14.2 Working DRG versus Final DRG Reconciliation

14.2 Working DRG versus Final DRG Reconciliation

Quick Answer: A working DRG is a preliminary grouping CDI assigns from current inpatient documentation while the stay is open. The final DRG is what coding assigns after discharge from the complete record using ICD-10-CM/PCS, the Official Guidelines, and Coding Clinic. When they mismatch, neither role automatically wins and you never bill the higher relative weight by default. Reconcile to the record and the rules. Patterns of mismatch are education opportunities, not a reason to silence disagreement.

The ACDIS handbook asks candidates to reconcile discrepancies between working DRG assignments from CDI staff and final, coded DRGs. This is a Domain V skill because it is a communication process between two professional roles, not a secret button in the grouper. Get the process wrong and you either under-report supported conditions or drift into DRG creep.

Working DRG: preliminary on purpose

A working diagnosis-related group (DRG) is the Medicare Severity DRG (MS-DRG) that concurrent CDI (and often case management) assigns during the stay from documentation that exists today. It is a planning and communication tool:

• Utilization review and case management use it to anticipate length of stay and post-acute needs.
• CDI uses it to see whether the record currently groups with a complication or comorbidity (CC), a major CC (MCC), or neither.
• Quality and physician-advisor workflows may glance at it, but it is not the claim.

Because the record is still moving, the working DRG should change. New imaging, a query response, a procedure, or a consultant note can regroup the case overnight. A working DRG that never changes on a 12-day stay is a red flag that nobody is rereading the chart—not a sign of excellence.

What a working DRG is not:

• It is not a physician order. Physicians document conditions and procedures; they do not assign MS-DRGs.
• It is not a promise of payment. Relative weights can shift at the annual IPPS update, and the grouper still needs complete, guideline-supported codes.
• It is not a license to park an unsupported MCC "so the working DRG looks right."
• It is not automatically more accurate than coding because CDI "saw the patient." Concurrent proximity is useful. Official coding rules still govern the bill.

Assign the working DRG from documented diagnoses and procedures, using the same principal-diagnosis and CC/MCC logic you expect coding to use. If the indicators suggest a diagnosis that is not in the record, query. Do not invent a working MCC from a CDI worksheet.

Final DRG: coded after the record is complete

The final DRG is assigned after discharge (or at the coding complete event) by a coder applying:

• ICD-10-CM diagnosis codes and ICD-10-PCS procedure codes
• The ICD-10-CM Official Guidelines for Coding and Reporting and the PCS guidelines
• American Hospital Association (AHA) Coding Clinic advice
• UHDDS definitions of principal and secondary diagnoses
• Present on admission (POA) indicators
• The MS-DRG grouper, including CC exclusion (IPPS Table 6K) logic that can neutralize a CC or MCC

The final DRG is what the hospital bills (subject to later amendment). It must be defensible to a Medicare Administrative Contractor (MAC), Recovery Audit Contractor (RAC), Comprehensive Error Rate Testing (CERT) review, or Office of Inspector General (OIG) lookback. "CDI's working DRG said so" is not a defense.

Why working and final DRGs legitimately differ

Mismatch is not automatically a failure. Common legitimate reasons include:

Illegitimate reasons also exist: CDI assigned a working MCC that was never documented; coding dropped a well-documented, treated diagnosis without a guideline basis; someone picked a DRG because of relative weight. Those are the mismatches reconciliation is designed to catch.

The mismatch process

Treat reconciliation as a structured conversation, not a scoreboard.

1. Identify the mismatch. Software may flag working versus final MS-DRG, CC/MCC presence, or principal diagnosis.
2. Open the record together (or in a documented handoff). CDI brings clinical indicators and query history. Coding brings guideline, Coding Clinic, and grouper logic, including exclusions.
3. Ask what the documentation actually supports. If a diagnosis is missing, unclear, or conflicting, the next step is a query, not a quiet regroup.
4. Apply official coding rules to the documented conditions. Clinical intuition does not override UHDDS or the Official Guidelines.
5. Record the outcome. Typical endings: coding revises the final DRG; CDI revises the working-DRG logic for next time; both agree the original final DRG was correct; a query goes out before the claim drops.
6. Escalate role-versus-role stalemates to CDI and coding leadership, then to a physician advisor when the issue is clinical interpretation. Escalation is still not "pick the higher weight."

