[The validity of data in registration of operations. A quality analysis]

In Denmark all operations are registered centrally and identified by a unique number describing the exact operation. The aim of this study has been to validate the data, which are registered. Within a period of four months all registrations were carried out in duplo in two independent registers. The contents of the two registers were compared manually. A total of 1568 operations were registered. In one of the registrations 102 (6.5%) were missing and in the other 112 (7.1%). In 70 cases (4.5%) a discrepancy was found between the registers according to type or date of operation. It is concluded, that about 10% of all registrations in such a register are wrong. A better validity of data might be obtained by a simplification of registration procedures. Double registration is an effective way to locate the mistakes.