Clinical–Management Collaboration for Improving TBI Principal Diagnosis Coding: A Quasi-Experimental Quality Improvement Study

Background Clinical coding in complex diseases depends on the accurate generation, interpretation, documentation, and translation of clinical information. Coding discrepancies may therefore arise across the clinical-information pathway rather than solely during final code assignment. This study examined whether implementation of a disease-specific quality improvement programme for TBI was accompanied by a differential change in principal diagnosis coding accuracy relative to a contemporaneous comparison group. Methods A single-centre, controlled quasi-experimental quality improvement study using a repeated cross-sectional design was conducted in a tertiary hospital in China. In January 2025, a FOCUS-PDCA programme integrating clinical documentation optimization, neuroimaging verification, clinical-to-classification mapping, and multidisciplinary quality control was implemented in the Department of Neurosurgery. The study included 400 patients: 200 TBI patients in the intervention group and 200 patients with nontraumatic intracranial haemorrhage in a contemporaneous comparison group, with 100 patients per group in each study year. Differential changes in principal diagnosis coding accuracy were evaluated using difference-in-differences analysis based on linear probability models (LPM-DID). Results Principal diagnosis coding accuracy increased from 53.0% to 74.0% in the TBI intervention group, while it changed from 89.0% to 86.0% in the contemporaneous comparison group. The adjusted LPM-DID estimate indicated a 22.2-percentage-point differential change in coding accuracy in the intervention group relative to the concurrent change in the comparison group (95% CI, 5.8–38.7 percentage points; P = 0.008). A logistic DID sensitivity analysis yielded findings consistent with the primary analysis. Conclusions The observed differential improvement in principal diagnosis coding accuracy was consistent with an information-pathway approach to coding quality management. The findings suggest that coding quality improvement in clinically complex diseases may benefit from extending quality control beyond final code review to multidisciplinary review across the clinical-information pathway. More reliable standardized coded information may improve the fitness of routinely collected hospital data for case-mix measurement, healthcare quality assessment, DRG/DIP-based management, and data-driven hospital management.