Do this before the claim drops whenever the account is still in a pre-bill window. After billing, corrections follow the facility's amendment and rebilling process. Do not "fix" a paid claim by informal agreement that never touches the coded record.

Neither role automatically wins

This is the exam sentence to memorize in substance. CDI does not automatically win because specialists are clinicians or because they assigned the working DRG at the bedside. Coding does not automatically win because they "own" the bill or because the grouper has already spun. The winner is the assignment that is supported by the health record and by official coding and grouping rules.

Practical implications:

• A coder who ignores a documented, treated MCC without a guideline reason needs to recode, not to lecture CDI about "final means final."
• A CDI specialist who wants an MCC that exists only as a suspected finding, or that a query never placed in the record, needs to stand down or query—not to override coding.
• If two MS-DRGs are both possible depending on principal-diagnosis selection, apply the Official Guidelines on two or more diagnoses that equally meet the definition of principal diagnosis. Do not break the tie with dollars.
• Physician advisors help with clinical questions (is this encephalopathy? was this POA?). They do not get a vote to increase relative weight for its own sake.

Never bill the higher weight by default

When working and final DRGs differ, one of them often has a higher relative weight. Billing the higher one by default is how programs wander into DRG creep: reporting a higher-weighted group than the documentation and rules support. That pattern shows up in OIG work-plan themes, RAC denials, and Program for Evaluating Payment Patterns Electronic Report (PEPPER) outliers. It is also a False Claims Act risk if the claim is false and submitted with knowledge or reckless disregard.

The correct default is accuracy, which might be the lower-weighted DRG. Examples:

• Working DRG assumed acute systolic heart failure as principal; the complete record shows the admission was for a hip fracture and the heart failure was a stable secondary. The lower-weighted orthopedic DRG can be the correct bill.
• Final coding missed documented acute hypoxic respiratory failure that met UHDDS secondary criteria and was not excluded. The higher-weighted DRG can be correct after recoding—because the record supports it, not because it pays more.
• Both assignments rest on an unspecified pneumonia versus a specified organism. If the culture and provider documentation do not support the specified code, you do not "split the difference" by taking the higher weight.

Trap: "We'll take CDI's DRG when it pays more and coding's DRG when that pays more." That sentence is a compliance scenario, not a reconciliation policy.

Education opportunities from patterns

A single mismatch is a case. A pattern is a curriculum.

Track mismatch reasons in categories the team can teach to: principal-diagnosis disagreement, missed or extra CC/MCC, procedure PCS differences, POA/hospital-acquired condition (HAC) effects, unanswered queries, CC exclusion surprises, and late documentation. Then:

• If providers repeatedly leave heart-failure type unspecified, that is physician education (and a concurrent query strategy), not a coding workaround.
• If CDI working DRGs routinely include diagnoses that never appear in the record, that is CDI education on assigning working groups only from documentation.
• If coding routinely omits well-documented secondary conditions that meet UHDDS, that is coder education and possibly a second-level review screen—not a reason for CDI to stop reconciling.
• If CC exclusion or the "with" convention drives most surprises, run a joint guideline huddle. That is cheaper than arguing case by case.

Do not hide mismatches to inflate an "agreement rate." Agreement after honest reconciliation is a quality signal. Forced agreement is a cooked metric. The same discipline Domain IV teaches for query rates applies here: measure to improve, not to punish a role into silence.

Worked mini-scenario

Concurrent CDI assigned a working MS-DRG with MCC based on documented severe protein-calorie malnutrition, with Academy of Nutrition and Dietetics / American Society for Parenteral and Enteral Nutrition (ASPEN)-consistent indicators and a provider diagnosis in the progress notes. Final coding grouped without an MCC, omitting the malnutrition code. Reconciliation opens the record. The diagnosis is present, treated with nutrition support, and not excluded by Table 6K with this principal diagnosis. Coding adds the diagnosis and the final DRG changes—because the record and the rules support it. If, instead, only a dietitian note existed and no provider diagnosis was documented, the correct path would be a query (or no MCC), not a demand that coding "honor the working DRG." Either outcome can generate education: query earlier for malnutrition, or recode documented MCCs more consistently.

Working versus final at a glance

Reconciliation is how CDI and coding stay on the same side of the record. The working DRG is a hypothesis. The final DRG is the coded conclusion. The health record and the coding rules decide which conclusion may be billed.