A new clinical decision support tool based on personalized evidence-based medicine improves outcomes of anticoagulation therapy in patients with atrial fibrillation: an analysis from the AF registry

Abstract

Abstract Background New clinical decision support system (CDSS) MedicBK use algorithms for presents treatment suggestions in frame of guideline-based therapy and personalized evidence-based medicine (EBM) therapy. Objective We hypothesized that clinicians with access to CDSS (MedicBK) would provide more personalized EBM anticoagulation treatment and improve outcomes. Methods 2683 patients at high risk for stroke and 1-year follow-up were enrolled in study and analyzed by CDSS and core laboratory. The primary endpoint was the quantify the performance of the CDSS (MedicBK) algorithm. Secondary endpoints include the following: outcomes for all-cause mortality, thromboembolism (TE), bleeding, and the composite endpoints. Results From 2683 patients, 1971 patients were core laboratory-classified as guideline adherent and 824 of them were personalized evidence-based therapy adherent. All 1971 (100%) patients were also detected as guideline adherent by CDSS, whereas only 748 (90.8%) were classified as personalized EBM adherent. The sensitivity for identifying patients with guideline and personalized EBM therapy adherence is therefore 100% and 90.8%. 712 and 1935 patients classified as guideline and personalized EBM therapy non-adherent by the core laboratory. All of them CDSS confirmed the absence of adherence, resulting in a specificity of 100%. The PPV and NPV of detecting with or without adherence to guideline, respectively, are 100% (1971 of 1971 pts) and 100% (712 of 712 pts). The PPV and NPV of detecting personalized EBM therapy adherence or non-adherence, respectively, are 100% (748 of 748 pts) and 96.1% (1859 of 1935 pts). The composite endpoint of cardiovascular death, any thromboembolism (TE) or bleeding was significantly lower in personalized EBM adherent patients during 1-year follow-up (P=0.02) and increased risk by >40% [HR 1.454 (95% CI 1.037; 2.040)], and non-guideline adherent and non-personalized EBM adherent treatment by >110% [HR 2.113 (95% CI 1.453; 3.074)]. Conclusion MedicBK appears to be promising to improve adherence to guideline and personalized EBM therapy in general practice. Personalized EBM anticoagulation management is associated with significantly better outcomes. Funding Acknowledgement Type of funding sources: None.