If they were available, we recorded the number of online pharmacies analyzed in each study, year of data collection, willingness to dispense pharmaceuticals with or without a prescription, availability of a physician’s assistance or online medical consultation, disclosure of contact details, geographical location, delivery conditions, types of medicines available, availability of drug information, prices of online drugs and overall costs, sales-promotion strategies, how long websites were accessible, privacy and disclaimer statements, date of last website update, and presence of quality certifications (for instance, Verified Internet Pharmacy Practice Sites [VIPPS] by the NABP; the Health on the Net Foundation [HON] code; and the Joint Commission on Accreditation of Healthcare Organizations [JCAHCO]).

We recorded studies in which the researchers bought prescription drugs online and evaluated the actual purchase and its characteristics. We summarized data regarding the type of drug ordered, the response rate, the quality of the process, and the drugs purchased. Regarding process characteristics, we recorded prescription requirements, management of the online questionnaire, money transactions, and subsequent advertising; with regard to drug quality, we recorded data about packaging and instructions, and chemical composition.

We described articles dealing with the number of people purchasing drugs online, which was estimated by means of questionnaires or interviews. Researchers attempted to list the most frequently requested drugs, the main reasons for buying pharmaceutical products online, the importance of the location of online pharmacies they bought from, and the perceived risks related to this practice. In addition, we classified in this section articles that reported clinical cases of adverse effects to active substances and drugs purchased via the Internet as an indicator of this phenomenon.

One of the most controversial and widely studied features is prescription requirement. For the first time, online pharmacies are providing easy access to traditionally controlled (according to each country’s regulations) substances, such as pharmaceutical products, which in regulated systems need an original medical prescription before they can be bought. The peculiar distribution chain for prescription drugs is due to the unique nature of pharmaceutical products; the philosophy behind prescription is that consumers are not skilled enough to make their own choice, but need to be given a prescription by a health professional trained to make a risk–benefit evaluation [2,24]. In addition, the medical follow-up of a drug treatment is crucial. Furthermore, there are problems regarding strict quality control for substances that can determine life or death. Last but not least, there are economic implications in ruled systems where drug prices are negotiated and partially funded by the health system [27].

Due to the different methods of sample selection, distribution varies greatly in the articles. Some online pharmacy samples were specifically selected to have only websites not asking for a prescription from the customer’s physician [57,67-70, 82,93]. In the samples that did not use prescription as a criterion for inclusion or exclusion, the prescription requirement varied widely: all websites asked for an original medical prescription in Wagner et al [95] (who selected only US-based online pharmacies), more than half did so in other samples [60,72,87], less than half [10,58,63,65,66,71,73,75,80,84,86,90] in still others, and none in a few [74,83,91]. In Bloom and Iannacone [61,62] it was not possible to determine whether the required prescription was an original one by the customer’s physician or replaced by an online health evaluation as explained below.

Some of the online pharmacies not asking for a prescription replaced it with an online health status evaluation, performed by means of an “online questionnaire.” Researchers reported a percentage of online pharmacies offering and/or requiring an online prescription in a range of 10% to 81% [91,93]. Those who looked deeper into this issue found that the identity of the professional who made the prescription was not usually provided [57], whereas they often declared that the questionnaire reviewer was a physician [57,61,62,66,84] or, more rarely, a pharmacist [27,84]. Gallagher and Chapman [71] reported that, of the 26 websites using a questionnaire, 1 involved multiple-choice check-box answers and drop-down menu answers, and the others used questionnaires that allowed users to type in an answer. Eysenbach [66]—who attempted to buy the products—found that the identity of the physician who made out the prescription was revealed in 2 out of 10 orders. The US Government Accountability Office (GAO) [72,73] found the collaboration of licensed physicians, although when they verified these licenses [72] almost all of them were nonexistent or not valid for all the US states where they were declared to be. No study found the use of a validated or standardized questionnaire. Orizio et al [85] did a subanalysis of a previous study [84] that focused on the characteristics of online questionnaires; it reported that the questionnaire was already filled in with negative answers in 70% of online pharmacies using this tool, and that in only 53% of cases were different questions asked for different products ordered. Since the questionnaires were often incomplete, the authors concluded that they appeared to aim more at giving the consumer a false sense of health assurance than at actually assessing health status. The questionnaire frequently investigated personal characteristics, allergies, medical conditions, current therapies, and medical history [57,61,62,66,72,73,85].

