The aim of this study was to determine the link between the characteristics of antimicrobial therapy and occurrence of antimicrobial resistance in Escherichia coli of clinically healthy pigs exposed to antimicrobial treatments. A total of 918 Escherichia coli isolates were obtained from faecal samples, collected from 50 pig herds at the end of the fattening period and susceptibility was tested towards 15 different antimicrobial agents, using the disk diffusion method. The Antimicrobial Resistance Index (ARI) of each isolate was calculated, as the number of antimicrobials to which resistance was found divided by the number of drugs tested. The antimicrobial resistance percentage per class (ARclass) was defined as the percentage of E. coli strains clinically resistant to that specific class. Data on group level antimicrobial use in the sampled herds was collected and quantified as treatment incidences (TI) based on the used daily dose pig (UDDpig) and the animal daily dose pig (ADDpig) (number of pigs treated with one ADDpig or UDDpig/1000 pigs at risk/day). The UDDpig/ADDpig ratio gives an indication of the correctness of dosing. The TIADDpig for group level use was 235.7 per 1000 pigs at risk per day, whereas the TIUDDpig equaled 200.7. This means that in reality, fewer pigs were treated with the same amount of antimicrobials than theoretically expected and thus antimicrobials were generally overdosed. Generalized linear regression analysis showed a significant relation between the TIADDpig and the ARI (p< 0.01), whereas there were no significant links for the TIUDDpig (p> 0.05). Analysis of the antimicrobial resistance for β-lactam antimicrobials and tetracyclines suggests that the effect of correct or incorrect dosing on resistance development was different for the different antimicrobial classes tested. Besides the amount of administered antimicrobial agents, the frequency of drug administration may play a role in the selection of antimicrobial resistance in commensal E. coli.