For older adults navigating an ever-shifting COVID-19 landscape, real-world evidence on antiviral effectiveness matters far more than trial-era data collected under very different immunological conditions. This Australian analysis addresses a critical gap: how well do molnupiravir and nirmatrelvir-ritonavir actually perform against severe outcomes when prescribed to high-risk older adults outside controlled trial settings, amid evolved variants and widespread hybrid immunity?
Using record-linked administrative health data from Victoria, researchers examined more than a year of outcomes — hospitalisation or death within 35 days of COVID-19 onset — among individuals aged 70 and older with confirmed infections between July 2022 and March 2023. Three distinct analytical frameworks were applied to the same dataset: multivariable logistic regression, Cox proportional hazards regression, and target trial emulation. Across all three, antiviral treatment was consistently associated with meaningfully reduced severe outcomes. Effect estimates ranged from a hazard ratio of 0.62 under the adjusted Cox model to an odds ratio of 0.28 in the logistic model, with target trial emulation yielding a hazard ratio of 0.45 — representing reductions in severe outcomes ranging from roughly 38% to 72% depending on method.
The substantial variation across methods is itself the study's most analytically instructive finding. Rather than reflecting genuine differences in drug performance, the divergent estimates stem from differing methodological assumptions, estimand definitions, and how each approach handles confounding and time-varying exposures. This highlights a persistent challenge in pharmacoepidemiology: real-world effectiveness estimates are not method-neutral. Target trial emulation — which explicitly mimics a randomised trial design within observational data — is increasingly considered the most rigorous approach and produced an intermediate estimate here. The study is limited to one Australian state, a narrow 9-month window, and older adults only, restricting generalisability. Still, the convergence across all three methods toward meaningful benefit strengthens confidence that oral antivirals retain protective value for high-risk older populations even as the pandemic has matured.