OASIS Accuracy Improvement Case Study
87% → 99%
OASIS Accuracy
$2.1M
Annual Revenue Recovered
73%
Fewer OASIS-Related Denials
45%
Faster Assessment Turnaround
How we improved OASIS assessment accuracy from 87% to 99%, directly increasing reimbursement accuracy and reducing audit risk for a large home health provider.
The Challenge
A large home health provider with over 3,200 active patients across five states was experiencing significant revenue leakage due to inaccurate OASIS assessments. Their internal quality audits revealed an 87% accuracy rate on OASIS responses—a figure that, while not unusual in the industry, was costing them approximately $2.8 million annually in under-coded reimbursements and denial-related rework. The Patient-Driven Groupings Model (PDGM) made OASIS accuracy more critical than ever. Under PDGM, case-mix weights are directly derived from OASIS responses, meaning even small inaccuracies in functional status, clinical severity, or comorbidity documentation could shift a patient into a lower payment group. The agency estimated that 18% of their episodes were being classified into lower HIPPS codes than the clinical documentation supported. Additionally, the agency had received a Targeted Probe and Educate (TPE) notice from their MAC, putting them under heightened scrutiny for OASIS-related billing accuracy. The urgency to improve was both financial and regulatory.
Our Solution
AI-Powered OASIS Cross-Referencing We deployed our AI review engine to cross-reference every OASIS assessment against the full clinical record—visit notes, physician orders, medication lists, and prior assessments. The system identifies discrepancies between documented clinical findings and OASIS responses, flagging items where the coded response does not align with the supporting documentation. This automated first-pass review catches 85% of accuracy issues without human intervention. Targeted Clinician Coaching Analysis of error patterns revealed that accuracy issues were concentrated in specific OASIS items: M1800-M1860 (functional status), M1311-M1322 (skin integrity), and GG items (functional abilities). We developed targeted coaching programs for clinicians based on their individual error profiles, focusing on the clinical reasoning behind each OASIS response rather than rote memorization of answer choices. Real-Time Assessment Validation We integrated validation logic directly into the assessment workflow. As clinicians complete OASIS items, the system checks for internal consistency (e.g., a patient documented as requiring maximum assistance with ambulation should not be coded as independent in transfer), and cross-references responses against historical data and the current plan of care. Continuous Quality Monitoring We implemented agency-wide dashboards tracking OASIS accuracy by clinician, branch, and OASIS item category. Monthly accuracy reports with trending data allowed clinical managers to identify emerging issues before they impacted revenue or compliance. The system also tracks the correlation between OASIS accuracy and case-mix weight to quantify the financial impact of each improvement.