Analytical flexibility in research can lead to variability in results and contribute to the replication crisis. Multiverse analysis addresses this issue by evaluating the robustness of findings across multiple plausible models. The Post-selection Inference approach to Multiverse Analysis (PIMA) further enhances this framework by providing valid statistical inference. In this study, we apply PIMA to assess whether SARS-CoV-2 infection increases the likelihood of disability onset, measured through the Global Activity Limitation Index, or mortality among initially non-disabled individuals. We construct a multiverse of models using longitudinal data from the Survey of Health, Ageing, and Retirement in Europe. Results highlight how significance strongly depends on specific modeling choices, emphasizing the value of PIMA in ensuring robust conclusions.