INTRODUCTION

Physicians play an important role in tobacco control by offering a point--of-care cessation counseling option, although studies have found that they often do not assess their patients’ smoking status, provide only limited advice, and do not offer cessation treatment as regularly as they should.1 – 4 However, physicians not only have the potential to help their patients quit smoking; they are also considered “role models.”5 Although training physicians on tobacco cessation improves their adherence to smoking treatment guidelines, their counseling performance may ultimately have little impact on the smoking status of their patients.6 – 10 Moreover, while positive attitudes towards cessation advice appear to be correlated with greater odds of providing such advice,11 physicians’ perceptions of the effectiveness of and recommendation of cessation counseling may be negatively affected by their personal smoking status.12 , 13 Therefore, the study of health professionals’ attitudes towards smoking and tobacco control offers an opportunity for analyses that may result in new strategies focused mainly on providing training.14

Considering that medical students are “soon to be” physicians, their smoking behavior and attitudes have been the object of multiple studies around the world.15 – 21 Smoking prevalence among physicians has declined in most high-income countries, but it remains highly variable in middle-income countries,22 – 24 including among medical students.25 Smoking prevalence among US students in the health profession is close to 7 %.26 In Europe, it ranges from 3.8 % in Wales27 to 22.1 % in Germany28; in Asia, from 2.1 % in Thailand29 to 66.9 % among Chinese men.30 Among Latin American countries, it was estimated that 16.9 % of Brazilian medical students were current smokers, while this proportion was reported to be 41.1 % in Bolivia.29

Argentina is a high-middle-income South American country with a smoking rate of 22.1 %.31 Previous research has shown that rates of smoking among medical students in Argentina are at least as high as those in the general population.19 , 32 , 33 The School of Medicine at the University of Buenos Aires (UBA) is the largest in the country; students enroll after graduating from high school, with completion in about 7 years.34 A previous survey conducted among its medical students found that more than 90 % thought that health professionals should receive specific training on smoking cessation techniques, but less than 10 % reported having any training in medical school.33 Although this survey partially addressed UBA medical students’ attitudes and knowledge about tobacco, these topics are in need of further exploration. Studies have shown that media images of smoking, particularly in movies, are linked to smoking and smoking initiation among adolescents.35 – 40 Therefore, media literacy—or the ability to analyze and evaluate media messages41—has been proposed as a tool to better interpret media messages related to smoking, potentially influencing attitudes towards smoking and smoking behavior. Prior studies have shown an association between the level of smoking media literacy and smoking status among adolescents in the USA and in Jujuy, Argentina, as well as in US college students.42 – 44 In a previous article, we analyzed access to tobacco company websites among medical students from UBA.45 The objective of the present paper is to describe the attitudes and knowledge of these medical students and recent graduates towards smoking and smoking behavior, including their level of smoking media literacy, based on their smoking status.

METHODS

Study Procedures

In 2011, current and former UBA medical students were invited through email communication to participate in a one-time self-administered online survey. Email addresses were obtained from the registry of the required 1-year rotating internship (the last practice-based year of medical school before graduation), the pharmacology department of the UBA School of Medicine, and the main university teaching hospital. Datstat Illume software was used to conduct the survey.46 Over the course of 11 weeks, one initial email and 14 reminder messages were sent requesting participation. The links allowing access to the questionnaire were disabled after the last email resulted in no new responses. Inclusion criteria were age of 18–30 years, responding to smoking status questions, and status as current or recent medical student at UBA School of Medicine. The institutional review committee of the Hospital de Clínicas, UBA, approved the study protocol. A more detailed description of data collection methods as well as the instrument development was described previously.45

Smoking Behavior Variables

The questionnaire included questions regarding participants’ demographic characteristics, cigarette consumption, tobacco dependence, knowledge about tobacco and tobacco cessation, attitudes towards smoking, and smoking media literacy. A participant was considered a current smoker if he/she reported to have smoked a cigarette in the 30 days prior to the survey. A former smoker was someone who reported to have smoked cigarettes in the past but not within the previous 30 days. A never smoker had never tried a cigarette. Among current smokers, nicotine dependence was assessed using the Heaviness of Smoking Index, through two items of the Fagerström Test for Nicotine Dependence. Item 1 was time from waking to first cigarette, and item 4 was the number of cigarettes smoked per day (each score from 0 to 3). A participant was considered to have high nicotine dependence when the score was ≥4.47 – 49

