Untitled

Searching the Internet for drug related websites; analysis of onlineavailable information on ecstasy (MDMA) Paolo Deluca1, PhD; Fabrizio Schifano1, MD, MRCPsych; and the Psychonaut research group2 1St. George’s, University of London, London (UK), 1 Division of Mental Health – Section of Addictive Behaviour,2Agosti Lisa (UK), Baldacchino Alex (UK), Bonan Caterina (I), Bothas Heikki (SF), Brigada Raffaella (I), Comacchio Anna (I), Di Furia Lucia (I),Eastwood Dorte (I), Farre Magi (E), Ferreira Susana (P), Flores Irene (P), Gastpar Markus (D), Ghodse Hamid (UK), Guionnet Claude (F), HarderLisbet (DK), Stokholm Jensen Lene (DK), Leoni Mauro (I), Littlejohn Christopher (UK), Majava Aino (SF), Peltoniemi Teuvo (SF), Pizza Milena (I),Prandin Cristian (I), Rawaf Salman (UK), Robert Damien (F), Rossi Maria Angela (I), Rota Deborah (I), Rovetto Francesco (I), Scherbaum Norbert(D), Siemann Holger (D), Tarrago Josep (E), Torrens Marta (E), Zambello Francesco (I) Corresponding Author:
Paolo Deluca, PhD
St George’s, University of London
Senior Researcher Fellow
Division of Mental Health – Section of Addictive Behaviour
Cranmer Terrace
London, SW17 0RE
Website: http://www.psychonaut2002.org
United Kingdom
Phone: +44 20 8266 6426
Fax: +44 20 8725 2914
Email:
Background: The emergence of the Internet as an unregulated source of controlled substances is an important development
that may have significant public health implications. Although the Internet is an essential source of information on MDMA, no
studies have addressed the issue of what is available to the average web user in terms of information on MDMA.
Objectives: Our study aimed at identifying, developing and evaluating a reliable methodology for sampling the Internet for
relevant information related to MDMA and to provide a description of the type of information available.
Methods: This study investigated 280 web pages generated by two search engines in running queries for MDMA (ecstasy).
Such queries were carried out in English using a single generic keyword (MDMA) on a week in June 2003 to overcome the
problem of the ever changing and expanding nature of the Internet. The Google™ and AltaVista™ search for MDMA (ecstasy)
related websites respectively provided a list of 71400 and 48747 web pages. The first 100 websites identified by both search
engines, together with a further random sample of 5% of the remaining sites (101 to 1000) were fully assessed. Since Google™
displayed less than 1000 web pages, the snapshot overall counted 280 web pages.
Results: Of the 280 web pages' links (i.e., 145 from AltaVista™ and 135 from Google™), 23 (i.e., 17 identified by Google™
and 6 by AltaVista™) were offline or with restricted access at the time of their assessment, thus reducing the sample to 257.
Overall, we identified 49 (19.1%) irrelevant websites (25 with Google™ and 24 with AltaVista™). Only Google™ showed a
statistically significant difference between relevant and irrelevant websites’ ranking, having the former an average value of 80.69
and the latter a value of 142.32 (t116 = -2.55, P < .02). In considering the first 100 results within each search engine, both Google™
and AltaVista™ produced 18 duplicate links. Moreover, thirty-three out of 164 websites were quoted by each search engine and
49 websites were unique to each search engine. The ranking of these 66 (33 per search engine) websites, in comparison to those
98 (49 per search engine) unique websites was significantly different (t162 = -6.315, P < .001), having the common websites a
lower ranking level (37.3) compared to the others (63.3). Regarding websites’ position towards drug use, 50.4% were anti drug,
16.2% harm reduction and 24.8% pro drug. The MDMA pro drug websites showed a lower ranking (mean = 40.34; SD = 32.5),
hence appearing significantly earlier in the results’ list, than both the anti drug (mean = 55.0; SD = 26.6) and the harm reduction
(mean = 54.25; SD = 22.9) websites (F (3; 159) = 3.288; P = .022).
Conclusions: This study represents the first systematic analysis of the information available online on ecstasy. Health professionals
may need to be aware of the web being a new drug resource for information where pro drug use messages are more likely to be
encountered by web users.
