Hospital infections continue to be an important problem. Changes in patient categories and in the examinations and treatments offered at the hospital, probably constitute the basis upon which new types of infections and new groups of infecting microorganisms continually emerge. The treatment of patients with hospital infections may become complicated by the emergence of drug resistance in the infecting microorganisms. The situation varies between hospitals, and this necessitates a local microbiological monitoring. Prerequisites of a surveillance system in general are described and illustrated by a system monitoring antimicrobial resistance in a Danish county. A specific computer system (software) for the processing of clinical microbiological results was developed by the author to make the local microbiological monitoring possible. Computer-assisted analyses for an evaluation of the statistical association of drug resistance traits (phenotype) were also developed. Data on local drug resistance patterns were systematically obtained and analysed in a seven-year study comprising several hundred thousands specimens. Drug resistance genotypes were determined for selected groups of microorganisms (e.g. coagulase-negative staphylococci) in various periods. Data from these studies emphasize that it is important to correlate information about resistance phenotypes of all clinical isolates with the resistance genotypes of selected strains in the monitoring of plasmid mediated resistance and its dissemination. These findings agree well with the results of other but generally less comprehensive studies in the literature. In a study of outpatients treated with a particular antimicrobial (tetracycline), it was shown that this drug selected for plasmid mediated multiple drug resistance in the normal flora of the patients. On the basis of these results and the literature, a general model for the emergence and spread of antimicrobial drug resistance is discussed. Examinations of quantitative relationships between antimicrobial usage and microbial resistance are also reviewed. Analysis of the statistical correlation between various measures of antimicrobial use and the prevalence of microbial resistance in the Danish county corroborates the observation that general changes in prevalence of a particular resistance trait seem to be best explained by taking the mechanisms of co-selection by other antimicrobials into consideration.