Tobacco Price Increases and Population Interest in Smoking Cessation in Japan Between 2004 and 2016: A Google Trends Analysis

Tobacco Price Increases and Population Interest in Smoking Cessation in Japan Between 2004 and... Abstract Introduction Tobacco price increases can generate increased public interest in smoking cessation, but it is not clear how long this interest lasts. Our objective was to evaluate the duration of the impact of cigarette price increase in Japan using Google search data. Methods Monthly or daily aggregated Google search volume for smoking cessation in Japan from 2004 to 2016 was collected in terms of relative search volume (RSV) ranging from 0 to 100. Using monthly RSV data, we evaluated possible relationships between the RSVs and tobacco control measures in Japan. Time periods within which the impact of search volume significantly increased were identified by cluster detection test, using daily RSV data. A spike in RSV preceding the enforcement of a cigarette price increase revealed an anticipation effect. Results Between 2004 and 2016, the three highest monthly RSV spikes were observed in July 2006 (RSV = 66), when cigarette prices increased by 11%, and in September (RSV = 90) and October 2010 (RSV = 100), when cigarette prices increased by 37%. Regarding daily RSV, the detected cluster size around the price increase in 2010 (52 days) was longer than that in 2006 (17 days). In 2010, a cluster period of 25 days before the date of the price increase was observed, suggesting an anticipation effect. After the onset of the price increase, a cluster of 27 days was detected. When the cigarette price increased due to consumption tax in April 2014, almost no anticipation effect was observed. Conclusions The population impact of tobacco price increases on smoking cessation may be assessed using Google Trends data. The cluster indicates that a higher cigarette price increase had a higher and longer lasting effect on population interest in cessation, but the impact may continue for a relatively short time. Implications To examine the duration of the impact of cigarette price increases on population interest in smoking cessation in Japan, Google search data for smoking cessation were analyzed. Between 2004 and 2016, the three highest spikes of monthly RSV were observed in October 2010, when cigarette prices increased by 37%. Analyzing daily RSV data, the detected cluster size around the price increase in 2010 was 52 days, and a cluster period of 25 days before the date of the price increase was observed, suggesting an anticipation effect. The cluster indicates that a higher cigarette price increase had a higher and longer lasting effect, but the population impact continues for a relatively short time. Further increases in the price of cigarettes are necessary. Introduction Evaluating the impact of tobacco control measures on smoking cessation and prioritizing the different measures are important steps to advance tobacco control activity and consequently reduce tobacco-related mortality and morbidity.1–5 In the most recent (2015) WHO Report on the Global Tobacco Epidemic (MPOWER),2 cigarette price rise was listed as the first priority, as this is the most effective way to reduce tobacco use. In the report, Japan was ranked as moderate level for tobacco price, using a relative coverage index of tobacco tax (WHO standard index); however, the absolute price is very low according to the affordability index.6,7 Japan was also ranked as moderate level for cessation support because nicotine patches (since 2006) and varenicline (since 2008) are already covered by medical insurance (generally used in outpatient departments for smoking cessation). However, Japan was ranked as the lowest level (no or weak policies) for other tobacco control measures such as smoke-free policy, warning labels, and antitobacco media campaigns. Mass media activity has been limited to poster and web-based campaigns in support of World No Tobacco Day (May 31) and a subsequent Smoking Cessation Week (May 31 to June 6). The impact of these campaigns has not been evaluated.8 It was therefore clear that objective assessment of tobacco control measures by various methods was necessary. However, very few previous studies have examined the duration of the impact of a control measure.9,10 This study aims to examine the impact of tobacco control measures—especially focusing on cigarette price increases, because of its priority level and because the price increased several times during the study period—on population interest in smoking cessation in Japan, using Google Trends search query data from January 2004 to August 2016. Examining the types and frequency of Internet queries can reveal population interests, thoughts, and engagement.11 Recently, Internet search query data from Google Trends have been used to evaluate policy, objectively determining the impact of the enforcement of a specific policy. Changes in searches for “smoking cessation” were monitored as possible precursors of interest in smoking cessation.12 Therefore, evaluations using Google Trends data have been conducted on the impact of tobacco control measures, such as tobacco taxation, smoke-free legislation, antitobacco media campaign, and reimbursement of smoking cessation support.13–16 The Internet was estimated to be accessible to 66% of the Japanese population in 2004, more than 70% since 2005 and 83% in 2015.17 From 2004 to 2015, this Internet access rate exceeded 90% in the young population (ie, 13–49 years old) whose smoking prevalence was higher than other age groups.17,18 This large, stable proportion enables researchers to evaluate long-term trends, using Google Trends search query data in Japan. Methods Data Google Trends (http://google.com/trends), a publicly available record of search activity, calculates the relative search volume (RSV) for specific search terms. It measures the fraction of searches that include a specific search term in a specific, user-chosen location (in this case, Japan) at a specific time (from January 2004 to August 2016; total of 152 months: because data were collected in September 2016) relative to the total number of searches at that time, normalized to a 0–100 scale (100, highest proportion of all searches; 50, 50% of the highest proportion). Monthly or daily aggregated search query trends originating in Japan were analyzed. The word for “smoking cessation” is “kin-en” in Japanese. This word is used as noun or verb (eg, “kin-en shi-en” or “kin-en suru”). Searches on these terms were monitored in Japan because other synonymous terms for “quit smoking” (eg, “tabako-wo-yameru” or “sotsu-en” in Japanese) were used considerably less frequently than the term “kin-en” (data not shown) probably because of the nature of the Japanese wording. In the main analysis, a query on “kin-en” was used to capture cessation-related population interest in the study.13–15 As a sensitivity analysis, combined RSVs from all searches for “smoking cessation”-related terms (eg, “kin-en,” “tabako-yametai,” “tabako-wo-yameru,” “sotsu-en,” “nicotine patch,” “nicotine gum,” “varenicline,”or others) were monitored (Supplementary Data).11 To validate the RSV trends for smoking cessation, the search for “smoking cessation” was compared with reference search terms “alcohol” and “tea” (Supplementary Data).11,19 Tobacco Control Measures in Japan Information on tobacco control measures in Japan during the period 2004–2016 was gathered from previous studies and reports2,20–22 based on the MPOWER framework (excluding monitoring [M]): that is, smoke-free policy (P), cessation support including cessation therapy (O), warning labels (W1), antitobacco media campaign (W2), ban on promotion and advertisement (E), and cigarette price increases (R).2 The information (two “P,” two “O,” one “W1,” no “E,” and three “R”) is shown according to the time course in Figure 1. Additionally, World No Tobacco Day (May 31) and Smoking Cessation Week (May 31 to June 6) (13 events for “W2”) for every year are indicated as white bars (for May and June) in the figure. Figure 1. View largeDownload slide Tobacco control measures and monthly RSV for smoking cessation from January 2004 to August 2016, Japan. White bars represent RSVs for the months including the World No Tobacco Day, May 31 and “smoking cessation week” (May 31 to June 6). Blue bars represent RSVs of the months when cigarette price increased. *1, from 270 to 300 yen (11% increase) for a pack of 20 cigarettes; *2, The legislation allows a partial smoking ban; *3, from 300 to 410 yen (37% increase) for a pack of 20 cigarettes; *4, The legislation allows a partial smoking ban; *5, The price of cigarettes increased from 410 to 420 yen (2% increase) for a pack of 20 cigarettes, when the consumption tax on consumer products including cigarettes was raised from 5% to 8%. RSV = relative search volume; WHO FCTC = the World Health