accuracy of procedure codes based on icd9cm

introduction: using hospital databases is extremely depended on an accurate classification that is based on clinical coding. we aimed to determine the validity of procedural coding in teaching hospitals. methods: in this cross-sectional study, we selected 246 medical records from kashan hospitals in 1386 and recoded procedures. procedures, coders’ information, and documentation principals were recorded in a valid checklist. accuracy was determined by agreement between original codes and recodes. we analyzed the data using spss through x2, fisher test, or, and ci 95% for or. results: there were 46 (18.7%) errors in procedure codes. using the coding book significantly accompanied more code errors; however, there were fewer errors in records which coders checked codes with tabular index. documenting more information decreases the errors (not significant). there were fewer errors in readable records. in addition, clear abbreviations reduced errors. conclusion: we can trust on procedure databases. better documenting the procedure details and factors related to coders can increase the quality of procedure coding and databases.