Intent To Treat Analysis In Statistics

Under some circumstances, researcher may wish to assess a treatment as if it was given in a real world setting. In such cases they may choose to use an intent to treat analysis designed study. This is an advantageous way of determining the effects of a treatment because it includes subjects that may for example drop out, incorrectly follow protocol or refuse treatment. In other words, the analysis takes into account the non adherent subjects and therefore gives a closer approximation to the results that would be seen in real life. The main concept of this type of analysis is that it is accepted that the treatment in the field may differ significantly from the tightly controlled treatment in a laboratory. In studies that do not show an intent to treat effect, a per-protocol analysis may be performed that excludes subjects who have not complied with the methodology, as this will increase the chances of finding a significant effect.

One practical use of an intent to treat analysis might be the administration of dietary advice to lose weight. The attention of researchers and desire to please in a closely monitored laboratory setting may cause a significant increase in the likelihood of subjects adhering to a weight loss programme. When the same subjects follow the same plan but in their own home with no study interference and without chaperones, the weight loss achieved may be significantly less. In particular, more subjects may drop out and there is less likelihood that the subjects will rigidly follow the diet protocol. This may cause an over-estimate of the real-world effectiveness of a weight loss protocol if performed in a laboratory. Understanding beneficial weight loss protocols in a non-laboratory setting, and allowing for natural failure rates using an intent to treat analysis, is therefore much more representative of the real benefits of the diet.

RdB intent to treat analysis