefing item that mentions Metolius Health carries an inline disclosure.
Where do user reviews fit?
Reviews submitted by readers are moderated and shown in their own block on each profile, labeled as user reviews. A review’s star rating never changes a composite directly.
How can a vendor reply or correct a listing?
Any vendor can send a correction or a reply through the vendors page, with a source. If the source changes a fact, we update the listing, re-run the score and log the change with the date and reason. A vendor can also ask us to publish a short reply on its profile. No vendor can pay us, and no vendor can have a listing removed or a score hidden.
How often is this updated?
Evidence is updated as vendors change and as we find new sources. NCQA statuses are re-pulled every quarter. Every vendor score change is logged in listing and score changes. Changes to this rubric are logged in the rubric changelog.
What the labels mean
| State | Label |
|---|---|
| Capability, full | Full |
| Capability, partial | Partial |
| Capability, explicitly unsupported | Not supported |
| Any dimension, no evidence | Not yet verified |
| Overall rating, threshold not met | Not published |
| Dimension not applicable to track | Doesn’t apply to this track |
“Not supported” requires an affirmative cited statement of absence. Omission from a feature list, an inaccessible page or missing research is not evidence of absence. Without a source, the label is “Not yet verified.”
Methodology questions
Can a vendor pay to improve its score?
No. We don’t sell placement, sponsorship or score changes, and we don’t accept payment from vendors for coverage.
Why is there no score for most vendors yet?
A documented capability score requires eight capabilities including the four core areas. Below that we show sourced scope and coverage. Overall ratings require additional quality evidence and remain separate; we do not invent it to fill the page.
Does the operator see scores before they publish?
The rubric and the code are the same for every vendor, and every score links to its sources, so anyone can check the result.
Can I use your scores in my own materials?
Link to the profile rather than copying a number, so readers see the date and the sources. Scores change as evidence changes.