The availability of accurate tobacco harm reduction (THR) information at locations where smokers purchase cigarettes is largely unknown but has great public health importance. THR, the substitution of lower risk sources of nicotine for smoking, is a promising intervention for smokers who will not quit nicotine or tobacco entirely [1–4] http://tobaccoharmreduction.org. Almost all the risk from smoking comes from inhaling chemicals produced during the combustion of organic matter, not from nicotine or the tobacco plant itself. It is because of this that non-combustion sources of nicotine, such as smokeless tobacco (ST) and pharmaceutical nicotine products cause roughly 1/100^th the risk of life-threatening disease from cigarettes [5]. Electronic cigarettes probably have approximately the same mortality risks (because users do not inhale combustion products) but have not been studied as extensively. The ability of smokers to make an informed, autonomous choice about whether to keep smoking, switch to less harmful nicotine products, or stop using nicotine entirely, should be based on accurate information about the products, including information about the relative health risks of the different products. Documented misperceptions about THR include the beliefs that: ST poses the same or greater health risks as smoking; ST has been shown to cause a measurable risk of oral cancer (typically colloquially phrased as "ST causes oral cancer"); and the smoke itself is not the source of most of the health risks from smoking [6–12]. Accurate knowledge about ST products is especially important for retailers who interact with customers purchasing tobacco products, and may prevent or contribute to the propagation of disinformation. This is particularly true in Canada because of the near prohibition on the manufacturers' ability to communicate health information to their customers other than in-person at the point of sale, and restrictions on the right to free speech that criminalize even private provision of accurate information about tobacco products.

The introduction of a new Swedish-style ST product, du Maurier snus (dMS), by Imperial Tobacco Canada Limited (ITC) (a subsidiary of British American Tobacco) in 2007 provided a unique opportunity to assess retailers' knowledge of THR and the sale of ST to young adults. Snus is the Swedish term for pasteurized moist snuff that is usually sold in small sachets that users place between their upper lip and gum and du Maurier is the brand name of one of ITC's premium cigarettes. Other ST products (mainly US Smokeless Tobacco Company's moist snuff products) were widely available in Edmonton prior to the launch of this product [13, 14]. The marketing strategy for dMS differs from that for other ST products because ITC is marketing it explicitly to their and other companies' cigarette customers as a harm reduction product. Around the time of the rollout retailers were educated about the product category and provided with a brochure, entitled "What is SNUS" to distribute to adult customers, particularly those purchasing tobacco products. They also received oral briefings by sales representatives of ITC and some of them attended an educational/social event at the time of the product rollout. The dMS product displays were quite prominent at the time of the rollout and data collection [13] before a provincial legal change mandated that no tobacco products could be visible to consumers. The display consists of a small refrigerator, usually located behind or beside the cashier.

Our study examined retailers' knowledge of the comparative risks of different tobacco products and other health information about ST; information that they received in oral briefings and written materials about dMS. In addition, we took advantage of the study to also examine compliance with recommendations regarding the sale of tobacco to young adults. According to recommendations from Operation I.D., which provides materials about the sale of tobacco products to youth, retailers should ask individuals who appear to be under the age of 25 for identification before selling any tobacco product [15].

A list of the 219 retail outlets in the Edmonton area where dMS was sold at the time of the study was obtained from ITC. Fifty-two outlets outside of the city of Edmonton were excluded to simplify the logistics of data collection so that the study could be completed in a timely manner. A random sample of 60 of the remaining 167 stores in the city of Edmonton was selected. Two male undergraduate students (two of the authors (FT and ZR)), hereafter referred to as researchers, aged 20 and 21 were trained to approach the retailers, ask questions about THR as part of a conversation about dMS, and attempt to purchase dMS. The dMS refrigerator was often near the cash register, allowing for a visual reference to the product. The researchers were greater than the legal age to purchase tobacco in Alberta but sufficiently young-looking that they should have triggered the "check identification if under 25" recommendation. No female students were included because males are much more likely to use ST (e.g., [9, 16]), and thus appeared more natural. The researchers dressed in casual clothes (e.g., jeans and sweatshirts).

