Automated severity classification of neurosurgical complications in routine clinical practice using a modified Landriel-Ibáñez framework

Background Standardized classification of surgical complications is essential for quality assurance, transparent complication reporting and patient safety. Although therapy-oriented grading systems such as the Landriel-Ibáñez classification are well established in neurosurgery, severity grading remains largely dependent on manual physician assessment, potentially limiting reproducibility and integration into routine clinical workflows. Methods A rule-based framework was developed and implemented within a prospective hospital-wide neurosurgical complication registry. Rather than requiring direct physician assignment of complication grades, the framework prospectively records predefined objective clinical descriptors, including treatment modality, complication-related intensive care unit (ICU) treatment, multiorgan failure, and complication-attributable death. Based on transparent decision rules, complication severity is assigned automatically according to a modified Landriel-Ibáñez classification. The framework was integrated into a hierarchical electronic registry combining automated hospital information system data import with structured manual documentation and continuous quality assurance procedures. Results The developed framework enables fully automated severity assignment from the documented clinical descriptors using predefined objective decision rules without requiring manual grading by the treating physician. Severity classification is performed in the background following completion of mandatory documentation fields and is fully integrated into the prospective documentation workflow. The algorithm assigns one of seven severity grades (Ia, Ib, IIa, IIb, IIIa, IIIb, and IV), with Grade IV representing death attributable to a documented complication. The framework combines standardized complication documentation, automated severity assignment, and continuous documentation quality monitoring within a single electronic registry architecture, providing a transparent rule-based approach to severity classification in routine clinical practice. Conclusions We present a rule-based framework for automated severity classification of neurosurgical complications based on a modified Landriel-Ibáñez classification. By replacing direct physician assignment of severity grades with predefined objective clinical descriptors, the framework standardizes the rule-based translation of documented clinical descriptors into severity grades while remaining integrated into routine clinical workflows. Its modular design may support prospective complication reporting, quality assurance, and future multicenter registry-based research.