Reading the ratings

Where the three categories come from

The scenario builder shows each test’s rating for the scenario you describe. Those ratings come from the appropriate use criteria (AUC). In an AUC, a writing group builds clinical scenarios, and an independent rating panel scores each test in each scenario in two rounds with a discussion in between, using a modified Delphi process. Each test ends up as Appropriate, May Be Appropriate or Rarely Appropriate Winchester et al., 2023, Rating Process and Scoring. The panel works from numerical scores, but this tool does not show them: it shows the categories only.

The authors also stress that the three groups are a simplification: appropriateness is most accurately viewed as a continuum, depending on the benefits and risks for the individual patient Winchester et al., 2023, Rating Process and Scoring.

What ratings are not

They are not rankings. The 2023 AUC says its ratings are explicitly not competitive rankings Winchester et al., 2023, Abstract, and the 2024 pre-operative AUC adds that within a category the numerical score is not a rank order, and that local expertise may favor one modality over another Doherty et al., 2024, Rating Process and Scoring. They are not mandates either: the AUC is meant to support, not replace, clinician judgment Winchester et al., 2023, Preface. A rating applies to the scenario as described, and it presumes the patient has no contraindications Winchester et al., 2023, Assumptions, item 23.

The tool’s other labels

Some cells do not carry one of the three categories. The tool uses three other labels, and each means something different:

• Not applicable: the source document deliberately did not rate that test for that scenario, most often because it is the test that was just performed. The AUC tables mark these cells as grayed out Winchester et al., 2023, Table 1.2.
• No data: no in-scope document rates the combination and no defensible inference exists. It is not a recommendation against the test.
• Not specified by the AUCs: the existing AUCs do not cover this combination. The main example is a patient with known coronary disease who has already been tested during the current evaluation. The decision is individualized with the patient’s cardiologist, and reference ratings are shown for context only.

Verbatim, High, Moderate and Low confidence

Not every cell is printed in an AUC table. Where the documents rate a neighbouring scenario or a related test, the tool derives the cell and marks it Inferred, shows the rule it used and attaches a confidence grade. The grade describes how direct the evidence for that cell is. It is the tool’s own scale, not an AUC concept:

• Verbatim: the document rates this exact scenario and modality.
• High: one input differs from a verbatim scenario and the document’s neighbouring ratings are consistent on both sides, or the document itself states the inference (for example, SPECT and PET rated together).
• Moderate: rests on a single analogous documented scenario, or combines an AUC rating with a protocol or guideline statement.
• Low: crosses indication families or borrows from a similar modality, or rests mainly on clinical reasoning.
• No data: no defensible analogue exists. It is not a recommendation against the test.

A grade says nothing about how appropriate the test is. A Low-confidence “Appropriate” and a Verbatim “Appropriate” are the same category, but the second is printed in a source table. One example of the documents’ own reasoning about grouping: the 2023 writing group judged that splitting a modality into its subtypes would not make a substantial difference to the ratings (for example, both SPECT and PET could be appropriate for recurrent symptoms after PCI), so the subtypes are not rated separately Winchester et al., 2023, Discussion.

How to use the explanation on each tile

Open any result tile, then work down the “Why this rating” panel:

1. Check the rating, the confidence grade and whether it is marked Inferred.
2. Read the justification and the rules applied. They explain why the cell has this rating.
3. Read the source passages, which show the document, table and page.
4. Look for a document discordance note. It appears where two source documents rated the same scenario differently.
5. Check the patient factors affecting this test. These are contraindications and cautions shown beside the rating. They do not change it.

Open the scenario builder → and click through the tiles for a scenario you know. The reasons behind each rating are shown there, not on this page.