Organization’s Framework Convention on Tobacco Control. Figure 1. View largeDownload slide Tobacco control measures and monthly RSV for smoking cessation from January 2004 to August 2016, Japan. White bars represent RSVs for the months including the World No Tobacco Day, May 31 and “smoking cessation week” (May 31 to June 6). Blue bars represent RSVs of the months when cigarette price increased. *1, from 270 to 300 yen (11% increase) for a pack of 20 cigarettes; *2, The legislation allows a partial smoking ban; *3, from 300 to 410 yen (37% increase) for a pack of 20 cigarettes; *4, The legislation allows a partial smoking ban; *5, The price of cigarettes increased from 410 to 420 yen (2% increase) for a pack of 20 cigarettes, when the consumption tax on consumer products including cigarettes was raised from 5% to 8%. RSV = relative search volume; WHO FCTC = the World Health Organization’s Framework Convention on Tobacco Control. Statistical Analysis To observe the impact of tobacco-related events on public interest in smoking cessation over time, projections of RSV were interpreted as the higher impact point.11 Thus, high spikes were detected by visual inspection, and RSVs ranging from 0 to 100 were evaluated as continuous values. Using those values with spikes, we evaluated possible relationships between the RSV and tobacco control measures in Japan. In addition to a nonstatistical method of visual inspection, a cluster detection test was used as a statistical approach to observe statistically significant clusters of RSV. Since the cigarette price in Japan increased several times during the study period, and the highest projected spike was observed at the time of the price increase in 2010 (Figure 1), we focused especially on cigarette price increases. To detect the time duration of the impact of the price increase from start to finish, daily RSV data from 12 months (365 days: 6 months before and 6 months after the time of the price increase) were analyzed for each price increase in 2006, 2010, and 2014. We adopted a cluster detection test widely used in spatial epidemiology that calculates flexible scan statistics under a Poisson model with a restricted likelihood ratio, using FleXScan software, version 3.1.2 (https://sites.google.com/site/flexscansoftware/download_e).23 The prespecified significance level for the restriction was set as α1 = 0.2. The significance level of the test was set as α = 0.05, and its p value was calculated from 999 replications of the Monte Carlo hypothesis testing. The maximum temporal length of a cluster was set to enough length (120 days). The average RSV during the 12 months was used as an expected value in the baseline to evaluate the impact of the price increases. Detected clusters of at least 10 days are shown as the results, although there is no logical cutoff point for the duration. In the analyses, 55 clusters were detected in 2006, 22 clusters in 2010, and 35 clusters in 2014. As for the duration of the clusters, two clusters had more than 10 days in 2006, one cluster in 2010, and two clusters in 2014; only one cluster was 7–9 days in total; and most clusters had less than 5 days (89% in 2006, 68% in 2010, and 86% in 2014). When we selected other cutoff points of duration, our interpretation of the results did not differ. The price increases may have an impact prior to its implementation in practice (eg, via anticipation effects).24 In fact, a survey of cessation intention among smokers found that many intended to quit by the day of the price increase (ie, before the tax increase).25 Therefore, the time duration when the impact of the price increase started and finished was evaluated as three time durations (before, after, and the total of the cigarette price increase). Since there are no local-level price variations and the regulated elevated cigarette price was applied concurrently at the time of tobacco taxation and industrial price increase within Japan under the Tobacco Business Law (ie, with no time delay in the market price increase throughout Japan), all subjects were assumed to be affected by the price increases. Results Monthly RSVs for smoking cessation with tobacco control measures in Japan from January 2004 to August 2016 are shown in Figure 1 (RSVs with reference search terms in Supplementary Figure 1). Visually inspecting the RSV over time, spikes of volume were observed according to the time of tobacco control measures in Japan. In October 2010, when the cigarette price increased by 37%,26 an RSV estimated at 100, the highest spike, was observed. An RSV of 90, the second highest spike, was observed on September 2010, a month before the price increase. In June 2006, when the cigarette price was due to increase the following month and “smoking cessation week” took place, an RSV of 66, the third highest spike, was observed. The fifth highest spike, an RSV of 60, was observed in July 2006 when the cigarette price increased by 11%. In June 2008, when Smoking Cessation Week took place, an RSV of 65, the fourth highest spike, was observed. It appears that several months (before and after events) form one spike. After the highest spike, an RSV of 100, following the price increase in October 2010, the next highest RSVs were observed in January 2014 (52) and April 2014 (49). The latter period was when the cigarette price increased due to the consumption tax increase (2% increase: from 410 to 420 yen). Generally, when the cigarette price increased (blue bars in Figure 1) and antitobacco media campaigns were conducted (white bars), relatively higher RSVs were observed over time. Compared with RSVs for “alcohol” and “tea” (the two reference search terms), RSVs for “smoking cessation” show larger up-and-down estimates with some projections (see Supplementary Figure 3). Figure 2 and Table 1 show detected time clusters (red lines) that showed significantly higher levels (p < .001 for all five clusters) of population interest before and after the cigarette price increase in 2006, 2010, and 2014. A cluster (no. 1) was found around the World No Tobacco Day and the campaign week in 2006 (11 days of duration). Another cluster (no. 2) was located around the price increase on June 1, 2006 with a duration of 17 days (5 days before the measure and 12 days after). The longest cluster (no. 3) was located around the price increase on October 1, 2010 with a duration of 52 days (25 days before the measure and 27 days after). A cluster (no. 4) was found in January 2014, although there was no known related tobacco control in this period. The cluster (no. 5) around the price increase on April 1, 2014 lasted 25 days (1 day before the measure and 24 days after). Figure 2. View largeDownload slide Cigarette price increases (2006, 2010, and 2014) and detected clusters (>10 days) of daily RSV. Note. Each 12-month period around a cigarette price increase was used. RSVs were standardized as highest RSV reached 100 during each 12 months. Therefore, RSVs should be cautiously interpreted especially the third row, that is, the variability of RSVs was large, so a small impact may result in a cluster, because there were no largely effective tobacco control measures during the period. Blue dotted vertical lines represent a period of cigarette price increases. Red lines with number (also used in Table 1) represent detected clusters. RSV = relative search volume. Figure 2. View largeDownload slide Cigarette price increases (2006, 2010, and 2014) and detected clusters (>10 days) of daily RSV. Note. Each 12-month period around a cigarette price increase was used. RSVs were standardized as highest RSV reached 100 during each 12 months. Therefore, RSVs should be cautiously interpreted especially the third row, that is, the variability of RSVs was large, so a small impact may result in a cluster, because there were no largely effective tobacco control measures during the period. Blue dotted vertical lines represent a period of cigarette price increases. Red lines with number (also used in Table 1) represent detected clusters. RSV = relative search volume. Table 1. Detected Clusters (≥10 Days) of Daily RSV and Tobacco Control Measures No. of cluster shown in Figure 2  Tobacco control measures and the periods  Cluster duration (days)  Period before date of measure (days)  Period after date of measure (days)  ①  The World No Tobacco Day with campaign week (May 31 to June 6, 2006)  May 29 to June 8, 2006 (11)  2  3a  ②  Cigarette price increase (11%) from July 1, 2006  June 19 to July 12, 2006 (17)  5  12  ③  Cigarette price increase (37%) from October 1, 2010  September 6 to October 27, 2010 (52)  25  27  ④  Unknown  January 5 to January 31, 2014 (27)  —  —  ⑤  Cigarette price increase (2%) from April 1, 2014  March 31 to April 24, 2014 (25)  1  24  No. of cluster shown in Figure 2  Tobacco control measures and the periods  Cluster duration (days)  Period before date of measure (days)  Period