Contact details were revealed in the majority (ranging from 100% [71,75,95], or a little less among online pharmacies when this item was analyzed [58,60,72,75,87,92,95], to 66% [60]). Only Tsai et al [93] found just 37% of pharmacies giving a telephone number for costumers’ enquiries, and the European Alliance for Access to Safe Medicines’ [65] analysis found that 42% provided a working telephone number. Bessel et al [60] pointed out that only 35% of their sample published the owners’ or the director’s names, and European Alliance [65] found that 94% of the sample did not have a named verifiable pharmacist to answer questions.

Declaration of geographical location is an important feature with regard to transparency. Apart from the studies that focused only on websites based in a specific area, the sampled websites declared a geographical location ranging from 11% to 100% [47,61]; according to the results of Armstrong et al [57], Bessel et al [60], Bloom and Iannacone [62], Gallagher and Chapman [71], and Tsai et al [93], more than half of the samples mentioned where the company was based; the others found that less than half did [10,61,65,84,90]. Orizio et al [84] found that online pharmacies asking for a prescription were significantly more likely to declare their geographical location than were online pharmacies that did not, which were more often just virtual interfaces. Online pharmacies were most frequently based in the United States in virtually all the studies that investigated samples that were not restricted geographically [10,57,60,61,62,66,73,84]. Other studies focused on online pharmacies based in specific countries, such as the United States [87,88,95], the United States and Europe [27], the United States and Canada [89], Canada [94], or Brazil [92]. Some researchers looked for where the websites were registered, and they found the United States to be the most frequent location of domain registration [10,67,68,84,93]. Orizio et al [84], who compared the declared physical location with the registration domain, found that in only 55% of online pharmacies declaring their physical location did it correspond to the area of domain registration. Armstrong et al [57] found that US-based websites were more likely to ask for medical information and provide information about risk of treatment or its efficacy, and the GAO [73] found a better quality in websites based in the United States.

The National Center on Addiction and Substance Abuse (CASA) [63] found that 24% of shipments were declared to be from the United States, and 36% of the sample gave no indications. In Orizio et al [84] 42% of the websites declared where they deliver from, the most frequent place being Asia (47%); agreement between declared physical location and delivery location was found in 31% of online pharmacies providing both details. The attempt by Bate and Hess [59] to purchase drugs showed several cases of shipments from a different location from what was indicated on the website.

Both Arruanda [58] and Bessel et al [60] found that two-thirds of the online pharmacy samples sold internationally, whereas all the websites sampled by Schifano et al [91] did. Raine et al [90] selected only online pharmacies delivering to the United Kingdom. Makinen et al [27] related the fact of shipping abroad to the prescription requirements: online pharmacies that asked for a prescription were more likely than those that did not to sell only in the country they were based in, and delivered anywhere in the world. Mainous et al [82] found that almost all the websites sold to the United States, followed by the United Kingdom (84%) and Canada (80%). None of the small sample of 4 US-based online pharmacies analyzed by Wagner et al [95] delivered internationally.

What do online pharmacies have to offer? From an analysis of the literature it appears that online pharmacies have become more and more complex as time passes; whereas 10 years ago they tended to sell principally lifestyle drugs such as sildenafil, it looks today as if they offer virtually anything. Back in 1999, Bloom and Iannacone [61] studied a sample of online pharmacies in which the majority sold 1 or 2 drugs only, one for erectile dysfunction and the other for alopecia; in the same year Armstrong et al [57] found an entire sample of 77 websites selling sildenafil. In 2003 Arruanda [58] found that their sample was equally divided into 3 groups: selling 1 drug only, between 2 and 29 drugs (average 7), and selling 30 or more. The various studies reported several specific drug offers: sildenafil [59,60,80,84,86,87], benzodiazepines [84,86,87], painkillers [80,84,88], antibiotics [87,88], insulin [87,88], female hormones [87,88], antidepressants [80,84,86], alopecia medications [14,27,80], and obesity medications [59,80]. Bloom and Iannacone [62] reported that 160 separate medications were offered in their sample, which gives an idea of how widely differentiated the drug offer is. Quon et al [89] compared the offers of 12 Canadian-based Internet pharmacies with 3 US-based drug chain pharmacies for 44 different drugs. The GAO [73] found that some drugs were more widely available and easier to purchase (Celebrex, Lipitor, Viagra, and Zoloft) than others, which were available from fewer sources or were more difficult to obtain (Accutane and Clozaril). As we will see in more detail in the next section on actual purchase, what is hard to obtain—despite appearing to be available—is US class II and III opioid analgesics, as found in Peterson and colleagues’ studies in 2001 [87] and 2003 [88] and confirmed by the GAO 2004 [73]. In the United States, “Under the Controlled Substances Act all substances that are regulated under existing federal law are placed in one of five schedules on the basis of the substances’ medicinal value, harmfulness, and potential for abuse or addiction. Schedule I is reserved for the most dangerous drugs that have no recognized medical use, while Schedule V is the classification used for the least dangerous drugs.” [72]. The NABP [10] survey showed that 14% of websites dispensed controlled substances as defined by the Act cited above, and 40% foreign or non-FDA-approved drugs.

