The intention-to-screen design of the trial depends heavily on the effectiveness of recruitment and the pre-existing “prevalence” of the test in the setting in which the study is conducted. I can see rejecting an unpleasant, rarely used screening test. But if it were presented in a less impersonal (but still unbiased) way, I imagine recruitment would have been higher, skewing the data towards cost-effectiveness, as re-analyses showed.