Attitudes and Knowledge About Tobacco Use

Tobacco-related attitudes and knowledge were evaluated using a five-item Likert scale (from “strongly agree” to “strongly disagree”) for 16 statements regarding the role of doctors in tobacco control, attitudes towards patients who smoked, effects of smoking, and effectiveness of counseling and smoking cessation treatment.9 Statements were adapted from a questionnaire used in a clinical study with Argentine physicians.9 Content validity was evaluated by cognitive interviews with ten UBA medical students. Factor analysis did not support acceptable psychometric properties for use as an attitudes and knowledge scale. Thus, the analysis was planned with individual items as outcomes.

The “agree” and “strongly agree” options and the “disagree” and “strongly disagree” options were combined into a single category (“agree” or “disagree”). The neutral option (“neither agree nor disagree”) was differentially assigned to the response group that did not indicate an anti-smoking attitude or that did not correctly answer a knowledge question. For example, when the anti-smoking attitude or the correct knowledge answer implied “agreeing” with the statement, the neutral option was included among the “disagree” responses; when it implied “disagreeing” with the statement, the neutral option was included among the “agree” responses. Participants were also asked about their level of agreement with banning indoor smoking in various public places.9 , 33 The questionnaire is in Online Appendix 1 in both Spanish and English.

Smoking Media Literacy

Smoking media literacy was assessed with a smoking media literacy (SML) scale,50 developed and used in the USA, and adapted for Argentine youth.42 – 44 The original scale has 18 items divided into three domains: authors and audiences (AA), messages and meanings (MM), and representation and reality (RR). As a result of cognitive interviews conducted for transcultural adaptation, five items were deemed not relevant; therefore, the final version of the scale consisted of 13 of the 18 items. Items used included the following: “Tobacco companies are very powerful, even outside of the cigarette business.” (AA); “When people make movies and TV shows, every camera shot is very carefully planned.” (MM); and “Advertisements usually leave out a lot of important information.” (RR). Each statement was evaluated using a four-point Likert scale, from “strongly disagree” (1 point) to “strongly agree” (4 points). Based on previous reports, high smoking media literacy was defined as an average score ≥3..43 , 44 After the survey was conducted, confirmatory factor analysis was used to verify adjustment to the original model. The complete list of all items used is available upon request.

Data Analysis

Data analysis was conducted using STATA software version 13 (StataCorp LP, College Station, TX, USA). Statistical significance was defined with a two-sided alpha of 0.05 (P values < 0.05). Descriptive analysis was used to compare respondents by sex. Chi-square tests were used for categorical variables. Logistic regression models were used to explore the association between the various attitude statements and opinions on banning indoor smoking and status as current smoker or recent graduate. Models using current smoking as outcome were designed to explore its association with the level of smoking media literacy (high or low). All models for both responses to attitudes, knowledge and full smoking ban, and SML were adjusted for age, gender, smoking status and student status.

RESULTS

Of the 4969 persons included in the database, 1743 (35.1 %) completed the survey, but 84 were excluded from analysis for not meeting the inclusion criteria. The final sample consisted of 1659 participants (33.4 % of the 4969 original); of these, 1212 (73.1 %) were women, 447 (26.9 %) were men and 55.7 % were current medical students; 12 participants (0.7 %) reported to have dropped out of medical school before graduation. The majority of the recent medical graduates (605/723, 83.7 %) had completed medical school in the 2 years prior to the survey (median: 2010; interquartile range: 1), and 99.2 % of graduates had completed medical school in the 5 years prior to the survey. Table 1 presents the demographic and tobacco consumption characteristics. Among respondents, 453 (27.3 %) were current smokers at the time of the survey; 311 (18.7 %) were daily smokers. Almost all current smokers reported low nicotine dependence (434; 95.8 %). A more detailed description of the population and of current smokers’ characteristics can be found in a previous publication.45