Some studies focused on specific classes of drugs. They looked for Parkinson disease medications [95], 3 types of controlled substances (opioids, and central nervous system depressants and stimulants) [63], opiates [64,67-70]), analgesics [90], dextropropoxyphene (a painkiller) [91], antibiotics [82], and specifically ciprofloxacin [93], contraceptives [83], the erectile dysfunction medication sildenafil [57,66,71], and psoriasis medications [81].

Littlejohn et al [80] and Makinen et al [27] linked the drugs offered to the type of online pharmacy: “legitimate pharmacies” did not supply opioids or ritalin [80], selling only over-the- counter products and herbal, hygiene, and cosmetic products [27]; “lifestyle pharmacies” supplied erectile dysfunction and alopecia medications; and “no-prescription pharmacies” supplied virtually everything, including opioids [80] and unapproved pharmaceuticals [27]. Both generic and brand drugs were available on the market [14,58,79,89,95].

In 1999 Armstrong et al [57] found that 55% of their sample included drug information. Several researchers who attempted to evaluate the presence of information on side effects found that a fairly consistent portion of online pharmacies, ranging from a quarter to a third, declared none of them [84,86,90,92,93]. Arruanda [58] found that 35% of his sample provided a service allowing buyers to consult experts about the use of medicine. In the GAO [72] sample, 61% of websites gave drug information. Wagner et al [95] found in all the online pharmacies in their small sample (4 US-based online pharmacies) more comprehensive information than received from the community store. Peterson [87] did not find a statistically significant difference between the types of pharmacies and the provision of drug information, although it should be noted that his sample was small for performing a group comparison (33 online pharmacies). Gernburd and Jadad’s [74] research based on spam offers found that all of the websites in their sample made benefit claims and warned about potential side effects. Interestingly, an attempt to count the declared side effects of 4 drugs tracked in online pharmacies that asked for a prescription and those that did not revealed that no-prescription online pharmacies declared more side effects for amytriptiline, fluoxetine, and tramadol, but fewer for sildenafil, which—probably not by accident—is one of the most widely offered “lifestyle drugs” in online pharmacies [84,86].

Holmes et al [76] evaluated the quality of online pharmacy “ask the pharmacy” services regarding adverse effects, drug interaction, risk factors, drug information, and directions for use. They found that answers were received to only 51% of all questions submitted to the websites, and the percentage of correct answers provided for each of 22 response components ranged from 7% to 96%, with few differences in quality between VIPPS-approved and -unapproved websites.

Arruanda [58] evaluated the presence of drug price lists (found in 96% of the sample) and of “price comparisons with competitive pharmacies” (1%).

A comparison of online pharmacy and retail pharmacy prices had different results. Drugs offered online were more expensive in Bloom and Iannacone’s studies [61,62], in Tsai and colleagues’ [93] analysis, and in Cicero and colleagues’ [64] purchase. Wagner et al [95], who compared prices of US retail versus US online pharmacies, found that the latter were cheaper for both generic and brand-name medications. Levaggi et al [79] found that drugs bought without a prescription cost more than with one. In Bate and Hess’s [59] purchase, larger orders generally tended to have lower per-tablet/per-capsule prices, and were more prevalent among online pharmacies that had not been approved by the NABP. Interestingly, they found that Viagra offered on noncredentialed websites was on average far more expensive than from credentialed ones, and for all the drugs purchased (Lipitor, Celebrex, Nexium, and Zoloft), except Viagra, prices were higher at physical-location pharmacies. Quon et al [89] compared US retail prices with those of Canadian online pharmacies, and they found savings when purchasing from Canadian online pharmacies for the majority of brand drugs (3 exceptions in the 44-drug sample were medications for erectile dysfunction), whereas generics were more expensive. The same result is confirmed by an aggregate macroeconomic analysis based on IMS Incorporated data on prices, patents, and cross-border Internet pharmacies between online Canada and retail US pharmacies [15]. Memmel et al [83] found lower prices for contraceptives purchased online. Mahé et al [81] found that, with the exception of tazarotene, the average price of all the online psoriasis medicines analyzed was higher than the French retail price.

