Prescription of heroin to treatment resistant heroin addicts

Editor—As mentioned in the discussion of the paper by van den Brink et al,¹ experiments with heroin maintenance cannot be double blind. This problem is much more serious than the authors acknowledge, particularly because considerable sanctions were connected to the participants' responses.

The participants in the control groups knew that the promise of heroin maintenance later on could be withdrawn if they improved during the control period without heroin on prescription. Moreover, the participants in the experimental groups knew that they could be expelled from the experiment if they deteriorated while receiving heroin. Finally, the participants who improved while receiving heroin were aware that they would have a fair chance of continued heroin on prescription provided that they deteriorated in an interim period without heroin provision. Even if improvement could have been measured fully unobtrusively rather than with self reports, this would have created serious problems.

However, it is not certain that the results are positively biased because of this. It means that heroin experiments are tests in the sense of examinations rather than scientific experiments. If much is at stake in examinations people might fail not because they lack skills but because the tests are too nerve racking. This may have suppressed the results.

Europe is currently flooded by a tidal wave of expensive and demanding heroin maintenance experiments. Even if these experiments could be conducted as double blind trials there are other reasons why experimental studies are inappropriate in such cases.²,³