1

The patients with depression had a significant improvement in their severtity of disease after 8 weeks antidepressant treatment and I found the serum BDNF levels of the patients with depression after 8 weeks antidepressant treatment were lower compared to those in baseline.

This study is a uncontrolled observational study, so we do not have a control group. Last month, a reviewer proposed a question that I should calculate the connection of the decrease of the BDNF and the course of treatment, and he thought using paired t test to analyze the data was not appropriate since there were many confounding factors needed to adjust such as length of disease, baseline severtity .

I want to know if I can adjust the mentioned confouding factors when we do not have a control group and how can I calculate the connection of the BDNF decrease and the course of treatment?

shen
  • 11
  • 2

1 Answers1

1

You can, and indeed should adjust for confounding variables in a non-experimental study like the one you're describing.

Some relevant questions and answers on this site: How exactly does one “control for other variables”? and Adjusting for Confounding Variables . You may simply be able to stratify your data, based on what confounders you think are important and whether they're categorical or continuous variables, but in all likelihood you need to be looking to a regression-based approach to account for the differences between your groups that are not due to your exposure of interest.

Giving advice on your specific study is beyond the scope of this site, for the most part, and definitely can't be answered in a single question with the amount of information you have provided. My recommendation would be to consult with a statistician or experienced researcher at your institution to see if they can provide some guidance to you.

Community
  • 1
Fomite
  • 21,264
  • 10
  • 78
  • 137
  • Thanks for your very helpful advice, and it benefits me a lot. However, I have still a question. You think I need to use a regression-based approach to account the differences between my groups that are not due to my exposure of interest. However, in fact, in our study, there is only one patients group. so I want to make sure whether your meaning is that I should stratify the patients group into two groups based on an important confounders first, and then I can use regression-based approach to adjust some other confounding factors between two groups. Looking forward to your help. Thanks! – shen May 07 '14 at 00:59
  • I am sorry that I did not express my meaning clearly. According to your advice, I want to use SPSS to conduct linear regression to adjust confounding factors. In general, in SPSS, I can use group as a independent (for example antidepressant treatment patients group= group 1, placebo treatment patients group= group 2), in the same time, and include some confouding factors in the independent. However, in fact, there is only one patients group and our study lacked a placebo treatment patients group as a control group. As a result, the group is not a variable. – shen May 07 '14 at 02:02
  • So, in this situation, can I still use regression to adjust confounders to calculte the connection of the BDNF decrease and the course of treatment? if so, how can? Thanks very much!! – shen May 07 '14 at 02:02
  • @shen No, you should not stratify and then do a regression analysis - that was an either/or suggestion. – Fomite May 07 '14 at 14:22
  • @shen As to the rest of your question, you actually *do* have two groups - Patients at Baseline and Patients After Treatment. The patients at baseline are essentially serving as their own controls. There are ways to analyze this, but some of them are relatively sophisticated. I **strongly** suggest going and finding a statistician or experienced methodologist who can advise you more directly. – Fomite May 07 '14 at 14:23