Additional costs have to be considered: the need to join clubs or member groups and pay a nonrefundable fee, with the risk of not finding the wanted drug after joining [64,80]; and delivery costs, with free delivery in some online pharmacies or for some purchases, but with charges depending on the type of shipping in other cases (standard, express, overnight) [57,58,61,62,66, 74,79,89,95]. Online prescription could be another added cost [57]. It is interesting to note that, although a prescription is not necessary in marketing messages, Levaggi et al [79] found in their price analysis that the prescription has a value on the market; indeed, when you buy a drug without a prescription, the drug costs more.

Only a few articles focused on the marketing strategies of online pharmacies. Levaggi et al [79] and Orizio et al [84] disclosed the persuasive statements more frequently used by websites to promote their products, and identified arguments regarding privacy, service and drug quality, price offers, reassurances that buying drugs online is legal, and the suggestion that you can obtain a drug while avoiding a visit to the doctor. In particular, privacy issues were about the use of personal data and discreet packaging; service quality statements regarded short delivery times, online tracking of the state of the orders, and—indirectly—displaying testimonials by people who had already bought online; and price offers referred to encouragement to buy bulk purchases (found by Armstrong et al [57], European Alliance for Access to Safe Medicines [65], and Mahé et al [81] as well), lower prices than in “brick and mortar” pharmacies, fidelity bonuses, free delivery, and special discounts [79]. Armstrong et al [57] reported that one-third of their sample pointed out the advantages of ordering online, including confidentiality, ease of ordering, and lower cost. Forman and Block [70] analyzed the implied legitimacy and credibility claims, which they divided into 3 types: medical legitimacy, found in 82% of online pharmacies (eg, pictures of lab coats, rx/health symbols, and pharmaceutical logos); legal legitimacy, found in 72% (eg, government logos, explicit and implicit claims of being legal, and FDA approval statements); and retailer legitimacy, found in 24% (eg, customer testimonials). They also reported claims about shipping regarding security (52%), secretiveness (70%), reshipment if seized (2%), risk of seizure by US customs (6%), free delivery (24%), and the delivery company (52%). All the websites in Gernburd and Jadad’s [74] sample put forward benefit claims. One website in Eysenbach’s [66] sample even offered to ship cimetidine together with a Viagra order because it causes “a 56% increase in plasma sildenafil concentration when co-administered”. Gurau [75] pointed out that providing contact information may be a way to reduce the perceived risk of online transactions; moreover, statements about price, convenience, choice, and discreetness of service are more frequently used by online pharmacies not asking for a prescription than by those asking for one. Advertising prescription-only medicines was also found by Bessel et al [60] in 20% of online pharmacies. Orizio et al [86] pointed out that the marketing strategies adopted by online pharmacies enhance consumers’ peripheral reflection: by analogically playing with the sale of other commodities, they magnify aspects of online trading that consumers might find convenient, but overshadow the nature and risks of the actual products they sell.

The presence of at least 1 quality certification was found to range from 12% to 13% [60,70,86] to 28% [58], but Tsai et al [93] did not find any websites displaying certificates. Researchers found certificates about the quality of the health contents and/or about security in the money transaction.

Regarding health content quality, Arruanda [58] found that the most frequent quality certificate was VIPPS [96] (9%), followed by the HON code [97] (5%). Bessel and colleagues’ [60] sample had quality certification on 12% of the websites: the national pharmacy authority, the HON, and the JCAHCO [98]. Makinen et al [27] reported finding the HON code and VIPPS seal. Peterson’s sample in 2001 [87] had 12% of websites with the VIPPS seal, rising to 18% in 2003 [88]. The European Alliance for Access to Safe Medicines [65] study found that 4% of the sample was licensed by a board of pharmacy or had an appropriate pharmacy listing, while 20% had a “stamp approval” from a recognized society or association, but they found when clicking on them that 86% gave a link to a bogus “approval” webpage. Soares Gondim and Borges Falcao [92] found that 15 of the 16 websites analyzed lacked the Brazilian National Health Surveillance Agency seal.