after date of measure (days)  ①  The World No Tobacco Day with campaign week (May 31 to June 6, 2006)  May 29 to June 8, 2006 (11)  2  3a  ②  Cigarette price increase (11%) from July 1, 2006  June 19 to July 12, 2006 (17)  5  12  ③  Cigarette price increase (37%) from October 1, 2010  September 6 to October 27, 2010 (52)  25  27  ④  Unknown  January 5 to January 31, 2014 (27)  —  —  ⑤  Cigarette price increase (2%) from April 1, 2014  March 31 to April 24, 2014 (25)  1  24  RSV = relative search volume. aDays after June 6, 2006. View Large Table 1. Detected Clusters (≥10 Days) of Daily RSV and Tobacco Control Measures No. of cluster shown in Figure 2  Tobacco control measures and the periods  Cluster duration (days)  Period before date of measure (days)  Period after date of measure (days)  ①  The World No Tobacco Day with campaign week (May 31 to June 6, 2006)  May 29 to June 8, 2006 (11)  2  3a  ②  Cigarette price increase (11%) from July 1, 2006  June 19 to July 12, 2006 (17)  5  12  ③  Cigarette price increase (37%) from October 1, 2010  September 6 to October 27, 2010 (52)  25  27  ④  Unknown  January 5 to January 31, 2014 (27)  —  —  ⑤  Cigarette price increase (2%) from April 1, 2014  March 31 to April 24, 2014 (25)  1  24  No. of cluster shown in Figure 2  Tobacco control measures and the periods  Cluster duration (days)  Period before date of measure (days)  Period after date of measure (days)  ①  The World No Tobacco Day with campaign week (May 31 to June 6, 2006)  May 29 to June 8, 2006 (11)  2  3a  ②  Cigarette price increase (11%) from July 1, 2006  June 19 to July 12, 2006 (17)  5  12  ③  Cigarette price increase (37%) from October 1, 2010  September 6 to October 27, 2010 (52)  25  27  ④  Unknown  January 5 to January 31, 2014 (27)  —  —  ⑤  Cigarette price increase (2%) from April 1, 2014  March 31 to April 24, 2014 (25)  1  24  RSV = relative search volume. aDays after June 6, 2006. View Large Discussion Summary of Findings We evaluated the impact of multiple tobacco control measures on population interest in smoking cessation in Japan, using a long-term Internet search query data. Using this alternative method, this study indicated that cigarette price rises increase interest in smoking cessation. The highest spike of monthly RSV for “smoking cessation” was observed around October 2010, when the highest cigarette price increase (37% increase) occurred in Japan. The spike was exceptionally high compared with others, indicating the large impact of cigarette price increases on population interest in cessation.3,10 Analyzing daily RSV, the detected cluster suggests that a higher cigarette price increase had a higher and longer lasting effect on population interest in smoking cessation. These findings are in line with previous studies that showed cigarette price increases are the best policy to increase smoking cessation.2,27,28 Interpretations From when to when will a cigarette price increase raise population interest in cessation? We examined this using daily RSV data. When the price increased substantially on October 1, 2010, a cluster period of 25 days before the date of the price increase was observed, suggesting an anticipation effect. After the price increase, a cluster of 27 days was observed, indicating the impact did not continue for a long time. The cigarette price increase due to consumption tax on April 1, 2014 might not have been recognized as such before the actual increase (almost no anticipation effect) as it affected all consumer products, not just tobacco. This result suggests that the effect of a cigarette price increase might be partly due to it being publicly announced and receiving attention in the media, although no previous studies have investigated the impact of this type of publicity in Japan. In 2006, the detected cluster size around the price increase (17 days) was smaller than that detected in 2010 (52 days). A possible explanation for the smaller cluster size was that the level of the price increase might be simply too small to have any meaningful impact.10 A previous epidemiological study in California revealed that the impact of a price increase on cessation lasted for 4 months after the cigarette price increased by 95 cents in 1998.9 In this study, in November 2010, a month after the highest RSV spike (100), the monthly RSV fell to less than half (47) and the average RSV of the next successive 6-month period (December 2010 to May 2011) was 35.2. (This was the lowest 6-month average during the whole period 2004–2016.) This study, using Internet search data, suggested that the population impact of cigarette price increases may not continue for a long time. Most smokers who intended to quit smoking might have searched on the term “smoking cessation” by November 2010, and thus, it would not have been searched thereafter. If the size of the smoker population decreased over time, the population volume who search “smoking cessation” may also decrease over time. As for monthly RSV for other tobacco control policies, no obvious increase of RSV was observed when prefecture-level, smoke-free legislation came into force. This is possibly because country-wide, smoke-free legislation was not implemented in Japan; in other words, the legislation only covered one prefecture out of 47. Furthermore, partial bans have been allowed and many excluded facilities like restaurants and bars, which may have resulted in a limited effect.21,29 However, it might be useful to use Google Trends’ functionality to calculate RSVs for these prefectures separately in a future study. Regarding antitobacco media campaigns, relatively high RSVs around World No Tobacco Day (May 31) and the subsequent Smoking Cessation Week (May 31 to June 6) were observed in some years, probably depending on the context of the annual campaign term and the media coverage. For example, the campaign term “smoking cessation support to help people who want to quit” was used and reported in some newspapers in 2006. These results were consistent with the results from a previous study that evaluated World No Tobacco Day in Latin America using Google Trends data.14 These campaigns might be effective for increasing population interest in cessation, although the impact was lower than that of a cigarette price increase. As for antitobacco media campaigns, our findings were consistent with the findings of other studies in that although there was a noticeable effect, this reduced rapidly when the campaign ended.30,31 This would suggest that continuing media campaigns may be effective, as the MPOWER defined long-lasting media campaign (at least 3 weeks) as the recommended approach.2 Since this study focused on price increases rather than media campaigns, further detailed evaluation of the effect of media campaigns is necessary in future research. The framework of this study (Google Trends analysis) enabled us to evaluate multiple tobacco policies simultaneously in one country. Based on the value of the RSV, it may be possible to prioritize different tobacco control measures. The cigarette price increase in 2010 led to the largest RSV, suggesting that cigarette price increases are the most effective tobacco control measure, although comprehensive smoke-free policies and long-term media campaigns have not been conducted in Japan. Since there are still very few studies using this analytic method, further research is necessary but should be pursued carefully (bearing in mind that search terms are specific for each language). Limitations There are several limitations to address. The first is that Google searches data are not a validated measure and, as pointed out by J. W. Ayers,11,15 its application currently has “face validity.” However, validation studies32,33 using Google search data have been conducted, suggesting the search may reflect real-world data (although there is no gold standard criterion). Second, various factors can affect RSV trends. For example, environmental factors and Google users’ characteristics may be different between 2004 and 2016. Therefore, this study is not able to establish a causal association between cigarette price and interest in cessation. Third, we only used Google search data. Because there are other forms of data, such as other search engines, Twitter, or newspaper databases; these data sources should be accounted for in our future research. Fourth, this study was conducted in Japan, using Japanese terms. To make our study results generalizable, similar research needs to be conducted using other languages. Despite these limitations, this study has strength in that these search data can be directly observed in near real time.34 Conclusions A significant impact of cigarette price increases on population interest in smoking cessation was observed using Google Trends query data. The time cluster detection statistic may help to reveal when the effect of a tobacco control measure on cessation started and how long it lasted. The detected cluster suggests that a higher cigarette price increase had a higher and longer lasting effect, but the population impact continues for a relatively short time. Therefore, we need to continue and advance tobacco control measures that increase public awareness of the need to stop smoking. Supplementary Material Supplementary Data and Figures 1–3 can be found online at http://www.ntr.oxfordjournals.org Funding This work was supported by the Ministry of Health, Labor, and Welfare (Grant; Comprehensive Research on Life-Style Related Diseases including Cardiovascular Diseases and Diabetes Mellitus [H28-junkankitou-ippan-002 and H28-junkankitou-ippan-008]) and the Ministry of Education, Culture, Sports, Science, and Technology (grant-in-aid for Young Scientists B: number 15K19256). The study sponsor had no role in study design; collection, analysis, and interpretation of the data; writing the report; or the decision to submit the report for publication. Declaration of Interests None declared. Acknowledgments We thank J. Mortimer for her English language editing. References 1. Eriksen M, Mackay J, Schluger N, Gomeshtapeh F, Drope J. The Tobacco Atlas, Fifth Edition: Revised, Expanded, and Updated . Atlanta, GA: American Cancer Society; 2015. 