210
European Journal of Public Health
......................................................................................................... European Journal of Public Health, Vol. 25, No. 2, 210–215 ß The Author 2014. Published by Oxford University Press on behalf of the European Public Health Association. All rights reserved. doi:10.1093/eurpub/cku203 Advance Access published on 8 December 2014
.........................................................................................................
Tobacco-free schools as a core component of youth tobacco prevention programs: a secondary analysis of data from 43 countries Israel T. Agaku1, Enihomo M. Obadan2, Oluwakemi O. Odukoya3, Olubode Olufajo4,5 1 Department of Social and Behavioral Sciences, Center for Global Tobacco Control, Harvard School of Public Health, Boston, MA 02115, USA 2 Department of Oral Health Policy and Epidemiology, Harvard School of Dental Medicine, Boston, MA 02115, USA 3 Department of Community Health and Primary Care, College of Medicine, University of Lagos, Idi-Araba, Lagos State, Nigeria 4 Department of Epidemiology, Harvard School of Public Health, Boston, MA 02115, USA 5 TIMI Study Group, Brigham and Women’s Hospital, Boston, MA 02115, USA
Background: Preventing tobacco use is a key aspect of health promotion during adolescence. We assessed prevalence and impact of school-based tobacco prevention programs in 43 countries. Methods: We performed a secondary analysis of national data of students aged 13–15 years (Global Youth Tobacco Surveys) from 43 countries during 2005–2011. National surveys of the corresponding school personnel (Global School Personnel Surveys) were performed in each country during the same year as the student surveys. Data on status of enforcement of national smoke-free school policies were obtained from the 2008 and 2009 WHO MPOWER reports. Logistic regression was used to measure ecologic-level associations between school-based tobacco prevention programs and tobacco-related knowledge and behaviour among students (P < 0.05). Results: The proportion of students who were taught in class about the dangers of tobacco use during the school year ranged from 31.4% (Georgia) to 83.4% (Papua New Guinea). For every 10% increase (country level) in the proportion of teachers who reported having a tobacco prevention curriculum in their school, the odds of students reporting exposure to education in class about the dangers of tobacco increased by 6.0% (AOR = 1.06; 95% CI: 1.04–1.08). However, didactic education in class about the dangers of tobacco use was not independently associated with student current cigarette smoking behavior. Conversely, the likelihood of being a current smoker was significantly lower among students in countries with moderate/strongly enforced national smoke-free school policies compared with those in countries with poorly enforced/no national smoke-free school policies (AOR = 0.59; 95% CI: 0.45–0.76). Conclusions: Comprehensive tobacco prevention programs that include well-enforced smoke-free school policies may help reduce youth smoking.
.........................................................................................................
Introduction reventing tobacco use among adolescents is fundamental to
Pcombating the global tobacco epidemic as tobacco initiation
primarily occurs during adolescence. Approximately 70% of adult smokers in the European Union, and 88% of those in the United States started cigarette smoking by age 18 years.1,2 Research however shows that population-level interventions primarily targeting adults may not necessarily lead to changes in adolescent tobacco use.3 Hence, interventions targeted at adolescents and young adults should be an integral part of any comprehensive tobacco prevention and control program.2 The school—an environment where adolescents congregate for a large part of the day—affords several opportunities to invest in tobacco prevention and cessation programs. Educators such as school teachers are authority figures who may be perceived by youths as being trusted sources of information on the health consequences of tobacco use. In line with this, Article 12 of the WHO Framework Convention on Tobacco Control (FCTC) calls for training of educators in tobacco prevention strategies in order to equip students with knowledge about the dangers of tobacco use, how the tobacco industry targets youths, as well as how to resist pressure to use tobacco.4,5 Combining tobacco prevention education with strong enforcement of smokefree school policies may help denormalize tobacco use, and
protect non-smokers from involuntary exposure to secondhand smoke.6 Very little cross-country data exists on the implementation of school-based tobacco prevention programs. Hence, the objectives of this study were to: (i) assess teacher-reported exposure to training and access to educational materials on tobacco prevention among youths (ii) explore ecologic-level associations between teacher-reported presence of classroom and non-classroom-based tobacco prevention programs in school and student-reported receipt of education about the dangers of tobacco use and (iii) assess the association between presence of well enforced smokefree school policies and student-reported smoking behaviour. Data were obtained from the Global Youth Tobacco Surveys (adolescents students aged 13–15 years) and the Global School Personnel Surveys (school personnel) from 43 countries during 2005–2011.
Methodology Data sources The current analysis was deemed IRB exempt as a secondary analysis of publicly available de-identified data. National samples of adolescent students aged 13–15 years from 43 countries were obtained from the 2005–2011 Global Youth Tobacco Surveys (GYTS).7 In addition, within each country, a nationally sample of
Downloaded from http://eurpub.oxfordjournals.org/ by guest on November 16, 2015
Correspondence: Israel Agaku, Department of Social and Behavioral Sciences, Harvard School of Public Health, Center for Global Tobacco Control, Boston, MA 02115, USA, Tel: +1 (770) 728-3220, e-mail: [email protected]
Tobacco-free schools as a core component of youth tobacco prevention programs
school personnel–the Global School Personnel Survey (GSPS)—was obtained, sampled from the schools selected for the GYTS.8 The 43 countries in the study were from all six WHO regions, and were selected based on two criteria: (i) data were available for both the GSPS and GYTS and (ii) both surveys were conducted within the same year/wave. Countries that had data for one survey but not the other, or those with surveys in discordant waves were excluded. For our study, the pooled sample size for school personnel within the GSPS was n = 43 186, while the pooled sample size for students within the GYTS was n = 114 889 (table 1).
