Sepsis, Severe Sepsis, and a 'Resolved' Discharge Summary
A 71-year-old man is admitted from the emergency department at 22:10 with two days of productive cough and fever. ED vitals: temperature 39.1 C, heart rate 118, blood pressure 84/48 mmHg, respirations 26, SpO2 90% on room air; lactate 4.1 mmol/L. Chest x-ray shows a right lower lobe infiltrate. Blood cultures are negative; sputum culture grows Streptococcus pneumoniae. He receives 35 mL/kg of crystalloid without a sustained response and is placed on norepinephrine at 01:15, weaned off after seven hours. Ceftriaxone and azithromycin continue through day 4. The day-one hospitalist note reads, 'Severe sepsis secondary to community-acquired pneumonia.' The discharge summary lists 'Pneumonia; sepsis, resolved.' The coder has drafted A41.9, J18.9, and R65.20, sequenced with sepsis first.
1. Evaluate the draft. Are the diagnosis codes and their sequence supportable, and does the positive sputum culture justify a more specific organism code?
Read the worked answer
Sequence holds, with one necessary caveat on organism.
- The physician linked sepsis to pneumonia in a single diagnostic statement, so the generalized guideline for sepsis applies: when sepsis is the reason for admission, the systemic infection is sequenced first and the localized infection follows. A41.9, R65.20, J18.9 is correct order (R65.20 carries a use additional code instruction for the underlying infection).
- Vasopressor-dependent hypotension after adequate fluid resuscitation is organ dysfunction, so severe sepsis (R65.20) is supported by the record. Do not assign R65.21 (septic shock) unless the physician documents shock; do not infer it from a vasopressor. Do not separately code the hypotension - it is the defining dysfunction captured by R65.20.
- Organism: a positive sputum culture is laboratory data, not a physician diagnosis. A40.3 (pneumococcal sepsis) and J13 (pneumococcal pneumonia) require the provider to document the organism. A coder may not assign them unilaterally.
- Judgement call on querying: organism specificity does not change the DRG family here (871 either way), so a query is optional rather than required. If sent, it must ask for clarification of the organism, not announce it.
2. The discharge summary says 'sepsis, resolved.' A colleague proposes deleting sepsis because it resolved. How do you handle it?
Read the worked answer
Do not delete it, and do not query for its presence.
- Resolution at discharge does not erase a condition that was present on admission, required ICU-level support (vasopressors for seven hours), and drove length of stay and resource use. UHDDS criteria are met.
- The day-one hospitalist note is a valid, timely diagnostic statement. Querying a diagnosis that is already clearly documented and treated is inappropriate and looks like a payment-driven query.
- The phrase 'resolved' describes clinical status, not codeability. The discharge summary is one data source, not the controlling one.
- If the attending later states in a signed addendum that sepsis was never present, that changes the analysis - clarify with a non-leading query that cites the conflicting documentation and offers options including 'other.'
- Record your rationale in the audit trail (which notes were reviewed and why sepsis was retained). This is the record you produce at a RAC or OIG review.
3. What is the DRG and reimbursement consequence of each version, and what compliance exposure does deletion create?
Read the worked answer
Materially different payment, plus real compliance risk.
- With sepsis: principal A41.9 plus R65.20 (an MCC) groups to the septicemia/severe sepsis family - DRG 871 when no mechanical ventilation exceeds 96 hours, 870 when it does.
- Without sepsis: principal J18.9 with no remaining MCC groups to DRG 195 (simple pneumonia and pleurisy without CC/MCC); it would rise to 193 or 194 only if another CC/MCC is documented. Payment is DRG weight x the hospital's adjusted base rate, so the gap is usually substantial.
- Deleting a documented, treated diagnosis that drives severity is under-coding. It distorts quality and severity reporting and creates False Claims Act exposure in the opposite direction from upcoding - auditors pursue both.
- The reverse error is equally risky: sepsis is one of the most heavily audited diagnoses. Never add it from a lactate or a positive culture alone.
- If a physician disputes the coder, escalate to the physician advisor rather than resolving it in the coding queue.