Surveying drug consumption: Assessing reliability and validity of the European Web Survey on Drugs questionnaire

Background: The European Web Survey on Drugs aimed to obtain in-depth data on consumption of cannabis, ecstasy/MDMA, cocaine, and amphetamines in different populations of drug users in 16 European countries. This paper examines test-retest reliability, the consistency and the comprehensibility of the prevalence and frequency of use questions in the Czech part of the survey. Methods: A baseline web survey was performed (N = 610) with follow-up data collection in a sub-sample of volunteers providing email addresses (N = 158). The baseline sample was self-selecting, responding to advertisements made available through multiple channels designed to attract diverse samples of drug users. Test-retest analysis was conducted for core questionnaire items. Results: Respondents to the follow-up were predominantly socially integrated; 91% reported last year cannabis use, 42% used Ecstasy/MDMA, 23% amphetamines, and 27% reported cocaine use. Test-retest reliability was rated moderate to good (reliability coefficients between 0.55-0.87) for most prevalence items with sufficient sample sizes. Items assessing frequency of use were more reliable for most substances when asking about the exact number of days used, compared to categorical items that implicitly assume a regular pattern of use and were interpreted differently by different respondents. Conclusions: Simplicity and unambiguity of questions increase the reliability of results. Tools measuring drug consumption need to take into consideration the irregularity of drug using patterns. Question testing is important to increase validity and support a correct interpretation of the data.