Basic Physician Trainee at Flinders Medical Centre, South Australia.
Clinical Coding Accuracy In A General Medicine Department And Its Effects On Hospital Funding
Teng Yuan Kang, Flinders Medical Centre, Adelaide, Australia
Udul Hewage, Flinders Medical Centre and School of Medicine, Flinders University, Adelaide, Australia
Patrick Russell, Royal Adelaide Hospital, Adelaide, Australia
Aim: Clinical coding is an important aspect of hospital administration and governance, with impacts on hospital resource allocation. Accurate clinical coding that reflects a department’s casemix is essential to ensure appropriate funding distribution. The Australian Refined Diagnosis Related Group (AR-DRG) is the coding scheme used in Australia. It aims to structure episodes of care, and is formulated by coding staff using the clinical documentation to establish diagnoses and procedures during a hospital admission in accordance with the International Statistical Classification of Diseases and Related Health Problems, 10th Revision, Australian Modification (ICD-10-AM). Coders rely on documentation most often completed by junior doctors. The aim of this audit was to determine the accuracy of clinical coding in a general medicine department of a tertiary hospital in South Australia.
Method: We reviewed a series of admissions under the general medical units whose episodes of care had already been coded by the hospital coding staff, and selected those patients attended by two of the authors. Blinded to the codes, 24 case files were reviewed by the relevant physician to compile a list of the pertinent medical issues during those admissions. With this additional information, the files were re-coded to determine if there was a change in the AR-DRG.
Results: We found an 18% error rate from misinterpreted or missing ICD-10-AM codes, resulting in a change in the AR-DRG of 6 out of 24 cases. In four cases the split indicator, a measure of resource consumption, increased. In the other two cases the partition, generally associated with the principle diagnosis, changed. When funding allocation rates were re-applied, two of these changes caused in a reduction in the amount of funding received, while four changes resulted in an increase. The net change for the 24 cases was an increase of AUD$9,913.64.
Conclusion: This audit suggests that greater clarity of documentation, and confirmation by the responsible consultant, could help improve clinical coding and allow more accurate resource distribution.