Accurate coding is essential for local and national data reporting and for contracting. It is also integral to clinical governance. This study aimed to assess the accuracy of coding in Morriston Hospital plastic surgery theatres and coding office, to reaudit and address poor practice. A third coding system, a computerised logbook developed by the senior author, was not analysed in this study. Fifty operations coded using OPCS-4 were compared with a gold standard for overall accuracy, primary and procedural codes. Results were discussed with all relevant staff and reaudit took glace 3 months later. The data were analysed using the paired Student's t-test for intergroup comparisons and the unpaired test for intragroup assessment. At initial audit, the coding office was significantly better than theatre staff in overall accuracy (78% vs 43% respectively P much less than 0.01) and in procedural codes (98% vs 42%, P much less than 0.01) but there was no difference in primary codes (62% vs 74%). At reaudit the only significant improvement was in overall accuracy of coding office records, although the clinical coders were now significantly better at recording primary codes than theatre staff (76% vs 56%, P < 0.05). The conclusions were that the quality of coding in theatre was poor and should stop. Clinical coders performed better but 1/3-1/4 of essential codes were inaccurate. This may have been due to limited understanding of terminology and techniques, difficulty reading operation notes and complexity of OPCS-4. Recommendations included closer cooperation between surgeons and coders to support and improve clinical coding performance (C) 2000 The British Association of Plastic Surgeons.
