Different methods of analysing a randomised controlled trial

	Explanation	Benefits	Negatives
As treated	Analyses by treatment received	Easy to calculate Tries to address patient oriented question	Results in non-random omission bias
Per protocol	Omits all participants who do not follow protocol	Easy to calculate Tries to address patient oriented question	Results in non-random omission bias
Intention to treat	Analyses by randomisation, ignores whether treatment received	Easy to assess Provides good estimate of the effect of recommending a treatment to study population	Underestimates value of receiving the treatment
Contamination adjusted intention to treat	Analyses by randomisation, adjusts for whether treatment received	Provides good estimate of an individual’s risks and benefits of receiving a treatment	Overestimates population level treatment benefits Somewhat more difficult calculation