Compliance with outpatient antibiotic regimens can now be measured by electronically monitoring the time history of dosing. This new approach reveals that many patients comply only partially with prescribed regimens in randomized, controlled, outpatient trials. Omitted or delayed doses and early cessation of dosing are commonly observed. Partial compliance converts a fixed-dose trial into a set of natural experiments in dose ranging, presenting a variety of patterns of dose timing that can be correlated with clinical outcomes for an estimate of minimum compliance needed for satisfactory outcome. Reliable measures indicate little difference in compliance between once- and twice-daily regimens, but considerably higher rates of omitted doses with three-times-daily or four-times-daily dosing. A key practical issue is to ensure continuity of drug action in the face of the most common compliance errors. Continuity of action is more likely when the prescribed interval between doses is considerably shorter, preferably half or less, than the drug's duration of action, allowing doses occasionally to be delayed or omitted without a gap in drug action. Thus, a twice-daily regimen may be expected to maintain crucial continuity of drug action better than a once-daily regimen, even if a few more doses are missed. Errors to avoid in compliance with outpatient antibiotic regimens are prolonged intervals between doses and early cessation of treatment.