Table 1 Demographic Characteristics and Smoking Behavior Among 1659 Current Students and Recent Graduates, by Gender, at the University of Buenos Aires Medical School, Argentina, 2011
Full size table

Attitudes and Knowledge

Table 2 presents the level of agreement with different statements regarding attitudes and basic tobacco control knowledge by smoking status. In addition, 84.6 % of participants answered that medical students need more training on how to counsel patients on smoking cessation. Current smokers were significantly less likely than non-smokers to indicate a desire for more training (79.9 % vs. 86.4 %, respectively; P < 0.05).

Table 2 Knowledge and Attitudes Towards Smoking Among 1659 Current Students and Recent Graduates, by Smoking Status, University of Buenos Aires Medical School, Argentina, 2011
Full size table

Recent medical graduates had more accurate knowledge about tobacco control strategies than did current medical students. These results are summarized in Online Appendix 2 (Table 2a).

Participants were also asked about their agreement with banning indoor smoking in various public places; Table 3 presents the results by smoking status. Overall, participants were less likely to agree with a full smoking ban in recreational venues, even though over 90 % agreed with a ban in health care settings, schools and public transportation. In addition, current smokers were significantly even less inclined to agree with a full smoking ban in seven different sites. Current medical students were also less likely to agree with full smoking bans in recreational venues. Analysis by student status can be found in Online Appendix 2 (Table 3a).

Table 3 Level of Agreement with Banning Smoking in Public Places Among 1659 Current Students and Recent Graduates, by Smoking Status, University of Buenos Aires Medical School, Argentina, 2011
Full size table

Multivariate Analysis

Adjusted logistic regression models were used to assess the association between current smoker or recent medical graduate status and agreement with the attitudinal or knowledge statements and with a full smoking ban in public places. Table 4 presents the items in which a significant difference between current smokers and non-smokers was observed. Analyses by student status showed that recent graduates were significantly less likely to have selected smoking and cessation misconceptions and more likely to support a full smoking ban in recreational venues.

Table 4 Odds of Agreeing with a Specific Attitude/Knowledge Statement or with a Full Smoking Ban Among 1659 Current Students and Recent Graduates, by Smoking Status and Student Status, University of Buenos Aires Medical School, Argentina, 2011
Full size table

Men were more likely than women to agree with the following statements: “Patients already have too many problems to add to quit smoking.” (OR: 1.85, 95 % CI 1.14–2.99, P = 0.013); “If a patient has smoked for a long time, it is too late to stop, because the patient won’t be able to do it.” (OR: 2.57, 95 % CI 1.26–5.22, P = 0.009); “It is best to use physician time on other things rather than advising patients to quit smoking.” (OR: 3.17, 95 % CI 1.86–5.41, P < 0.001); and “If a doctor wishes, he/she should be able to refuse to care for a patient just because he does not quit smoking.” (OR: 1.34, 95 % CI 1.03–1.76, P = 0.031). No differences between genders were found when asked their opinions on banning smoking in public places.

Smoking Media Literacy

Most participants reported high SML scores (1458, 87.9 %). A statistically significant difference between current smokers and non-smokers was found only for the following individual items:

  • “Tobacco companies are very powerful, even outside of the cigarette business.” (total agreement: 87.2 % vs. 98.8 % respectively, P = 0.02)

  • “There are often hidden messages in cigarette ads.” (total agreement: 80.9 % vs. 85.2 % respectively, P = 0.04)

  • “Most movies and TV shows that show people smoking make it look more attractive than it really is.” (total agreement: 83.7 % vs. 90.6 % respectively, P < 0.01)

Logistic regression models did not show a significant association between SML scores and smoking status (current vs. non-smokers): OR 0.82, 95 % CI 0.57–1.18, P = 0.29 in the unadjusted model; and OR 0.80, 95 % CI 0.55–1.16, P = 0.24 in the adjusted model.