As regards the security of the money transaction, Arruanda [58] found the VeriSign seal [99] (8%), Forman and Block [70] found Verified by Visa and Master Card securicode logos, and Makinen et al [27] reported finding the TRUSTe seal [100]. As well, 24% of Peterson’s sample in 2001 [87] incorporated secure socket layer technology, rising to 33% in 2003 [88].

A special study was performed by Kuzma [78] about the Web vulnerability of a random sample of 60 online pharmacies. She chose as her testing tool the N-Stalker Web Application Security Scanner 2009 Free Edition 7.0, which showed that a majority of worldwide online pharmacies do not provide adequate protection for their consumers, especially in cross-site scripting. The NABP [10] found that 17% of the websites in their sample did not have a secure site.

Kunz and Osborne [77] assessed the readability of 16 online pharmacies using Storytoolz; they found that the majority of the information provided on direct-to-consumer pharmaceutical websites is written at a level far higher than that which the average consumer can understand.

Arruanda [58] found that the sample online pharmacies had been online for a length of time ranging from just over 1 year to 7 years. Armstrong et al [57] reported that in the 10 days between website identification and data collection 9 (12%) ceased operating. In the CASA study [63] only 2% of the 152 non-VIPPS anchor sites identified in 2004 were still operating in 2008. When Orizio et al [86] rechecked in 2008 the 118 websites found in 2007, 75% were working. The follow-up of Peterson’s [87] study found that 1 year later 88% of the initial 33 online pharmacies were still working [88]. Gernburd and Jadad [74], who worked on links addressed by advertising spam, found that only 58% of the active link in health-related messages received during the first week of the study remained active at the end of the second week, whereas 26% were active at the end of the month.

In several of the papers analyzed we found that some websites required the consumer to relieve the companies from all liability, from 100% of websites [62,66], to 68% [57], 50% [65], and 33% [72]. Bessel et al [60] found privacy statements, information disclaimers, and return policies on 40%, 31%, and 37% of the websites, respectively, while 37% of websites displayed none of these policies. The GAO [72] found privacy statements on 23% of the sample. Gurau [75] found that websites displayed their privacy policy more frequently when they asked for a prescription (98%) than if they asked for an online questionnaire to be filled out (86%) or asked for nothing (56%). Peterson [87] found that a privacy policy was present in a significantly different way in the various types of pharmacies that he identified (chain extension 100%, mail order 16%, online 64%, independent extensions 25%); the same author reported in 2003 [88] that several websites had added a privacy policy.

Only 1 study investigated the date of the online pharmacies’ last update, which was displayed by 4% of them (data not shown in Table 1) [60].

The actual purchase of drugs without having an original medical prescription was verified by all authors except for Dean et al [102], Veronin and Youan [103], Veronin et al [94], and Westenberger et al [104], who did not specify whether a prescription was used. Two studies evaluated the effectiveness of online questionnaires for assessing health status. The results highlighted the low performance of this tool, which allowed the purchase of contraindicated products by fictitious consumers in some or all orders [66,83].

Several authors found a lack of reliability in the business practices of some online pharmacies, linked to completed money transactions without actual receipt of the drugs [66,73,74,101] or having to join clubs, only to find no drugs available after having paid the necessary fee [64]. One study reported subsequent advertising by email and phone for more than 4 months after ordering [64].

Regarding drug quality characteristics, researchers evaluated different features: the packaging of the drugs purchased, the instructions included, and the chemical composition.

As to packaging characteristics, packaging showed problems in more than half of the drug samples in the GAO [73] study. Memmel et al [83] found that all packaging appeared legitimate and products were within the expiration date, but patient education material (instructions) did not always match the product. GAO [73] researchers found better quality instructions in samples from US and Canadian Internet pharmacies, which in all 47 samples included dispensing pharmacy labels that generally provided patient instruction for use and in 87% of cases included warning information, compared with other foreign pharmacies, whose labeling and facility of manufacture were not FDA approved in the majority (19/21) of cases. Westenberger et al [104] found that only 1 of 20 samples had final packaging, including package insert, similar to that of the US products. The final packaging of the remainder consisted of bubble wrap inside a paper envelope, a Styrofoam sheet inside a paper envelope, loose blister packs, capsules or tablets in clear plastic bags without labels, and capsules or tablets in opaque plastic containers or boxes with labels. Bate and Hess [59] reported that many drugs, including some that did not fail in spectrometry testing, had an unclear or problematic origin or problematic packaging. In the European Alliance for Access to Safe Medicines study [65], 47% of the drugs purchased had no packaging, and in 6% the packaging had been tampered with; half of the sample supplied a patient leaflet, which was in English in 80% of cases. Mainous and colleagues’ [82] purchase arrived with no instructions.

