Statistics on misuse of drugs have been misused

Editor—The government’s anti-drugs strategy places great emphasis on reducing drug related crime.¹ Part of its concern is based on evidence of high drug use among offenders and the assumption that crime is driven by the need to finance drug use and may be prevented if problem drug users are diverted into treatment. The white paper reports that “latest indications from a random sample of suspected offenders arrested by the police suggest that over 60% of arrestees have traces of illegal drugs in their urine.” It also emphasises the importance of evidence and information in developing the strategy.² The notion that 60% of all arrested people give urine samples that are positive for drugs has entered policy debate as “evidence.” But this evidence is based on shaky ground if we examine the report from which it is derived.³

Urine testing was conducted in five police areas for different periods over two years. Sites were chosen for convenience: Cambridge because the police quickly participated, Hammersmith and Manchester because they were considered to be areas of high drug use, and Nottingham and Sunderland because they were thought to differ from other sites. In the first three surveys samples were non-random, with arrested people being selected at the interviewer’s discretion (mostly during the daytime). In the last two surveys all eligible candidates were approached whatever the time of day. The first three surveys have no data on those not sampled. In the other two, 803 of 1416 arrested people were deemed eligible, 557 were approached, 480 interviewed, and 342 had a urine test.

The time for which drugs remain testable in urine varies (amphetamine, two days; opiates, cocaine, and benzodiazepines, three days; and cannabis, up to a month with chronic use). Figures should be adjusted for differences in half life; there were high rates of cannabis use and it was responsible for over four tenths of the “60% of arrestees” found positive for drugs. The above factors suggest that the samples are not representative of those obtained from people who have been arrested and that the results are not generalisable.

These and other problems were acknowledged by the report’s author. The research was designed to refine methodology because such a study had not been done before in the United Kingdom. Though the author is cautious in the interpretation of the data, the white paper is not. If the government is committed to evidence based policymaking, greater caution should be exercised in the selection and use of drug statistics.