The effect of an antibiotic policy on bacterial resistance in patients in geriatric medical wards.

In an effort to reduce levels of trimethoprim resistance amongst urinary isolates and faecal organisms two antibiotic policies (policy 1, erythromycin, nitrofurantoin substituted for trimethoprim and ampicillin; policy 2, Augmentin (ampicillin + clavulinic acid) substituted for trimethoprim and ampicillin) were used for 50 weeks on two geriatric acute/rehabilitation wards. The policies were evaluated by comparison with the pre-policy period and between policies. The policies were applied successfully and both were associated with a fall in the proportion of trimethoprim-resistant faecal coliforms and urinary isolates. Policy 1 appeared to affect plasmid-mediated resistance to a greater extent than policy 2. There were no differences in outcome for patients during policy periods. Control of antibiotic usage by formal policies is a viable means of controlling bacterial resistance in geriatric wards. The mode of action of such policies requires further elucidation.