211
We classified the 43 countries into four categories, based on their 2012 gross national incomes (GNI) as reported by the World Bank.9 These were; low-income (n = 9 countries: Eritrea, Malawi, Niger, Rwanda, Togo, Uganda, Kyrgyzstan, Bangladesh, and Myanmar); lower-middle-income (n = 15 countries: Republic of Congo, Ghana, Mauritania, Senegal, Swaziland, Morocco, Sudan, Syria, Yemen, Georgia, Moldova, India, Sri Lanka, Mongolia and Papua New Guinea); upper-middle-income (n = 11 countries: Mauritius, Seychelles, South Africa, Islamic Republic of Iran, Lebanon, Bulgaria, Serbia, Panama, Maldives, Tuvalu and Tunisia); and
Table 1 Characteristics of study participants in the Global School Personnel Surveys (school teachers) and Global Youth Tobacco Surveys (students aged 13–15 years) in 43 countries, 2005–2011 Region/country
School teachers (Global School Personnel Survey) Survey year
Percentage of female students (%)
% of students who reported being taught in class about the dangers of tobacco use during the school year (95% CI)
1141 4813 4171 1293 1769 1286 1307 688 1611 844 3947 1239 1947 2251
48.9 40.0 50.6 49.7 51.6 53.9 51.8 52.8 52.8 50.9 57.7 59.8 40.4 50.6
45.5 44.7 61.8 77.6 54.3 76.4 33.1 51.3 62.4 75.8 61.7 71.2 46.5 73.9
(35.0–56.0) (41.6–47.8) (58.1–65.5) (71.0-84.2) (50.0–58.5) (70.9–81.8) (27.5–38.8) (43.8–58.8) (56.8–67.9) (71.7–80.0) (54.9–68.5) (65.5–76.9) (36.3–56.6) (70.5–77.2)
46.8 43.1 61.0 35.8 58.5 50.7 42.3 73.9 56.3 58.2 88.1
1153 2213 3314 2106 905 1797 950 1210 1294 10 821 650
47.9 55.9 52.4 47.0 59.1 50.3 52.8 50.5 52.5 52.7 37.5
35.3 71.7 59.0 44.7 77.5 52.7 44.2 64.7 70.4 53.3 47.1
(29.2–41.5) (68.1–75.4) (52.2–65.8) (39.4–50.0) (73.2–81.8) (49.1–56.4) (39.6–48.9) (55.9–73.5) (65.6–75.2) (51.1–55.5) (41.1–53.1)
2044 2224 423 797 283 155 1090 1599 1525
78.6 58.7 81.6 85.2 91.7 76.1 86.2 84.7 69.2
3241 12 477 3350 2330 891 3004 2476 3501 3350
48.3 51.2 50.9 50.7 52.8 52.6 56.6 55.6 54.6
82.2 57.2 82.7 72.6 31.4 71.8 77.3 82.0 70.3
(80.1–84.3) (55.3–59.1) (78.5–86.8) (67.3–77.8) (23.3–39.5) (67.0–76.6) (72.8–81.8) (78.7–85.2) (64.8–75.9)
2,357
53.2
2716
52.9
79.0 (76.6–81.5)
380 2463 789 2515 3518
28.6 45.9 65.4 88.9 75.6
2135 10 112 1494 1652 4031
59.2 41.7 47.6 49.3 52.6
64.4 71.9 47.7 73.4 78.0
1482 548 120
75.9 58.3 63.6
1499 1368 542
53.4 51.9 60.6
45.3 (35.0–55.5) 83.4 (80.4–86.5) 74.0 (70.3–77.8)
Total number of teachers
Percentage of female teachers (%)
344 1359 346 253 412 837 591 184 593 389 1408 447 461 472
23.2 36.1 36.1 41.6 53.8 56.2 23.1 21.8 19.4 56.3 64.1 59.4 12.8 29.0
327 1318 214 533 2052 477 233 935 65 4329 295
Total number of students sampled
(56.8–72.1) (69.5–74.3) (42.7–52.6) (68.4–78.4) (75.2–80.9)
Notes: Samples were not weighted while percentages were weighted to account for the complex survey design. CI = confidence interval.
Downloaded from http://eurpub.oxfordjournals.org/ by guest on November 16, 2015
African Region Republic of Congo 2009 Eritrea 2006 Ghana 2009 Malawi 2009 Mauritania 2009 Mauritius 2008 Niger 2009 Rwanda 2008 Senegal 2007 Seychelles 2007 South Africa 2011 Swaziland 2009 Togo 2007 Uganda 2007 Eastern Mediterranean Region Islamic Republic of Iran 2007 Kuwait 2009 Lebanon 2005 Morocco 2010 Oman 2010 Saudi Arabia 2010 Sudan 2009 Syria 2010 Tunisia 2010 United Arab Emirates 2005 Yemen 2008 European Region Bulgaria 2008 Cyprus 2005 Czech Republic 2011 Estonia 2007 Georgia 2008 Kyrgyzstan 2008 Latvia 2007 Republic of Moldova 2008 Serbia 2008 Region of the Americas Panama 2008 South-East Asian Region Bangladesh 2007 India 2009 Maldives 2011 Myanmar 2011 Sri Lanka 2011 Western Pacific Region Mongolia 2007 Papua New Guinea 2007 Tuvalu 2005/2006
Students aged 13–15 years (Global Youth Tobacco Survey)
212
European Journal of Public Health
Survey measures School personnel-reported exposure to tobacco prevention programs (GSPS) Within the GSPS, school personnel were asked to respond ‘Yes’ or ‘No’ to the following questions: ‘Have you ever received training to prevent tobacco use among youths?’; ‘Do you have access to teaching and learning materials about tobacco use and how to prevent its use among youth?’; ‘Is tobacco use prevention included somewhere in your school curriculum?’; and ‘Are non-classroom programs or activities (such as an assembly) used to teach tobacco use prevention to students in your school?’
Student-reported exposure to education in class about the dangers of tobacco use (GYTS) Within the GYTS, students’ receipt of education in class about the harmfulness of tobacco use was defined as a ‘Yes’ response to the question: ‘During the school year, were you taught in any of your classes about the dangers of smoking?’ Data were also collected on students’ age, gender and current cigarette smoking status, with the latter defined as having smoked cigarettes on at least one day within the past 30 days.
Analyses Prevalence estimates were calculated as percentages and compared using the chi-squared statistic (P < 0.05). Analyses within the GYTS were restricted to students aged 13–15 years to ensure a comparable cross-country denominator regardless of any heterogeneity between countries in school systems. Analyses within the GSPS were restricted to school personnel who reported their primary role as being a ‘teacher’ (vs. ‘school health services personnel’; ‘Administrator’; or ‘clerical staff’) since an outcome measure assessed in our study in-class student education about the dangers of tobacco use. A binary logistic regression model was fitted to measure the ecologic-level association between teacher-reported presence of tobacco-prevention curricula or non-classroom educational programs and students’ exposure to education about the harmfulness of tobacco use. The dependent variable was a report by a student that they had been taught in any of their classes about the harmfulness of tobacco use during the school year. For our independent variables, we computed as ecological covariates the following: the proportion of school teachers within each country
who reported that tobacco use prevention was included in their school curriculum, the proportion of school teachers who reported having access to teaching and learning materials about tobacco use and prevention, and finally, the proportion of school teachers who reported that non-classroom programs or activities existed in their school to teach tobacco use prevention to students. Adjustments were made for survey year, students’ sex, WHO region and GNI. A second binary logistic regression model was fitted to assess the association between students’ current cigarette smoking behaviour and the degree of enforcement of national smoke-free school policies, adjusting for exposure to in-class education about the dangers of tobacco use, survey year, students’ sex, WHO region and GNI. All data were weighted and analysed with Stata 11 (StataCorp 2009, TX).
