MDM: data, risk and the final level

Where you are

You have a supported inventory of the problems addressed. Now examine data and management risk, then determine the overall MDM level. Keep your current authorized E/M guidance open. This lesson supplies an original worksheet and worked comparisons; the full category definitions and scoring table remain in your book.

Use three separate headings. Under each, record the qualifying evidence and the corresponding level. Do not carry the answer from one column into another just because the clinical story feels complicated.

Make a data inventory before counting

For every possible data item, record what it is, where it came from and what the reporting clinician did with it. Was an external note reviewed? Was a test ordered or a result reviewed? Did another person supply history? Was a test independently interpreted? Was there a qualifying discussion with another professional?

Then compare those facts with the current book's data instructions. The categories have different requirements. A long list of laboratory values is not automatically a long list of unique tests, and a copied document is not proof that its contents were reviewed.

AMA's FAQ notes that serial results of the same unique test do not become multiple unique tests, and overlapping tests require care. It also distinguishes a unique source from the number of documents received from that source. [em-faq-ama, Amount and complexity of data]

Worked data sorting exercise

Original fictional office record: the clinician reviews three pages from the same external physician in one specialty and group, compares several results of the same laboratory test and documents an assessment. The learner's first inventory lists each page and each repeated result as a separate unique item.

Before assigning a data level, correct that inventory. Page count is not source count, and repeated results are not automatically different unique tests. Write the source once, identify the test and record the actual review. Then apply the current category instructions to the supported items.

This exercise deliberately stops before assigning a final data level. The source and test relationships need to be established before you can safely count them. If a question supplies additional distinct tests or a qualifying separate category, add only what the documentation and instructions support.

Separate result review from independent interpretation

A clinician reading another person's report is doing something different from personally interpreting a test. Your inventory should preserve that distinction. Locate the current definition and exclusions for independent interpretation in your authorized book, including how separately reported work is handled.

Original fictional comparison: Note A says an outside imaging report was reviewed. Note B describes the clinician's own interpretation and findings. These notes provide different evidence. Do not rewrite Note A into Note B merely to reach another data category.

Also distinguish sending a referral from a documented discussion of management or test interpretation. A routing action and an interactive clinical exchange are not interchangeable descriptions. Use the full applicable definition to decide whether the documented event qualifies.

Risk concerns the management decision for this patient

AMA's FAQ explains that risk depends on the patient's circumstances and the clinician's management decisions. A procedure name alone does not establish the same risk for every patient. The MDM meaning of major or minor surgery is not defined by the surgical global-period classification. [em-faq-ama, Risk; Additional E/M questions]

Return to the assessment and plan. Identify the decision being made, the documented patient factors and the clinician's assessment. If the note does not establish the relevant facts, do not create them from a generic list of possible complications.

Original fictional comparison: two patients are considered for a similar intervention. Their records describe different comorbidities and different assessments of the management options. A single risk label copied from the procedure name would discard those differences. Preserve the individual assessment before applying the current definition.

A medication list is not a management explanation

A record may display many current medications. For your worksheet, identify what the clinician decided about treatment today. Was there a documented decision to initiate, continue, adjust or stop therapy? What evidence supports the applicable risk classification in the current guidance?

Do not replace that review with “there is a prescription somewhere in the chart.” The worksheet needs the encounter's management work. Equally, do not assume that a continued plan means no decision occurred; read the documented assessment and action.

This is another reason to keep diagnosis abstraction and MDM separate. The diagnosis may explain the condition being treated, while the management entry shows what was decided at this encounter.

Combine the elements with the two-of-three rule

Under the applicable MDM table, two of the three elements must meet or exceed the selected level. One high element does not make the whole encounter high. [em-faq-ama, Risk of patient management; em-guidelines-ama-2023, page 7]

The following original exercises supply already-verified element classifications so you can practice the combination step. They do not replace the work required to classify a real record.

Record A has moderate problems, low data and moderate risk. Two elements meet moderate, so the overall MDM is moderate. The low data element does not force the whole result to low.

Record B has high problems, low data and low risk. Only one element reaches high. Two meet low, so the overall MDM is low. Do not choose high solely because the problem element is high.

Record C has high problems, moderate data and high risk. Two meet high, so the overall MDM is high. Record D has moderate problems, high data and low risk. Problems and data meet at least moderate, so moderate is supported.

Return to the family before choosing the code

Once the MDM level is established, return to the correct family and status. Use its current entry instructions to select the service. A moderate MDM conclusion does not identify one universal E/M code across every setting.

Check the whole chain: family, status, three element classifications, overall MDM and final entry. If an earlier step changes, revisit the later ones. The worksheet should let another learner see how you reached the answer without guessing what you assumed.

Book drill

Locate the data categories and risk definitions in your current book. Annotate the fictional data inventory with the distinction between unique test, unique source and independent interpretation. Then solve the four supplied element combinations without looking at the explanation, and check the two elements supporting each result.

Checkpoint

Explain why repeated results do not automatically increase unique-test count, why a procedure's global period does not decide MDM risk, and why high problems with low data and low risk does not produce high overall MDM.

Check the combination step

Separate the evidence before combining it. Count only items that meet the applicable data definitions, preserve patient-specific management decisions, and require two elements at the selected MDM level. Then use the correct E/M family and status to complete code selection.

Sources

1. American Medical Association — CPT Evaluation and Management revisions FAQs. Current 2026 page explaining E/M revisions; use current authorized book. Accessed 2026-09-10.
2. American Medical Association — 2023 E/M descriptors and guidelines — historical foundation for the two-of-three rule. 2023 foundational rule, corroborated by January 2026 AMA FAQ; not a 2026 descriptor reference. Accessed 2026-09-12.