Health Authorities receive vast quantities of data from providers relating to patients treated. Ibis data is used to survey the health of the resident population and to determine future healthcare services. It is therefore essential that the quality of this data is measured. North Staffordshire Health Authority already monitor, to a certain extent, the quality of data received. However, accuracy is one attribute of quality not monitored. This thesis proposes a method to measure the accuracy of patient data, in particular clinical coding. The traditional method of measuring accuracy determines whether a data item is correct or incorrect. The definition of accuracy, however, is the measure of agreement with an identified source. The proposed measure ranks incorrect clinical codes by their level of inaccuracy. Concepts from measurement theory are used to ensure that this measure adhered to the rules of the theory. This alternative method of measuring data accuracy was tested on a sample of inpatient data. From the results, the most appropriate way to analyse clinical data whilst still maintaining a level of accuracy satisfactory for the intended information purposes could be identified. Managers at North Staffordshire Health Authority were surveyed for their views on the usefulness of this alternative method of measuring data accuracy compared with the traditional method. Auditing a sample of data like this does not help to prevent errors occurring. Therefore, to identify how data accuracy could be improved in the long term, the source of the errors were discovered by examining the data life cycle.