Recall of Tobacco Corrective-statements advertisements and Effects on Health Information-seeking Behavior

Background : In the 2006, landmark ruling, US District Judge Gladys Kessler instructed tobacco companies to disseminate corrective-statements (CSs) against their products through media advertisements. This study objectives were to (1) examine the proportion of adults who were exposed to each of the five CS messages ; and to (2) describe the association between exposure to CSs and health-information seeking behavior among the US adult population. Methods : Data, settings, participants, outcomes, and statistical approach. We analyzed the most recent nationally representative data from the population-based cross-sectional survey of US adults, the Health Information National Trends Survey (HINTS5-Cycle2,2018). Data collection began in January 2018 and concluded in May 2018, and analysis took place from May 2019 to October 2019. Statistical significance was defined as a P-value less than 0.05, and all tests were 2-tailed. All data were weighted to be nationally representative. Results : Key findings.


INTRODUCTION
Over the past few decades the world has evolved in dramatic ways with regard to Mass Media Communication (MMCs) even as it continues to be a powerful tool among public health practitioners to project health advocacy messages and advance public health. 1 Especially when it comes to anti-tobacco educational campaigns; MMCs are a key component of comprehensive tobacco control programs; 1,2 they are also an evidence-based intervention for promoting cigarette cessation. 3 Tobacco control MMCs are composed of paid and earned media disseminated through television, radio, out-of-home placements (eg, billboards, bus shelters), magazines, newspapers, and digital platforms. A recent review of MMCs showed that their effects on tobacco use are more numerous than for any other healthrelated issue. 4 That could be attributed to the fact that exposure to advertisements (ads) occurs during routine media use, rather than being explicitly sought out, there is high potential for widespread and repeated population exposure. 5 Previous studies of mass media anti-tobacco campaigns have revealed that when implemented with sufficient reach, intensity, and duration, they can effectively promote quitting and reduce smoking prevalence. 3,6 In addition, the National Cancer Institute Corrective Statements study highlighted how media exposure was associated with modifications in health attitudes, knowledge, and behavior; with accurate information being a crucial element for effective health promotion. 7 Moreover, contemporary healthcare systems can help improve health literacy outcomes by motivating people to educate themselves and seek health information which could be the first step they take in order to break bad habits and meet behavioral goals. Targeted approaches are particularly important given that the risks associated with tobacco consumption are not evenly distributed among the population, nor is the awareness of the risks of tobacco use. 8 In particular, individuals with low socioeconomic status are significantly more likely to believe myths about smoking and hold inaccurate beliefs about the risks of smoking. 7 Another concern is that public health messages are usually developed by subject matter experts by using models and theories available from behavior change research, using technical language. 9 Developing messages sometime cannot accurately describe that consider the intended recipients' characteristics that may affect respondent's can enhance message comprehension, recall, or predict barriers to and encourage behavior change.
In the United States, for over 50 years, public health has battled against smoking. 10 The release of the first U.S. Surgeon General's Report on smoking and health prompted Congress to act and within a year they passed the 'Federal Cigarette Labeling and Advertising Act of 1965'. 11,12 This first Surgeon General's report and the 31 that followed it persuasively documented the adverse health effects of smoking, and subsequently influenced public health policy-breaking the silence around this insidious killer and fundamentally changing the way Americans view tobacco use. 13 Since the 1965 'Federal Cigarette Labeling and Advertising Act', several regulatory policies and other public health interventions have been enacted over the succeeding decades leading to ultimate reduction of smoking rates. 14,15,16 Effective public health messaging against tobacco product use is however complicated by the fact that the tobacco industry far outspends public health in advertising; in 2016 alone, the leading cigarette companies spent $8.7 billion dollars on advertising and promoting cigarettes within the United states. 17  proportion of adults who were exposed to each of the five CS messages ; and to (2) describe the association between exposure to CSs and health-information seeking behavior among the US adult population.

