Impact of auto-text documentation on ICD-10-AM code assignment in cardiothoracic surgery: An Australian...

Clinical documentation facilitates the delivery of safe, effective, and timely patient care. Failures in communication or inadequate documentation can contribute to poor patient outcomes. Multiple barriers exist to producing clinically meaningful and coding-compatible documentation. To address this, a collaborative approach was undertaken between clinicians, coders and clinical documentation specialists (CDS) at our institution.

To evaluate whether condition-specific auto-text improves documentation clarity and increases the capture of ICD-10-AM coded postoperative conditions.

A retrospective cross-sectional comparative audit was conducted at a tertiary Australian hospital. The analysis compared all patients discharged from a cardiothoracic unit in May 2023 (post-implementation) to those discharged in May 2022 (pre-implementation). We designed and implemented 12 condition-specific auto-texts, and a targeted clinician education programme. Documentation clarity was assessed using an ambiguity scoring framework.

A total of 80 patients were analysed (45 pre-implementation; 35 post-implementation). Documentation ambiguity improved significantly following the introduction of auto-text, with a mean reduction of 0.8 points on a 5-point scale (p < 0.01). Lower ambiguity scores were associated with increased likelihood of coding, with each one-point increase in ambiguity reducing coding odds by 66% (p < 0.01). The odds of clinical coding increased by 110% with the introduction of auto-text (odds ratio 2.1, 95% Confidence interval 1.3 to 3.3; p < 0.01); the mean number of diagnoses coded per patient increased from 9.36 to 13.3 (p < 0.01). While the mean Queensland Weighted Activity Units increased from 7.32 to 9.58, this was not statistically significant (p = 0.5).

This preliminary evaluation suggests that clinician-led auto-text tools may reduce documentation ambiguity and improve the capture of coded patient complexity. As a low-cost and scalable intervention, this approach has the potential to support more accurate clinical documentation and hospital reporting. Future research should explore the sustainability of these strategies and their interaction with emerging artificial intelligence tools in clinical documentation.Implications for health information management practice:Clinician-led auto-text tools, supported by CDSs, may provide a sustainable, low-cost practical strategy to reduce ambiguity, improve the capture of patient complexity and optimise hospital data integrity within existing clinical workflows.