15.2 Principal Diagnosis Selection

15.2 Principal Diagnosis Selection

Quick Answer: Principal diagnosis (PD) is defined in the Uniform Hospital Discharge Data Set (UHDDS) as the condition established after study to be chiefly responsible for occasioning the admission of the patient to the hospital for care. Circumstances of admission always govern. If two diagnoses equally meet that definition, either may be sequenced first unless the Alphabetic Index, Tabular List, or another coding guideline directs otherwise. PD selection is not the same decision as medical necessity of the inpatient setting.

Section II of the ICD-10-CM Official Guidelines is the inpatient (non-outpatient) rule set for this choice. CDI specialists live in Section II even when they never open an encoder. A working MS-DRG, a query about “why the patient is here,” and a medical-necessity review all collapse if the chart’s headline diagnosis is the wrong condition.

The UHDDS sentence, unpacked

Three phrases in the UHDDS definition do almost all of the exam work:

After study. The principal diagnosis is not automatically the emergency-department chief complaint, the first diagnosis the intern typed in the history and physical (H&P), or the reason listed on the inpatient order. Study includes the workup that happens after arrival: imaging, laboratories, consultant notes, operative findings, and the attending’s synthesis at discharge. Chest pain can be the presenting problem; non–ST-elevation myocardial infarction (NSTEMI) established after serial troponins and electrocardiograms can be the condition that occasioned admission once study is complete.

Chiefly responsible. Many conditions are present. Only one occupies the principal-diagnosis slot. The test is which condition occasioned—caused the need for—the admission for care, not which condition is the most interesting comorbidity, the best teaching case, or the highest relative weight.

Occasioning the admission. This is a reason-for-being-in-the-hospital test. A patient may have severe protein-calorie malnutrition that is an MCC and still have been admitted because of acute hypoxemic respiratory failure. Malnutrition can be a reportable secondary diagnosis (15.3). It is not PD unless it is the condition that brought the patient in.

The guidelines tell you to review the entire record to determine the specific reason for the encounter. “Encounter” in this usage includes hospital admissions. Pulling PD from a single note—especially a copy-forward problem list—is how incidental conditions become fake headlines.

Circumstances of admission always govern

Section II opens with a sentence CDI should be able to quote in a reconciliation: the circumstances of inpatient admission always govern the selection of principal diagnosis. Circumstances include what was going on at the time of the admission decision, what diagnostic workup was directed at that problem, and what therapy the team actually provided for it.

That sentence blocks several bad habits:

• Sequencing from problem-list order (“pneumonia is line 4, so it cannot be PD”).
• Sequencing from consultant prestige (“cardiology wrote more notes, so heart failure must be PD”).
• Sequencing from CMI (“the MCC-tier family pays more, so that diagnosis is PD”).
• Sequencing from day-1 working DRG as if it were locked. A working DRG is a concurrent hypothesis. Final PD follows the completed study and the circumstances of admission, not the first grouper run.

Symptoms, then the established cause

When a patient is admitted with a symptom, sign, or ill-defined condition, and study establishes a related definitive diagnosis, the definitive diagnosis is sequenced as PD—not the symptom. Admission for syncope that is then shown to be complete heart block is an admission for the block. Admission for abdominal pain that is then shown to be acute cholecystitis is an admission for cholecystitis. Chapter 18 symptom codes are not PD when a related definitive diagnosis has been established. Uncertain diagnoses, integral signs and symptoms, and the “with” convention continue in the next Domain VI chapter; the PD point here is simpler: do not leave the presenting symptom in the principal slot after study has named the cause.

Original treatment plan not carried out

If the patient is admitted for a procedure or treatment that is not carried out, the condition that occasioned the admission remains PD. Cancelled surgery for personal reasons, or a planned intervention abandoned because of a new finding, does not orphan the original reason for admission. The new finding may be an additional diagnosis; it becomes PD only if it, not the original condition, is what actually occasioned admission after study.

Two diagnoses that equally meet the principal-diagnosis definition

The guidelines address the unusual case in which two or more diagnoses equally meet the UHDDS test as determined by the circumstances of admission, the diagnostic workup, and/or the therapy provided. In that situation, any one of those diagnoses may be sequenced first unless the Alphabetic Index, Tabular List, or another coding guideline provides sequencing direction.