2. World Health Organization. MPOWER. 2015. www.who.int/tobacco/mpower/en/. Accessed June 14, 2015. 3. US Department of Health and Human Services. The Health Consequences of Smoking—50 Years of Progress. A Report of the Surgeon General . Atlanta, GA: Centers for Disease Control, Office on Smoking and Health; 2014. 4. Shang C, Chaloupka F, Kostova D. Who quits? An overview of quitters in low- and middle-income countries. Nicotine Tob Res . 2014; 16( suppl 1): S44– S55. Google Scholar CrossRef Search ADS PubMed  5. Wilson LM, Avila Tang E, Chander Get al.   Impact of tobacco control interventions on smoking initiation, cessation, and prevalence: a systematic review. J Environ Public Health . 2012; 2012: 961724. Google Scholar CrossRef Search ADS PubMed  6. Tabuchi T, Nakamura M, Nakayama T, Miyashiro I, Mori J, Tsukuma H. Tobacco price increase and smoking cessation in Japan, a developed country with affordable tobacco: a national population-based observational study. J Epidemiol . 2016; 26( 1): 14– 21. Google Scholar CrossRef Search ADS PubMed  7. Eriksen M, Mackay J, Ross H. The Tobacco Atlas . 4th ed. Atlanta, GA: American Cancer Society; 2012. 8. Tabuchi T. Anti-tobacco media-campaigns in Japan. In: The Study Group on Health Effects of Smoking, ed. Smoking and Health: A Report of the Study Group on Health Effects of Smoking . Tokyo, Japan: Ministry of Health, Labour and Welfare; 2016: 536– 43. (in Japanese) 9. Reed MB, Anderson CM, Vaughn JW, Burns DM. The effect of cigarette price increases on smoking cessation in California. Prev Sci . 2008; 9( 1): 47– 54. Google Scholar CrossRef Search ADS PubMed  10. International Agency for Research on Cancer. IARC Handbooks of Cancer Prevention Tobacco Control Volume 14: Effectiveness of Tax and Price Policies for Tobacco Control . Lyon, France: International Agency for Research on Cancer; 2011. 11. Ayers JW, Althouse BM, Allem JP, Leas EC, Dredze M, Williams RS. Revisiting the rise of electronic nicotine delivery systems using search query surveillance. Am J Prev Med . 2016; 50( 6): e173– e181. Google Scholar CrossRef Search ADS PubMed  12. Ayers JW, Westmaas JL, Leas ECet al.   Leveraging big data to improve health awareness campaigns: a novel evaluation of the great American smokeout. JMIR Public Health Surveill . 2016; 2( 1): e16. Google Scholar CrossRef Search ADS PubMed  13. Troelstra SA, Bosdriesz JR, de Boer MR, Kunst AE. Effect of tobacco control policies on information seeking for smoking cessation in the Netherlands: a Google Trends study. PLoS One . 2016; 11( 2): e0148489. Google Scholar CrossRef Search ADS PubMed  14. Ayers JW, Althouse BM, Allem JP, Ford DE, Ribisl KM, Cohen JE. A novel evaluation of World No Tobacco day in Latin America. J Med Internet Res . 2012; 14( 3): e77. Google Scholar CrossRef Search ADS PubMed  15. Ayers JW, Althouse BM, Emery S. Changes in Internet searches associated with the “Tips from Former Smokers” campaign. Am J Prev Med . 2015; 48( 6): e27– e29. Google Scholar CrossRef Search ADS PubMed  16. Ayers JW, Althouse BM, Ribisl KM, Emery S. Digital detection for tobacco control: online reactions to the 2009 U.S. cigarette excise tax increase. Nicotine Tob Res . 2014; 16( 5): 576– 583. Google Scholar CrossRef Search ADS PubMed  17. Ministry of Internal Affairs and Communication. Communications Usage Trend Survey. 2015. www.soumu.go.jp/johotsusintokei/statistics/data/160722_1.pdf. Accessed September 9, 2016. 18. Tabuchi T, Kondo N. Educational inequalities in smoking among Japanese adults aged 25-94 years: nationally representative sex- and age-specific statistics. J Epidemiol . 2017; 27( 4): 186– 192. Google Scholar CrossRef Search ADS PubMed  19. Nghiem le TP, Papworth SK, Lim FK, Carrasco LR. Analysis of the capacity of google trends to measure interest in conservation topics and the role of online news. PLoS One . 2016; 11( 3): e0152802. Google Scholar CrossRef Search ADS PubMed  20. Kanda H, Osaki Y, Ohida T, Kaneita Y, Munezawa T. Age verification cards fail to fully prevent minors from accessing tobacco products. Tob Control . 2011; 20( 2): 163– 165. Google Scholar CrossRef Search ADS PubMed  21. Yamada K, Mori N, Kashiwabara Met al.   Industry speed bumps on local tobacco control in Japan? The case of hyogo. J Epidemiol . 2015; 25( 7): 496– 504. Google Scholar CrossRef Search ADS PubMed  22. Ministry of Health, Labour and Welfare, The Study Group on Health Effects of Smoking. Smoking and Health: A Report of the Study Group on Health Effects of Smoking. 2016. www.mhlw.go.jp/stf/shingi2/0000135586.html. Accessed September 26, 2016. 23. Takahashi K, Kulldorff M, Tango T, Yih K. A flexibly shaped space-time scan statistic for disease outbreak detection and monitoring. Int J Health Geogr . 2008; 7: 14. Google Scholar CrossRef Search ADS PubMed  24. Abadie A, Diamond A, Hainmueller J. Synthetic control methods for comparative case studies: estimating the effect of California’s tobacco control program. J Am Stat Assoc . 2010; 105( 490): 493– 505. Google Scholar CrossRef Search ADS   25. Pfizer Japan Inc. Internet survey for Japanese smokers before tobacco tax increase in 2010. Tokyo: Pfizer Japan Inc.; 2010. (in Japanese) 26. Tabuchi T, Fujiwara T, Shinozaki T. Tobacco price increase and smoking behaviour changes in various subgroups: a nationwide longitudinal 7-year follow-up study among a middle-aged Japanese population. Tob Control . 2017; 26( 1): 69– 77. Google Scholar CrossRef Search ADS PubMed  27. Joossens L, Raw M. The Tobacco Control Scale: a new scale to measure country activity. Tob Control . 2006; 15( 3): 247– 253. Google Scholar CrossRef Search ADS PubMed  28. Levy DT, Yuan Z, Luo Y, Mays D. Seven years of progress in tobacco control: an evaluation of the effect of nations meeting the highest level MPOWER measures between 2007 and 2014. Tob Control . 2018; 27( 1): 50– 57. Google Scholar CrossRef Search ADS PubMed  29. Tabuchi T, Hoshino T, Nakayama T. Are partial workplace smoking bans as effective as complete smoking bans? A national population-based study of smoke-free policy among Japanese employees. Nicotine Tob Res . 2016; 18( 5): 1265– 1273. Google Scholar CrossRef Search ADS PubMed  30. Wakefield MA, Durkin S, Spittal MJet al.   Impact of tobacco control policies and mass media campaigns on monthly adult smoking prevalence. Am J Public Health . 2008; 98( 8): 1443– 1450. Google Scholar CrossRef Search ADS PubMed  31. Wakefield MA, Coomber K, Durkin SJet al.   Time series analysis of the impact of tobacco control policies on smoking prevalence among Australian adults, 2001–2011. Bull World Health Organ . 2014; 92( 6): 413– 422. Google Scholar CrossRef Search ADS PubMed  32. Santillana M, Zhang DW, Althouse BM, Ayers JW. What can digital disease detection learn from (an external revision to) Google Flu Trends? Am J Prev Med . 2014; 47( 3): 341– 347. Google Scholar CrossRef Search ADS PubMed  33. Goel S, Hofman JM, Lahaie S, Pennock DM, Watts DJ. Predicting consumer behavior with Web search. Proc Natl Acad Sci USA . 2010; 107( 41): 17486– 17490. Google Scholar CrossRef Search ADS PubMed  34. Ayers JW, Althouse BM, Dredze M. Could behavioral medicine lead the web data revolution? JAMA . 2014; 311( 14): 1399– 1400. Google Scholar CrossRef Search ADS PubMed  © The Author(s) 2018. Published by Oxford University Press on behalf of the Society for Research on Nicotine and Tobacco. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Nicotine and Tobacco Research Oxford University Press

Tobacco Price Increases and Population Interest in Smoking Cessation in Japan Between 2004 and 2016: A Google Trends Analysis

Loading next page...
