Quality of urological cancer diagnoses in the Danish National Registry of Patients.

Valid and up-to-date data on cancer diagnoses are needed for clinical quality monitoring and epidemiological research. The Danish National Registry of Patients (DNRP) is continuously updated, recording all Danish hospital contacts including cancer diagnoses. The Danish Cancer Registry (DCR) is updated once a year and includes quality control of diagnoses. We compared the quality of urological cancer diagnoses in the DNRP with the DCR to assess whether data in an administrative hospital registry are valid compared with data from a well-established cancer registry. We identified 60 434 incident urological cancer cases in the DNRP and/or the DCR from 2001 to 2009. Completeness and the positive predictive value (PPV) of urological cancer registration in the DNRP were estimated using the DCR as the reference standard. To examine the impact of potential misclassification, we computed mortality estimates for urological cancer patients in each registry. Because DCR registration procedures changed in 2004, the periods 2001-2003 and 2004-2009 were analyzed separately. In 2004-2009, the overall completeness and PPV of urological cancer registration in the DNRP were 94.9% (95% confidence interval: 94.7-95.2%) and 86.6% (95% confidence interval: 86.3-86.9%), respectively, compared with diagnoses recorded in the DCR. Completeness and PPV of cancer registration in the DNRP varied between cancer subgroups. In 2001-2003, both completeness and PPV in the DNRP were slightly lower compared with 2004-2009. Mortality estimates in patients registered in the DNRP and the DCR varied slightly. The DNRP could be a valuable source of data for clinical quality monitoring and epidemiological research for some urological cancers, especially when current data are essential.