Who should code orthopaedic inpatients? A comparison of junior hospital doctors and coding clerks

Our principal aim was to determine whether coding for billing purposes is adequately completed by hospital doctors. We also wanted to determine whether events during the inpatient stay prompted coding. We investigated the completeness and accuracy of ICD9 and OPCS4 coding at two departments of orthopaedics. Coding was for billing purposes only; retrieval of clinical data was limited to individual cases. Junior Hospital Doctors (JHDs) undertook coding at one department, coding clerks at the other. At each side 100 sets of notes concerning inpatient admissions were reviewed. Coding clerks returned coding data of higher completeness and quality (completeness x accuracy) than JHDs, but of lower accuracy. Completeness of data from JHDs reflected motivation to code and this was influenced by events during the course of admission. JHDs coded best for patients undergoing elective surgery, and worst for patients managed non-operatively. Coding clerks coded for all groups equally well. In the current study coding data returned by JHDs was of lower completeness and accuracy than has been described for other coding systems. We attribute the lower completeness to the low motivation of JHDs when coding for billing purposes and the lower accuracy to difficulty in using the system. Motivation could be increased by incorporating coding into departmental audit, the generation of discharge summaries and record collection for personal logbooks.