The French registry for complex bone and joint infections (C-BJIs) was created in 2012 in order to facilitate a homogeneous management of patients presented for multidisciplinary advice in referral centres for C-BJI, to monitor their activity and to produce epidemiological data. We aimed here to present the genesis and characteristics of this national registry and provide the analysis of its data quality. A centralized online secured database gathering the electronic case report forms (eCRFs) was filled for every patient presented in multidisciplinary meetings (MM) among the 24 French referral centres. Metrics of this registry were described between 2012 and 2016. Data quality was assessed by comparing essential items from the registry with a controlled dataset extracted from medical charts of a random sample of patients from each centre. Internal completeness and consistency were calculated.Aims
Methods
Bone and joint infections (BJI) are associated with a heavy morbidity and high health costs. Comorbidities, device associated infections and complicated journeys are associated with increased mortality, treatment failures and costs. For this reason, 24 referral centers (RC) have been created in 2009 in order to advise about management of “complex” BJI in weekly multidisciplinary meetings (MM). Since end of 2012, data from these meetings are gathered in a national database. We aimed to describe the data from this French registry of BJI and determine factors associated with the definition of “complex” BJI. Demographic, clinical, microbiologic and therapeutic characteristics of patients are systematically recorded in the database. Data from the first presentation in RC for each adult patients are presented. Complexity of BJI is recorded after each meeting according to 4 criteria (first failure, complex antibiotic therapy, precarious underlying conditions or complex surgical procedure). Part of unavailable data have been completed by pattern extraction from text-encoded commentaries. Factors associated with complexity were determined by multivariate logistic regression.Aim
Method
