Principles of research: limitations of non-randomized studies

I will not consider case-control, cross-sectional or uncontrolled studies (i.e. case series) or non-randomized cohort studies for other purposes (e.g. for studying aetiology), prognosis or evaluating diagnostic accuracy.