• SUNDAY, NOVEMBER 18, 2018
  • ISSN 2399-1623
Chandrashekhar T Sreeramareddy, Joanne Hon, Anshad Mohamed Abdulla, and Sam Harper

Abstract

Background

“Hardcore smokers” (HCS) who do not want to quit make it more difficult for tobacco control efforts to further reduce smoking prevalence. We aimed to quantify the burden of HCS among daily smoking adult males and females in 27 countries.

Methods

We used Global Adult Tobacco Survey (GATS) data to estimate the prevalence of HCS ie, daily smokers who smoke within 30 minutes after waking up, smoke ≥10 cigarettes per day, have not made any quit attempts during previous 12 months or have no intention to quit at all during the coming 12 months. For each GATS country, we estimated sex-wise, weighted and age-adjusted prevalence of daily smoking and HCS.

Results

Overall weighted population prevalence (%) of HCS was highest in Greece (21.0), followed by Russia (13), Poland (9.4), Romania (9.0), and Ukraine (8.9) and lowest in Nigeria (0.4). Estimated number of HCS (in millions) was highest in China (35.8) followed by India (28.2), Russia (18.5), Indonesia (16.1) and lowest in Panama (0.03). The proportion (%) of daily smokers classified as HCS was highest in Greece (56.2%) followed by Russia (42.2%), Ukraine (37.2) and Poland (36.2) and lowest in Mexico (8.29). Overall, proportion of HCS was higher among males in all countries. However, in Greece, Russia, Romania, Ukraine and Poland both population prevalence of HCS among women and proportion of HCS among daily smoking women was higher than in other countries.

Conclusions

At the country-level, a higher daily smoking rates also suggest a higher proportion of HCS. Countries with greater burden of HCS pose greater challenges to tobacco control efforts specifically towards tobacco cessation interventions. Interventions to reduce tobacco use and smoking-related mortality may need to be altered in populations with high proportions of HCS.


Tobacco use, particularly cigarette smoking including environmental tobacco smoke is still a substantial contributor to the global disease burden ( 1 ). Smoking prevalence is decreasing worldwide, particularly in developed countries, although rate of decline has slowed down in recent years ( 2 ). The sustained presence of heavy smokers who are more addicted and less able to quit ( 3 ) has led to emergence of “hardening hypotheses” ( 4 ). Light smokers are more likely to quit than the heavy smokers, thus leading over time to a decrease in light smokers and a relative increase in heavy smokers. Over time these heavy smokers become much harder to reach, and it becomes more difficult for them to quit ( 5 ), leading to a “hardened” population of smokers who are more resistant to quit. However, it has also been argued that the process of ‘hardening’ may not occur due to dynamic nature of cohorts of smokers arising from changes in subpopulations of quitters, smokers who die and new smokers ( 6 ).

The ‘hardening hypothesis’ has been tested using data on smoking behaviours in United States ( 7 - 9 ), Australia ( 10 ) England ( 11 , 12 ), Norway ( 13 ) and Italy ( 14 ). The ‘hardening hypothesis’ has also been evaluated in several studies using varying definitions for ‘hardcore smoker’ (HCS). In general, HCS was defined based on the constructs of duration and intensity of smoking, time to smoke the first cigarette of the day, number of quit attempts made, intention to quit, and knowledge about harms of smoking ( 6 , 15 ). Additionally, empirical evidence for the “hardening hypothesis “ has been evaluated using repeated cross-sectional survey data ( 16 , 17 ) and national population monitoring data ( 16 ). Based on Global Adult Tobacco Survey (GATS), HCS was defined using different constructs and prevalence, and factors associated HCS were reported ( 18 , 19 ). A study from Poland assessed factors associated with HCS including the constructs listed above to define HCS ( 18 ) while another study from Bangladesh, India and Thailand tested the association of HCS with socio-demographic factors only ( 19 ).