Results Teacher-reported tobacco prevention programs and activities in schools The proportion of school teachers that reported ever receiving training in tobacco prevention among youths ranged from 4.5% (Bangladesh) to 53.7% (Kyrgyzstan). The proportion of school teachers that reported having access to educational materials about tobacco prevention among youths ranged from 4.2% (Mauritania) to 88.0% (Republic of Congo). The proportion of school teachers that reported that tobacco use prevention material was included in their school curriculum ranged from 20.3% (the Islamic Republic of Iran) to 95.7% (Kyrgyzstan). The proportion of school teachers that reported having non-classroom programs for tobacco prevention among youths in their schools ranged from 8.7% (Mauritania) to 97.8% (Republic of Moldova) (table 2). Detailed country-specific estimates are shown in table 2.
Relationship between school-based tobacco prevention programs and student-reported exposure to education about the dangers of tobacco use The proportion of adolescent students aged 13–15 years who reported being taught in class about the dangers of tobacco use during the school year ranged from 31.4% (Georgia), to 83.4% (Papua New Guinea) (table 1). Tobacco prevention curricula and non-classroom educational programs were both significantly associated with increased odds of student-reported exposure to education about the dangers of tobacco use. Within our multivariate analyses, for every 10% increase (country level) in the proportion of teachers that reported having a tobacco prevention curriculum in their school, the odds of students reporting that they were taught about the dangers of tobacco use during the school year increased by 6.0% (AOR = 1.06; 95% CI: 1.04–1.08). Similarly, for every 10% increase in the proportion of school teachers who reported the presence of a non-classroom tobacco prevention program in their school, the odds of students reporting they were taught about the dangers of tobacco use increased by 2.0% (AOR = 1.02; 95% CI: 1.01–1.04).
Relationship between enforcement of smoke-free school policies and student smoking behaviour Within our multivariate analyses, the odds of current smoking behaviour among adolescent students were significantly lower in countries with moderate/strongly enforced national smoke-free school policies than those with poorly enforced policies/no smokefree school policies (AOR = 0.59; 95%CI: 0.45–0.76). Conversely, didactic education in class about the harmfulness of tobacco use during the school year was not independently associated with current cigarette smoking among adolescent students (table 4).
Downloaded from http://eurpub.oxfordjournals.org/ by guest on November 16, 2015
high-income countries (n = 8 countries: Kuwait, Oman, Saudi Arabia, United Arab Emirates, Republic of Cyprus, Czech Republic, Estonia and Latvia). Data on the status of smoke-free policies in ‘educational institutions’ for each of the 43 countries were obtained from the 2008 and 2009 WHO MPOWER reports.10,11 The 2008/2009 reports were used rather than more recent MPOWER reports in order to ensure that the reported status of smoke-free school policies accurately reflected the situation as of the time of data collection in the GYTS-GSPS surveys during 2005–2011. Within each country, data were obtained from the MPOWER reports regarding both the presence of national smoke-free school policies (measured as ‘Yes’ or ‘No’), as well as the overall degree of enforcement of smoke-free policies (measured on an increasing scale of 1–10). Using these data, we categorized all countries into two groups: (i) those with reported moderate/ strong enforcement of smoke-free policies (scores 5–10) and (ii) those with either no national smoke-free school policies or poorly enforced policies (scores 1–4). In total, 9 of 43 countries assessed were reported to have moderate/strongly enforced smoke-free policies, namely: Eritrea, Niger, Seychelles, South Africa, Uganda, Islamic Republic of Iran, Estonia, Panama and Sri Lanka.
Tobacco-free schools as a core component of youth tobacco prevention programs
213
Table 2 Teacher-reported exposure to tobacco prevention training and presence of school-based tobacco prevention programs, Global School Personnel Survey in 43 countries, 2005–2011 Region/Country
% that reported having access to educational materials on tobacco use and prevention (95% CI)
% that reported presence of non-classroom tobacco prevention programs (95% CI)
% that reported that tobacco prevention was included in school curriculum (95% CI)
14.6 22.4 14.3 7.8 22.0 6.1 32.9 12.0 9.7 15.9 13.1 8.6 6.3 16.2
(5.9–23.4) (15.9–28.9) (8.9–19.7) (4.1–11.5) (16.8–27.3) (4.0–8.3) (23.5–42.2) (6.6–17.4) (3.8–15.5) (12.8–18.9) (10.2–16.1) (4.4–12.7) (3.1–9.4) (10.8–21.5)
88.0 33.6 24.1 42.6 4.2 38.1 14.4 86.7 81.6 38.9 49.0 34.5 9.2 30.4
(82.7–93.3) (27.3–40.0) (16.6–31.6) (30.9–54.3) (2.4–5.9) (30.5–45.8) (10.0–18.9) (81.6–91.9) (75.1–88.1) (32.9–45.0) (44.5–53.5) (27.1–41.9) (5.2–13.2) (24.3–36.4)
10.8 31.0 32.4 51.4 8.7 54.1 30.8 22.5 26.3 44.6 46.6 53.4 52.4 58.2
(6.1–15.4) (24.0–38.0) (21.9–42.9) (31.6–71.3) (6.2–11.2) (41.8–66.3) (22.2–39.5) (10.9–34.1) (17.2–35.5) (30.9–58.4) (30.9–62.3) (45.7–61.2) (44.2–60.6) (52.4–63.9)
59.4 55.2 56.3 65.2 26.7 56.1 55.2 76.1 47.5 71.3 69.0 68.6 71.9 52.4
(51.7–67.1) (46.9–63.4) (47.6–65.0) (55.0–75.4) (20.3–33.1) (47.9–64.2) (41.6–68.8) (66.9–85.2) (38.4–56.7) (62.4–80.2) (63.8–74.1) (60.3–76.9) (65.4–78.3) (44.4–60.3)
