4.2 Working vs. Final DRG Assignment & Optimization vs. Maximization

4.2 Working vs. Final DRG Assignment & Optimization vs. Maximization

A working DRG is a provisional classification based on the record available at a particular review. The final coding DRG reflects the completed coding review used for billing. A difference between them is a signal to investigate: it can reflect new information, an input error, a coding interpretation, or an unresolved documentation issue. The higher of the two is not automatically correct.

Give the working DRG an explicit information date

During concurrent review, the CDI practitioner can record a working DRG, the diagnoses and procedures supporting it, the grouper version, and the time of review. The organization defines who performs this work and its review schedule. An initial review within a chosen number of hours is an operational policy, not a universal coding requirement.

A useful concurrent record distinguishes three things:

Keeping these separate prevents a pending query's hoped-for answer from becoming a documented diagnosis in a forecast. Update the working result when the record changes, and retain enough information to understand the earlier result.

Working classifications may support utilization discussions and financial forecasts. A group's geometric mean length of stay is an aggregate benchmark; it does not prescribe an individual patient's discharge date or establish what care that patient needs. Discharge planning follows the patient's condition, treatment, function, and safe follow-up needs. Similarly, an interim forecast based on incomplete active stays must be labeled as a forecast, rather than presented as a finalized discharge-cohort CMI or actual reimbursement.

Understand what final coding does and does not establish

The coding team reviews the relevant encounter documentation, applies the applicable code set and guidelines, resolves necessary clarification through the organization's process, and groups the resulting inputs for billing. Job assignments, credentials required by the employer, and release authority are organizational matters; calling a result final does not create a universal requirement that only a named certification can assign it.

The full record matters because the principal diagnosis is established after study, secondary diagnoses must meet reporting requirements, and procedure coding depends on what was actually performed. Discharge does not make every record complete at the same instant, nor does an operational chart-closure label prohibit all subsequent clarification.

The 2026 AHIMA/ACDIS query practice brief addresses retrospective queries and the incorporation of authenticated responses into the health record. Organizations should define approved response locations, time frames, escalation, and handling of unanswered queries. A compliant response in the permanent record need not universally be copied into the discharge summary; a later conflict may still require clarification.

A final coding result is also subject to the applicable correction, review, and claim processes. It is not an unchangeable legal judgment. As discussed in the preceding section, a DRG and its weight are inputs to payment. They do not alone establish a remittance amount, a quality score, or the outcome of a payer review.

Explain a variance using its actual cause

The ACDIS reconciliation practice example describes comparing CDI and coding results and reviewing code, POA, and DRG assignments. It also describes different organizations' review approaches; it does not prescribe one universal workflow. A practical reconciliation sequence is:

1. Compare the same encounter and grouper release. Identify the exact time and input set represented by each result.
2. Compare principal diagnosis, secondary codes, procedures, POA, and other relevant inputs. A matching DRG can conceal a coding disagreement, and a different DRG does not prove an error.
3. Determine whether the record changed or whether the two reviewers applied the same information differently.
4. Resolve the specific documentation or coding question using the applicable source. Regroup the corrected inputs and record the reason for the final decision.

Examples illustrate different causes:

• New documentation: A provider answers a supported query after the last concurrent review. If the authenticated answer supports a reportable diagnosis, final coding can differ. The answer might increase, decrease, or leave the DRG unchanged.
• Late diagnostic findings: A pathology report returns after the concurrent review. The finding prompts the appropriate provider-significance and coding review; it does not automatically authorize every diagnosis in the report. A colon finding changing from benign to malignant does not, by itself, prove that the case moves to a different MDC.
• Same record, different interpretation: Both reviewers see the same diagnoses but select different principal diagnoses. Revisit admission circumstances and the applicable sequencing provision. Resolve the rule rather than choosing whichever result pays more.
• Technical input difference: A completed procedure was omitted from one abstract, or one result used an inapplicable release. Correct the input or version before treating the difference as physician-documentation impact.

These are possible causes, not a ranked list of the most frequent causes at every hospital. Establish local frequencies from reviewed cases and clearly defined categories.

Define optimization by accuracy

The goal of compliant optimization is to represent the encounter completely and accurately under applicable requirements. A correct change can increase payment, reduce payment, or have no payment effect. Intentional addition of unsupported codes, inappropriate omission, or distorted sequencing undermines that goal even when an organization's metric improves.

The AHIMA Standards of Ethical Coding call for supported coding, clarification when needed, and refusal to misrepresent data for reimbursement or reported outcomes. Apply current coding and query guidance to the specific encounter. Practical boundaries include:

Query quality is assessed from the entire record and communication. A particular oxygen flow rate or test result alone does not establish a universal diagnostic or fraud threshold. The query must be clinically supported and neutral, without directing a response through reimbursement or quality incentives.

Separate a professional violation from a legal finding

A noncompliant query needs correction and may warrant compliance review. That finding alone does not establish that a false claim was submitted or that every element of a civil or criminal statute is satisfied.

The False Claims Act, 31 USC 3729, addresses specified conduct including knowingly presenting a false claim and knowingly using a false record material to a false claim. Its knowledge definition includes actual knowledge, deliberate ignorance, and reckless disregard; specific intent to defraud is not required. Liability depends on the applicable statutory provision and facts, rather than the word optimization or a DRG mismatch alone.

The DOJ overview explains damages and inflation-adjusted civil penalties, as well as private qui tam actions. The statute provides a general three-times-damages rule and specified circumstances for reduced damages. For an actual penalty question, use the applicable violation and assessment dates and 28 CFR 85.5; an undated dollar range is inadequate.

18 USC 1347 is a separate criminal healthcare-fraud provision addressing knowing and willful execution or attempted execution of the specified schemes. It provides a general maximum of 10 years' imprisonment, with greater potential penalties when the violation results in serious bodily injury or death. These are statutory possibilities, not automatic sentences for a coding disagreement.

Professional discipline is another separate process. The AHIMA certification disciplinary policy distinguishes specified automatic grounds from discretionary action and provides review and appeal procedures. An ethical violation does not automatically mean permanent credential revocation in every case. A CDI practitioner should preserve facts, use the appropriate reporting channel, and correct the documentation or coding problem without declaring an unadjudicated legal outcome.