Studying the population prevalence of HCS and the proportion of current daily smokers who are HCS helps design tobacco control policies and set up support services for smoking cessation targeted at the ‘hardened’ smokers ( 20 ). In this paper, we use a five-item construct for HCS based on definitions used in previous studies. We report sex-specific, country-wise proportions of HCS among the current daily smokers, population prevalence of HCS and estimated total number of HCS in 27 GATS countries.

METHODS

We used publicly available data of the Global Adult Tobacco Surveys (GATS) (http://nccd.cdc.gov/gtssdata/Ancillary/DataReports.aspx?CAID=2), a series of nationally representative, cross-sectional household surveys done as a part of worldwide Global Tobacco Surveillance System (GTSS) to monitor tobacco use among various population groups. The GATS uses a standardised questionnaire to assess tobacco use behaviours among civilian, non-institutionalised individuals aged 15 years and above ( 21 ). In each GATS country, the residents of all regions were eligible to be sampled by a stratified multi-stage probability sampling technique. Within a selected geographic location, households were selected at random and within a selected household, all eligible persons were interviewed using a handheld device used for rostering and data collection, and one household member was selected at random for the interview. The implementing agencies in each country adapted the core questionnaire to suit the local tobacco use context. In all countries interviews were done privately by either a male or female interviewers, except for India, Bangladesh, Indonesia and Qatar interviewers were of same gender as the respondent. Further details about survey instrument, methodology etc. are published elsewhere ( 21 ). The sample sizes, response rates by sex are given in Table 1 .

