Data analysis methods
Posted on 14th July 2015 by Dabean Faraj
Data Analysis: Introduction
Conducting good clinical trials is hard. It means you need to overcome several problems along the way. Collecting data is only one part of the story. You also need to consider data interpretation and data analysis. After you have collected all of your data, it can be interpreted in a huge number of ways. In this post, I will describe two types of data analysis – “As Treated” (AT) and “Intention to Treat” (IT)
Well as a study progresses, you may need to adjust for different scenarios. I will explain AT and IT using the following hypothetical study:
- Trial involves comparison of cancer treatments
- One drug is a new medication
- It is compared to the current medication used
- This is a study that follows the sample population over a long time period
- This is a double blind trial
- Participants are randomly allocated the different treatments
Now some patients will not follow the trial until the end. Some will drop out along the way. The problem with this is that you don’t know why they dropped out. They may have dropped out because:
- They thought the medication was not working
- The medication was intolerable
- It was hard to maintain compliance
- It was too inconvenient
- They dropped out just as an anomaly
- … and many other reasons
You can never really know! The only thing you do know is the data you’ve already collected about them. The aim of your trial is to work out if the new drug is better physiologically (in terms of treating lung cancer). Or if the new medication is more acceptable (more tolerable side effect profile, easier to remain compliant etc.).
You can adjust for this though using either AT or IT analysis. But, like everything, they have their pros and cons.
As Treated Data Analysis (AT)
At the end of the study you can you do an AT Data Analysis. This is when you include the data from only the participants of your trial that stayed until the end. By doing so, you can actually compare the two drugs entirely. This means you can determine which of the two drugs treats cancer more effectively.
That said, it means that you have excluded the participants who dropped out. What this means is that your sample population is smaller. It also means you lose the randomisation of the trial. And of course, you cannot determine why those that dropped out, did so.
Intention to Treat Data Analysis (IT)
In contrast you can use the IT Data Analysis method. This is when at the end of your study you include the data taken from every single participating, regardless of if they completed the study or not.
The advantage of this is that you will be able to determine which of the two drugs is more acceptable. It will therefore give you an idea of how the public may respond to using the new drug, if the new medication were to become available to individuals.
Conclusion: which one is better?
Well. This is a tough question to answer! I think it really does depend on what you want to achieve in your study. Maybe you just want to know which treatment is better. Then obviously AT is superior. Or maybe you’re interested in how the population will respond to the medication in terms of compliance. In this case, IT is the way to go.
My personal opinion? I would do both. By doing both, you can see the whole picture. This can be quite useful because it allows you to weigh up the pros and cons. What if the new treatment is so much better at dealing with the condition? In this case, is it worth the extra side effects you can get? Consider chemotherapy. It is not a pleasant treatment. Its side effect profile can be quite horrible. But it is quite effective against some cancers.
These are all questions that you will need to answer.
Check out this article
Data analysis methods by Dabean Faraj is licensed under a Creative Commons Attribution-NoDerivatives 4.0 International License. All images used within the blog are not available for reuse or republication as they are purchased for Students 4 Best Evidence from shutterstock.com.