(page number not for citation purposes) KEYWORDS
MDMA; Ecstasy; Internet; Psychonaut 2002 project; drug abuse; Google™; AltaVista™; search engine amphetamines, cannabis, cocaine and crack cocaine, dancedrugs (i.e., ketamine, LSD, GHB), ecstasy, herbs and plants, The Internet with all its web pages, databases, libraries, forums, heroin and opiates, inhalants and solvents, precursors to illicit blogs and online shops, has become an enormous resource where drugs, prescription drugs, and tobacco. Search queries were most of the knowledge is stored and accessible. Accessibility, however, in terms of usefulness of this extensive amount of The present study focused exclusively on the findings generated information is always a problem. Searching for relevant by a search carried out in English on ecstasy (MDMA).
information on the World Wide Web is often a laborious and On the week starting on June 23rd, 2003, a “snapshot” of the frustrating task for casual and experienced users alike. Word of available websites with contents related to MDMA (ecstasy) mouth might be a way of identifying the right direction towards was carried out, to overcome the problem of the ever changing the right website/resource. However, search engines are both and expanding nature of the Internet. Search engine queries the quickest and most popular way for finding information were carried out using a single generic keyword (ie, MDMA).
online []. Indeed, new algorithms and technologies are It is important to note here that different keywords or constantly under development to improve their reliability in combination of words would have inevitably generated different retrieving relevant information and this will be a major outputs. However, it was felt that the choice of a single generic contribution on the development of Internet in the coming years keyword was less biased than for instance searching for ‘buying or manufacturing MDMA’. Moreover, MDMA was preferred to the term ‘ecstasy’ because during pilot testing on various legal and illicit drugs are no exceptions. In the last 5 years there keywords it emerged that the term MDMA generated more has been a growing interest on the availability of both drug Data was collected and stored in an online, password-protected, database. Searches for MDMA (ecstasy) related websites provided a list of 71400 and 48747 web pages respectively with search engines, both are however designed for displaying only Among these substances, ecstasy has received limited attention up to the first 1000 pages found []. The sampled websites, [], notwithstanding that it is now a very popular drug after therefore, were selected using a purposeful sampling technique.
a global escalating trend in the second half of the 1990s and The first 100 websites identified by Google™ and AltaVista™ that currently around 2.6 million adults in the European Union with the chosen keyword were fully assessed, together with a (0.8% of all adults) are estimated to have used ecstasy recently ] sample of 5% of the remaining sites (101 to 1000). Since Google™ displayed less than 1000 web pages and essential source of information on how to use/consume listed, the snapshot overall counted 280 web pages.
Website’s assessment
[]. However, if the information about a drug contribute to], monitoring the information available to the Although Google™ and AltaVista™ results would point to a average web user becomes crucial. To the best of our knowledge single webpage, we mapped and evaluated the whole website nothing has yet been published on the content of what is by accessing all its sections. Survey data of sampled websites available to the average web user in terms of MDMA included 64 variables, including: website relevance, ranking, information on the Internet. Moreover, none of previous studies country of origin, website position towards drug use, and have addressed comprehensively the methodology they have used for finding and evaluating identified websites.
Relevance referred to the content of the information offered bythe website, which was not always pertinent to the aims of the search; ranking referred to the link’s numerical order ofpresentation in the engine results’ page. The country of origin This study aimed at identifying, developing and evaluating a of a website was not routinely identified through the URL reliable methodology for both sampling and assessing the domain extension (i.e., .uk, .com, .it, etc) but with the help of Internet for relevant information related to MDMA and to ] and other information available (i.e., postal provide a description of the type of information available. This address, telephone numbers etc). The categorization of the exercise was part of a larger European Collaborative Study, the website position towards drug use included four main groups: Psychonaut 2002 project that investigated the availability of anti drug, pro drug, harm reduction approach, and not stated.
Anti drug websites typically advocated against the use of any The Psychonaut 2002 project investigated thousands of web drug; most usually, they offered either prevention and/or pages generated by two search engines in querying for 11 treatment information. Conversely, pro drug websites were psychoactive substances, or groups of substances, including: actively promoting or facilitating the use of psychoactive drugs.