Table 1.  Survey characteristics, sex-wise and country-wise distribution of smoking characteristics and weighted prevalence of daily smoking in 27 GATS countries
Smoking status (number and %) Weighted prevalence and 95% CI of daily smoking
Country Survey year Sample size (% of total sample) Median age (IQR) Daily smoker Non-daily smoker Non-smoker
Men (≥15 years):
Argentina 2012 2985 (44.9) 56 (37, 70) 681 (22.8) 271 (9.1) 2033 (68.1) 21.89 (17.32-26.47)
Bangladesh 2009 4468 (46.4) 35 (25, 45) 1972 (44.1) 185 (4.1) 2311 (51.7) 40.69 (38.50-42.87)
Brazil 2008 18039 (45.8) 38 (27, 51) 3513 (19.5) 524 (2.9) 14002 (77.6) 18.91 (18.18-19.63)
Cameroon 2013 4049 (60.9) 31 (23, 42) 79 (2.0) 6 (0.1) 3969 (97.9) 9.11 (7.63-10.61)
China 2010 6603 (49.4) 46 (36, 59) 3303 (50.0) 469 (7.1) 2831 (42.9) 45.35 (42.73-47.97)
Egypt 2009 10062 (48.1) 34 (25, 47) 3904 (38.8) 189 (1.9) 5969 (59.3) 35.85 (34.52-37.17)
Greece 2013 2092 (48.0) 48 (35, 67) 1037 (49.6) 27 (1.3) 1028 (49.1) 49.7 (46.42-52.99)
India 2009 33767 (48.7) 35 (26, 45) 8164 (24.2) 2084 (6.2) 23519 (69.7) 18.35 (17.43-19.27)
Indonesia 2011 3948 (47.5) 38 (28, 52) 2335 (59.1) 385 (9.8) 1228 (31.1) 56.74 (53.82-59.65)
Kazakhstan 2014 2085 (47.1) 39 (28, 52) 795 (38.1) 118 (5.7) 1172 (56.2) 36.91 (34.20-39.63)
Kenya 2014 2077 (47.1) 35 (26, 46) 341 (16.4) 82 (3.9) 1654 (79.6) 11.60 (9.71-13.50)
Malaysia 2011 2104 (49.5) 41 (29, 55) 873 (41.5) 82 (3.9) 1149 (54.6) 39.87 (36.51-43.23)
Mexico 2009 6160 (45.2) 36 (27, 50) 654 (10.6) 744 (12.1) 4762 (77.3) 11.81 (10.61-13.02)
Nigeria 2012 5058 (51.8) 31 (24, 45) 323 (6.4) 91 (1.8) 4644 (91.8) 5.55 (4.60-6.51)
Pakistan 2014 3782 (48.2) 35 (24, 46) 811 (21.4) 67 (1.8) 2904 (76.8) 20.59 (18.88-22.30)
Panama 2013 7679 (45.3) 39 (27, 54) 298 (3.9) 477 (6.2) 6904 (89.9) 4.37 (2.98-5.76)
Philippines 2009 4740 (48.9) 37 (27, 50) 1887 (39.8) 420 (8.9) 2433 (51.3) 38.24 (36.35-40.13)
Poland 2009 3867 (49.3) 47 (32, 60) 1309 (33.9) 116 (3.0) 2442 (63.1) 33.52 (31.58-35.46)
Qatar 2013 4237 (50.5) 33 (26, 41) 799 (18.9) 178 (4.2) 3260 (76.9) 16.51 (14.95-18.07)
Romania 2011 2070 (45.8) 56 (40, 70) 657 (31.7) 50 (2.4) 1363 (65.8) 34.85 (32.27-37.43)
Russia 2009 6217 (54.5) 46 (32, 58) 3486 (56.1) 300 (4.8) 2431 (39.1) 54.98 (53.15-56.80)
Thailand 2011 8781 (42.6) 47 (35, 60) 3525 (40.1) 368 (4.2) 4888 (55.7) 42.04 (40.17-43.90)
Turkey 2012 4470 (45.4) 43 (30, 59) 1593 (35.6) 189 (4.2) 2688 (60.1) 37.33 935.38-39.27)
Uganda 2013 3853 (45.3) 31 (24, 42) 400 (10.4) 79 (2.1) 3374 (87.6) 8.69 (7.49-9.88)
Ukraine 2010 4076 (50.0) 51 (35, 66) 1866 (45.8) 168 (4.1) 2042 (50.1) 45.48 (43.55-47.420