Study population, Design, Setting
We analyzed the most recent nationally representative data from the Health Information addresses located in counties comprising Central Appalachia, regardless of minority population. The second stage was equal probability sampling, consisted of selecting one adult within each sampled household using the next-birthday method. The overall household response rate using the next-birthday method was 32.39% calculated using the American Association for Public Opinion Research response rate 2 (RR2) formula. 22 A detailed description of survey methodology has been published. 23 Sample size for this study was

Health Information Seeking Behavior
The primary outcome in the study was self-reported health information seeking behavior among US adults. This was defined as an affirmative response to the question "Have you ever looked for information about health or medical topics from any source?" Exposure to corrective statements in the past 6 months The main exposure was a report from the participant that they had seen the corrective statements. Within the survey, this was defined as an affirmative response to the question, "In the past 6 months, have you seen messages in newspapers or on television that say that a Federal Court has ordered tobacco companies to make statements about the dangers of smoking cigarettes?" Respondents who reported exposure to court-ordered corrective statements messages were

Statistical Analyses
Data was cleaned prior to the analysis, and about 13% (n=480) of the population had missing and smoking status. They were found to be missing at random and therefore deleted from the analysis.
Prevalence of healthy behavior among US adults exposed and non-exposed to CSs messages was calculated for the overall sample as well as by sociodemographic characteristics such as; age, sex, race/ethnicity, level of education, rural-urban residence, household annual income, and tobacco use status. The proportion of participants who were exposed to each type of antismoking message was calculated for the general population and by sociodemographic characteristics. Group differences were assessed using chi-square. Associations between exposure to CSs and health information seeking behavior was assessed by logistic regression.
Regression models were fitted, after comparing the independent association of each variable to the outcome (health information seeking behavior) using bivariate analysis at P value less than 0.1 level, followed by assessment of the dependent variables collinearity. Then we advanced with backwards stepwise regression with non-overlapping variables. The final model was adjusted for 4 demographic confounding variables (Sex, education, Income and race).
Because category of smoking-related advertisements seen could affect the health information seeking behavior, we examined prevalence of information seeking behavior by message category (Message 1-5) reported by the participants that they were exposed to. The proportion of participants who were exposed to each message category was evaluated among overall and by sociodemographic characteristics to assess message salience. Statistical significance was defined as a P value less than 0.05, and all tests were 2-tailed. All data were weighted to be nationally representative and analyzed with STATA version 14. found to be independently associated with health-information seeking behavior. Table 1 further describes the prevalence of health information seeking behavior among US adults exposed and non-exposed to Federal Court-Ordered Antismoking Advertisements by different sociodemographic characteristics. Group differences in educational attainment was found to be statistically significant; those with less than high school education sought out health information significantly less (70.2%, 95%CI=53.8-86.5) when compared to people who were college graduates or more (93.3%, 95%CI=90.8 -95.7) among those who were exposed to CSs (p<0.0002). In addition, there was a significant difference between sexes, with females reporting higher prevalence at 88.4% (95%CI= 85.9-90.96) compared to males at 75.4% (95%CI =67.3-83.6) among those who were exposed (p<0.0001) ( Table 1).
While assessing the impact of all five corrective statement messages language and advertisement framing on participants' message recall, we found that majority reported Furthermore, Figure 1 shows that the vast majority (72.62%) of US Adults who reported exposure to CSs advertisements reported exposure to more than one federal Court-Ordered Messages. Among those exposed, 28 Figure).
Moreover, among exposed adults who reported seeing multiple messages (72.62%), the most In Table 2, we explored the prevalence of health information seeking behavior among US adults exposed to CSs antismoking advertisements stratified by category of message reported.
Among the overall sample, no significant difference in health seeking behavior was observed by type of CSs. Significant variation was seen by education; people with high-school education were the least subgroup to report seeking health information behavior; with prevalence as low as 63.5% (95%CI= 40.

