资讯内容 Content

——爱因斯坦大学医学院内分泌科Jeffrey S. Gonzalez教授

作者:J.S.Gonzalez 2012/6/13 16:28:00    加入收藏

  <International Diabetes>: How can collecting adherence data throughout a study bridge the gap between intention-to-treat and per-protocol? Can you explain that for us?
  Dr Gonzalez: Collecting careful adherence data throughout the trial and monitoring adherence over the course of the investigation in both conditions can allow me to then look at your intention-to-treat analysis where you compare the effect of being assigned to receive one or the other intervention (and that has public health implications), but then I think the benefit compared to a per-protocol analysis is that you don’t necessarily have to throw out any of your data. You can look at whether ones level of adherence impacts the effects that are obtained by the drug under investigation or maybe it’s adherence to a particular behavior protocol or whatever it. If you can quantify someone’s level of adherence and use that as a variable that may moderate the treatment’s effectiveness, I think that can add a lot of information about what effect can be expected out in the real world where adherence is going to be a bigger problem than it was in your trial where you have tightly controlled for factors that might limit a patient’s ability to adhere.


  <International Diabetes>: Whenever you are looking at statistics for people who had adhered though, they are always going to be tainted numbers because you are always going to be looking at the “healthy adherers”.
  Dr Gonzalez: That’s right. I think that is part of the problem with a per-protocol analysis. The way a lot of people handle a per-protocol analysis is that they throw the non-adherent people out of their experimental trial and only focus on the adherers. But in the control condition, they compare them to everyone so this analysis I am presenting is an alternative approach where you don’t throw anyone out but you ask the question of whether one’s adherence influenced the effect that was obtained.
  In the DPP project I mentioned in my talk, within the metformin arm they looked at whether adherence was associated with benefit in terms of decreased risk of developing diabetes and then they were able to show that that effect was not found in the placebo group. You have to be worried about or cautious about this “healthy adherer” effect, but that was a nice demonstration that said adherence is the same thing in both of these conditions and people are randomly assigned and we are keeping all of them in there so we can be confident about balancing but now you see it, and now you don’t. In metformin there is an adherence effect and in placebo there is not. It depends what you are adhering to. The more you adhere, the better benefit but it is dependent on what you are taking.


  <International Diabetes>: Do we see this kind of result a lot?
  Dr Gonzalez: Not a lot. Part of the reason for this whole symposium was to encourage people engaged in clinical trials and research who may not be focusing on behavioral factors, adherence or treatment satisfaction. If you take these factors into account, even if your objective is just does my drug beat that competitor’s drug or are they equivalent, for instance, taking account of this information will give you a better evaluation of that question than the two choices of intention-to-treat or per-protocol. If you have detailed information on adherence, you can use it in the evaluation of your effects.


[1]  [2]  下一页


网友评论仅供其表达个人看法,并不表明国际糖尿病同意其观点或证实其描述。 发表评论需登陆

关于本站 | 设为首页 | 加入收藏 | 站长邮箱 | 友情链接 | 版权申明


国际糖尿病 版权所有  2008-2018 idiabetes.com.cn  All Rights Reserved