Interpreting wastewater data alongside other population-level indicators of stimulant use

Part A: Measures based on analysis of communal wastewater in many European cities, based on 7 days of sampling per year, are used by the EMCDDA as a national indicator of drug consumption. It is unclear how representative a week of sampling might be of a year. Few studies have compared long term trends in wastewater data with other indicators of drug consumption.

Part B: The size of the European drugs market has previously been estimated using General Population Survey (GPS) data. This likely provides an under-estimate of the market size and relies on a number of assumptions about typical consumption. Wastewater data could provide an alternative or complementary basis for market size estimates.

Methods:

Part A: We fitted hierarchical regression models to wastewater data (population-normalised quantities of stimulants or their metabolites) in three locations with frequent, long term sampling: Oslo (Norway), South-East Queensland (Australia) and Eindhoven (the Netherlands). We compared evidence for trends over time with trends in other epidemiological indicators of stimulant use, e.g. treatment demand, driving under the influence of drugs statistics and police seizure data. In South East Queensland, we explored whether changes in methamphetamine in wastewater could be explained by changes in purity. We also compared estimates of annual averages from our regression model with estimates based on one week of data.

Part B: We explored the potential for wastewater data as the basis for ‘sizing the market’ of stimulant use in a city or country. We compared wastewater-based estimates of the size of the Norwegian MDMA market in Norway from 2013 to 2016 with estimates based on GPS data.

Results and Conclusions:

Part A: We found long-term trends in stimulants or their metabolites in communal wastewater to be broadly consistent with trends in other indicators of stimulant use in two of the three cases we examined. However, all three wastewater data sets exhibited a large amount of day-to-day variation. In Oslo, estimates of annual average measures based on long term data were inconsistent with estimates based on a single week of sampling in half of cases examined. Apparent changes over time based on such a sampling scheme should therefore be interpreted cautiously. It is important also to interpret wastewater data in the context of other evidence, particularly with regard to purity. In South East Queensland, increases in methamphetamine in communal wastewater between 2009 and 2016 could be partly, but not entirely, explained by increases in purity.

Part B: Estimation of market size from wastewater data presents its own difficulties, particularly with regard to extrapolating from a wastewater treatment plant catchment area to the rest of the country. If wastewater sampling occurs in higher prevalence areas, the approach could provide an upper bound on the plausible market size, to be interpreted together with a lower bound provided by the GPS-based approach.