The present analyses highlight that the share of substance use treatment entrants who seek help for cannabis use problems has increased slightly between 2013 and 2020 on average in European countries. It is estimated that cannabis problems are the main reason to seek help for about one-third of treatment entrants. While the rise in cannabis-related treatment entrants is larger than for other drugs, the treatment demand increases at a slower pace than the prevalence of (near) daily cannabis use. If the number of (near) daily users is a reasonable proxy for changes in treatment needs over time, then the treatment gap is growing in most European countries.

Analyses of routine data can only be as accurate as the documentation and reporting system. The restricted temporal data availability for some countries (e.g., no data from the UK after 2019) has constrained the analyses to subsets of countries, limiting the generalisability of the presented findings. Analysing treatment demand for cannabis use problems requires a common definition of treatment and cannabis use problems to be applied across all countries, but it is unclear to what degree existing guidelines [12] are put into practice. Moreover, the treatment systems largely differ between countries, with some mostly reporting data from outpatient facilities and other referring mostly to interventions delivered in prisons, low-threshold treatment facilities or hospitals [11]. Based on the data available, it cannot be ascertained what kind of services are effectively delivered in which settings to people seeking help for their cannabis use problems. Similarly, further research is required to assess to which degree the increasing trend in the CATF is due to declining use of other drugs (for changes in any drug treatment trends, see Supplementary Fig. 1) or changes in treatment availability.

In absence of more accurate information, treatment need was measured by survey data on the number of people who used cannabis (near) daily. However, (near) daily cannabis use is not a clinical diagnosis but indicates a potential risk for developing a cannabis use disorder [5, 30]. These prevalence data were derived from (general population) surveys which may be distorted by sampling bias (e.g., under-sampling people who use drugs, not surveying persons with no fixed address), response bias and stigma (e.g., not acknowledging frequent cannabis use because of its illegal status). One recent study showed that common survey techniques largely underestimate (frequent) cannabis use in Sweden [1]. Lastly, the frequent use of cannabis may lead to problems later in life, but this potential time lag is not considered in these analyses.

The fraction of cannabis to any-drug treatment demand (expressed as CATF) varies considerably across the European continent. Generally, the CATF is rather low in many Eastern European countries, suggesting that other substances are responsible for a greater share of treatment demand in this region. Conversely, cannabis makes up a sizeable share of entrants in Central and Western European countries. This pattern generally compares to the prevalence of use, which is highest in France, Spain, and neighbouring countries, while cannabis use remains rather uncommon in Eastern European countries [20].

While more information on those entering treatment and the characteristics of treatment systems will be required to fully understand the identified country differences, the two indicators CATF and TUR together provide important insights. For instance, the CATF in Portugal is slightly above the European average, with about 1 out of 3 treatment entrants receiving support for cannabis problems. However, only about 1 in 20 people who use cannabis (nearly) daily are estimated to have received treatment—lower than in most other countries. This gap could be explained in various ways. First, a sizeable number of people in treatment may not be covered in the TDI due to the limited coverage of the TDI documentation system in Portugal. While current information on treatment demand coverage in Portugal is not available, the share of cannabis treatment entrants referred by legal sources is relatively high in this country (Portugal: between 32 and 53%; European average: between 19 and 27%; see Supplementary Figs. 2 and 3). It should be noted that cannabis possession is decriminalized in Portugal but violations against drug laws can still be sanctioned with coerced treatment. Considering the liberal cannabis policies, the high share of legal referrals could indicate that not all treatment settings, especially medical settings with high rates of self-referrals, are appropriately covered in the TDI data from Portugal. Second, the prevalence of (near) daily cannabis use is higher in Portugal than in most other European countries. A higher prevalence of (near) daily cannabis use could first indicate a higher treatment need in this country. Alternatively, it is also conceivable that the comparably liberal drug policy environment in Portugal has normalized cannabis use, which may have minimized reporting biases in surveys. Assuming that users in Portugal are more willing to disclose sensitive behaviour than users in countries with more repressive environments, this would introduce a bias in the treatment need estimator as employed in this study.

This example highlights the need to view treatment demand from different perspectives. With both the CATF and the TUR, key (but not all) determinants of treatment demand are controlled so each indicator provides a different perspective on the same problem. The limitation of each indicator is that their interpretation depends on secondary data, i.e., consumption and treatment for other drugs (CATF; see also Supplementary Fig. 1) and cannabis treatment need (TUR). However, the combination of both indicators provides for a more complete picture of treatment demand for cannabis problems.