(page number not for citation purposes) Typically, they provided information on how to synthesize, how to synthesize MDMA were also providing information on purchase or consume a variety of substances. Harm reduction websites neither condemned nor promoted the use of drugs; Interestingly, the websites’ ranking levels did significantly differ they aimed instead at presenting evidence and facts, giving the as a function of the website position towards drug use in both visitors the possibility to draw their own conclusions. In those search engines (analysis here was limited to the first 100 links cases in which no statement was found and/or when the website in both search engines, F (3; 159) = 3.288; P = .022). Pro drug aims were unclear, the website was given a not stated position.
websites showed a lower ranking level (mean = 40.34; SD = Data analysis
32.5), hence appearing significantly earlier in the results’ list, Descriptive statistics were used to assess a number of than both the anti drug (mean = 55.0; SD = 26.6) and the harm parameters, including: information on country of origin, reduction (mean = 54.25; SD = 22.9) websites. The Tukey’s relevance of findings, type of information, and position towards HSD post-hoc test confirmed that MDMA pro drug websites drug use. Chi-square analyses were carried out to test association had a lower (P < .05) ranking than MDMA anti drug websites.
between different types of information provided and websites Distribution of relevant websites between and within
position towards drug. Differences in terms of findings between the two search engines
two search engines were tested for statistical significance by The second set of analyses focused on the quality of results generated by both search engines. We analyzed the distribution To check for websites’ ranking number as a possible function of irrelevant websites (i.e., those not referring to drugs but for of their position towards drug use, the one-way ANOVA and example to music, computer viruses etc.) within and between the Tukey’s HSD post-hoc tests were both applied. The the search engines. Overall, we identified 49 (19.1%) irrelevant St.George’s, University of London, Local Research Ethics websites. Irrelevant websites were almost equally distributed Committee granted the ethical approval of the project.
between the two search engines (25 in Google™, and 24 inAltavista™). By means of t-test analysis we investigated where these websites were more likely to appear. It emerged thatGoogle™ had a statistical significant difference between Of the 280 web pages links, 145 were from AltaVista™ and Relevant and Irrelevant websites, having the former a ranking 135 from Google™, 23 (ie, 17 identified by Google™ and 6 by mean of 80.69 and the latter a ranking mean of 142.32 (t116 = AltaVista™) were either not available online (offline) or access -2.55, P < .02). On the other hand, AltaVista™ did not show was password protected at the time of their assessment. An additional 49 (19.1%) websites were not relevant for MDMA 137 = 0.475, P > .05), although the Irrelevant websites showed a ranking mean of 185.79 and the content, thus reducing the sample to 208, 115 and 93 for relevant ones a ranking mean of 216.89.
Google™ and AltaVista™ respectively.
Repetitions and overlapping between the two search
Although for 21 (10.1%) websites it was not possible to identifytheir country of origin, 133 (63.9%) were from the United States, followed by 23 (11.1%) from the United Kingdom. With respect The third and final set of analyses focused on the type of to the websites’ position towards drug use, it emerged that 98 findings reported by AltaVista™ and Google™. In considering (47.3%) of them showed a clear anti-drug position, 20 (9.7%) the first 100 results within each search engine, in AltaVista™ did not state clearly their view, 34 (16.4%) websites showed a we identified 18 duplicate links, resulting in 82 unique websites.
harm reduction approach and 55 (26.6%) a pro-drug stand.
Similar results were found in Google™, where 18 links were Moreover, twenty (9.6%) websites provided detailed information duplicates. In comparing those 164 unique websites, it was on how to manufacture/synthesize MDMA (18 Pro drug and 2 found that 33 were quoted in both search engines databases leaving the other 49 websites as unique to each search engine.
3 = 47.6, P < .001). Further, thirty-five (17.2%) websites informed their visitors on how to use MDMA In comparing the ranking of these 66 (33 per search engine)websites with that of those 98 (49 per search engine) unique (32 Pro drug, 11 Harm reduction, and 1 Anti drug, x23 = 52.6, websites, we found a significant difference between them (t162 P < .001) and just 57 (27.7%) had clear warnings on the = -6.315, P < .001), having the common websites a lower mean consequences of using MDMA (19 Pro drug, 8 Harm reduction, (37.3) compared to unique websites (63.3). Therefore popular 26 Anti drug, and 4 Not stated, x23 = 2.24, P > .05). However, websites, which were quoted in both search engines, were also warnings were slightly more likely to appear on websites appearing on top of the ranking compared to those websites that providing information on how to use MDMA. Indeed, 51.4% were instead unique for each search engine.
of these websites provided health warnings (x21 = 12.52, P <.001). Similarly, 80% of websites providing instructions on how to manufacture/synthesize MDMA did also provide their The results showed that even though anti drug web sites were customers with warnings/disclaimers (x21 = 31.21, P < .001).