DISCUSSION

This study analyzed the attitudes towards smoking, knowledge about smoking, and the level of smoking media literacy among students of the largest and most prestigious medical school in Argentina. In contrast to what has happened in most high-income countries, where a decline in smoking prevalence among physicians preceded a decline in the general population, the smoking rate among medical students and recent graduates was similar to that in the general population. This high prevalence is particularly concerning considering that the potential positive impact of health professionals on public policies may be diminished by their own high smoking prevalence. We found that only half of participants agreed that doctors should act as an example to their patients, around one third thought that medical advice had little effect on patients’ cessation behavior, and nearly 20 % believed that physicians could decline care for patients who were smokers and failed to quit.

Current smokers were generally less likely to reveal an anti-smoking attitude when asked about physicians’ role in tobacco control, especially with regard to being role models, while non-smokers and men were more likely to stigmatize patients who smoked and were unable to quit. Current medical students were less likely to correctly answer some of the knowledge questions and to recognize the importance of health professionals in helping patients to successfully quit smoking, which may be due to less clinical experience in observing the effects of tobacco use.

Our results were largely consistent with those found in other studies, in which medical students expressed a need for more training, and current smokers were less likely to have anti-smoking attitudes.15 , 19 – 21 , 33 Even though most participants answered the knowledge-related questions correctly, most expressed a need for more training on tobacco cessation. Although a majority of respondents reflected an anti-smoking attitude, these results indicate that a significant number of medical students and young physicians do not believe that they have a greater responsibility in tobacco control just by virtue of being in a health-related profession.

Although indoor smoking bans have become widely accepted worldwide, including in Argentina, roughly one third to one half of respondents did not think that smoking should be fully banned in recreational venues such as cafes, which is similar to findings in other studies.20 , 21 This contradicts the fact that over 96 % of respondents correctly disagreed that second-hand smoke is harmful only to young children, and implies that though most understand that second-hand smoking is harmful, exposure by consenting adults in recreational venues may be acceptable. Another possible explanation could be a belief that the rights of nonsmokers should not prevail in places where their presence is purely recreational and voluntary. This contrasted with more than 80–90 % agreement among participants with a full smoking ban at work, health care and educational settings. Even though the reasons for these observations deserve to be further explored, these findings do show a gap in knowledge among health professionals regarding the harm of second-hand smoke that needs to be addressed.

Our study did not find a significant difference in the level of smoking media literacy between current smokers and non-smokers. This result is probably best explained by the fact that almost 90 % of participants had high smoking media literacy, making it very difficult to find any difference between subgroups. The measure may also lose discriminatory power among older, more highly educated respondents such as medical students. Nevertheless, current smokers had a tendency to be less receptive to the notion that tobacco companies disguise certain messages, as shown in the individual items in which a difference was found.

There are limitations associated with this study. First, the lower response rate may affect the generalizability of the results, although this is consistent with other online surveys,51 – 54 and there was no obvious non-response bias, with gender distribution in our sample similar to that found in UBA.34 Second, the items about tobacco-related attitudes were not possible to use as a scale. Lastly, the cross-sectional design of the study does not allow us to make any causal inference. Nevertheless, our study suggests that, while overall most medical students and recent medical graduates have strong anti-tobacco attitudes, a substantial proportion do not recognize the importance of their role as health professionals in tobacco control strategies. It is disappointing that even though the prevalence of tobacco use has decreased in Argentina, the rate among medical students has not declined at a faster pace. Considering that health professionals should be at the front line of anti-tobacco strategies,55 – 57 our findings imply that the opportunity to help smoking patients quit is probably being missed. Since most students reported a need for more training on smoking cessation treatment, increasing the time that UBA’s medical school dedicates to tobacco control is desirable. An initial evaluation of the current curricular situation and a survey of the expectations of both students and faculty would guide the changes that are needed. Multimodal training strategies, including role-playing scenarios as well as seeing preceptors as positive role models, may also help in changing attitudes and teaching appropriate cessation treatments.26 , 58 Because physicians continue to be viewed with great respect by the Argentine public, their lack of awareness of this fact may partially explain the high smoking rate in this population, and may project them as negative role models. Medical schools not only need to implement changes in the medical curricula, but also need to develop cessation programs among their students.