9.4 Clinical Indicators and How to Source Them

9.4 Clinical Indicators and How to Source Them

Quick Answer: List patient-specific findings with their source in the record—assessments, diagnostics, treatments, medications, trends, and ancillary notes. The number and type vary. You do not diagnose; the provider decides significance. The 2026 brief allows nationally recognized clinical definitions or criteria only if they sit neutrally at the end of the template.

A query without clinical indicators is a suggestion. A query with indicators that you interpreted into a diagnosis is a leading suggestion. The 2026 ACDIS/AHIMA brief treats clinical indicators as documentation elements that support a more complete or accurate diagnosis or procedure, or that identify a reported diagnosis that lacks support and needs clinical validation. They must be specific to the patient and episode of care, free from subjective interpretation by the query professional, and accompanied by a citation of where they were found.

This is Domain III skill, not filler. Exam items will ask whether a listed finding is an indicator, whether your wording crossed into diagnosis, and whether a copied criteria table at the bottom of a template is allowed. Get the sourcing habits right and the formats in the previous section become much easier to keep nonleading.

Where indicators come from

Relevant indicators may be sourced from anywhere in the current encounter and, when clinically pertinent, from a prior record—but prior documentation cannot be the sole basis for a query, and code assignment still follows the current encounter. Typical inpatient sources include:

• Assessments and provider impressions: ED notes, H&P, progress notes, consults, operative notes.
• Diagnostics: laboratory values, imaging reports, pathology, cultures, tracings.
• Treatments and procedures: oxygen devices, ventilation, dialysis, debridement, transfusions, emergency interventions.
• Medications: drugs and dosing that imply a clinical problem (vasopressors, insulin infusions, broad-spectrum antibiotics, diuretics), cited as ordered or administered facts, not as your conclusion.
• Trends: serial creatinine, hemoglobin, oxygen saturations, weights, or mental-status scores that change across the stay.
• Ancillary notes: nursing assessments, nutrition, wound care, respiratory therapy, rehabilitation, care management. These notes are legitimate indicator sources even when those professionals cannot independently establish the medical diagnosis you hope to code.

Quote marks belong on verbatim pulls (a radiology sentence copied unchanged). Paraphrase is acceptable when you still cite the location and do not change the meaning. Preserve source formatting of an already-emphasized abnormal lab if policy allows; do not add new highlighting to steer the eye toward the finding you like.

Sourcing rules that show up on test items

Cite location. "Nursing admission assessment 09/12" or "progress note, Dr. Lee, hospital day 3" lets the provider go look. An unsourced stack of numbers looks like a CDI opinion.

Do not subjectively interpret. Report the finding as recorded. If the flowsheet lists a heart rate of 132, write that rate and the source; do not insert "tachycardia" as if you diagnosed it. If a hemoglobin is reported as 7.4 g/dL, list the value and date; do not write "anemia" as an indicator you coined. The query professional must not insert diagnoses or private wording into the body of the query for findings not yet identified by a provider.

Keep indicators relevant to the question. A query about the type of heart failure does not need an unrelated skin finding unless that finding actually supports the clarification. Dumping the entire problem list is not sourcing; it is noise that can look like pressure.

No reimbursement or quality bait in the indicator block. Relative weights, CC/MCC labels, PSI numbers, and mortality-index language are not clinical indicators. They are influencing information. Leave them out of titles, stems, and option lists.

Number and type vary; the provider decides

There is no official quota of indicators. A single definitive pathology interpretation may support a confirmation query. A validation query on a heavily documented but weakly supported diagnosis may need a cluster of contradictory findings. Organizations, payers, and specialty societies may publish clinical criteria, but the treating provider makes the final determination of what indicators define a diagnosis. CDI does not "meet criteria" on the provider's behalf and then demand a signature.

That rule is why indicator lists should look like evidence, not like a completed scoring sheet with the answer circled. You may still need enough facts that a reasonable clinician can see why the record is unclear. Thin queries (one mildly abnormal vital sign and a hoped-for MCC) fail both compliance and credibility. Bloated queries that interpret every abnormality as the same diagnosis fail the no-subjective-interpretation test.

If the provider answers that the findings are not clinically significant, that is an answer. If the provider selects unable to determine, consider escalation case by case. If later notes conflict with a query response that already entered the record, query again rather than silently picking the version that groups better.

Neutral clinical definitions at the end of the template

A 2026 change candidates must use: it is permissible and nonleading to include clinical definitions or diagnostic criteria in the query template. Nationally recognized or standardized organizational criteria—including hyperlinks—may be placed at the bottom or end of the template so the provider can refer to widely accepted language when choosing a response. Examples of the kind of criteria organizations place there include Kidney Disease: Improving Global Outcomes (KDIGO) AKI staging, Academy of Nutrition and Dietetics / American Society for Parenteral and Enteral Nutrition (ASPEN) malnutrition characteristics, chronic kidney disease staging, and pressure-injury staging, each with a source citation.

The organization decides whether to include criteria at all. When they appear, they must be neutral: not highlighted, not bolded as a preferred stage, and not captioned as the answer you want. Criteria at the end are a reference appendix, not a scored rubric the provider is instructed to match. Putting KDIGO language in the stem as "your creatinine meets stage 3, agree?" is not what the brief allows.

Prior records, technology, and inpatient discipline

Prior-encounter facts (a last-year echocardiogram, a chronic steroid dose, a previous amputation) may support a query when they help explain the current picture and when the current record also contains supporting information. They do not let you code last year's diagnoses into this MS-DRG without current provider documentation. They do not let you skip sourcing. They do not turn a prospective outpatient problem-list cleanup into an inpatient concurrent review.

Automated CDI suggestions must source indicators the same way a human would. An alert that dumps an unsourced criteria table beside a highlighted MCC option is still a query, and it is still noncompliant. Human reviewers who click "send" on that alert own the brief's requirements.

Build the habit: pull the facts, cite them, resist labeling them, ask a format-appropriate question, and let the provider decide. That is how clinical indicators earn their place in the health record rather than in a leading worksheet.