65.
The Aarhus Sarcoma Registry: Validity and Completeness in a
Population of Bone and Soft Tissue Sarcomas
Katja Maretty-Nielsen, Ninna Aggerholm-Pedersen, Johnny Keller, Akmal
Safwat, Alma B. Pedersen
Department of Experimental Clinical Oncology Sarcoma Centre of Aarhus
University Hospital; Department of Experimental Clinical Oncology, Sarcoma
Centre of Aarhus University Hospital; Department of Orthopaedics, Sarcoma
Centre of Aarhus University Hospital; Department of Oncology, Sarcoma
Centre of Aarhus University Hospital; Department of Epidemiology, Sarcoma
Centre of Aarhus University Hospital
Background:
Clinical databases are important sources for clinical studies, but
data incompleteness and incorrect data are major problems.
Purpose / Aim of Study:
The aim of this study was to validate the data in the
Aarhus Sarcoma Registry (ASR).
Materials and Methods:
The Aarhus Sarcoma Centre serves a population of
2.5
million. All patients diagnosed with a bone and soft tissue sarcoma during
the period 1979-2008 have been prospectively registered in the ASR. Medical
files on all ASR patients were reviewed by an independent medical doctor and
a medical student and used as gold standard to assess the completeness and
accuracy of key variables in the ASR. Missing patients were identified through
the local pathology department and registered in the ASR. The completeness
of registration of patients in the ASR was compared to the Danish Cancer
Registry (CR).
Findings / Results:
There were 1474 non-visceral sarcoma patients in the
ASR before, compared to 1829 after validation. In the same time period 1853
patients were registered in the CR with non-visceral sarcoma. 357 patients
were registered in ASR, but not CR. The completeness of the ASR compared
to CR was 79.2% (1467/1853). Important dates (e.g. admission, operation)
were missing in 339 patients before validation and in 216 patients, the dates
were incoherent (i.e. admission before operation). After validation, all dates
were correctly registered. A number of other variables in the ASR had missing
values. After validation, missing values in each of 39 variables registered in
the ASR was reduced to less than 4.4%.
Conclusions:
The validation process substantially improved completeness as
well as and accuracy of all variables included in the ASR. After validation the
ASR is an important source for the studies answering a number of clinical and
research questions on patients with bone and soft tissue sarcoma.
