The Sepsis Query That Arrived After Discharge
A 71-year-old woman was admitted with a two-day history of productive cough and confusion. Admission vitals: temp 38.9 C, HR 112, BP 92/58, RR 24, SpO2 91% on room air. Admission lactate 3.1 mmol/L, WBC 17.2, and a chest X-ray reading 'right lower lobe infiltrate.' The hospitalist documented 'pneumonia, community-acquired; rule out sepsis' and ordered fluids and IV ceftriaxone plus azithromycin. The discharge summary lists principal diagnosis 'pneumonia' with no mention of sepsis, and the coder has already submitted the claim (DRG for pneumonia without MCC). Two weeks later, the CDI nurse reviews the record and sends a post-bill physician query asking whether sepsis should be added, noting the SIRS criteria and lactate. The physician responds 'yes, sepsis present on admission' and the query is added to the chart.
1. Should the post-discharge query result be used to change the coded DRG on this already-submitted claim? Walk through the reasoning.
Read the worked answer
No - not on the strength of this query as constructed. Three problems compound here:
- Timing and authority. A query generated after the claim is submitted, based on a retrospective chart review, is not a substitute for clinical documentation created during the encounter. If the physician genuinely treated sepsis (fluids, cultures, antibiotics, lactate), the compliant path is a clarifying query tied to documented clinical indicators - not a leading query that names the answer.
- Leading structure. 'Should sepsis be added?' telegraphs the desired response. A compliant query presents the clinical indicators (lactate 3.1, SIRS vitals, hypotension) and asks an open question: 'Please clarify the clinical significance of the lactate and hypotension in the context of this admission.'
- POA and MCC logic. Even if sepsis is clinically supportable, the query must establish present-on-admission status and the causal link to the principal diagnosis. 'Rule out sepsis' documented at admission, by itself, is not a confirmed diagnosis.
Practical answer: The claim should not be reopened solely because a retrospective query was answered. If the facility believes the original coding was wrong, the correct route is a formal coding audit with physician attestation based on contemporaneous documentation, then a corrected claim if supported. Answering a leading post-bill query and silently rebilling is a compliance risk, not a coding fix.
2. The physician pushes back: 'The patient clearly had sepsis - the lactate was 3.1. Just code it.' How should the CDI reviewer respond, and what documentation would actually support the change?
Read the worked answer
The reviewer should separate clinical plausibility from codeable documentation.
- Acknowledge the clinical picture. Lactate 3.1 with hypotension and SIRS vitals is a legitimate sepsis concern. The reviewer is not disputing the medicine.
- Explain the coding standard. A diagnosis must be documented by the provider and supported by the record. The provider's own words at the time - 'rule out sepsis' - are not a confirmed diagnosis. Retrospective agreement does not retroactively create documentation that existed at the time of care.
- State what would support the change:
- A provider attestation that explicitly states the patient had sepsis present on admission, with the supporting indicators named.
- Evidence the sepsis was treated (IV fluids, cultures, antibiotics, repeat lactate).
- A clear link between sepsis and the principal diagnosis (pneumonia as the source).
- A properly worded, non-leading query documented in the record.
- Offer the compliant alternative. If the physician will not attest, the reviewer should document the query, the response, and the rationale, and leave the coding as submitted. The reviewer's job is to get accurate documentation - not to manufacture a DRG upgrade.