In each store, one researcher approached a cash register and asked the nearest employee a series of questions about the health risks of dMS and THR. The researcher then purchased one container of dMS, showing his Alberta driver's license if he was asked for identification. As it was crucial for the researchers to appear as normal customers rather than researchers, they did not follow the script exactly, but rather rehearsed following naturally flowing conversations and asking their questions at appropriate opportunities. The researchers completed a data collection form as soon as possible after leaving each retail outlet, as doing so inside the store might have affected the employee's interactions with the researcher. The script, data collection form and a de-identified version of the data are available at http://tobaccoharmreduction.org/research/retailer.htm

After data collection was completed, the responses to each question were categorized based on the correct answers. These categories are described in the discussion section to provide the necessary regulatory and ideological framework for the retailers' responses. SAS (version 9.1, SAS Institute, Cary, North Carolina) was used for the sample selection and data analysis.

Retailers' consent was not obtained for this study. Our goal was to observe the retailers' behavior during the course of their normal jobs, and asking for consent would have prevented this. Asking for consent would have necessitated limiting the study to an assessment of the retailers' responses to what they knew was, in effect, an exam, and would have prohibited any assessment of whether retailers' appropriately asked seemingly underage customers for identification. The retailers were later sent a letter and fact sheet describing the study. The study protocol was reviewed and approved by the Health Research Ethics Board at the University of Alberta.

Retailers' misperceptions were consistent with the confusing and often inaccurate information about ST in the mandated health warnings, the dMS brochure, popular press and on the internet. Additional information may have come from public statements to the community by our research group, particularly a presentation by one of us (CVP) at an educational/social event organized by ITC prior to the product launch. In particular the specific estimate of 99% risk reduction and the fact that smoking is much more likely to cause oral cancer than snus use are common in our communications but not ITC's, and thus most likely trace specifically to us.

Retailers may have a rapport with cigarette customers who they see frequently and may respond differently to those individuals. However, it is possible that the researchers' experiences may be similar to those of young smokers who are interested in reduced harm nicotine products. Many of the retailers were hesitant to speak with the researchers or did not answer their questions. Retailers working alone were more likely to engage in a conversation with the researchers than if there were other employees or customers in the vicinity. Obviously, if retailers suspected that the researchers were "secret shoppers" (underage youths attempting to purchase tobacco to see if retailers asked them for identification), their interactions with the researchers might have been different than with other young adults. We had no clear indications that this was the case, but it was possible.

The promotion of low-risk nicotine products as an alternative to smoking may depend largely on information provided by retailers. This is the case because the environment is characterized by manufacturers having limited opportunities to communicate to customers, there is limited communication of accurate information from the scientific community and inaccurate and misleading information is often issued by anti-tobacco groups and governmental and non-governmental organizations. Our study suggests that despite efforts to educate retailers, they lacked some combination of the time, knowledge, or analytic sophistication to provide several of the key bits of information needed to explain the value of THR. While some retailers provided useful and accurate information, many did not. Lack of accurate information about THR is not surprising given the misinformation in the popular press[17], and on the internet [18, 19]. It is somewhat disappointing, though not necessarily surprising, that retailers who either received directed education or could have been educated by other staff members on the point shared the popular misperceptions. The misleading or unclear warning statements on ST packages probably contributed to this, and the equivocal claims in the dMS brochure may have also contributed.

Regulatory changes occurred subsequent to the introduction of dMS (which we detail elsewhere [14]), including prohibiting the display of snus or informational brochures. It is unlikely that current customers would seek information like our researchers did, and if they did, the printed material would not be available. Thus, this study is probably more informative for markets where free speech at point-of-sale is still protected than it is about the current situation in Edmonton. The results from this study suggest that retailers in Edmonton may be contributing to public misperceptions about THR as much as they are reducing them. This suggests that other restrictions on free speech about THR -- advertising, package inserts, etc. - may be detrimental to the public health, since smokers' major remaining potential source of information is inadequate. The result is that even where actively providing accurate point-of-sale information is not criminalized and retailers are actively encouraged to provide the information, many smokers who might have quit by switching products will never learn about this potentially lifesaving option.