Experienced drug users assess the relative harms and benefits of drugs: a web-based survey

Earlier expert-led efforts to rank the harms of drugs have used Delphic or multicriteria approaches and found poor correspondence between expert assessments and existing legal classifications. Those studies typically engaged psychiatrists and other drug experts and produced a harm matrix that weighted physical harm, dependence, and social harm, but they did not directly consult experienced drug users despite users having intimate knowledge of drug effects and harms. Lester and colleagues set out to fill that gap by collecting the opinions of highly experienced drug users about the relative harms and benefits of a range of commonly used illicit substances, together with views on how each substance should be classified under the UK Misuse of Drugs Act and which information sources most influence perceptions of harm. The study therefore aims to compare user-derived rankings with prior expert assessments and to capture qualitative reasons why particular drugs are seen as harmful or beneficial.

The study received ethical approval from Imperial College Research Ethics Committee and used a web-based survey built with Bristol Online Survey. The anonymous questionnaire ran from April to September 2010, took about 20–30 minutes to complete, and included 56 questions with a mix of categorical and open responses; only items relevant to users' opinions on harms, benefits, legal classification, and information sources are reported here. A total of 1,506 submitted forms were received; the analysed sample was filtered to include only respondents who reported having taken each of 11 specified illicit drugs at least once: heroin, cocaine, amphetamine, cannabis, GHB, magic mushrooms, ketamine, benzodiazepines, mephedrone, MDMA, and LSD. The extracted text does not clearly report the precise recruitment webpages by name (some site names are missing in the extraction). Ninety-three respondents met the inclusion criterion and are the focus of the analyses. Respondents were asked to rank the listed drugs by perceived general level of harm and by perceived level of benefit, to state how each drug should be classified under the Misuse of Drugs Act (Class A = 1 through unclassified = 4), to explain choices for the most harmful and most beneficial drugs, and to rank up to three influential sources of information on drug harms. Statistical analysis comprised Spearman rank correlations where appropriate, with two-tailed p values reported.

Ninety-three respondents (mean age 29, SD = 8.1, range 16–58) provided the analysed data; 83% were male and 71% lived in Britain. Recruitment sources among the analysed sample included forum sites and other web sources, though some site names are not clear in the extracted text. Thirty-two respondents ranked alcohol as the drug with the greatest potential for harm, citing addiction potential, acute and chronic toxicity including fatal risk, promotion of aggression and risky behaviour, and harmful withdrawal effects. Fifteen respondents ranked tobacco as most harmful, emphasising addiction, carcinogenicity, and mortality. Twenty-six respondents ranked heroin as the most harmful, noting high dependence potential, risks from intravenous use, impurity-related harms, social harms, and overdose; heroin rankings showed greater variance than tobacco. On potential benefits, respondents' rankings correlated significantly with a prior user-based analysis (Morgan et al.), with Rho = 0.83 (p = .002) for acute benefits and Rho = 0.73 (p = .01) for chronic benefits. Twenty respondents rated MDMA as having the greatest potential for benefit, emphasising empathogenic effects and usefulness in facilitating openness in psychotherapy. Eighteen ranked cannabis highest for benefit, noting stress relief, anxiety and pain reduction, and sleep improvement. Sixteen respondents prioritised LSD for benefit because of its capacity to promote insight and self-understanding. Seven respondents each ranked magic mushrooms and benzodiazepines as most beneficial, the former for well-being and insight and the latter for recognised roles in acute psychiatric care and anxiety reduction. When asked how each drug should be classified under the Misuse of Drugs Act, average chosen classifications did not correlate with respondents' own harm rankings (Rho = .41, p = .16), nor with the Act's current classifications (Rho = .05, p = .87). Thirty-two respondents (34%) stated that all listed drugs should be unclassified. Cross-study comparisons showed strong positive correlations between the users' harm rankings and several prior expert and user assessments: for example, correlations reported include Rho = .90 (harms to users) and Rho = .93–.94 for van Amsterdam and van den Brink's harms-to-individual/population analyses (all p < .001). Finally, when ranking sources of information that influenced their perceptions of drug harms, 93% placed "mine and my friends' own experiences" in their top three and 70% ranked it as the single most influential source.

Lester interprets the findings as providing novel insight from a rare sample of experienced drug users: their harm and benefit rankings align closely with previous expert and user assessments, yet these rankings diverge from current legal classifications. The researchers highlight that three drugs commonly placed in the highest legal harm bracket in the UK (MDMA, LSD, and magic mushrooms) were rated among the least harmful and among the most beneficial by this sample; when explaining perceived benefits respondents frequently referred to potential therapeutic applications rather than only hedonic effects. The mismatch between perceived harm and preferred legal classification is emphasised. For example, although respondents often judged alcohol and tobacco to be especially harmful, these substances did not receive correspondingly severe suggested classifications. Lester and colleagues suggest this indicates respondents consider factors beyond perceived harm when recommending legal status—possible influences include familiarity, reluctance to criminalise widely used legal substances, or the influence of existing classifications themselves. The authors also note that a substantial minority (34%) favoured declassification of all listed substances. Respondents prioritised personal and peer experience as the dominant source shaping their harm perceptions, while formal information channels such as government campaigns and advice services were ranked less influential. The study authors take this as evidence that user experience meaningfully informs lay perceptions of drug harms. Several limitations are acknowledged. The chosen method asked respondents to rank drugs relative to one another rather than to rate them on fixed, weighted criteria; the authors note more sophisticated multicriteria approaches exist and recommend engaging users on which parameters to weight. The sample was self-selected and male-dominated, which may introduce bias and the possibility of inaccurate self-reporting cannot be excluded. The authors also acknowledge they did not distinguish cocaine powder from crack cocaine and that they could not independently verify respondents' drug histories or capture finer-grained details such as order of drug initiation. Despite these caveats, the authors conclude the data support the view that certain Class A drugs may present relatively lower potential for harm yet some therapeutic potential, while alcohol and tobacco appear to carry comparatively high potential for harm.