AIMS We aimed to determine whether simultaneous use of two tracer compounds with long and differing elimination half-lives may permit more precise recognition of percent compliance and discrimination among remote, intermediate, and recent patterns of daily dosing. METHODS We have derived, through computer modelling, the potential utility of digitoxin and phenobarbitone as simultaneous tracers. From kinetic data on these reliably measured compounds with long half-lives we defined the percentage error below which this pair of tracers can be expected to discern the precise, and the approximate, pattern and number of daily doses taken over a mouth. RESULTS The longer the half-lives of the tracers the greater the sensitivity of our model to percentage compliance. At extremely long half-lives, relative to the period of dosing being monitored, sensitivity approaches 100% if measurement error is in the 1-2% range. The ratio of the half-lives between the two tracers that is optimal for detection of percentage compliance is 0.2 to 0.8. The detection of pattern does not depend importantly either on the tracer half-lives or on their ratio to each other. The greater the percentage compliance, the more accurately our two-tracer model determines the pattern of dosing. CONCLUSIONS Precise characterization of compliance using tracers must probably await the development of compounds with half-lives measured in weeks or months. The probability of compliance of a given degree can be derived from our two-tracer technique with best-fit analysis such as we describe.