12.1 15.1 21.5 5.9 8.5 18.9 5.9 14.6 6.2 9.7 7.2
(7.3–16.9) (n/a) (15.3–27.8) (2.1–9.7) (5.9–11.2) (9.6–28.2) (1.9–10.0) (11.3–17.8) (n/a) (8.4–11.0) (2.9–11.4)
26.0 44.4 44.6 30.9 58.8 56.9 48.0 47.2 47.7 35.3 19.6
(16.9–35.2) (n/a) (35.0–54.1) (21.6–40.2) (55.1–62.5) (50.1–63.8) (38.0–58.0) (42.1–52.3) (n/a) (32.8–37.8) (11.3–27.8)
14.2 29.3 26.2 22.0 51.0 41.4 26.5 27.1 13.9 43.6 28.7
(7.8–20.5) (n/a) (18.9–33.5) (16.0–27.9) (46.8–55.1) (34.5–48.3) (17.0–35.9) (21.6–32.5) (n/a) (38.9–48.2) (17.2–40.3)
20.3 64.8 –a 61.6 84.0 73.6 86.8 74.6 79.7 –a 76.5
(8.9–31.6) (n/a)
12.3 12.3 21.2 25.0 34.9 53.7 31.4 18.9 14.6
(10.2–14.4) (10.4–14.2)
64.3 26.6 57.2 78.1 56.5 43.2 81.6 63.6 65.1
(60.2–68.3) (24.6–28.5)
58.9 50.4 43.4 66.9 49.0 55.6 62.7 97.8 74.6
(53.3–64.6) (46.4–54.3)
73.8 48.0 –a 94.1 56.8 95.7 86.6 80.0 83.1
(69.5–78.1) (44.3–51.6)
(19.3–30.7) (23.0–46.7) (38.7–68.7) (25.3–37.4) (14.5–23.3) (9.4–19.8)
(73.5–82.7) (43.6–69.4) (31.2–55.2) (77.5–85.8) (58.7–68.5) (59.1–71.0)
(57.4–76.5) (38.8–59.1) (39.5–71.7) (54.6–70.8) (96.6–99.0) (67.0–82.3)
(48.3–74.9) (80.7–87.3) (69.1–78.2) (76.4–97.2) (70.6–78.6) (n/a) (66.0–87.0)
(91.5–96.6) (40.2–73.5) (90.9–100) (81.6–91.7) (73.9–86.2) (75.7–90.5)
24.0 (20.7–27.3)
36.9 (32.4–41.4)
25.7 (19.6–31.9)
53.3 (48.4–58.2)
4.5 10.1 14.6 26.3 10.9
–a 37.8 28.8 76.4 60.1
45.0 49.1 51.5 74.0 35.4
76.3 46.4 66.5 87.6 73.3
(1.9–7.2) (7.0–13.2) (n/a) (21.0–31.6) (9.4–12.5)
25.7 (21.3–30.2) 12.7 (7.6–17.9) 31.7 (15.1–48.4)
(32.0–43.6) (n/a) (70.8–82.1) (56.6–63.6)
46.1 (38.3–53.9) 42.2 (36.6–47.8) 51.8 (34.5–69.0)
(33.9–56.0) (43.7–54.4) (n/a) (66.2–81.7) (31.1–39.6)
37.2 (28.3–46.1) 49.4 (41.4–57.3) 10.5 (2.2–18.8)
(68.3–84.2) (42.1–50.6) (n/a) (83.7–91.6) (70.1–76.6)
59.4 (50.1–68.8) 68.9 (59.5–78.3) 73.5 (57.0–90.0)
Note: CI = Confidence Intervals; n/a = not applicable (both upper and lower confidence limits identical to point estimate since a complete census of school personnel was taken, yielding final sampling survey weight = 1 for all respondents). a Data not available because question was not asked in survey.
Discussion This study showed that students from countries with well-enforced national smoke-free school policies were less likely to be current cigarette smokers compared with those from countries with no national smoke-free school policies or with poorly enforced smoke-free school policies. This highlights the need for enhanced and sustained efforts to implement and enforce smoke-free school policies as a vital component of tobacco prevention and control programs to reduce tobacco use among youths. Smoke-free school policies can be implemented at the national, state, local or school
level and can change social norms and attitudes about tobacco use among youths, and also protect nonsmokers from involuntary secondhand smoke exposure.12 However, enforcement of such policies is necessary to ensure compliance by students, school personnel and visitors.11–13 We found that exposure to didactic education in class about the harmfulness of tobacco use was not independently associated with reduced likelihood of smoking among students. Indeed, schoolbased tobacco educational programs have yielded inconsistent results in previous research with regards their effectiveness in eliciting long-term behavioural change among tobacco users.14–18
Downloaded from http://eurpub.oxfordjournals.org/ by guest on November 16, 2015
African region Republic of Congo Eritrea Ghana Malawi Mauritania Mauritius Niger Rwanda Senegal Seychelles South Africa Swaziland Togo Uganda Eastern Mediterranean Region Islamic Republic of Iran Kuwait Lebanon Morocco Oman Saudi Arabia Sudan Syria Tunisia United Arab Emirates Yemen European region Bulgaria Cyprus Czech Republic Estonia Georgia Kyrgyzstan Latvia Rep. of Moldova Serbia Region of the Americas Panama South-East Asia Bangladesh India Maldives Myanmar Sri Lanka Western Pacific Region Mongolia Papua New Guinea Tuvalu
% that reported being trained in tobacco use prevention among youths (95% CI)
214
European Journal of Public Health
Table 3 The impact of classroom and non-classroom tobacco prevention programs on exposure to didactic education about the dangers of tobacco use among students aged 13–15 years, GYTS-GSPS, 2005–2011 Characteristic
Category
Adjusted odds ratio (95% CI)
P-value
Teacher-reported inclusion of tobacco prevention material in curriculum Teacher-reported access to educational materials on tobacco prevention Teacher-reported presence of non-classroom tobacco prevention activities Survey year WHO Region
(Per 10% increase in proportion of teachers) (Per 10% increase in proportion of teachers) (Per 10% increase in proportion of teachers) (Per unit increase) African region (referent) Eastern Mediterranean region European region Region of the Americas Southeast Asian region Western Pacific region Low (referent) Lower-Middle Upper-Middle High Male (referent) Female
1.06 1.01 1.02 1.02
(1.04–1.08) (0.99–1.03) (1.01–1.04) (0.95–1.10)
European Journal of Public Health
......................................................................................................... European Journal of Public Health, Vol. 25, No. 2, 210–215 ß The Author 2014. Published by Oxford University Press on behalf of the European Public Health Association. All rights reserved. doi:10.1093/eurpub/cku203 Advance Access published on 8 December 2014
.........................................................................................................