 
/lp/ou_press/tobacco-price-increases-and-population-interest-in-smoking-cessation-0GwsV34T5B
Publisher
Oxford University Press
Copyright
© The Author(s) 2018. Published by Oxford University Press on behalf of the Society for Research on Nicotine and Tobacco. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com.
ISSN
1462-2203
eISSN
1469-994X
D.O.I.
10.1093/ntr/nty020
Publisher site
See Article on Publisher Site

Abstract

Abstract Introduction Tobacco price increases can generate increased public interest in smoking cessation, but it is not clear how long this interest lasts. Our objective was to evaluate the duration of the impact of cigarette price increase in Japan using Google search data. Methods Monthly or daily aggregated Google search volume for smoking cessation in Japan from 2004 to 2016 was collected in terms of relative search volume (RSV) ranging from 0 to 100. Using monthly RSV data, we evaluated possible relationships between the RSVs and tobacco control measures in Japan. Time periods within which the impact of search volume significantly increased were identified by cluster detection test, using daily RSV data. A spike in RSV preceding the enforcement of a cigarette price increase revealed an anticipation effect. Results Between 2004 and 2016, the three highest monthly RSV spikes were observed in July 2006 (RSV = 66), when cigarette prices increased by 11%, and in September (RSV = 90) and October 2010 (RSV = 100), when cigarette prices increased by 37%. Regarding daily RSV, the detected cluster size around the price increase in 2010 (52 days) was longer than that in 2006 (17 days). In 2010, a cluster period of 25 days before the date of the price increase was observed, suggesting an anticipation effect. After the onset of the price increase, a cluster of 27 days was detected. When the cigarette price increased due to consumption tax in April 2014, almost no anticipation effect was observed. Conclusions The population impact of tobacco price increases on smoking cessation may be assessed using Google Trends data. The cluster indicates that a higher cigarette price increase had a higher and longer lasting effect on population interest in cessation, but the impact may continue for a relatively short time. Implications To examine the duration of the impact of cigarette price increases on population interest in smoking cessation in Japan, Google search data for smoking cessation were analyzed. Between 2004 and 2016, the three highest spikes of monthly RSV were observed in October 2010, when cigarette prices increased by 37%. Analyzing daily RSV data, the detected cluster size around the price increase in 2010 was 52 days, and a cluster period of 25 days before the date of the price increase was observed, suggesting an anticipation effect. The cluster indicates that a higher cigarette price increase had a higher and longer lasting effect, but the population impact continues for a relatively short time. Further increases in the price of cigarettes are necessary. Introduction Evaluating the impact of tobacco control measures on smoking cessation and prioritizing the different measures are important steps to advance tobacco control activity and consequently reduce tobacco-related mortality and morbidity.1–5 In the most recent (2015) WHO Report on the Global Tobacco Epidemic (MPOWER),2 cigarette price rise was listed as the first priority, as this is the most effective way to reduce tobacco use. In the report, Japan was ranked as moderate level for tobacco price, using a relative coverage index of tobacco tax (WHO standard index); however, the absolute price is very low according to the affordability index.6,7 Japan was also ranked as moderate level for cessation support because nicotine patches (since 2006) and varenicline (since 2008) are already covered by medical insurance (generally used in outpatient departments for smoking cessation). However, Japan was ranked as the lowest level (no or weak policies) for other tobacco control measures such as smoke-free policy, warning labels, and antitobacco media campaigns. Mass media activity has been limited to poster and web-based campaigns in support of World No Tobacco Day (May 31) and a subsequent Smoking Cessation Week (May 31 to June 6). The impact of these campaigns has not been evaluated.8 It was therefore clear that objective assessment of tobacco control measures by various methods was necessary. However, very few previous studies have examined the duration of the impact of a control measure.9,10 This study aims to examine the impact of tobacco control measures—especially focusing on cigarette price increases, because of its priority level and because the price increased several times during the study period—on population interest in smoking cessation in Japan, using Google Trends search query data from January 2004 to August 2016. Examining the types and frequency of Internet queries can reveal population interests, thoughts, and engagement.11 Recently, Internet search query data from Google Trends have been used to evaluate policy, objectively determining the impact of the enforcement of a specific policy. Changes in searches for “smoking cessation” were monitored as possible precursors of interest in smoking cessation.12 Therefore, evaluations using Google Trends data have been conducted on the impact of tobacco control measures, such as tobacco taxation, smoke-free legislation, antitobacco media campaign, and reimbursement of smoking cessation support.13–16 The Internet was estimated to be accessible to 66% of the Japanese population in 2004, more than 70% since 2005 and 83% in 2015.17 From 2004 to 2015, this Internet access rate exceeded 90% in the young population (ie, 13–49 years old) whose smoking prevalence was higher than other age groups.17,18 This large, stable proportion enables researchers to evaluate long-term trends, using Google Trends search query data in Japan. Methods Data Google Trends (http://google.com/trends), a publicly available record of search activity, calculates the relative search volume (RSV) for specific search terms. It measures the fraction of searches that include a specific search term in a specific, user-chosen location (in this case, Japan) at a specific time (from January 2004 to August 2016; total of 152 months: because data were collected in September 2016) relative to the total number of searches at that time, normalized to a 0–100 scale (100, highest proportion of all searches; 50, 50% of the highest proportion). Monthly or daily aggregated search query trends originating in Japan were analyzed. The word for “smoking cessation” is “kin-en” in Japanese. This word is used as noun or verb (eg, “kin-en shi-en” or “kin-en suru”). Searches on these terms were monitored in Japan because other synonymous terms for “quit smoking” (eg, “tabako-wo-yameru” or “sotsu-en” in Japanese) were used considerably less frequently than the term “kin-en” (data not shown) probably because of the nature of the Japanese wording. In the main analysis, a query on “kin-en” was used to capture cessation-related population interest in the study.13–15 As a sensitivity analysis, combined RSVs from all searches for “smoking cessation”-related terms (eg, “kin-en,” “tabako-yametai,” “tabako-wo-yameru,” “sotsu-en,” “nicotine patch,” “nicotine gum,” “varenicline,”or others) were monitored (Supplementary Data).11 To validate the RSV trends for smoking cessation, the search for “smoking cessation” was compared with reference search terms “alcohol” and “tea” (Supplementary Data).11,19 Tobacco Control Measures in Japan Information on tobacco control measures in Japan during the period 2004–2016 was gathered from previous studies and reports2,20–22 based on the MPOWER framework (excluding monitoring [M]): that is, smoke-free policy (P), cessation support including cessation therapy (O), warning labels (W1), antitobacco media campaign (W2), ban on promotion and advertisement (E), and cigarette price increases (R).2 The information (two “P,” two “O,” one “W1,” no “E,” and three “R”) is shown according to the time course in Figure 1. Additionally, World No Tobacco Day (May 31) and Smoking Cessation Week (May 31 to June 6) (13 events for “W2”) for every year are indicated as white bars (for May and June) in the figure. Figure 1. View largeDownload slide Tobacco control measures and monthly RSV for smoking cessation from January 2004 to August 2016, Japan. White bars represent RSVs for the months including the World No Tobacco Day, May 31 and “smoking cessation week” (May 31 to June 6). Blue bars represent RSVs of the months when cigarette price increased. *1, from 270 to 300 yen (11% increase) for a pack of 20 cigarettes; *2, The legislation allows a partial smoking ban; *3, from 300 to 410 yen (37% increase) for a pack of 20 cigarettes; *4, The legislation allows a partial smoking ban; *5, The price of cigarettes increased from 410 to 420 yen (2% increase) for a pack of 20 