e

‘censored’ ’ in the more ial deemed inappropriate, al analysis. In the clinical nt might still be alive at d thus his actual post-s unknown (though, of e time from treatment to say the patient’s survival g the catch censored, we nds for the catch, but not tch-modelling approaches 998) found that, in terms atch error model provided PA-derived model) for els and survey index survey-only method. This ompare approaches, but odels rather than age-y to date comparing th