Quality control in drug measurement.

Therapeutic drug monitoring of serum concentrations has become a rapidly expanding area of clinical pharmacology and is likely to remain an important growth area in the future, despite that the evidence for the efficacy of such monitoring is inconclusive except for a handful of drugs and findings that the reliability of much drug assay work is far from optimal. Nonetheless, in view of the likelihood that the use of routine monitoring will increase, the institution of quality control programs seems not only desirable, but necessary. We describe the design of Bartscontrol (formerly known as St. Bartholomew's Hospital Quality Control Scheme for Antiepileptic Drugs) and discuss the experience of this program. Although the feedback provided to the participants of such a program seems to encourage an overall increase in precision, our findings also indicate that some laboratories perform consistently badly. While the data generally do not reveal that variations in precision are due to analytical method employed, a scheme such as the one described can occasionally make the important finding that a particular technique is inappropriate, as was found for the case of spectrophotometry for phenytoin.