Understanding digital drug markets through the geography of postal drug deliveries in Scotland

Research study link: https://journals.sagepub.com/doi/full/10.1177/1477370821997323

Method – Strengths

- Real-world data are used for this study (obtained from the National Crime Agency), so the results are rich in ecological validity.

- The authors are open about their methodological approaches and the limitations of their data.

- The independent variables are informative and sensible predictors of drug use.

Method – Weaknesses

- Scotland as a case study – the study focuses on Scotland as a case study to address its research question. Scotland has a rather unique geography, with a low population (5.52 million) and low population density (67.2 people per km2). For context, the population of England is 57 million with a population density of 275.27 people per km2. For this reason, using Scotland as a case study could be considered a weakness of the study, given that its atypical characteristics lead the findings to be of limited applicability to other countries.

- The study was published in 2023, yet the data in question were extracted from the NCA database and cover the time period 2011 through to 2016. The extent to which these findings remain relevant to 2023 Scotland is difficult to predict, but one might reasonably expect that many of the independent variables utilised here (e.g., internet connectivity, access, deprivation) are substantially different now.

- The data are very limited in terms of actual numbers – 1,374 packages divided across a population of 5 million people and over a time period of 5 years. The authors also do not fully report the breakdown of the package contents (9% were MDMA, 27% were cannabinoids, and 26% were benzodiazepines, meaning that 38% remain unaccounted for).

Results – Strengths

- Novel/innovative use of figures to convey the findings and address the research questions.

- The claim that built up/high-density urban areas generally saw more drug package rates seems logical – these areas had higher internet connectivity and greater scope for deprivation zones.

- The claim/observation that if an area had a high rate of package delivery, then the neighbouring areas (data zones) were also likely to have higher rates also make good sense and seem logical. If drug package acquisition is indicative of social deprivation and internet connectivity, then it is entirely conceivable that neighbouring regions will show similar patterns, given that deprivation does not occur in pockets, but rather is continuously distributed.

Results – Weaknesses

- The passage of time is not considered as a predictor in the results. Given that the NCA data upon which the findings are based span a 5-year period, changes in drug package delivery over time would have been a relevant factor to consider.

- The type of drug package is not factored into the results. Whilst we know that similar proportions of packages contain benzodiazepines and cannabinoids, the results are not stratified based on this aspect, which might have produced different patterns and outcomes.

- The data in Table 1 reflect that, on average, 0.2 packages were identified per 1,000 people, with a standard deviation around that mean of 0.6. Given that the standard deviation is 3 times greater than the mean, this indicates that the mean is not particularly representative of the dataset. The authors do acknowledge this in the method section, however, as they comment on the dispersion of the data.