A prospective survey was conducted in patients admitted to 11 randomly selected general practices and eight hospitals located in six provinces of Poland. For each patient meeting the international acute gastrointestinal infection (AGI) case definition criteria, information was collected on healthcare resources used. Antibacterial drug consumption was assessed using defined daily doses (DDD) and extrapolated to the national level using results from a parallel study of AGI incidence in the community. Additionally, a logistic multivariable model was fitted assessing determinants of antibacterial drug administration. Valid questionnaires were collected from 385 general practitioner (GP) consultations and 504 hospital admissions. Antibacterials for systemic use were prescribed during 60 (16%) GP consultations and 179 (36%) hospital admissions. The estimated societal AGI-related consumption of antibacterials amounted to 5·48 million DDD (95% uncertainty interval 1·56–14·12 million DDD). Antibacterial prescription was associated with work in large practices [adjusted odds ratio (aOR) 3·16] and hospital wards (aOR 2·87), compared to small general practices, referral for microbiological testing (aOR 2·88), presence of fever (aOR 2·50), presence of mucus or blood in stool (aOR 1·94), age >65 years vs. <5 years (aOR 1·88), and rural vs. urban residence (aOR 1·53). Despite the fact that antibacterials were prescribed to a minority of consulted AGI patients, their consumption in society was not negligible due to the high prevalence of AGI symptoms. Prescription of antibacterial drugs should be restricted to cases with specific indications, preferably following microbiological investigation of AGI aetiology. To achieve this, clear national recommendations should be widely disseminated to physicians, and included in medical training curricula.