Testing for drugs of abuse in urine is usually performed in two steps; after initial screening positive findings are confirmed with specific chromatographic techniques. The use of different methods with different cut-off levels, may lead to variable results in a test sample containing a drug. Sixty-eight Swedish laboratories were enrolled in an external quality study and received three control samples. The laboratories were instructed to perform only in-house tests for amphetamines, benzodiazepines, cannabinoids, cocaine and opiates. Nineteen out of the 68 laboratories reported fully correct results. Thirty reported one o more false positive drug findings. One laboratory performing confirmation with gas chromatography--mass spectrometry reported a false positive finding of amphetamines in the sample containing ephedrine. Since testing for drugs of abuse in urine is a delicate matter, with both medical and legal implications, such testing should be performed with a zero rate of false positive results. Routine use of specific methods for confirmation should be used since performance of these laboratories was better in this study. It is concluded that there is a need to continuously measure the level of quality of laboratories by a program for external quality control.