Anyone with good knowledge of medical research - question on study type.Watch
Case control - Using archival tissue samples, 100 samples of premalignant tissue where the patient went on to develop the cancer (case), and 100 samples of premalignant tissue where the patient did not develop cancer (control). Samples would be tested for mutation and then the rate of mutation (exposure) in each group compared to see if it is higher in the cancer group.
Retrospective cohort - 100 samples of premalignant tissue which present with mutation upon analysis (exposed), and 100 samples of premalignant tissue which do not present with mutation (non exposed). See using medical records if the rate at which each group developed cancer is different (outcome).
Hopefully I haven't misunderstood something in the above descriptions of each study. So my question is, which type of study would be better for the sake of this type of question, or are both equally valid?
It depends on the characteristics of the population you're testing which of these will be most useful. E.g. if this mutation is very rare, study a) might not have sufficient sample size to draw a firm conclusion. Equally, if most of these samples do not go on to get cancer then study b might not be useful either. I'd choose the type that will illustrate the difference best (this should be possible - you will know the frequency of cancer amongst your samples even if you don't know the frequency of your mutation).
I see, so are you saying if the mutation is very rare then we would have to examine a huge number of samples to find enough mutations to have a large enough "exposed" group, so a case control would be better? And vice versa if the incidence of cancer is rare?