Sie befinden Sich nicht im Netzwerk der Universität Paderborn. Der Zugriff auf elektronische Ressourcen ist gegebenenfalls nur via VPN oder Shibboleth (DFN-AAI) möglich.
mehr Informationen...
476-P: Predicting the Risk of Chronic Kidney Disease in People with Diabetes (PWD) from Electronic Health Care Records (EHRs): A Validation Study
Ist Teil von
Diabetes (New York, N.Y.), 2020-06, Vol.69 (Supplement_1)
Ort / Verlag
New York: American Diabetes Association
Erscheinungsjahr
2020
Link zum Volltext
Quelle
EZB Electronic Journals Library
Beschreibungen/Notizen
Chronic kidney disease (CKD) is one of the most frequent complications faced by people with diabetes mellitus.
The ability to accurately predict the onset of this disease at an early stage and adjust the individual therapy and lifestyle appropriately has the potential to prevent or at least slow down the progression of the disease significantly.
We have previously published a 3-year CKD risk engine based on longitudinal U.S. EHRs data (Indiana Network for Patient Care and IBM Explorys) from more than 600,000 PwDs and have demonstrated a superior performance compared to previously published risk algorithms developed solely on clinical trial data (AUC = 0.794 vs. AUC = 0.718 for the best benchmark algorithm). Nevertheless, our previous analysis was limited to healthcare data from the U.S. and the risk prediction was only tested for patients at the time point of the initial diabetes diagnosis. To prove that our algorithm has a broader applicability, we have independently validated our risk engine on a dataset of general practitioners from the UK (Clinical Practice Research Datalink) including longitudinal data from more than 140,000 PwDs, thereby showing the transferability of our algorithm into a European context (AUC = 0.763 +/- 0.005 vs. AUC = 0.594 +/- 0.006 for the best benchmark). Moreover, we used this analysis to apply our risk engine at variable time points after the initial diabetes diagnosis within the U.S. and European data and could thereby demonstrate that our algorithm is not limited to the initial time of diagnosis (average AUC = 0.79 irrespective of the time after diagnosis).
This validation analysis further corroborates the external validity of our approach to use Real-World-Data from EHRs in the development of CKD risk prediction algorithms and thereby supports a broader use of those algorithms in standard clinical practice.
Disclosure
T. Huschto: Employee; Self; Roche Diabetes Care. H. König: Employee; Self; Roche Diabetes Care. C.E. Marriott: Employee; Self; Roche Diabetes Care. D. Militz: Employee; Self; Roche Diabetes Care. C. Ringemann: Employee; Self; Roche Diabetes Care.