Uruguay 2009 2634 (47.2) 47 (32, 65) 677 (25.7) 134 (5.1) 1823 (69.2) 24.85 (22.45-27.25)
Viet Nam 2010 4356 (43.9) 40 (29,53) 1794 (41.2) 377 (8.7) 2185 (50.2) 38.74 (36.88-40.61)
Women (≥15 years):
Argentine 2012 3660 (55.1) 37 (27, 54) 505 (13.8) 193 (5.3) 2962 (80.9) 12.74 (9.79-15.69)
Bangladesh 2009 5161 (53.6) 37 (27, 48) 66 (1.3) 10 (0.2) 5085 (98.5) 1.33 (0.87-1.79)
Brazil 2008 21386 (54.2) 38 (27,54) 2560 (12.0) 406 (1.9) 18420 (86.1) 11.52 (10.98-12.06)
Cameroon 2013 2594 (39.1) 30 (23, 44) 18 (0.7) 8 (0.3) 2568 (99.0) 0.46 (0.20-0.71)
China 2010 6751 (50.6) 47 (37, 60) 197 (2.9) 41 (0.6) 6513 (96.5) 2.0 (1.44-2.55)
Egypt 2009 10862 (51.9) 38 (28, 49) 53 (0.5) 9 (0.1) 10800 (99.4) 0.45 (0.24-0.67)
Greece 2013 2267 (52) 47 (34, 65) 573 (25.3) 30 (1.3) 1664 (73.4) 23.91 (21.05-26.76)
India 2009 35529 (51.3) 36 (27, 48) 1059 (3.0) 289 (0.8) 34181 (96.2) 2.44 (2.06-2.82)
Indonesia 2011 4357 (52.5) 38 (29, 50) 90 (2.1) 45 (1.0) 4222 (96.9) 1.82 (1.29-2.34)
Kazakhstan 2014 2340 (52.9) 41 (29, 56) 77 (3.3) 30 (1.3) 2233 (95.4) 3.17 (2.24-4.10)
Kenya 2014 2331 (52.9) 31 (24, 45) 20 (0.9) 6 (0.3) 2305 (98.9) 0.60 (0.14-1.06)
Malaysia 2011 2146 (50.5) 40 (28, 53) 25 (1.2) 9 (0.4) 2112 (98.4) 0.69 (0.32-1.06)
Mexico 2009 7457 (54.8) 37 (26, 52) 202 (2.7) 221 (3.0) 7034 (94.3) 3.67 (2.88-4.47)
Nigeria 2012 4707 (48.2) 34 (25, 45) 12 (0.3) 3 (0.1) 4692 (99.7) 0.28 (0.099-0.46)
Pakistan 2014 4049 (51.8) 32 (24, 44) 79 (2.0) 6 (0.1) 3964 (97.9) 1.99 (0.02-2.52)
Panama 2013 9283 (54.7) 42 (28, 58) 69 (0.7) 152 (1.6) 9062 (97.6) 1.24 (0.69-1.79)
Philippines 2009 4961 (51.1) 36 (27, 48) 360 (7.3) 102 (2.1) 4499 (90.7) 6.86 (5.94-7.78)
Poland 2009 3973 (50.7) 44 (31, 57) 870 (21.9) 121 (3.0) 2982 (75.1) 21.0 (19.54-22.47)
Qatar 2013 4161 (49.5) 38 (29) (45) 52 (1.2) 50 (1.2) 4059 (97.5) 1.66 (1.02-2.300
Romania 2011 2447 (54.2) 50 (36, 65) 301 (12.3) 48 (2.0) 2098 (85.7) 14.49 (12.70-16.29)
Russia 2009 5189 (45.5) 43 (29, 55) 786 (15.1) 235 (4.5) 4168 (80.3) 16.26 (14.38-18.14)
Thailand 2011 11825 (57.4) 46 (33, 58) 328 (2.8) 70 (0.6) 11427 (96.6) 2.11 (1.75-2.47)
Turkey 2012 5381 (54.6) 42 (31, 56) 520 (9.7) 110 (2.0) 4751 (88.3) 10.67 (9.59-11.75)
Uganda 2013 4655 (54.7) 30 (23, 43) 71 (1.5) 19 (0.4) 4565 (98.1) 1.31 (0.91-1.7)
Ukraine 2010 4082 (50.0) 46 (32, 59) 289 (7.1) 74 (1.8) 3719 (91.1) 8.88 (7.60-10.15)
Uruguay 2009 2947 (52.8) 45 (31, 61) 488 (16.6) 95 (3.2) 2364 (80.2) 16.42 (14.77-18.07)
Viet Nam 2010 5569 (56.1) 40 (28, 52) 73 (1.3) 16 (0.3) 5480 (98.4) 1.21 (0.75-1.68)

