Background
A specific use of coded inpatient data is the identification of hospital acquired complications (HACs), among them procedural (surgical) complications.1 The accuracy of coded data is contentious where funding mechanisms reduce hospital reimbursement when HACs occur.2
Aims
The aim is to assess conceptual aspects of coding accuracy, specifically the impact on accuracy of clinical coding guidelines: the Australian Coding Standards (ACS) 1904 (Procedural Complications) and 0048 (Condition Onset Flag [COF]).
Methodology
All clinical coders in Western Australian (WA) were invited to participate in a voluntary anonymous survey where they were asked to translate diagnostic statements into code: the statements focus on concepts in ACS 1904/ 0048. Answers were scored against benchmark answers provided by the WA Clinical Coding Authority (WACCA).
Results
The response rate was approximately 30.0%, final sample size (N)=59. Mean accuracy is significantly lower than a hypothesised benchmark of 90.0% (M=62.7%, SD=16.2, 95% CI=58.5%, 66.9%, p < .001). Diagnosis coding accuracy, governed by ACS 1904, is lower still (M=48.4%, SD=22.0, 95% CI=42.7%, 54.1%). Misconceptions of ACS 1904 are evident. On a single common complication, postoperative haemorrhage, clinical coding practice is significantly divided (63% to 37%). Similarly high rates of divergence were observed for other complications.
Conclusion
Clinical coders have different understandings of the rules for interpreting common diagnostic terms. These conceptual issues emanate from interpretations of ACS 1904 and intrinsic ambiguities in national coding advice. Resulting coding practices are strikingly divergent. The underlying conceptual issues require resolution at national, state jurisdictional, and local hospital levels.