Tobacco-free schools as a core component of youth tobacco prevention programs: a secondary analysis of data from 43 countries Israel T. Agaku1, Enihomo M. Obadan2, Oluwakemi O. Odukoya3, Olubode Olufajo4,5 1 Department of Social and Behavioral Sciences, Center for Global Tobacco Control, Harvard School of Public Health, Boston, MA 02115, USA 2 Department of Oral Health Policy and Epidemiology, Harvard School of Dental Medicine, Boston, MA 02115, USA 3 Department of Community Health and Primary Care, College of Medicine, University of Lagos, Idi-Araba, Lagos State, Nigeria 4 Department of Epidemiology, Harvard School of Public Health, Boston, MA 02115, USA 5 TIMI Study Group, Brigham and Women’s Hospital, Boston, MA 02115, USA
Background: Preventing tobacco use is a key aspect of health promotion during adolescence. We assessed prevalence and impact of school-based tobacco prevention programs in 43 countries. Methods: We performed a secondary analysis of national data of students aged 13–15 years (Global Youth Tobacco Surveys) from 43 countries during 2005–2011. National surveys of the corresponding school personnel (Global School Personnel Surveys) were performed in each country during the same year as the student surveys. Data on status of enforcement of national smoke-free school policies were obtained from the 2008 and 2009 WHO MPOWER reports. Logistic regression was used to measure ecologic-level associations between school-based tobacco prevention programs and tobacco-related knowledge and behaviour among students (P < 0.05). Results: The proportion of students who were taught in class about the dangers of tobacco use during the school year ranged from 31.4% (Georgia) to 83.4% (Papua New Guinea). For every 10% increase (country level) in the proportion of teachers who reported having a tobacco prevention curriculum in their school, the odds of students reporting exposure to education in class about the dangers of tobacco increased by 6.0% (AOR = 1.06; 95% CI: 1.04–1.08). However, didactic education in class about the dangers of tobacco use was not independently associated with student current cigarette smoking behavior. Conversely, the likelihood of being a current smoker was significantly lower among students in countries with moderate/strongly enforced national smoke-free school policies compared with those in countries with poorly enforced/no national smoke-free school policies (AOR = 0.59; 95% CI: 0.45–0.76). Conclusions: Comprehensive tobacco prevention programs that include well-enforced smoke-free school policies may help reduce youth smoking.
.........................................................................................................
Introduction reventing tobacco use among adolescents is fundamental to
Pcombating the global tobacco epidemic as tobacco initiation
primarily occurs during adolescence. Approximately 70% of adult smokers in the European Union, and 88% of those in the United States started cigarette smoking by age 18 years.1,2 Research however shows that population-level interventions primarily targeting adults may not necessarily lead to changes in adolescent tobacco use.3 Hence, interventions targeted at adolescents and young adults should be an integral part of any comprehensive tobacco prevention and control program.2 The school—an environment where adolescents congregate for a large part of the day—affords several opportunities to invest in tobacco prevention and cessation programs. Educators such as school teachers are authority figures who may be perceived by youths as being trusted sources of information on the health consequences of tobacco use. In line with this, Article 12 of the WHO Framework Convention on Tobacco Control (FCTC) calls for training of educators in tobacco prevention strategies in order to equip students with knowledge about the dangers of tobacco use, how the tobacco industry targets youths, as well as how to resist pressure to use tobacco.4,5 Combining tobacco prevention education with strong enforcement of smokefree school policies may help denormalize tobacco use, and
protect non-smokers from involuntary exposure to secondhand smoke.6 Very little cross-country data exists on the implementation of school-based tobacco prevention programs. Hence, the objectives of this study were to: (i) assess teacher-reported exposure to training and access to educational materials on tobacco prevention among youths (ii) explore ecologic-level associations between teacher-reported presence of classroom and non-classroom-based tobacco prevention programs in school and student-reported receipt of education about the dangers of tobacco use and (iii) assess the association between presence of well enforced smokefree school policies and student-reported smoking behaviour. Data were obtained from the Global Youth Tobacco Surveys (adolescents students aged 13–15 years) and the Global School Personnel Surveys (school personnel) from 43 countries during 2005–2011.
Methodology Data sources The current analysis was deemed IRB exempt as a secondary analysis of publicly available de-identified data. National samples of adolescent students aged 13–15 years from 43 countries were obtained from the 2005–2011 Global Youth Tobacco Surveys (GYTS).7 In addition, within each country, a nationally sample of
Downloaded from http://eurpub.oxfordjournals.org/ by guest on November 16, 2015
Correspondence: Israel Agaku, Department of Social and Behavioral Sciences, Harvard School of Public Health, Center for Global Tobacco Control, Boston, MA 02115, USA, Tel: +1 (770) 728-3220, e-mail: [email protected]
Tobacco-free schools as a core component of youth tobacco prevention programs
school personnel–the Global School Personnel Survey (GSPS)—was obtained, sampled from the schools selected for the GYTS.8 The 43 countries in the study were from all six WHO regions, and were selected based on two criteria: (i) data were available for both the GSPS and GYTS and (ii) both surveys were conducted within the same year/wave. Countries that had data for one survey but not the other, or those with surveys in discordant waves were excluded. For our study, the pooled sample size for school personnel within the GSPS was n = 43 186, while the pooled sample size for students within the GYTS was n = 114 889 (table 1).
