Data-driven development of the nationwide hip fracture registry in the Netherlands

Abstract

Abstract Summary Additional variables for a nationwide hip fracture registry must be carefully chosen to prevent unnecessary registry load. A registry pilot in seven hospitals resulted in recommending polypharmacy, serum hemoglobin at admittance, and questions screening for risk of delirium to be used in case-mix correction and for development of quality indicators. Purpose Clinical registries help improve the quality of care but come at the cost of registration load. Datasets should therefore be as compact as possible; however, variables are usually chosen empirically. This study aims to evaluate potential variables with additional value to improve the nationwide Dutch Hip Fracture Audit (DHFA). Methods An expert panel selected eleven new variables for the DHFA, which were tested in a prospective cohort of all hip fracture patients treated in 2018 and 2019 in seven pilot hospitals participating in the DHFA. The association of these eleven variables with complications, mortality, and functional outcomes at 3 months was analyzed using multivariable logistic regression analysis. Based on the results, a proposal for variables to add to the dataset of the DHFA was made. Results In 4.904 analyzed patients, three tested variables had significant associations (p < 0.01) with outcomes: polypharmacy with complications (aOR 1.34), serum hemoglobin at admittance with complications (aOR 0.63) and mortality (aOR for 30-day mortality 0.78), and a set of questions screening for risk of delirium with complications in general (aOR 1.55), e.g., delirium (aOR 2.98), and decreased functional scores at three months (aOR 1.98). Conclusion This study assesses potential new variables for a hip fracture registry. Based on the results of this study, we recommend polypharmacy, serum hemoglobin at admittance, and questions screening for risk of delirium to be used in case-mix correction and for the development of quality indicators. Incorporating these variables in the DHFA dataset may contribute to better and clinically relevant quality indicators.