B54-41 Improving Risk Adjustment Through Present-On-Admission Comorbidity Documentation in Pulmonary Inpatients: A Qi Initiative

Abstract Rationale Observed to Expected (O:E) ratios for length of stay (LOS) are key Health System quality metrics tied to quality of care, reimbursement and benchmarking. These rely on accurate documentation of present-on-admission (POA) comorbidities, which are often undercaptured in pulmonary patients with complex trajectories and inter-service transfers. Missing POA data underestimates expected LOS and inflates O:E ratios, misrepresenting care quality. Methods We partnered with Michigan Medicine’s Revenue Cycle, Clinical Documentation Integrity, and key pulmonary stakeholders to develop a pulmonary-specific POA documentation tool. Using Vizient®’s LOS model, we reviewed MS-DRGs frequently encountered on our pulmonary medicine service that exhibited high O:E LOS. We then identified diagnoses that had a substantial impact on expected LOS within these DRG when documented as POA. This list was refined through iterative discussion to ensure clinical relevance. The finalized panel was embedded into Epic as a pre-populated SmartList within the Discharge Navigator, allowing clinicians to efficiently document applicable POA comorbidities as part of their standard admission and discharge workflow. We initially piloted this tool in April 2025 with inconsistent utilization. We launched a formal QI initiative on October 1, 2025. This included: faculty and APP education at monthly meetings, workflow document and Epic visualization, direct email outreach to clinicians, assignment of navigator responsibility to APPs for all discharges from our service, and monitoring navigator usage pre- and post-intervention via dashboard (Tableau) analytics. Results In preparation for this QI initiative, we retrospectively applied the Vizient® risk adjustment model to 10 randomly selected patients admitted to our pulmonary medicine service with an O:E LOS ≥ 1. We compared expected LOS using usual documentation alone versus the addition of appropriate POA comorbidities. The addition of 4-6 relevant POA diagnoses per patient led to a statistically significant increase in expected LOS, from a mean of 9.82 days to 13.23 days. This represents an average increase of 3.41 days (SD = 3.25, p = 0.0094). Following the formal QI rollout on October 1, 2025, use of the pulmonary-specific comorbidity navigator increased from 13% (8/61 discharges) in September 2025 to 79% (38/48 discharges) in October 2025. Quarterly quality data metrics for LOS remain pending. Discussion Our intervention shows that provider education, EHR tools, and workflow alignment can improve POA comorbidity documentation. Targeted diagnosis selection and role-based responsibility enabled sustained adoption with minimal burden. We aim to expand to the medical ICU and evaluate downstream effects on LOS and mortality. This abstract is funded by: None