The test-negative design (TND) is an efficient form of case-control study commonly applied to influenza vaccine effectiveness (VE) estimation. TND validity is predicated on the core assumption that the intervention (vaccine) has no effect on other non-targeted aetiologies resulting in similar illness/disease. Here we verify this core assumption and compare efficacy estimates derived by the TND versus classical per-protocol analysis of four datasets obtained from randomised placebo-controlled clinical trials (RCT) of the live attenuated influenza vaccine (LAIV) in children ≤7 years-old and the elderly ≥60 years-old. We further assess generalisability of the TND approach in two other RCT datasets to evaluate monoclonal antibody in the prevention of respiratory syncytial virus (RSV) hospitalisation. Efficacy estimates and their confidence intervals were virtually identical for per-protocol RCT versus TND analyses of LAIV and also for RSV monoclonal antibody. Neither LAIV nor monoclonal antibodies affected the risk of disease aetiologies that were not specifically targeted by the respective interventions (e.g. other respiratory viruses). This study validates the core assumption of the TND approach for influenza vaccine efficacy estimation and confirms the accuracy and precision of its estimates compared to the gold standard of classic per-protocol RCT analysis of the same data sets. The TND approach is generalisable for other conditions such as RSV for which the core assumption is also met. However, when used in observational studies, the TND, like all designs, still requires assessment for bias and confounding that may exist in the absence of randomised participation and blinded follow-up.