That rule is frequently inverted on exam items and in hallway arguments:

When the classification is not silent, follow it. Many chapter-specific guidelines and “code first” / “use additional code” notes are sequencing instructions. Those notes take the case out of the “either may be first” bucket. CDI’s job in an equal-meeting chart is to confirm both conditions truly occasioned admission (both treated, both part of the reason for coming in) and then to look for an actual guideline or Index/Tabular instruction before anyone starts talking about case mix.

Worked encounter: equal meeting

Ms. Chen, 68, is admitted in respiratory distress. After study, the attending documents acute exacerbation of chronic obstructive pulmonary disease (COPD) and acute decompensated heart failure with reduced ejection fraction. Both conditions were present at the time of the inpatient order. Both received hospital-level therapy (bronchodilators, steroids, diuresis, oxygen titration). Both explain why she could not go home. No chapter-specific note forces one ahead of the other for this pairing. Either may be sequenced as PD. Choosing the family with the higher relative weight because “finance is watching CMI” is not a guideline. Choosing one after the attending clarifies that only the heart-failure decompensation actually drove the admission is a guideline—because they are no longer equal.

Worked encounter: not equal

Mr. Diaz is admitted for a ground-level fall with a femoral neck fracture requiring arthroplasty. He also has chronic atrial fibrillation on his home beta blocker, continued without a new workup. The fracture, not the atrial fibrillation, occasioned admission. Atrial fibrillation may be a reportable secondary diagnosis if it was evaluated or treated. It is not a second PD candidate, and it does not “equally meet” the definition just because it is serious.

Principal diagnosis versus medical necessity of setting

Domain I already separated these drawers; Domain VI is where the coding definition has to sit next to that payment rule without merging them.

Principal diagnosis answers: Which condition, after study, occasioned this inpatient admission?

Medical necessity of setting (Medicare’s Two-Midnight Rule and related inpatient-order policy) answers: Was inpatient hospital care the correct setting—generally, did the physician expect medically necessary hospital care spanning at least two midnights, with the record supporting that expectation, unless a CMS-recognized exception such as an inpatient-only procedure applies?

They talk to each other because reviewers read PD as the headline reason the patient needed a hospital. A PD that is a brief, resolved event with no ongoing hospital-level care makes the two-midnight story harder to defend. A PD that clearly required hospital-level monitoring makes that story easier to write. They are still different decisions:

• A correctly sequenced PD does not automatically prove the Two-Midnight expectation was documented or met.
• A supported two-midnight expectation does not tell you which of two equal conditions is PD.
• An MCC on a secondary diagnosis does not repair a PD that never occasioned admission, and it does not repair a missing setting expectation.
• Observation versus inpatient is a status question. Section II PD rules apply to the inpatient encounter. Do not import outpatient “first-listed” logic to fix an IPPS chart, and do not use PD selection as a back-door status change.

CDI contributes by making the after-study reason for admission explicit, querying when two headlines compete and only one occasioned admission, and refusing to treat MS-DRG severity as a substitute for setting support. Utilization review and the physician advisor own the status determination; CDI owns the documentation that makes both the UHDDS test and the setting story readable.

What CDI should do when PD is unclear

1. Reconstruct circumstances of admission: why the inpatient order was written, what the ED and admitting notes were targeting.
2. Read after-study documents: imaging, pathology, consultant conclusions, discharge summary.
3. Separate the condition that occasioned admission from comorbidities that were along for the ride.
4. If two conditions truly share that role, check Index, Tabular, and chapter guidelines before calling it a free choice.
5. Query with clinical indicators when the record is conflicting, incomplete, or uses a symptom as if it were the established cause. Do not query “please make X principal so the DRG is Y.”

Traps: treating admitting diagnosis as PD; treating the highest-weight diagnosis anywhere in the chart as PD; treating equal-meeting as a CMI election; treating PD as proof of Two-Midnight compliance; treating a day-1 working DRG as final sequencing. Principal diagnosis is a reason-for-admission definition with a published tie rule. Everything else is a different chapter.