Page not availableThis non-medical use of AAS is associated with significant cardiovascular, endocrine, psychological and psychosocial morbidity Anabolic trend and Vorona, aNieschlag anaoblic Vorona, b. The public health significance of AAS injecting becomes more significant with the identification of rising rates of blood borne virus prevalence in this anabloic Hope et al. Anabolic trend reliance of self-reporting of AAS use from a secretive group, uncertainty among users about the consumed anabolic trend and indeed what is defined anabolic trend AAS limits the usefulness of this data Home Office, Increasing numbers of AAS users presenting to Anabolic trend, alongside arrests, drug seizures, media reports, internet forum monitoring and case reports venta anabolicos buenos aires a picture of increasing prevalence Evans-Brown et al. Nevertheless, current data sources do not appear to provide a real-time indication of variation in prevalence, potentially limiting the responsiveness of services and public health policy. Furthermore, it is increasingly identified as a anabolic trend driver for the dissemination of media representations of physical ideals, with consistent evidence associating social media usage with body dissatisfaction Sassatelli,
Tracking internet interest in anabolic-androgenic steroids using Google Trends
This non-medical use of AAS is associated with significant cardiovascular, endocrine, psychological and psychosocial morbidity Nieschlag and Vorona, a , Nieschlag and Vorona, b. The public health significance of AAS injecting becomes more significant with the identification of rising rates of blood borne virus prevalence in this group Hope et al.
The reliance of self-reporting of AAS use from a secretive group, uncertainty among users about the consumed substances and indeed what is defined as AAS limits the usefulness of this data Home Office, Increasing numbers of AAS users presenting to NSP, alongside arrests, drug seizures, media reports, internet forum monitoring and case reports provide a picture of increasing prevalence Evans-Brown et al.
Nevertheless, current data sources do not appear to provide a real-time indication of variation in prevalence, potentially limiting the responsiveness of services and public health policy. Furthermore, it is increasingly identified as a key driver for the dissemination of media representations of physical ideals, with consistent evidence associating social media usage with body dissatisfaction Sassatelli, Internet searches, for example through Google, are used to access information and consumer products online These searches provide data streams which can then be analysed through online services such as Google Trends GT.
GT is a free, open access online portal which allows users to analyse part of 3. This internet tool provides data on geographical and temporal patterns in user-specified search terms.
Epidemiologists are increasingly cognisant of the value of this data to complement traditional sources, providing actionable intelligence to policy-makers Nuti et al. Examples of GT research includes the tracking of flu epidemics Eysenbach, , and the use of drugs such as novel psychoactive substances Deluca et al. The latter two examples highlight the use of GT in situations where traditional systems may be lacking and where the Internet plays an important role in consumption.
Consequently, this study is intended to assess the feasibility of using GT to supplement what we know about AAS related behaviour. Reference was made to Nuti et al. An important issue in performing the analysis was to identify search terms which corresponded closely with what was available for sale online. Where an AAS was known by a generic and brand name, the term that produced the greater search volume was chosen. This approach has its limitations; in particular as systematic variations may exist depending on whether a generic or brand name was used.
Combining the searches was avoided in this study due to the risk of double counting as the charts for generic and brand names often produced similar time trends. When performing a GT query, the user interface allows for the narrowing down of the scope of the search to set categories such as business, health or sports. All categories were allowed in the search as non-medical AAS use may be viewed in different ways, for example as a health issue or part of fitness and beauty consumerism.
The GT data were accessed and downloaded on 14 January The time period selected was from 1 January to 31 December This start date was selected as the GT interface reported an improvement in geo-locating data at this time. A significant limitation of GT for research is that Google maintains a high level of secrecy around their algorithms Eysenbach, , and more details on this improvement is unavailable. The GT searches were restricted to the UK as this was the area of interest for this project.
It is important to note that a worldwide trend analysis would be difficult to interpret, due to differences in legislation and socio-cultural variations in AAS use. This interface allows for a simultaneous comparison of 5 search terms. The least popular search term was removed each time such that the top five were identified. The process was repeated in a second GT page and the next five, less popular search terms were identified.
This process was necessary as all data points in each time series are normalised by GT such that they are relative to the highest data point designated Google, The data points produced represent the relative search volume RSV , defined as the total number of searches for that term divided by the total number of Google searches for the specified location and time Google, Of the 15 AAS identified by Cordaro et al.
The data points were downloaded as. Visual inspection was used to determine months of upward and downward inflection for subsequent statistical tests. The Wilcoxon signed-rank test was applied to determine if there were significant differences between search volumes in the peak April-July and trough September-December months identified from the time series decomposition.
