【求助】初学meta分析,投稿后审稿人意见求解
审稿意见:
According to your description, this is a rather practical way to display the data. Could you make it more scientific, e.g. by additionally calculating the correlation coefficient or put the data into a regression model? What happens if you exclude the case on the right (4000/500)? It seems to be an outlier which may significantly affect the correlation of dosage and blood loss. Please revise the methods, results and possibly the figure.
关于意见中其中提到的一个图:

我原本的意图,是将文献中提及的药物剂量作为横坐标,总失血量(平均值)作为纵坐标,作的一个图。其中直线代表了大概趋势。作图是用GraphPad Prism直接得到。注:总失血量有两种数据可以搜集到:Mean+SD及Mean+可信区间。
因为刚接触meta,菜鸟一个,意见中提到的相关系数和回归模型,我在丁香园里搜了下,貌似和我遇到的问题都有点不同,所以也不知道怎么处理了,还望高手指点一二。