The GAO [73] found that the chemical composition of 4 samples out of 68 was not comparable with the product ordered: in 2 cases a counterfeit version contained less active ingredients, and in 2, a significantly different chemical composition from the product ordered. An analysis of counterfeit dapoxetine (a short-acting selective serotonin-reuptake inhibitor), used against premature ejaculation, found that the tablets contained undisclosed sildenafil [102]. Bate and Hess’s [59] analysis showed that 2.5% (3/121) of the tablet sample failed Raman spectrometry. Westenberger et al [104] found that 2 out of 20 samples failed in terms of dissolution and chromatographic purity; Veronin and Youan in 2004 [103] found differences in simvastatin tablet formulation, obtained from 5 countries via the Internet, using near-infrared spectroscopy chemical imaging methods to assess blend uniformity. In contrast, Veronin et al in 2007 [94] found a quality standard comparable with that of the American manufacturer and the Canadian generic drug product tested. The European Alliance for Access to Safe Medicines study [65] study found that 62% of the products received were counterfeit, substandard, or unapproved medications (68% of these were generic and 32% were branded).

Table 3 shows the 15 articles [12,64,75,105-116] dealing with consumer characteristics and based on surveys. Most studies were US based, except for 4, which were conducted in Europe [75,105,106] and South America [107]. The findings were published between 2003 [108] and 2010 [105-107,109-112]. Seven studies investigated the general population [12,75,106,108,109,113,114], while the remainder were about specific groups, described in Table 3 [64,105,107,109,111,112,115,116].

In studies about the general population, the percentage of people buying drugs online was between 4% and 6% in the United States in the studies by Fox [12], Baker et al [108], and Cohen and Stussman [109] (who analyzed the health information technology questions of the National Health Interview Survey). Atkinson et al [113] (who analyzed the data of the US national representative sample of the Health Information National Trends Survey [HINTS] by the National Cancer Institute) found a higher percentage (13%), probably because the purchase of vitamins was included in the estimate. The UK-based survey considered both having bought and the intention to buy, so a third of the sample gave a positive response [75].

As in the general population studies, Mazer et al [112] found that 5.4% patients at an emergency department had bought drugs online.

Both surveys on prescription drug abusers [64,115] reported that 6% of the interviewees had used the Internet to purchase prescription medications for their addiction. The 2 papers by Inciardi et al [111,116] were based on several sources of information investigating different populations, including Internet-savvy high school and college students, chronic drug users, and members of the general population. Inciardi et al [116] tried to reveal the “black box” of drug diversion (the transfer of a prescription drug from a lawful to an unlawful channel of distribution or use) using several sources of information: the Researched Abuse Diversion and Addiction-Related Surveillance (RADARS) System, the National Survey of Drug Use and Health (NSDUH), the Delaware School Survey, several Miami street studies, and 2 qualitative studies. The authors concluded that, although the Internet is indeed a source for prescription drugs, the overwhelming volume of purchases is probably at the wholesale level, since few end users report accessing the Internet for drugs. In the various population groups analyzed in this study, the percentage of people accessing prescription drugs via the Internet ranges from 1% to 6%. The results of the 2010 work by Inciardi et al [111] seem to confirm the findings of the previous paper: based on the RADARS System, the NSDUH, and the Monitoring The Future (MTF) survey, they still found that the Internet is a source of prescription opioid acquisition for 0.5%–3% of the investigated populations.

A US survey on the recreational use of erectile dysfunction medications in undergraduate men showed that 12% of users had bought them on the Internet [110]. Bechara et al [107] in Argentina investigated the recreational use of PDE5Is by healthy young men and found that 2.9% of the interviewees had bought the drug through the Internet. A cross-sectional Europe-based observational study examining the purchasing patterns for the same substances found that 11% of the 11,889 subjects reported current use of PDE5Is; 32% of these reported obtaining their PDE5Is from sources outside the health care system without prior health care personnel interaction (eg, Internet, friends) [105].

