Observational studies often guide treatment recommendations in infectious diseases when controlled trials are not available. Nevertheless, insufficient methodological rigour can compromise their validity and reproducibility. The Assessment of Methodological Quality and Bias in Observational Studies for Infectious Diseases (ASSURE-ID) questionnaire was created via a Delphi process involving 34 international experts. Three Delphi rounds granted consensus on the assessment tool comprising five domains, 42 items, and 48 assessment questions. A systematic literature review (PROSPERO number CRD420251065957) identified eligible observational studies comparing new drugs for bloodstream infections due to carbapenem-resistant Enterobacterales. After piloting, its final version was applied. Of the 5268 records screened, 14 studies were included. Tool application showed clear reporting of study objectives; however, only one (7%) study documented protocol pre-registration, two (14%) documented adherence to official reporting guidelines, and no studies documented a pre-established statistical analysis plan. Missing data were acknowledged in four (29%) studies and addressed in one (7%). Immortal time bias was addressed in two (14%) studies. Confounders were adjusted in 12 (86%) studies, but causal models and interaction testing were uncommon (in four [29%] studies, and one [7%] study, respectively). ASSURE-ID offers a structured, consensus-based framework to appraise methodological rigour in observational infectious diseases studies comparing treatments. Application to studies on bloodstream infections due to carbapenem-resistant Enterobacterales highlighted substantial limitations in the analysed studies.