211
We classified the 43 countries into four categories, based on their 2012 gross national incomes (GNI) as reported by the World Bank.9 These were; low-income (n = 9 countries: Eritrea, Malawi, Niger, Rwanda, Togo, Uganda, Kyrgyzstan, Bangladesh, and Myanmar); lower-middle-income (n = 15 countries: Republic of Congo, Ghana, Mauritania, Senegal, Swaziland, Morocco, Sudan, Syria, Yemen, Georgia, Moldova, India, Sri Lanka, Mongolia and Papua New Guinea); upper-middle-income (n = 11 countries: Mauritius, Seychelles, South Africa, Islamic Republic of Iran, Lebanon, Bulgaria, Serbia, Panama, Maldives, Tuvalu and Tunisia); and
Table 1 Characteristics of study participants in the Global School Personnel Surveys (school teachers) and Global Youth Tobacco Surveys (students aged 13–15 years) in 43 countries, 2005–2011 Region/country
School teachers (Global School Personnel Survey) Survey year
Percentage of female students (%)
% of students who reported being taught in class about the dangers of tobacco use during the school year (95% CI)
1141 4813 4171 1293 1769 1286 1307 688 1611 844 3947 1239 1947 2251
48.9 40.0 50.6 49.7 51.6 53.9 51.8 52.8 52.8 50.9 57.7 59.8 40.4 50.6
45.5 44.7 61.8 77.6 54.3 76.4 33.1 51.3 62.4 75.8 61.7 71.2 46.5 73.9
(35.0–56.0) (41.6–47.8) (58.1–65.5) (71.0-84.2) (50.0–58.5) (70.9–81.8) (27.5–38.8) (43.8–58.8) (56.8–67.9) (71.7–80.0) (54.9–68.5) (65.5–76.9) (36.3–56.6) (70.5–77.2)
46.8 43.1 61.0 35.8 58.5 50.7 42.3 73.9 56.3 58.2 88.1
1153 2213 3314 2106 905 1797 950 1210 1294 10 821 650
47.9 55.9 52.4 47.0 59.1 50.3 52.8 50.5 52.5 52.7 37.5
35.3 71.7 59.0 44.7 77.5 52.7 44.2 64.7 70.4 53.3 47.1
(29.2–41.5) (68.1–75.4) (52.2–65.8) (39.4–50.0) (73.2–81.8) (49.1–56.4) (39.6–48.9) (55.9–73.5) (65.6–75.2) (51.1–55.5) (41.1–53.1)
2044 2224 423 797 283 155 1090 1599 1525
78.6 58.7 81.6 85.2 91.7 76.1 86.2 84.7 69.2
3241 12 477 3350 2330 891 3004 2476 3501 3350
48.3 51.2 50.9 50.7 52.8 52.6 56.6 55.6 54.6
82.2 57.2 82.7 72.6 31.4 71.8 77.3 82.0 70.3
(80.1–84.3) (55.3–59.1) (78.5–86.8) (67.3–77.8) (23.3–39.5) (67.0–76.6) (72.8–81.8) (78.7–85.2) (64.8–75.9)
2,357
53.2
2716
52.9
79.0 (76.6–81.5)
380 2463 789 2515 3518
28.6 45.9 65.4 88.9 75.6
2135 10 112 1494 1652 4031
59.2 41.7 47.6 49.3 52.6
64.4 71.9 47.7 73.4 78.0
1482 548 120
75.9 58.3 63.6
1499 1368 542
53.4 51.9 60.6
45.3 (35.0–55.5) 83.4 (80.4–86.5) 74.0 (70.3–77.8)
Total number of teachers
Percentage of female teachers (%)
344 1359 346 253 412 837 591 184 593 389 1408 447 461 472
23.2 36.1 36.1 41.6 53.8 56.2 23.1 21.8 19.4 56.3 64.1 59.4 12.8 29.0
327 1318 214 533 2052 477 233 935 65 4329 295
Total number of students sampled
(56.8–72.1) (69.5–74.3) (42.7–52.6) (68.4–78.4) (75.2–80.9)
Notes: Samples were not weighted while percentages were weighted to account for the complex survey design. CI = confidence interval.
Downloaded from http://eurpub.oxfordjournals.org/ by guest on November 16, 2015
African Region Republic of Congo 2009 Eritrea 2006 Ghana 2009 Malawi 2009 Mauritania 2009 Mauritius 2008 Niger 2009 Rwanda 2008 Senegal 2007 Seychelles 2007 South Africa 2011 Swaziland 2009 Togo 2007 Uganda 2007 Eastern Mediterranean Region Islamic Republic of Iran 2007 Kuwait 2009 Lebanon 2005 Morocco 2010 Oman 2010 Saudi Arabia 2010 Sudan 2009 Syria 2010 Tunisia 2010 United Arab Emirates 2005 Yemen 2008 European Region Bulgaria 2008 Cyprus 2005 Czech Republic 2011 Estonia 2007 Georgia 2008 Kyrgyzstan 2008 Latvia 2007 Republic of Moldova 2008 Serbia 2008 Region of the Americas Panama 2008 South-East Asian Region Bangladesh 2007 India 2009 Maldives 2011 Myanmar 2011 Sri Lanka 2011 Western Pacific Region Mongolia 2007 Papua New Guinea 2007 Tuvalu 2005/2006
Students aged 13–15 years (Global Youth Tobacco Survey)
212
European Journal of Public Health
Survey measures School personnel-reported exposure to tobacco prevention programs (GSPS) Within the GSPS, school personnel were asked to respond ‘Yes’ or ‘No’ to the following questions: ‘Have you ever received training to prevent tobacco use among youths?’; ‘Do you have access to teaching and learning materials about tobacco use and how to prevent its use among youth?’; ‘Is tobacco use prevention included somewhere in your school curriculum?’; and ‘Are non-classroom programs or activities (such as an assembly) used to teach tobacco use prevention to students in your school?’
Student-reported exposure to education in class about the dangers of tobacco use (GYTS) Within the GYTS, students’ receipt of education in class about the harmfulness of tobacco use was defined as a ‘Yes’ response to the question: ‘During the school year, were you taught in any of your classes about the dangers of smoking?’ Data were also collected on students’ age, gender and current cigarette smoking status, with the latter defined as having smoked cigarettes on at least one day within the past 30 days.
Analyses Prevalence estimates were calculated as percentages and compared using the chi-squared statistic (P < 0.05). Analyses within the GYTS were restricted to students aged 13–15 years to ensure a comparable cross-country denominator regardless of any heterogeneity between countries in school systems. Analyses within the GSPS were restricted to school personnel who reported their primary role as being a ‘teacher’ (vs. ‘school health services personnel’; ‘Administrator’; or ‘clerical staff’) since an outcome measure assessed in our study in-class student education about the dangers of tobacco use. A binary logistic regression model was fitted to measure the ecologic-level association between teacher-reported presence of tobacco-prevention curricula or non-classroom educational programs and students’ exposure to education about the harmfulness of tobacco use. The dependent variable was a report by a student that they had been taught in any of their classes about the harmfulness of tobacco use during the school year. For our independent variables, we computed as ecological covariates the following: the proportion of school teachers within each country
who reported that tobacco use prevention was included in their school curriculum, the proportion of school teachers who reported having access to teaching and learning materials about tobacco use and prevention, and finally, the proportion of school teachers who reported that non-classroom programs or activities existed in their school to teach tobacco use prevention to students. Adjustments were made for survey year, students’ sex, WHO region and GNI. A second binary logistic regression model was fitted to assess the association between students’ current cigarette smoking behaviour and the degree of enforcement of national smoke-free school policies, adjusting for exposure to in-class education about the dangers of tobacco use, survey year, students’ sex, WHO region and GNI. All data were weighted and analysed with Stata 11 (StataCorp 2009, TX).