cigarettes, when the consumption tax on consumer products including cigarettes was raised from 5% to 8%. RSV = relative search volume; WHO FCTC = the World Health Organization’s Framework Convention on Tobacco Control. Figure 1. View largeDownload slide Tobacco control measures and monthly RSV for smoking cessation from January 2004 to August 2016, Japan. White bars represent RSVs for the months including the World No Tobacco Day, May 31 and “smoking cessation week” (May 31 to June 6). Blue bars represent RSVs of the months when cigarette price increased. *1, from 270 to 300 yen (11% increase) for a pack of 20 cigarettes; *2, The legislation allows a partial smoking ban; *3, from 300 to 410 yen (37% increase) for a pack of 20 cigarettes; *4, The legislation allows a partial smoking ban; *5, The price of cigarettes increased from 410 to 420 yen (2% increase) for a pack of 20 cigarettes, when the consumption tax on consumer products including cigarettes was raised from 5% to 8%. RSV = relative search volume; WHO FCTC = the World Health Organization’s Framework Convention on Tobacco Control. Statistical Analysis To observe the impact of tobacco-related events on public interest in smoking cessation over time, projections of RSV were interpreted as the higher impact point.11 Thus, high spikes were detected by visual inspection, and RSVs ranging from 0 to 100 were evaluated as continuous values. Using those values with spikes, we evaluated possible relationships between the RSV and tobacco control measures in Japan. In addition to a nonstatistical method of visual inspection, a cluster detection test was used as a statistical approach to observe statistically significant clusters of RSV. Since the cigarette price in Japan increased several times during the study period, and the highest projected spike was observed at the time of the price increase in 2010 (Figure 1), we focused especially on cigarette price increases. To detect the time duration of the impact of the price increase from start to finish, daily RSV data from 12 months (365 days: 6 months before and 6 months after the time of the price increase) were analyzed for each price increase in 2006, 2010, and 2014. We adopted a cluster detection test widely used in spatial epidemiology that calculates flexible scan statistics under a Poisson model with a restricted likelihood ratio, using FleXScan software, version 3.1.2 (https://sites.google.com/site/flexscansoftware/download_e).23 The prespecified significance level for the restriction was set as α1 = 0.2. The significance level of the test was set as α = 0.05, and its p value was calculated from 999 replications of the Monte Carlo hypothesis testing. The maximum temporal length of a cluster was set to enough length (120 days). The average RSV during the 12 months was used as an expected value in the baseline to evaluate the impact of the price increases. Detected clusters of at least 10 days are shown as the results, although there is no logical cutoff point for the duration. In the analyses, 55 clusters were detected in 2006, 22 clusters in 2010, and 35 clusters in 2014. As for the duration of the clusters, two clusters had more than 10 days in 2006, one cluster in 2010, and two clusters in 2014; only one cluster was 7–9 days in total; and most clusters had less than 5 days (89% in 2006, 68% in 2010, and 86% in 2014). When we selected other cutoff points of duration, our interpretation of the results did not differ. The price increases may have an impact prior to its implementation in practice (eg, via anticipation effects).24 In fact, a survey of cessation intention among smokers found that many intended to quit by the day of the price increase (ie, before the tax increase).25 Therefore, the time duration when the impact of the price increase started and finished was evaluated as three time durations (before, after, and the total of the cigarette price increase). Since there are no local-level price variations and the regulated elevated cigarette price was applied concurrently at the time of tobacco taxation and industrial price increase within Japan under the Tobacco Business Law (ie, with no time delay in the market price increase throughout Japan), all subjects were assumed to be affected by the price increases. Results Monthly RSVs for smoking cessation with tobacco control measures in Japan from January 2004 to August 2016 are shown in Figure 1 (RSVs with reference search terms in Supplementary Figure 1). Visually inspecting the RSV over time, spikes of volume were observed according to the time of tobacco control measures in Japan. In October 2010, when the cigarette price increased by 37%,26 an RSV estimated at 100, the highest spike, was observed. An RSV of 90, the second highest spike, was observed on September 2010, a month before the price increase. In June 2006, when the cigarette price was due to increase the following month and “smoking cessation week” took place, an RSV of 66, the third highest spike, was observed. The fifth highest spike, an RSV of 60, was observed in July 2006 when the cigarette price increased by 11%. In June 2008, when Smoking Cessation Week took place, an RSV of 65, the fourth highest spike, was observed. It appears that several months (before and after events) form one spike. After the highest spike, an RSV of 100, following the price increase in October 2010, the next highest RSVs were observed in January 2014 (52) and April 2014 (49). The latter period was when the cigarette price increased due to the consumption tax increase (2% increase: from 410 to 420 yen). Generally, when the cigarette price increased (blue bars in Figure 1) and antitobacco media campaigns were conducted (white bars), relatively higher RSVs were observed over time. Compared with RSVs for “alcohol” and “tea” (the two reference search terms), RSVs for “smoking cessation” show larger up-and-down estimates with some projections (see Supplementary Figure 3). Figure 2 and Table 1 show detected time clusters (red lines) that showed significantly higher levels (p < .001 for all five clusters) of population interest before and after the cigarette price increase in 2006, 2010, and 2014. A cluster (no. 1) was found around the World No Tobacco Day and the campaign week in 2006 (11 days of duration). Another cluster (no. 2) was located around the price increase on June 1, 2006 with a duration of 17 days (5 days before the measure and 12 days after). The longest cluster (no. 3) was located around the price increase on October 1, 2010 with a duration of 52 days (25 days before the measure and 27 days after). A cluster (no. 4) was found in January 2014, although there was no known related tobacco control in this period. The cluster (no. 5) around the price increase on April 1, 2014 lasted 25 days (1 day before the measure and 24 days after). Figure 2. View largeDownload slide Cigarette price increases (2006, 2010, and 2014) and detected clusters (>10 days) of daily RSV. Note. Each 12-month period around a cigarette price increase was used. RSVs were standardized as highest RSV reached 100 during each 12 months. Therefore, RSVs should be cautiously interpreted especially the third row, that is, the variability of RSVs was large, so a small impact may result in a cluster, because there were no largely effective tobacco control measures during the period. Blue dotted vertical lines represent a period of cigarette price increases. Red lines with number (also used in Table 1) represent detected clusters. RSV = relative search volume. Figure 2. View largeDownload slide Cigarette price increases (2006, 2010, and 2014) and detected clusters (>10 days) of daily RSV. Note. Each 12-month period around a cigarette price increase was used. RSVs were standardized as highest RSV reached 100 during each 12 months. Therefore, RSVs should be cautiously interpreted especially the third row, that is, the variability of RSVs was large, so a small impact may result in a cluster, because there were no largely effective tobacco control measures during the period. Blue dotted vertical lines represent a period of cigarette price increases. Red lines with number (also used in Table 1) represent detected clusters. RSV = relative search volume. Table 1. Detected Clusters (≥10 Days) of Daily RSV and Tobacco Control Measures No. of cluster shown in Figure 2  Tobacco control measures and the periods  Cluster duration (days)  Period before date of measure (days)  Period after date of measure (days)  ①  The World No Tobacco Day with campaign week (May 31 to June 6, 2006)  May 29 to June 8, 2006 (11)  2  3a  ②  Cigarette price increase (11%) from July 1, 2006  June 19 to July 12, 2006 (17)  5  12  ③  Cigarette price increase (37%) from October 1, 2010  September 6 to October 27, 2010 (52)  25  27  ④  Unknown  January 5 to January 31, 2014 (27)  —  —  ⑤  Cigarette price increase (2%) from April 1, 2014  March 31 to April 24, 2014 (25)  1  24  No. of cluster shown in Figure 2  Tobacco