Warfarin dosing prediction algorithm assessments in Asian patients after aortic surgery

crossref(2022)

引用 0|浏览0
暂无评分
摘要
Abstract Background: To evaluate the accuracy and predictive performance of multiple linear regression algorithms, Gage algorithm and International Warfarin Pharmacogenetics Consortium (IWPC) algorithm, and Bayesian algorithm for the maintenance dose of warfarin in Asian patients after aortic surgery. Methods:The predictive performance of Gage, IWPC and Bayesian algorithm were compared by calculating mean prediction error (MPE), mean squared error (MSE), root-mean-squared error (RMSE), and the percentage of patients whose predicted warfarin dose fell within 20% of the maintenance dose. Then the predictive performance of each algorithm was calculated separately in conventional dose group (dose of warfarin between 3mg and 3.75mg) and unconventional dose group (dose less than 3mg or more than 3.75mg). Results: Among three algorithms, the Gage algorithm predicted the warfarin dose with the highest percentage within 20% (58.1%). Bayesian-priori (dose estimation before starting therapy) algorithm had the highest RMSE (1.45 mg/day) and the lowest percentage within 20% (40.9%). The algorithm of Gage method tended to perform better in the conventional dose group (percentage within 20%: 70.0%), and Bayesian-priori method performed better in the unconventional dose group (percentage within 20%: 38.5%). And as the number of international normalized ratio (INR) observations increases, the predictive performance of Bayesian-posteriori (dose estimation after initiating therapy) algorithm showed an increasing trend in both groups and a significantly higher advantage, especially in the unconventional dose group. Conclusion: For Asian patients after aortic surgery, the Gage algorithm was appropriate for patients requiring conventional dose (2.25-3.75mg). Bayesian-posteriori algorithm might be more appropriate for patients requiring high dose or low dose (<2.25mg or>3.75mg).
更多
查看译文
AI 理解论文
溯源树
样例
生成溯源树,研究论文发展脉络
Chat Paper
正在生成论文摘要