To gain a comprehensive understanding of the differences in the CATF and TUR among countries, it is necessary to gather more detailed information regarding the population who are registered as treatment entrants. Specifically, it would be of interest to ascertain the number of individuals who fulfil the diagnostic requirements for a cannabis use disorder and the settings (e.g., hospitals, general practitioners) where treatment is provided. Additionally, it is crucial to comprehend the range of services available to individuals and the duration of treatment. Lastly, policy approaches will need to be considered, as discussed for Portugal above. The large variations in legal drug referrals across countries (see Supplementary Fig. 2) as well as the increasing trend over time (Supplementary Fig. 3) may explain country-specific trends, which should be followed up in future research.

There are several hypotheses for the growing significance of cannabis in substance treatment facilities in Europe. First, the treatment demand follows the trends in treatment need, approximated using the (near) daily use prevalence. Importantly, the rising treatment demand does not seem to keep up with the rising treatment need, resulting in a possibly widening treatment gap. Second, changes in treatment referrals could contribute to an apparent increase in treatment demand. According to the latest available EMCDDA data, 26% of cannabis treatment entrants were legal referrals, i.e., from police, probation, or courts (see also Supplementary Fig. 2). Across all substances, only 15% were legal referrals [10]. The rising importance of legal referrals for cannabis (see Supplementary Fig. 3) may be one of the drivers for increased cannabis treatment demand observed in many countries. Third, the rising treatment demand could be explained in part by the increased use of synthetic cannabinoids. Overall, only a small minority of people report using synthetic cannabinoids on purpose [18], however, the risks for users are likely greater when exposed to this heterogeneous class of substances [4]. Robust data on treatment entrants for synthetic cannabinoids is not available, disallowing testing this hypothesis. Based on the available data from 2021, there were 685 treatment entrants for synthetic cannabinoids recorded in 15 countries, contrasting to 14,777 any-cannabis entrants (4.6%) [10]. The share of synthetic among any cannabis treatment entrants exceeds 5% in Bulgaria, Malta, and Turkey.

I can only speculate on reasons why CATF is increasing while TUR is decreasing. In US-based household surveys, a declining trend in the prevalence of cannabis use disorder among (near) daily cannabis users was observed between 2002 and 2016 [32], suggesting declining treatment demand among frequent users. It has been hypothesized that policy changes, risk perception, and reporting biases could explain these findings and it cannot be excluded that similar trends are occurring in European countries. I want to add additional thoughts on the discrepancy of trends in CATF and TUR. First, the prevalence of (near) daily cannabis use may increasingly capture groups that are considerably less likely to require treatment for cannabis use problems. This includes people using cannabis (primarily) for medical purposes and those using (primarily) low THC/high CBD products. Following legislative changes in several countries (e.g., Czechia 2013; Germany 2017), medical cannabis use may rise considerably in European countries but empirical data is scarce. General population surveys I am aware of do not explicitly distinguish medical from recreational use (e.g., Germany: [28]; Austria: [35]; Spain: [34]). As for low-THC/high-CBD products, the rising popularity has resulted in about 1 in 10 adults in Germany reporting using those products, with a sizeable share (18%) reporting daily consumption [16]. In Austria, the prevalence of high-THC and low-THC/high-CBD products was comparable in 2020 (3–5%: [35]). If population surveys do not explicitly remove those users from prevalence estimates, then the increasing prevalence of (near) daily use may be driven by increased medical use or use of low-THC/high-CBD products, diluting the robustness of the (near) daily use prevalence as an indicator for treatment need. A last reason why the TUR is decreasing during times of increasing CATF may be related to changes in documentation of treatment demand. As shown in Supplementary Fig. 1, there has been a substantial drop in the number of documented treatment entrants for cannabis and other drugs in Germany after 2015, driven by the introduction of a new documentation manual (“Kerndatensatz 3.0”, see [6]). With nearly 26,000 treatment entrants in 2021, Germany makes up a sizeable share of all treatment entrants documented across Europe, so, likely, the drop in the number of treatment entrants observed in 2017 across all countries (see Table 2) is driven by the documentation change in Germany. With the prevalence of near-daily use rising and the absolute number of treatment entrants falling, this results in a declining TUR for Germany. Considering these methodological problems, it is important to have robust and comparable estimates of treatment need, i.e., the prevalence of cannabis use disorders, in European countries.

Overall, my findings suggest that European countries also observe a declining trend of absolute treatment entrants for cannabis and other drugs as observed in the US [24]. However, in Europe, the relative importance of cannabis in substance use treatment settings continues to rise, while it decreases in the US, especially in regions where cannabis has been legalised [23]. It appears likely that the policy environment impacts treatment demand, directly or indirectly. While an increasing number of European countries are liberalising their cannabis regulations, we are only beginning to understand how the policy environment impacts consumption, risk perception, and treatment for cannabis problems [22].