Some reports suggest widespread use of online pharmacies in people over 35 years of age [75,113] and in women [109]. Mazer et al [112] found no difference in age between those who bought drugs online and those who did not, and no difference in student status, but patients on multiple medications and those with prescription plans used online pharmacies more frequently. Fox [12] found that the most frequently bought products were drugs for chronic conditions (75%), followed by weight loss and sexual performance substances (25%). The most frequent reasons quoted by interviewees for buying or intending to buy online were convenience and saving money [12,75], followed by information anonymity and choice [75]. Regarding location of online pharmacies, the majority of buyers chose or would have chosen sites based in their own countries or in economically developed countries [12,75]. Regarding risk perception, Gurau’s [75] interviewees reported being worried by a lack of a license on the part of the pharmacy (31%), privacy issues (27%), security of online payment (26%), additional charges, drug quality, and superficial prescription. It is interesting to note that health-related risks (drug quality and prescription requirement) rank last in consumers’ perception. In Fox’s [12] survey 68% agreed that online purchasing makes it too easy to obtain drugs illegally.

Rajamma and Pelton [114] explored the potential impact of consumers’ cognitive characteristics on their decision making as it relates to procuring pharmaceutical products via online retail channels. They sent 350 consumers a self-administered electronic questionnaire. Their analysis showed that male, higher-educated, and higher-income consumers had a greater propensity to procure medication online, whereas insurance status did not have any influence. Wiedmann et al [106] investigated the consumer-perceived values and risks related to online shopping attitude and behavior in an e-pharmacy context; they used their model to identify 4 clusters labeled as “enthusiastic experts” (29%), “risk-averse traditionalists” (24%), “convenience-oriented rationalists” (20%), and “inexperienced opponents” (28%).

We found 12 published papers on clinical cases related to prescription drugs obtained via the Internet, which are shown in Multimedia Appendix 3. One regarded the Internet purchase and injection of gamma-butyrolactone by an 18-year-old woman that led to admission to a pediatric intensive care unit [117]. Romero et al [118] reported florid withdrawal delirium following the discontinuation of Fioricet (a combination of butalbital, which is a barbiturate, with caffeine and acetaminophen indicated for muscle contraction headaches) that a 37-year-old woman purchased online and self-administered in escalating doses for the treatment of chronic headaches in a context of a history of depression. A 40-year-old woman was treated at the emergency department for severe tetany-like spasms, probably due to the ingestion of haloperidol and bentazepam purchased on the Internet, which were displayed on the online pharmacy under “sleep aids” [119]. Neuberg et al [120] reported a case of life-threatening thyroid hormone abuse in a 56-year-old woman encouraged and enabled by unconventional health advice and nonprescribed medication obtained via the Internet. Carisoprodol (a muscle relaxant) withdrawal after Internet purchase was reported by Eleid et al [121]. Another report described a 43-year-old woman who underwent surgery for brain cancer; after hospital discharge she researched adjunctive treatments of cancer on the Internet and self-initiated a 10-day course of cesium chloride, ending up with an acquired long QT syndrome [122]. A 55-year-old man with squamous cell carcinoma of the maxillary sinus, who declined to undergo surgery, radiation, and chemotherapy, decided to treat his cancer using hydralazine sulfate, obtained online; he died from fatal hepatorenal failure, probably caused by the hydralazine sulfate [123]. A series of case studies was reported by Lineberry and Bostwick [124] regarding a suicide attempt by a 35-year-old man using amitriptyline purchased online unbeknownst to his psychiatrist, who had prescribed him paroxetine, and 3 stories of opiate and painkiller addiction in a 37-year-old man, a 42-year-old man, and a 29-year-old woman facilitated by online purchase. Levesque [125] reported tardive dyskinesia in a 67-year-old man associated with the online purchase of the older antipsychotic drugs he probably received when he requested a tranquilizer. All the cases reported so far occurred in the US, although other evidence shows that the phenomenon exists in Europe as well. A case of prolonged hypogonadotropic hypogonadism caused by anabolic steroids purchased on the Internet by a 34-year-old Italian man was reported by Pirola et al [126]. In the United Kingdom, orlistat-induced subacute liver failure was reported in a 57-year-old woman [127]. Also in the United Kingdom, acute coronary syndrome was diagnosed in a 41-year-old man who had taken Viagra for erectile dysfunction [128].