The Mann-Kendall test was used to detect overall trends significantly larger than the variance in the data for each search term. We modified significance points using the Bonferroni correction to adjust for multiple testing.
All p -values and tests are reported. The individual decomposition graphs for each AAS is available in the Supplementary material. This seasonality was statistically significant for all the AAS except Testosterone enanthate when tested with the Wilcoxon signed-rank test. Statistically significant trends were identified in all the AAS except Boldenone.
Anavar, Trenbolone, and Masteron had statistically significant upward trends. Stanazolol, Dianabol, Sustanon, Nandralone, Metenolone and Testosterone enanthate had statistically significant downward trends.
With Google having The standout finding in this study is the presence of clear and consistent seasonal variation with 9 of the 10 compounds reaching statistical significance. This observed seasonality may support the notion of steroid use being part of health and beauty consumerism Brennan et al. Health promotion and prevention messages, as well as harm reduction services may be more effective if targeted during these periods of increased internet search interest.
It is important to note however that harm reduction policies do not address the underlying neoliberal and consumerism driven risk normalisation underlying steroid use, which also needs to be addressed. Our findings do need to be interpreted cautiously for a number of reasons.
Firstly, it is not possible to know the motivation behind the Google searches. Compounds such as Trenbolone for example are used as a veterinarian steroid, and a portion of search interest may reflect this. Searches may be performed by researchers, journalists and out of curiosity and does not indicate intention to use or purchase.
Furthermore, it is possible that a Google search for one AAS provides access to websites with information on the full range of compounds, reducing the need for further searches. This study did not use colloquial terms for various steroid products, rather names derived from an earlier published study Cordaro et al.
The lack of demographic data on who is performing the Google searches limits the inferences we can make. Trends may also be influenced by a number of confounding factors such as a general web interest in other issues or incidental news stories creating a spike in search interest in specific compounds.
These limitations have been identified in other GT work for example in the prediction of influenza outbreaks Eysenbach, For the reasons above, interpretation of the ranking by RSV and the upward and downward trends is problematic. Trenbolone was traditionally considered high risk and shunned by experienced bodybuilders in the s Monaghan, and its rise in internet search popularity may therefore warrant further investigation.
The large number of searches occurring for Dianabol and Anavar may have significance with regards to current health service provision.
It is worth noting that internet searches are far less likely to be met with sanctioned harm reduction or health promotion messages with such websites being substantially overwhelmed by pro-steroid content Brennan et al. The presence of fluctuating internet search volume of individual compounds, in the context of a perceived rise in AAS use overall, may be an area worth further investigation.
While it is not possible to draw conclusions on how search interest relates to levels of use, time trend variation in levels of interest in different AAS may gives some indication of what is occurring within the wider AAS using community. More detailed web-analytics work on internet vendor sites could link web searches to purchases Curtis et al.
This is of particular interest if prospective users seem to show greater interest in substances which are associated with higher risks. In conclusion, with current limited real-time, detailed AAS use surveillance, alternative data sources such GT may provide useful additional information. Cautious interpretation of GT data is recommended and future work involving triangulation with other data sources may make this method useful as a near real-time barometer of AAS interest, and variations over time.
This project was carried out as part of a dissertation for the MSc in Drug and Alcohol studies, University of Glasgow. The authors would like to thank Joyce Nicholson, lecturer for the MSc programme for her support in this project. Appendix A Supplementary data associated with this article can be found, in the online version, at https: National Center for Biotechnology Information , U. The International Journal on Drug Policy.
Int J Drug Policy. Joseph Tay Wee Teck: This is an open access article under the CC BY license http: Open in a separate window. Table 1 Statistical tests for Google Trends data. Conflict of interest statement No conflict of interest declared. Footnotes Appendix A Supplementary data associated with this article can be found, in the online version, at https: Supplementary data The following are Supplementary data to this article: Click here to view.
The new self-help culture of ethnopharmacology. Anabolic-androgenic steroids trafficking in the UK. European Journal of Criminology. Performance-enhancing drugs on the web: A growing public-health Issue. The American Journal on Addictions. The injecting use of image and performance-enhancing drugs IPED in the general population: Health and Social Care in the Community. Selling androgenic anabolic steroids by the pound: Identification and analysis of popular websites on the Internet. A cross-sectional analysis of seasonal differences in sexual behaviour and sexually transmissible diseases in Melbourne, Australia.
Identifying emerging trends in recreational drug use; outcomes from the Psychonaut Web Mapping Project. Progress in Neuro-Psychopharmacology and Biological Psychiatry. North West Public Health Observatory; The emerging challenges to public health. American Journal of Preventive Medicine. A deducer plug-In for econometrics.
Paper presented at the r user conference, useR; Albacete, Spain;