control measures and the periods  Cluster duration (days)  Period before date of measure (days)  Period after date of measure (days)  ①  The World No Tobacco Day with campaign week (May 31 to June 6, 2006)  May 29 to June 8, 2006 (11)  2  3a  ②  Cigarette price increase (11%) from July 1, 2006  June 19 to July 12, 2006 (17)  5  12  ③  Cigarette price increase (37%) from October 1, 2010  September 6 to October 27, 2010 (52)  25  27  ④  Unknown  January 5 to January 31, 2014 (27)  —  —  ⑤  Cigarette price increase (2%) from April 1, 2014  March 31 to April 24, 2014 (25)  1  24  RSV = relative search volume. aDays after June 6, 2006. View Large Table 1. Detected Clusters (≥10 Days) of Daily RSV and Tobacco Control Measures No. of cluster shown in Figure 2  Tobacco control measures and the periods  Cluster duration (days)  Period before date of measure (days)  Period after date of measure (days)  ①  The World No Tobacco Day with campaign week (May 31 to June 6, 2006)  May 29 to June 8, 2006 (11)  2  3a  ②  Cigarette price increase (11%) from July 1, 2006  June 19 to July 12, 2006 (17)  5  12  ③  Cigarette price increase (37%) from October 1, 2010  September 6 to October 27, 2010 (52)  25  27  ④  Unknown  January 5 to January 31, 2014 (27)  —  —  ⑤  Cigarette price increase (2%) from April 1, 2014  March 31 to April 24, 2014 (25)  1  24  No. of cluster shown in Figure 2  Tobacco control measures and the periods  Cluster duration (days)  Period before date of measure (days)  Period after date of measure (days)  ①  The World No Tobacco Day with campaign week (May 31 to June 6, 2006)  May 29 to June 8, 2006 (11)  2  3a  ②  Cigarette price increase (11%) from July 1, 2006  June 19 to July 12, 2006 (17)  5  12  ③  Cigarette price increase (37%) from October 1, 2010  September 6 to October 27, 2010 (52)  25  27  ④  Unknown  January 5 to January 31, 2014 (27)  —  —  ⑤  Cigarette price increase (2%) from April 1, 2014  March 31 to April 24, 2014 (25)  1  24  RSV = relative search volume. aDays after June 6, 2006. View Large Discussion Summary of Findings We evaluated the impact of multiple tobacco control measures on population interest in smoking cessation in Japan, using a long-term Internet search query data. Using this alternative method, this study indicated that cigarette price rises increase interest in smoking cessation. The highest spike of monthly RSV for “smoking cessation” was observed around October 2010, when the highest cigarette price increase (37% increase) occurred in Japan. The spike was exceptionally high compared with others, indicating the large impact of cigarette price increases on population interest in cessation.3,10 Analyzing daily RSV, the detected cluster suggests that a higher cigarette price increase had a higher and longer lasting effect on population interest in smoking cessation. These findings are in line with previous studies that showed cigarette price increases are the best policy to increase smoking cessation.2,27,28 Interpretations From when to when will a cigarette price increase raise population interest in cessation? We examined this using daily RSV data. When the price increased substantially on October 1, 2010, a cluster period of 25 days before the date of the price increase was observed, suggesting an anticipation effect. After the price increase, a cluster of 27 days was observed, indicating the impact did not continue for a long time. The cigarette price increase due to consumption tax on April 1, 2014 might not have been recognized as such before the actual increase (almost no anticipation effect) as it affected all consumer products, not just tobacco. This result suggests that the effect of a cigarette price increase might be partly due to it being publicly announced and receiving attention in the media, although no previous studies have investigated the impact of this type of publicity in Japan. In 2006, the detected cluster size around the price increase (17 days) was smaller than that detected in 2010 (52 days). A possible explanation for the smaller cluster size was that the level of the price increase might be simply too small to have any meaningful impact.10 A previous epidemiological study in California revealed that the impact of a price increase on cessation lasted for 4 months after the cigarette price increased by 95 cents in 1998.9 In this study, in November 2010, a month after the highest RSV spike (100), the monthly RSV fell to less than half (47) and the average RSV of the next successive 6-month period (December 2010 to May 2011) was 35.2. (This was the lowest 6-month average during the whole period 2004–2016.) This study, using Internet search data, suggested that the population impact of cigarette price increases may not continue for a long time. Most smokers who intended to quit smoking might have searched on the term “smoking cessation” by November 2010, and thus, it would not have been searched thereafter. If the size of the smoker population decreased over time, the population volume who search “smoking cessation” may also decrease over time. As for monthly RSV for other tobacco control policies, no obvious increase of RSV was observed when prefecture-level, smoke-free legislation came into force. This is possibly because country-wide, smoke-free legislation was not implemented in Japan; in other words, the legislation only covered one prefecture out of 47. Furthermore, partial bans have been allowed and many excluded facilities like restaurants and bars, which may have resulted in a limited effect.21,29 However, it might be useful to use Google Trends’ functionality to calculate RSVs for these prefectures separately in a future study. Regarding antitobacco media campaigns, relatively high RSVs around World No Tobacco Day (May 31) and the subsequent Smoking Cessation Week (May 31 to June 6) were observed in some years, probably depending on the context of the annual campaign term and the media coverage. For example, the campaign term “smoking cessation support to help people who want to quit” was used and reported in some newspapers in 2006. These results were consistent with the results from a previous study that evaluated World No Tobacco Day in Latin America using Google Trends data.14 These campaigns might be effective for increasing population interest in cessation, although the impact was lower than that of a cigarette price increase. As for antitobacco media campaigns, our findings were consistent with the findings of other studies in that although there was a noticeable effect, this reduced rapidly when the campaign ended.30,31 This would suggest that continuing media campaigns may be effective, as the MPOWER defined long-lasting media campaign (at least 3 weeks) as the recommended approach.2 Since this study focused on price increases rather than media campaigns, further detailed evaluation of the effect of media campaigns is necessary in future research. The framework of this study (Google Trends analysis) enabled us to evaluate multiple tobacco policies simultaneously in one country. Based on the value of the RSV, it may be possible to prioritize different tobacco control measures. The cigarette price increase in 2010 led to the largest RSV, suggesting that cigarette price increases are the most effective tobacco control measure, although comprehensive smoke-free policies and long-term media campaigns have not been conducted in Japan. Since there are still very few studies using this analytic method, further research is necessary but should be pursued carefully (bearing in mind that search terms are specific for each language). Limitations There are several limitations to address. The first is that Google searches data are not a validated measure and, as pointed out by J. W. Ayers,11,15 its application currently has “face validity.” However, validation studies32,33 using Google search data have been conducted, suggesting the search may reflect real-world data (although there is no gold standard criterion). Second, various factors can affect RSV trends. For example, environmental factors and Google users’ characteristics may be different between 2004 and 2016. Therefore, this study is not able to establish a causal association between cigarette price and interest in cessation. Third, we only used Google search data. Because there are other forms of data, such as other search engines, Twitter, or newspaper databases; these data sources should be accounted for in our future research. Fourth, this study was conducted in Japan, using Japanese terms. To make our study results generalizable, similar research needs to be conducted using other languages. Despite these limitations, this study has strength in that these search data can be directly observed in near real time.34 Conclusions A significant impact of cigarette price increases on population interest in smoking cessation was observed using Google Trends query data. The time cluster detection statistic may help to reveal