MIPS Quality Measures for Radiology: What Actually Applies to You

Most MIPS guidance assumes you see patients in a clinic, manage chronic disease, and control an EHR. Radiologists do none of those things, which is why so much of the standard advice reads as irrelevant. The programme does accommodate this - but only if you know which accommodations to claim.

This is the radiology-specific view of MIPS quality measures: what applies, what gets reweighted away, and which measures are realistically reportable from a reading list.

Payment adjustment percentages, performance thresholds and the measure inventory change every year in the CMS final rule. Treat the structure below as durable and verify the current year's numbers before you plan around them.

The four categories, and why two of them mostly disappear

MIPS scores clinicians across four categories: Quality, Promoting Interoperability, Improvement Activities, and Cost.

For most radiologists, the second one largely vanishes. CMS recognises a non-patient-facing designation, defined by the volume of patient-facing encounters billed. Diagnostic radiologists typically qualify. Non-patient-facing clinicians generally have Promoting Interoperability reweighted to zero, because the category assumes control over certified EHR technology that a radiologist reading from a PACS worklist does not have.

That weight does not evaporate - it is redistributed, mostly onto Quality. The practical consequence is counterintuitive: because a category was removed, the remaining ones matter more. A radiology group that shrugs at Quality because "MIPS does not really apply to us" is optimising exactly backwards.

Interventional radiologists may not qualify as non-patient-facing. The designation is determined by encounter volume, not by specialty label, and it is worth confirming rather than assuming.

Quality measures that fit radiology

The Quality inventory runs to hundreds of measures, the overwhelming majority of which a radiologist can never satisfy. The radiology-relevant subset clusters into three groups.

Radiation dose and safety. Measures covering dose indices for CT and fluoroscopy, use of dose-lowering techniques, and tracking of potential high-dose studies. These are attractive because modern scanners and dose-monitoring software already capture the data - the reporting burden is largely a data-extraction problem rather than a change in practice.

Appropriate follow-up for incidental findings. This group asks whether your report gave a follow-up recommendation for an incidental finding, where one is warranted. Incidental abdominal lesions and incidental thyroid nodules both have dedicated measures.

Communication and report completeness. Measures around communicating actionable findings to the referring clinician within a defined interval.

Why the incidental-findings measures reward good reporting

The follow-up measures are the ones worth dwelling on, because they are the rare case where the compliance requirement and the right clinical behaviour are the same thing.

The incidental thyroid nodule measure, in substance, asks: when you described a thyroid nodule that meets the threshold, did you say what should happen next? A report that reads "1.4 cm hypoechoic nodule in the right lobe" and stops has described a finding and left the decision to someone with less information than you had. A report that reads "1.4 cm solid hypoechoic nodule, wider-than-tall, smooth margins, no echogenic foci - ACR TI-RADS 3 points, TR3. Below the 25 mm FNA threshold; ultrasound follow-up at 12 months recommended" has closed the loop.

The second version scores. It is also just the better report. That alignment is unusual in quality programmes and worth taking advantage of.

The same logic applies to incidental abdominal lesions - the measure is satisfied by a recommendation consistent with a published guideline, which is what you should be writing anyway.

The reporting burden is the real problem

Radiologists rarely fail these measures because they practise badly. They fail because the documentation is inconsistent.

A recommendation phrased three different ways across three reports is three different strings to a quality-abstraction engine, even when the clinical content is identical. Structured, consistently-worded recommendations are what make the measure abstractable. This is the single highest-leverage change most groups can make: not doing anything new clinically, but saying it the same way every time.

MIPS Value Pathways

CMS has been steering clinicians toward MIPS Value Pathways (MVPs) - smaller, specialty-coherent bundles of measures intended to replace the pick-from-hundreds model. A diagnostic radiology pathway exists, and the direction of travel is clear even though the traditional route remains available.

If your group is choosing measures from scratch, it is worth looking at the relevant MVP first rather than assembling a set you will have to rebuild later.

Where reporting software helps, and where it does not

Software cannot make you compliant. What it can do is remove the variance.

MyRadAgent applies the relevant ACR classification systems as you dictate and writes the follow-up recommendation into the impression with the guideline named and the version cited - the same phrasing, every time, for every reader in the group. The implementation lives in mips_compliance.py and is gated by modality and clinical context, so a recommendation only fires where the finding actually warrants one.

That addresses the consistency problem. It does not address measure selection, attribution, or submission, which remain a practice-management exercise.

Related

• ACR TI-RADS calculator - score a thyroid nodule and get the FNA threshold
• Fleischner Society calculator - incidental pulmonary nodule follow-up
• Lung-RADS v2022 calculator - screening LDCT categories
• PI-RADS v2.1 calculator - prostate MRI assessment
• Breast MRI screening eligibility - ACS/ACR high-risk criteria
• All radiology guideline calculators
• How MyRadAgent writes recommendations into your reports

The short version

Confirm your non-patient-facing status, expect Promoting Interoperability to be reweighted onto Quality, and pick measures from the dose-safety and incidental-follow-up clusters. Then make your recommendations boringly consistent - that is what gets abstracted, and it is what makes the report better regardless of the programme.