almost double in number compared to those promoting the use Information on how to use and how to make MDMA were also of MDMA, anti drug websites were statistically more likely to more likely to appear together on the same website (x21 = 83.71, appear after pro drug websites in the results’ list of both search P < .001). In addition, ninety percent of websites informing on engines. In other words, information at risk related to MDMA (page number not for citation purposes) seemed to be available significantly earlier to the average user.
probability of identifying irrelevant sites significantly increases This finding, together with what has already emerged from other in the 101-1000 list. Google™ seemed here to perform better studies on the quality of the information available online, which than AltaVista™, at least with respect of identifying earlier is often a mis-information, may contribute to the growing relevant MDMA-related websites. Popular websites, quoted by evidence of Internet being an unregulated source on controlled both search engines, were appearing on top of the ranking substances and this may have significant public health compared to those websites that were instead unique to each search engine. In other words, if a site is popular it is likely bepicked up by all the search engines and will be offered at the Since search engines are the main gateways to the information top of their ranking. However, it was surprising that 49% of the available online, one might think that they should be accountable websites were unique to each search engine. This might confirm for the information they provide. Alternatively, they might want the underlying assumption that search engines crawl, index and to filter out those websites that have a clear pro drug stance rank the web in different ways and may also be following quite and/or sell drugs online. Deliberate, localized, search results’ different criteria when reporting links related to the same query.
exclusions are already occurring for sensitive or illegal This would support the need of using multiple search engines items/topics (eg, Nazi memorabilia in France and Germany) to is not apparently the case for online drug shops selling illicit In conclusion, it is important to note that only one fixed keyword and licit drugs thus apparently breaching the law in most nations.
was used in this study therefore statistical data about the One might argue that these websites should not be listed in underlying nature of each website (ie, pro drug vs. anti drug, search engines, so that it would become less and less profitable ranking positions etc) may be merely a reflection of the bias of setting up websites to promote and sell illicit drugs.
the keyword used in the search. Percentages reported are whollydependent on what was searched. However, we aimed here at Apart from being used for searching for relevant terms, search exploring what a generic drug name would generate from those engines are also a valuable source for discovering and analyzing search engines. Finally, due to the rapidly changing nature of search patterns, trends and sudden events that occur around the the web any report on Internet issues, including the present one, globe. This could be a valuable source of information for might be considered somewhat out of date. Further studies identifying the emergence of new drugs or new trends in drug should be carried out to shed more light on what information use. This information could be used in setting up an Early is available online and how it might change overtime and by using different keywords or combination of words.
It also appeared from this study that searching relevant datawithin the first 100 websites may be informative enough, as the Acknowledgments
This paper has been supported by the grant number SPC 2002 306 (2002-2004) of the EU DG – Public Health and Risk Assessment.
The conclusion and interpretation of the findings of this study reflects the authors’ views and the Commission is not liable forany use that may be made of the information contained in this publication.
PD collected the data and wrote the manuscript, FS is the grant holder and contributed to and commented early drafts of themanuscript.
Conflicts of Interest
References
1.
Holscher C, Strube G. Web search behavior of Internet experts and newbies. Comput Netw 2000 33(1-6):337-346.
O'Leary DE. Internet-based information and retrieval systems. Decis Support Syst 1999 27(3):319-327.
Search Engine Watch. Search Engine Ratings and Reviews. URL: http://searchenginewatch.com/reports/index.php [accessed2006 January 13] Baujard O, Baujard V, Aurel S, Boyer C, and Appel RD. Trends in medical information retrieval on Internet. Comput BiolMed 1998 Sep;28(5):589-601. Medline:9861514 Boyer EW, Shannon M, and Hibberd PL. Web sites with misinformation about illicit drugs. N Engl J Med 2001 Aug9;345(6):469-71. Medline:11496870 Schechter MD. 'Candyflipping': synergistic discriminative effect of LSD and MDMA. Eur J Pharmacol 1998 Jan12;341(2-3):131-4. Medline:9543229 M. Gordon S, Forman RF, and Siatkowski C. Knowledge and use of the internet as a source of controlled substances. JSubst Abuse Treat 2006 30(3):271-274.
Drug Enforcement Administration. DEA announces arrests of website operators selling illegal designer drugs. URL:http://www.dea.gov/pubs/pressrel/pr072204.html [accessed August 19 2004] (page number not for citation purposes) Siemann H, Specka M, Schifano F, Deluca P, and Scherbaum N. [Salvia divinorum - representation of a new drug in theinternet.]. Gesundheitswesen 2006 May;68(5):323-7. Medline:16773554 Bucheler R, Gleiter CH, Schwoerer P, and Gaertner I. Use of nonprohibited hallucinogenic plants: increasing relevancefor public health? A case report and literature review on the consumption of Salvia divinorum (Diviner's Sage).