when the effect of a tobacco control measure on cessation started and how long it lasted. The detected cluster suggests that a higher cigarette price increase had a higher and longer lasting effect, but the population impact continues for a relatively short time. Therefore, we need to continue and advance tobacco control measures that increase public awareness of the need to stop smoking. Supplementary Material Supplementary Data and Figures 1–3 can be found online at http://www.ntr.oxfordjournals.org Funding This work was supported by the Ministry of Health, Labor, and Welfare (Grant; Comprehensive Research on Life-Style Related Diseases including Cardiovascular Diseases and Diabetes Mellitus [H28-junkankitou-ippan-002 and H28-junkankitou-ippan-008]) and the Ministry of Education, Culture, Sports, Science, and Technology (grant-in-aid for Young Scientists B: number 15K19256). The study sponsor had no role in study design; collection, analysis, and interpretation of the data; writing the report; or the decision to submit the report for publication. Declaration of Interests None declared. Acknowledgments We thank J. Mortimer for her English language editing. References 1. Eriksen M, Mackay J, Schluger N, Gomeshtapeh F, Drope J. The Tobacco Atlas, Fifth Edition: Revised, Expanded, and Updated . Atlanta, GA: American Cancer Society; 2015. 2. World Health Organization. MPOWER. 2015. www.who.int/tobacco/mpower/en/. Accessed June 14, 2015. 3. US Department of Health and Human Services. The Health Consequences of Smoking—50 Years of Progress. A Report of the Surgeon General . Atlanta, GA: Centers for Disease Control, Office on Smoking and Health; 2014. 4. Shang C, Chaloupka F, Kostova D. Who quits? An overview of quitters in low- and middle-income countries. Nicotine Tob Res . 2014; 16( suppl 1): S44– S55. Google Scholar CrossRef Search ADS PubMed  5. Wilson LM, Avila Tang E, Chander Get al.   Impact of tobacco control interventions on smoking initiation, cessation, and prevalence: a systematic review. J Environ Public Health . 2012; 2012: 961724. Google Scholar CrossRef Search ADS PubMed  6. Tabuchi T, Nakamura M, Nakayama T, Miyashiro I, Mori J, Tsukuma H. Tobacco price increase and smoking cessation in Japan, a developed country with affordable tobacco: a national population-based observational study. J Epidemiol . 2016; 26( 1): 14– 21. Google Scholar CrossRef Search ADS PubMed  7. Eriksen M, Mackay J, Ross H. The Tobacco Atlas . 4th ed. Atlanta, GA: American Cancer Society; 2012. 8. Tabuchi T. Anti-tobacco media-campaigns in Japan. In: The Study Group on Health Effects of Smoking, ed. Smoking and Health: A Report of the Study Group on Health Effects of Smoking . Tokyo, Japan: Ministry of Health, Labour and Welfare; 2016: 536– 43. (in Japanese) 9. Reed MB, Anderson CM, Vaughn JW, Burns DM. The effect of cigarette price increases on smoking cessation in California. Prev Sci . 2008; 9( 1): 47– 54. Google Scholar CrossRef Search ADS PubMed  10. International Agency for Research on Cancer. IARC Handbooks of Cancer Prevention Tobacco Control Volume 14: Effectiveness of Tax and Price Policies for Tobacco Control . Lyon, France: International Agency for Research on Cancer; 2011. 11. Ayers JW, Althouse BM, Allem JP, Leas EC, Dredze M, Williams RS. Revisiting the rise of electronic nicotine delivery systems using search query surveillance. Am J Prev Med . 2016; 50( 6): e173– e181. Google Scholar CrossRef Search ADS PubMed  12. Ayers JW, Westmaas JL, Leas ECet al.   Leveraging big data to improve health awareness campaigns: a novel evaluation of the great American smokeout. JMIR Public Health Surveill . 2016; 2( 1): e16. Google Scholar CrossRef Search ADS PubMed  13. Troelstra SA, Bosdriesz JR, de Boer MR, Kunst AE. Effect of tobacco control policies on information seeking for smoking cessation in the Netherlands: a Google Trends study. PLoS One . 2016; 11( 2): e0148489. Google Scholar CrossRef Search ADS PubMed  14. Ayers JW, Althouse BM, Allem JP, Ford DE, Ribisl KM, Cohen JE. A novel evaluation of World No Tobacco day in Latin America. J Med Internet Res . 2012; 14( 3): e77. Google Scholar CrossRef Search ADS PubMed  15. Ayers JW, Althouse BM, Emery S. Changes in Internet searches associated with the “Tips from Former Smokers” campaign. Am J Prev Med . 2015; 48( 6): e27– e29. Google Scholar CrossRef Search ADS PubMed  16. Ayers JW, Althouse BM, Ribisl KM, Emery S. Digital detection for tobacco control: online reactions to the 2009 U.S. cigarette excise tax increase. Nicotine Tob Res . 2014; 16( 5): 576– 583. Google Scholar CrossRef Search ADS PubMed  17. Ministry of Internal Affairs and Communication. Communications Usage Trend Survey. 2015. www.soumu.go.jp/johotsusintokei/statistics/data/160722_1.pdf. Accessed September 9, 2016. 18. Tabuchi T, Kondo N. Educational inequalities in smoking among Japanese adults aged 25-94 years: nationally representative sex- and age-specific statistics. J Epidemiol . 2017; 27( 4): 186– 192. Google Scholar CrossRef Search ADS PubMed  19. Nghiem le TP, Papworth SK, Lim FK, Carrasco LR. Analysis of the capacity of google trends to measure interest in conservation topics and the role of online news. PLoS One . 2016; 11( 3): e0152802. Google Scholar CrossRef Search ADS PubMed  20. Kanda H, Osaki Y, Ohida T, Kaneita Y, Munezawa T. Age verification cards fail to fully prevent minors from accessing tobacco products. Tob Control . 2011; 20( 2): 163– 165. Google Scholar CrossRef Search ADS PubMed  21. Yamada K, Mori N, Kashiwabara Met al.   Industry speed bumps on local tobacco control in Japan? The case of hyogo. J Epidemiol . 2015; 25( 7): 496– 504. Google Scholar CrossRef Search ADS PubMed  22. Ministry of Health, Labour and Welfare, The Study Group on Health Effects of Smoking. Smoking and Health: A Report of the Study Group on Health Effects of Smoking. 2016. www.mhlw.go.jp/stf/shingi2/0000135586.html. Accessed September 26, 2016. 23. Takahashi K, Kulldorff M, Tango T, Yih K. A flexibly shaped space-time scan statistic for disease outbreak detection and monitoring. Int J Health Geogr . 2008; 7: 14. Google Scholar CrossRef Search ADS PubMed  24. Abadie A, Diamond A, Hainmueller J. Synthetic control methods for comparative case studies: estimating the effect of California’s tobacco control program. J Am Stat Assoc . 2010; 105( 490): 493– 505. Google Scholar CrossRef Search ADS   25. Pfizer Japan Inc. Internet survey for Japanese smokers before tobacco tax increase in 2010. Tokyo: Pfizer Japan Inc.; 2010. (in Japanese) 26. Tabuchi T, Fujiwara T, Shinozaki T. Tobacco price increase and smoking behaviour changes in various subgroups: a nationwide longitudinal 7-year follow-up study among a middle-aged Japanese population. Tob Control . 2017; 26( 1): 69– 77. Google Scholar CrossRef Search ADS PubMed  27. Joossens L, Raw M. The Tobacco Control Scale: a new scale to measure country activity. Tob Control . 2006; 15( 3): 247– 253. Google Scholar CrossRef Search ADS PubMed  28. Levy DT, Yuan Z, Luo Y, Mays D. Seven years of progress in tobacco control: an evaluation of the effect of nations meeting the highest level MPOWER measures between 2007 and 2014. Tob Control . 2018; 27( 1): 50– 57. Google Scholar CrossRef Search ADS PubMed  29. Tabuchi T, Hoshino T, Nakayama T. Are partial workplace smoking bans as effective as complete smoking bans? A national population-based study of smoke-free policy among Japanese employees. Nicotine Tob Res . 2016; 18( 5): 1265– 1273. Google Scholar CrossRef Search ADS PubMed  30. Wakefield MA, Durkin S, Spittal MJet al.   Impact of tobacco control policies and mass media campaigns on monthly adult smoking prevalence. Am J Public Health . 2008; 98( 8): 1443– 1450. Google Scholar CrossRef Search ADS PubMed  31. Wakefield MA, Coomber K, Durkin SJet al.   Time series analysis of the impact of tobacco control policies on smoking prevalence among Australian adults, 2001–2011. Bull World Health Organ . 2014; 92( 6): 413– 422. Google Scholar CrossRef Search ADS PubMed  32. Santillana M, Zhang DW, Althouse BM, Ayers JW. What can digital disease detection learn from (an external revision to) Google Flu Trends? Am J Prev Med . 2014; 47( 3): 341– 347. Google Scholar CrossRef Search ADS PubMed  33. Goel S, Hofman JM, Lahaie S, Pennock DM, Watts DJ. Predicting consumer behavior with Web search. Proc Natl Acad Sci USA . 2010; 107( 41): 17486– 17490. Google Scholar CrossRef Search ADS PubMed  34. Ayers JW, Althouse BM, Dredze M. Could behavioral medicine lead the web data revolution? JAMA . 2014; 311( 14): 1399– 1400. Google Scholar CrossRef Search ADS PubMed  © The Author(s) 2018. Published by Oxford University Press on behalf of the Society for Research on Nicotine and Tobacco. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com.

Journal

Nicotine and Tobacco ResearchOxford University Press

Published: Jan 31, 2018

There are no references for this article.

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off