DISCUSSION
This study, to our knowledge, is the first to explore the association between the federal courtordered antismoking corrective statements advertising campaign and health information seeking behavior within the US Adult population. Our results suggested that even though a large proportion of US adults reported exposure to CSs messages, this was not significantly associated with health information seeking behavior. This is an important finding since it underscores the need for a multi-pronged approach since any single intervention may have Despite the success of several large-scale public health campaigns, there is also evidence that some public health campaigns, including those with mass media components, have had ambiguous or no effects. 7 Possible explanations to why we did not see a significant association between CSs exposure and seeking information were that we could not specifically assess tobacco-related information seeking behaviors as such information was not available in the survey. Exposed individuals who smoked were possibly at the contemplation or preparation stages (transtheoretical model), but these did not lead to health info seeking behaviors. 21 Future research could explore temporality of positive behavior after established exposure to health advertisements. In this study, we saw that women had stronger responses to CSs compared to men; consistent with prior research, where women were found to be more likely to report that CSs messages were serious and that it motivated them to quit. 24 On the other hand, our finding regarding how US adults with lower educational attainment (high school or less) experienced weaker responses to CSs, contradicted what was reported in previous studies, where people with a high school education or less were found to be more responsive to anti-smoking warning labels. 25 Our report explores the reach of the five topics which the tobacco industry was obligated to issue CSs about. We saw variations in proportions that reported exposure to CSs by message topic. While the majority of adults 85.8% (95%CI= 82.9 -88.6) reported exposure to Given that each of the five messages equally appeared on major TV programs and newspapers during the campaign 26 ; this variation might reflect differences in message salience: simpler messages such as Message 1 may be easier to recall, compared to messages with more technical terms such as Message 4. One previous study reported that CSs message novelty was associated with higher reported information relevance, anger at the industry, and motivation to quit among exposed. 24 Our findings were consistent with (Chido-Amajuoyi 2019) report, comparing the exposure prevalence of CSs with previous federal and state sponsored antismoking campaigns, reporting that the impact of CSs advertisements on health behaviors was suboptimal. 27 This could be due to the fact that for decades, misinformation about tobacco products was being circulated in the public communication environments as part of the marketing strategies of tobacco manufacturers. 7 On the other hand, corrective statements ads ran only for a few months; it also originated from the tobacco industry, which could produce suspicion among consumers. 24 Health advertisement relative salience is crucial, because people were more likely to respond positively to its' message and perceive it as more important.

Strengths and Limitations
This report is the first to assess the proportion of adults who were exposed to CSs by CS message type. Our finding suggests that some CSs messages might be more salient than others and the majority of the exposed population were able to recall more than one CSs messages.
There are a number of limitations inherently in cross-sectional studies. First, as information on the source of CS exposure (television or newspaper) were not available in the survey, this prevented us from more nuanced analyses of the reach of CSs by different media type. In addition, though HINTS data are nationally representative it is still cross sectional, and causal inferences cannot be determined. Finally, HINTS was self-reported which is subjected to recall and social desirability bias.

CONCLUSION
This study found that the court ordered national-level antismoking advertising campaign had different exposure and recall patterns in subgroups depending on the message category.
While some messages were easier to recall others, perhaps more technical ones, were less likely to make an impact on participants' memory and prompt change to health behavior.
This study highlights the important role of well-designed mass media campaigns in helping

Declarations
• Ethics approval and consent to participate 'Not applicable' • Consent for publication 'Not applicable' • Availability of data and materials: The data that support the findings of this study are publicly available from The National Cancer Institute's Health Information National Trends Survey 5 (HINTS 5-cycle 2) • Competing interests: The authors declare that they have no competing interests.
• Funding 'Not applicable' • Authors' contributions: SA designed the analytic strategy, conducted the statistical analysis and wrote the manuscript. IA conceived the study, supervised the writing of the manuscript. The Manuscript has been read and approved by all authors.
• Acknowledgements: Authors sincerely acknowledge Ms.Satomi Odani for for her valuable comments on the draft of the manuscript.