Results from the mixed-effects meta-analysis based on 1000 simulations.

Estimates and standard errors for the relative log-potency of the insulin analogue compared to HI for study 060906 (insulin analogue 32) and 040801 (insulin analogue 7) are shown for different ways of re-fitting the simulated data: none and all indicate keeping and removing all the control rats, respectively, rem1 indicate removing one control rat per dose value, rem2 indicate removing two and so on.

针对研究060906（胰岛素类似物32）与040801（胰岛素类似物7），胰岛素类似物相较于HI的相对对数效价（relative log-potency）的估计值与标准误（standard errors），已按模拟数据的不同重拟合（re-fitting）方案进行展示：其中“none”与“all”分别代表保留全部对照大鼠与移除全部对照大鼠；“rem1”指移除每个剂量值对应的1只对照大鼠，“rem2”指移除2只，依此类推。