Validation
•Compared registry data with data from
–Norwegian Patient Register
•98-99% data completeness (2006)
–National Institute for Hospital Research
•95-97% data completeness (1999)
–local hospitals2,3
•93% data completeness (small hospital,1998)
•99.6% data completeness (large hospital, 2005)
–questionnaires to patient cohorts (1999)
•
1Espehaug B et al. Registration completeness in the Norwegian Arthroplasty Register… 2006. In press.
2Arthursson et al. Validation of data in the Norwegian Arthroplasty Register… 2005.
3Froen JF and Lund-Larsen F. Ten years of the Lubinus Interplanta…1998.
The cooperation of surgeons, the simple forms and the local routines for registration are essential for receiving valuable data. Furthermore, one of the most important reasons for good quality of data in the register database is probably the well-qualified and stable secretarial manpower. Systematic and continous efforts have been made to minimize the occurence of missing or erroneous data in the register database. When errors are identified, the forms are returned to the hospital for completion. KLIKK The number of operations reported to the arthroplasty register and to the patient register, are compared on a regular basis. Comparing the two databases for the period 1999-2002, we found a 98-99% completeness in registry database on hip and knee implants. Registration of revisions was even higher. (KLIKK) Validation of results has also been done by comparing data with data from the National Institute for Hospital Research (KLIKK) and with local data from reporting hospitals. Finally, data directly from patients through quetionnaires have been compared to the registry data.