Pharmacopsychiatry 2005 Jan;38(1):1-5. Medline:15706458 Dennehy CE, Tsourounis C, and Miller AE. Evaluation of herbal dietary supplements marketed on the internet for recreationaluse. Ann Pharmacother 2005 Oct;39(10):1634-9. Medline:16159994 Falck RS, Carlson RG, Wang J, and Siegal HA. Sources of information about MDMA(3,4-methylenedioxymethamphetamine): perceived accuracy, importance, and implications for prevention among youngadult users. Drug Alcohol Depend 2004 Apr 9;74(1):45-54. Medline:15072806 Schifano F, Deluca P, Agosti L, et al. New trends in the cyber and street market of recreational drugs? The case of 2C-T-7('Blue Mystic'). J Psychopharmacol 2005 Nov;19(6):675-9. Medline:16272191 Schifano F, Deluca P, Baldacchino A, et al. Drugs on the web; the Psychonaut 2002 EU project. Prog NeuropsychopharmacolBiol Psychiatry 2006 Jun;30(4):640-6. Medline:16458404 Micke MM. The case of hallucinogenic plants and the Internet. J Sch Health 1996 Oct;66(8):277-80. Medline:8899584 Halpern JH, Pope HJ. Hallucinogens on the Internet: a vast new source of underground drug information. Am J Psychiatry2001 Mar;158(3):481-3. Medline:11229993 Halpern JH. Hallucinogens and dissociative agents naturally growing in the United States. Pharmacol Ther 2004May;102(2):131-8. Medline:15163594 Compton WM, Volkow ND. Major increases in opioid analgesic abuse in the United States: concerns and strategies. DrugAlcohol Depend 2006 Feb 1;81(2):103-7. Medline:16023304 Forman RF. Availability of opioids on the Internet. Jama 2003 Aug 20;290(7):889. Medline:12928464 Wax P, Reynolds N. Just a click away: student Internet surfing for recreational drug information. J Toxicol Clin Toxicol200038(5):531.
Wax PM. Just a Click Away: Recreational Drug Web Sites on the Internet. Pediatrics 2002 June 1, 2002;109(6):e96-.
Henney JE, Shuren JE, Nightingale SL, and McGinnis TJ. Internet purchase of prescription drugs: buyer beware. AnnIntern Med 1999 Dec 7;131(11):861-2. Medline:10610633 Binns R, Driscoll B. The Internet, pharmaceuticals and the law. Drug Discov Today 2001 May 1;6(9):452-453.
Medline:11344025 Littlejohn C, Baldacchino A, Schifano F, and Deluca P. Internet pharmacies and online prescription drug sales: across-sectional study. Drugs-Educ Prev and Polic 2005 12:75-80.
EMCDDA, Annual report 2005: the state of the drugs problem in Europe. 2005.
Scholey AB, Parrott AC, Buchanan T, et al. Increased intensity of Ecstasy and polydrug usage in the more experiencedrecreational Ecstasy/MDMA users: a WWW study. Addict Behav 2004 Jun;29(4):743-52. Medline:15135556 Fishbein M, Middlestadt SE. Using the theory of reasoned action to develop educational interventions: Applications toillicit drug use. Health Educ Res 1987 2:361–371.
Google. Google. URL: http://www.google.com [accessed 2006 May 22] AltaVista. AltaVista. URL: http://www.altavista.com [accessed 2006 May 22] Research Randomizer. Reseach Randomizer: Instant Random Sampling and Random Assignment. URL: www.randomizer.org[accessed 2004 May 6] ARIN. ARIN WHOIS Database Search. URL: http://www.arin.net/whois/ [accessed 2005 June 6] St George BN, Emmanuel JR, and Middleton KL. Overseas-based online pharmacies: a source of supply for illicit drugusers? Med J Aust 2004 180:118–9.
BBC. Google censors itself for China. URL: http://news.bbc.co.uk/2/hi/technology/4645596.stm [accessed 2006 January25] Webcite:5GpdhR5Sn Zittrain J, Edelman B. Localized Google search result exclusions. URL: http://cyber.law.harvard.edu/filtering/google/[accessed 2006 May 22] (page number not for citation purposes)

Source: http://www.mednetcongress.org/fullpapers/MEDNET-226_DelucaPaoloA1_e.pdf