Results Teacher-reported tobacco prevention programs and activities in schools The proportion of school teachers that reported ever receiving training in tobacco prevention among youths ranged from 4.5% (Bangladesh) to 53.7% (Kyrgyzstan). The proportion of school teachers that reported having access to educational materials about tobacco prevention among youths ranged from 4.2% (Mauritania) to 88.0% (Republic of Congo). The proportion of school teachers that reported that tobacco use prevention material was included in their school curriculum ranged from 20.3% (the Islamic Republic of Iran) to 95.7% (Kyrgyzstan). The proportion of school teachers that reported having non-classroom programs for tobacco prevention among youths in their schools ranged from 8.7% (Mauritania) to 97.8% (Republic of Moldova) (table 2). Detailed country-specific estimates are shown in table 2.
Relationship between school-based tobacco prevention programs and student-reported exposure to education about the dangers of tobacco use The proportion of adolescent students aged 13–15 years who reported being taught in class about the dangers of tobacco use during the school year ranged from 31.4% (Georgia), to 83.4% (Papua New Guinea) (table 1). Tobacco prevention curricula and non-classroom educational programs were both significantly associated with increased odds of student-reported exposure to education about the dangers of tobacco use. Within our multivariate analyses, for every 10% increase (country level) in the proportion of teachers that reported having a tobacco prevention curriculum in their school, the odds of students reporting that they were taught about the dangers of tobacco use during the school year increased by 6.0% (AOR = 1.06; 95% CI: 1.04–1.08). Similarly, for every 10% increase in the proportion of school teachers who reported the presence of a non-classroom tobacco prevention program in their school, the odds of students reporting they were taught about the dangers of tobacco use increased by 2.0% (AOR = 1.02; 95% CI: 1.01–1.04).
Relationship between enforcement of smoke-free school policies and student smoking behaviour Within our multivariate analyses, the odds of current smoking behaviour among adolescent students were significantly lower in countries with moderate/strongly enforced national smoke-free school policies than those with poorly enforced policies/no smokefree school policies (AOR = 0.59; 95%CI: 0.45–0.76). Conversely, didactic education in class about the harmfulness of tobacco use during the school year was not independently associated with current cigarette smoking among adolescent students (table 4).
Downloaded from http://eurpub.oxfordjournals.org/ by guest on November 16, 2015
high-income countries (n = 8 countries: Kuwait, Oman, Saudi Arabia, United Arab Emirates, Republic of Cyprus, Czech Republic, Estonia and Latvia). Data on the status of smoke-free policies in ‘educational institutions’ for each of the 43 countries were obtained from the 2008 and 2009 WHO MPOWER reports.10,11 The 2008/2009 reports were used rather than more recent MPOWER reports in order to ensure that the reported status of smoke-free school policies accurately reflected the situation as of the time of data collection in the GYTS-GSPS surveys during 2005–2011. Within each country, data were obtained from the MPOWER reports regarding both the presence of national smoke-free school policies (measured as ‘Yes’ or ‘No’), as well as the overall degree of enforcement of smoke-free policies (measured on an increasing scale of 1–10). Using these data, we categorized all countries into two groups: (i) those with reported moderate/ strong enforcement of smoke-free policies (scores 5–10) and (ii) those with either no national smoke-free school policies or poorly enforced policies (scores 1–4). In total, 9 of 43 countries assessed were reported to have moderate/strongly enforced smoke-free policies, namely: Eritrea, Niger, Seychelles, South Africa, Uganda, Islamic Republic of Iran, Estonia, Panama and Sri Lanka.
Tobacco-free schools as a core component of youth tobacco prevention programs
213
Table 2 Teacher-reported exposure to tobacco prevention training and presence of school-based tobacco prevention programs, Global School Personnel Survey in 43 countries, 2005–2011 Region/Country
% that reported having access to educational materials on tobacco use and prevention (95% CI)
% that reported presence of non-classroom tobacco prevention programs (95% CI)
% that reported that tobacco prevention was included in school curriculum (95% CI)
14.6 22.4 14.3 7.8 22.0 6.1 32.9 12.0 9.7 15.9 13.1 8.6 6.3 16.2
(5.9–23.4) (15.9–28.9) (8.9–19.7) (4.1–11.5) (16.8–27.3) (4.0–8.3) (23.5–42.2) (6.6–17.4) (3.8–15.5) (12.8–18.9) (10.2–16.1) (4.4–12.7) (3.1–9.4) (10.8–21.5)
88.0 33.6 24.1 42.6 4.2 38.1 14.4 86.7 81.6 38.9 49.0 34.5 9.2 30.4
(82.7–93.3) (27.3–40.0) (16.6–31.6) (30.9–54.3) (2.4–5.9) (30.5–45.8) (10.0–18.9) (81.6–91.9) (75.1–88.1) (32.9–45.0) (44.5–53.5) (27.1–41.9) (5.2–13.2) (24.3–36.4)
10.8 31.0 32.4 51.4 8.7 54.1 30.8 22.5 26.3 44.6 46.6 53.4 52.4 58.2
(6.1–15.4) (24.0–38.0) (21.9–42.9) (31.6–71.3) (6.2–11.2) (41.8–66.3) (22.2–39.5) (10.9–34.1) (17.2–35.5) (30.9–58.4) (30.9–62.3) (45.7–61.2) (44.2–60.6) (52.4–63.9)
