INTRODUCTION: To validate the coding and local registration of vaginal prolapse surgery in 1996-1998, and reporting of such data to the Danish National Patient registry (LPR). MATERIAL AND METHODS: Comparison of different sets of data from the local data base (GS) and LPR, and retrospective study of 296 patient files. RESULTS: The concordance between two GS-data sets was 96.4%, and 0-100% between different LPR-data sets. The validity of reporting of data from GS to LPR was 96.2% on the level of patient ID (cpr), and 99.3% concerning length of hospital stay. If a procedure was registered as a vaginal prolapse procedure the validity of this being true was 93.6%. 18 non-prolapse procedures were registered as vaginal procedures. The registration was incomplete in 12.5% due to typing or hearing errors or missing codes. The coding was misleading in 11.8%, due to lack of consensus on coding. 86.6% of errors were passed through from the patient file to GS and LPR. Through simple guidelines misleading use of combination of codes was reduced from 7% to 0.5%. CONCLUSION: The validity of the two sets of GS data, and the reporting from GS to LPR of vaginal surgical procedures is satisfactory on the cpr-level. However, the validity of such register data is affected by the risk of errors when drawing such data, and by coding errors. The main problems are lack of consensus on which codes to use, and that codes for very different procedures are very much alike. To improve the quality and usefulness of register data in monitoring vaginal prolapse procedures consensus on coding is desirable. Different methods to increase the quality of coding are discussed.
INTRODUCTION: To validate the coding and local registration of vaginal prolapse surgery in 1996-1998, and reporting of such data to the Danish National Patient registry (LPR). MATERIAL AND METHODS: Comparison of different sets of data from the local data base (GS) and LPR, and retrospective study of 296 patient files. RESULTS: The concordance between two GS-data sets was 96.4%, and 0-100% between different LPR-data sets. The validity of reporting of data from GS to LPR was 96.2% on the level of patient ID (cpr), and 99.3% concerning length of hospital stay. If a procedure was registered as a vaginal prolapse procedure the validity of this being true was 93.6%. 18 non-prolapse procedures were registered as vaginal procedures. The registration was incomplete in 12.5% due to typing or hearing errors or missing codes. The coding was misleading in 11.8%, due to lack of consensus on coding. 86.6% of errors were passed through from the patient file to GS and LPR. Through simple guidelines misleading use of combination of codes was reduced from 7% to 0.5%. CONCLUSION: The validity of the two sets of GS data, and the reporting from GS to LPR of vaginal surgical procedures is satisfactory on the cpr-level. However, the validity of such register data is affected by the risk of errors when drawing such data, and by coding errors. The main problems are lack of consensus on which codes to use, and that codes for very different procedures are very much alike. To improve the quality and usefulness of register data in monitoring vaginal prolapse procedures consensus on coding is desirable. Different methods to increase the quality of coding are discussed.
Authors: Morten Schmidt; Sigrun Alba Johannesdottir Schmidt; Jakob Lynge Sandegaard; Vera Ehrenstein; Lars Pedersen; Henrik Toft Sørensen Journal: Clin Epidemiol Date: 2015-11-17 Impact factor: 4.790