Appropriateness of timing of drug administration in electronic prescriptions.

Abstract

For a small number of drugs circadian variability has been shown to modify efficacy, safety, or pharmacokinetics.
We aimed to develop a database containing optimum timing of drug administration and to test how well such information is considered in daily practice.
University hospital providing primary and tertiary care.
We included data of randomised controlled trials collected from Embase and Medline studying the impact of the timing of drug administration on pharmacodynamics, pharmacokinetics, and adverse events. Data were analysed and weighed according to an algorithm considering trial design and assessed endpoints. Each branch of the algorithm led to a specific recommendation as to the time of the day the drug should be administered. A second algorithm was used to establish a recommendation if studies differed in their conclusion. Subsequently we retrospectively analysed the dosing time in consecutive electronic prescriptions issued at our institution in 2007.
For 30 active compounds randomised controlled trials were published assessing optimum timing of their administration. In 33% of them timing had no impact on clinical endpoints while the administration at a certain time of the day significantly improved the outcome of another 64% no clear statement was made for one drug (ketoprofen). We then analysed 478,864 electronic prescriptions. Two percent of them contained drugs with known circadian variability. Only in 14% the suggested time was considered with a range between 0% for telmisartan (bedtime administration) and 85% for perindopril (morning administration).
Thus far, dedicated studies on circadian responsiveness to drugs are sparse and for the few drugs with unequivocal evidence this information is only rarely considered in daily practice. Integration of circadian dosing information into a clinical decision support system linked to electronic prescribing may be one promising way to make this information widely accessible.