59.4 55.2 56.3 65.2 26.7 56.1 55.2 76.1 47.5 71.3 69.0 68.6 71.9 52.4
(51.7–67.1) (46.9–63.4) (47.6–65.0) (55.0–75.4) (20.3–33.1) (47.9–64.2) (41.6–68.8) (66.9–85.2) (38.4–56.7) (62.4–80.2) (63.8–74.1) (60.3–76.9) (65.4–78.3) (44.4–60.3)
12.1 15.1 21.5 5.9 8.5 18.9 5.9 14.6 6.2 9.7 7.2
(7.3–16.9) (n/a) (15.3–27.8) (2.1–9.7) (5.9–11.2) (9.6–28.2) (1.9–10.0) (11.3–17.8) (n/a) (8.4–11.0) (2.9–11.4)
26.0 44.4 44.6 30.9 58.8 56.9 48.0 47.2 47.7 35.3 19.6
(16.9–35.2) (n/a) (35.0–54.1) (21.6–40.2) (55.1–62.5) (50.1–63.8) (38.0–58.0) (42.1–52.3) (n/a) (32.8–37.8) (11.3–27.8)
14.2 29.3 26.2 22.0 51.0 41.4 26.5 27.1 13.9 43.6 28.7
(7.8–20.5) (n/a) (18.9–33.5) (16.0–27.9) (46.8–55.1) (34.5–48.3) (17.0–35.9) (21.6–32.5) (n/a) (38.9–48.2) (17.2–40.3)
20.3 64.8 –a 61.6 84.0 73.6 86.8 74.6 79.7 –a 76.5
(8.9–31.6) (n/a)
12.3 12.3 21.2 25.0 34.9 53.7 31.4 18.9 14.6
(10.2–14.4) (10.4–14.2)
64.3 26.6 57.2 78.1 56.5 43.2 81.6 63.6 65.1
(60.2–68.3) (24.6–28.5)
58.9 50.4 43.4 66.9 49.0 55.6 62.7 97.8 74.6
(53.3–64.6) (46.4–54.3)
73.8 48.0 –a 94.1 56.8 95.7 86.6 80.0 83.1
(69.5–78.1) (44.3–51.6)
(19.3–30.7) (23.0–46.7) (38.7–68.7) (25.3–37.4) (14.5–23.3) (9.4–19.8)
(73.5–82.7) (43.6–69.4) (31.2–55.2) (77.5–85.8) (58.7–68.5) (59.1–71.0)
(57.4–76.5) (38.8–59.1) (39.5–71.7) (54.6–70.8) (96.6–99.0) (67.0–82.3)
(48.3–74.9) (80.7–87.3) (69.1–78.2) (76.4–97.2) (70.6–78.6) (n/a) (66.0–87.0)
(91.5–96.6) (40.2–73.5) (90.9–100) (81.6–91.7) (73.9–86.2) (75.7–90.5)
24.0 (20.7–27.3)
36.9 (32.4–41.4)
25.7 (19.6–31.9)
53.3 (48.4–58.2)
4.5 10.1 14.6 26.3 10.9
–a 37.8 28.8 76.4 60.1
45.0 49.1 51.5 74.0 35.4
76.3 46.4 66.5 87.6 73.3
(1.9–7.2) (7.0–13.2) (n/a) (21.0–31.6) (9.4–12.5)
25.7 (21.3–30.2) 12.7 (7.6–17.9) 31.7 (15.1–48.4)
(32.0–43.6) (n/a) (70.8–82.1) (56.6–63.6)
46.1 (38.3–53.9) 42.2 (36.6–47.8) 51.8 (34.5–69.0)
(33.9–56.0) (43.7–54.4) (n/a) (66.2–81.7) (31.1–39.6)
37.2 (28.3–46.1) 49.4 (41.4–57.3) 10.5 (2.2–18.8)
(68.3–84.2) (42.1–50.6) (n/a) (83.7–91.6) (70.1–76.6)
59.4 (50.1–68.8) 68.9 (59.5–78.3) 73.5 (57.0–90.0)
Note: CI = Confidence Intervals; n/a = not applicable (both upper and lower confidence limits identical to point estimate since a complete census of school personnel was taken, yielding final sampling survey weight = 1 for all respondents). a Data not available because question was not asked in survey.
Discussion This study showed that students from countries with well-enforced national smoke-free school policies were less likely to be current cigarette smokers compared with those from countries with no national smoke-free school policies or with poorly enforced smoke-free school policies. This highlights the need for enhanced and sustained efforts to implement and enforce smoke-free school policies as a vital component of tobacco prevention and control programs to reduce tobacco use among youths. Smoke-free school policies can be implemented at the national, state, local or school
level and can change social norms and attitudes about tobacco use among youths, and also protect nonsmokers from involuntary secondhand smoke exposure.12 However, enforcement of such policies is necessary to ensure compliance by students, school personnel and visitors.11–13 We found that exposure to didactic education in class about the harmfulness of tobacco use was not independently associated with reduced likelihood of smoking among students. Indeed, schoolbased tobacco educational programs have yielded inconsistent results in previous research with regards their effectiveness in eliciting long-term behavioural change among tobacco users.14–18
Downloaded from http://eurpub.oxfordjournals.org/ by guest on November 16, 2015
African region Republic of Congo Eritrea Ghana Malawi Mauritania Mauritius Niger Rwanda Senegal Seychelles South Africa Swaziland Togo Uganda Eastern Mediterranean Region Islamic Republic of Iran Kuwait Lebanon Morocco Oman Saudi Arabia Sudan Syria Tunisia United Arab Emirates Yemen European region Bulgaria Cyprus Czech Republic Estonia Georgia Kyrgyzstan Latvia Rep. of Moldova Serbia Region of the Americas Panama South-East Asia Bangladesh India Maldives Myanmar Sri Lanka Western Pacific Region Mongolia Papua New Guinea Tuvalu
% that reported being trained in tobacco use prevention among youths (95% CI)
214
European Journal of Public Health
Table 3 The impact of classroom and non-classroom tobacco prevention programs on exposure to didactic education about the dangers of tobacco use among students aged 13–15 years, GYTS-GSPS, 2005–2011 Characteristic
Category
Adjusted odds ratio (95% CI)
P-value
Teacher-reported inclusion of tobacco prevention material in curriculum Teacher-reported access to educational materials on tobacco prevention Teacher-reported presence of non-classroom tobacco prevention activities Survey year WHO Region
(Per 10% increase in proportion of teachers) (Per 10% increase in proportion of teachers) (Per 10% increase in proportion of teachers) (Per unit increase) African region (referent) Eastern Mediterranean region European region Region of the Americas Southeast Asian region Western Pacific region Low (referent) Lower-Middle Upper-Middle High Male (referent) Female
1.06 1.01 1.02 1.02
(1.04–1.08) (0.99–1.03) (1.01–1.04) (0.95–1.10)
