The validity of hospital administrative data for outcome measurement after hip replacement

Background: Because of the increasing availability of clinical information on the basis of electronically processed data obtained through the hospital discharge records in the HIS, large databases are being set up to develop risk adjustment models for outcome assessment. This study is aimed at assessing the validity of hospital discharge data from the Hospital Information System (HIS) of patients with hip arthroplasty. Methods: 677 records were extracted from the database of the pilot project “Lazio Region Hip Arthroplasty Register (Ripa-L)” and were compared to the corresponding HIS discharge records. The Ripa-L dataset was used as a reference to evaluate the completeness and accuracy of the socio-demographic and clinical HIS data. Results: Data such as the patients’ age and sex, principal diagnosis, and surgical procedures, showed a very high level of agreement. By contrast, clinical information about comorbidities on admission and in-hospital complications mostly showed unacceptable variances in the datasets. The sensitivity of hospital data reporting was generally very low for almost all conditions, with the highest value being observed for diabetes (58%) and the lowest for endocrine and peripheral venous diseases (4%). Conclusions: Gaps in clinical information may compromise the ability to carry out high quality appraisals. In particular, the underreporting of comorbidities in hospital administrative data may lead to misestimation of the providers’ skill and quality of care, as a consequence of imperfect risk-adjustment. Stakeholders should highlight the potentialities related to the use of high quality administrative datasets also in clinical evaluations by stimulating health professionals to further improve the quality of the collected data.