CI – confidence interval, IQR – interquartile range



Main outcome variable

Based on previous studies respondents were defined as HCS, if they satisfied the following five criteria ( 8 ): 1) is a current daily smoker; 2) smokes 10 or more cigarettes per day; 3) smokes their first cigarette within 30 minutes after waking up; 4) has not made any quit attempts during 12 months prior to the date surveyed; and 5) has no intention to quit smoking at all or during the next 12 months.

Statistical analyses

All the analyses were done on Stata/IC version 10 (Stata Corp LLC, College Station, TX, USA). Descriptive statistics were calculated for current daily smoking and five constructs of HCS including average of number of cigarettes smoked per day. The proportion of current daily smokers who were HCS was calculated for men, women and both sexes. To adjust for complex sampling design used in GATS, we used country sample weights to estimate prevalence rates and 95% confidence intervals (95% CI). We used total and sex-wise population aged 15 years above from United Nations census data to estimate the number (in millions) male and female HCS in each GATS country. As a sensitivity analysis, we also estimated age-adjusted rates of current daily smoking and HCS, adjusted to the world population standard ( 22 ).

RESULTS

Table 1 shows the year of survey followed by sex-specific sample size, median age of respondents and smoking status for each country. A total of 344,329 adults (164259 men and 180070 women) were surveyed, median age of the respondents varied between 31-55 years for men and 30-61 years for women. The weighted prevalence of current daily smoking among men was highest in Indonesia (57%) and Russia (55%), and lowest in Panama (4%) and Nigeria (6%). Among women current daily smoking prevalence was considerably lower: <1% in Nigeria, Egypt, Cameroon, Kenya and Malaysia and less than 5% in 12 countries; over 15% in Russia, Uruguay Poland and Greece ( Table 1 ).

In all countries, the weighted population prevalence HCS was considerably higher among men than women, and in nine countries HCS prevalence was higher than 10% ( Table 2 ). Among men, Greece (30.9%) had the highest prevalence, followed by Russia (23.8%) while in Argentina, Qatar, Panama, Cameroon, Kenya, Uganda, Mexico and Nigeria HCS prevalence was much lower (2.71-0.81%). However, in 25 countries the weighted population prevalence of HCS among women was less than 5.0% and in 17 of these 25 countries HCS prevalence was <1.0%; Poland (6.8%), and Greece (11.4%) had higher prevalence for women. ( Figure 1 shows weighted prevalence of current daily smoking and HCS in each country).

Table 2.  Weighted prevalence rates of HCS and distribution (number and %) of the HCS constructs among the daily smokers in 27 GATS countries
Weighted prevalence of HCS (95% CI) Mean sticks per day ≥10 sticks per day No quit attempt in last 12 months No intention to quit at all or next 12 months Smoke within 30 minutes after waking up
Men (≥15 years)
Argentina 2.31 (1.54-3.08) 15.26 499 (73.3) 554 (58.2) 721 (75.7) 219 (32.2)
Bangladesh 7.46 (6.21-8.71) 12.33 1220 (61.9) 1126 (52.2) 1358 (63) 881 (44.7)
Brazil 4.67 (4.29-5.05) 14.96 2316 (68.1) 2546 (60.8) 1977 (80.5) 2195 (62.5)
Cameroon 1.55 (0.86, 2.23) 8.76 95 9 (33.9) 218 (58.0) 258 (68.6) 83 (28.9)
China 5.12 (4.40-5.85) 17.24 2694 (81.6) 1044 (69.6) 3172 (84.1) 1840 (55.7)
Egypt 6.4 (5.72-7.08) 16.74 3114 (79.8) 2541 (62.1) 3004 (73.4) 1402 (35.9)
Greece 30.91 (27.78-34.04) 20 866 (83.5) 906 (85.2) 927 (87.1) 795 (76.7)
India 4.46 (4.03-4.90) 10.62 3721 (45.6) 7159 (69.9) 8004 (78.1) 5370 (65.8)
Indonesia 13.09 (10.95-15.24) 12.38 1545 (66.2) 1948 (71.6) 2427 (89.2) 906 (38.8)
Kazakhstan 10.52 (8.83, 12.21) 16.73 552 (77.0) 675 (73.9) 775 (84.9) 422 (53.1)
Kenya 1.97 (1.07, 2.86) 9.36 142 (42.8) 193 (45.6) 241 (57.0) 256 (75.1)
Malaysia 8.15 (6.46-9.85) 14.38 654 (74.9) 564 (59.1) 842 (88.2) 417 (47.8)
Mexico 1.15 (0.78-1.52) 10.02 268 (41.0) 789 (56.4) 922 (66.0) 180 (27.5)
Nigeria 0.81 (0.38-1.23) 8.9 105 (32.5) 235 (56.8) 284 (68.6) 168 (52.0)
Pakistan 5.54 (4.60, 6.48) 13.69 521 (65.4) 668 (76.1) 740 (84.3) 399 (49.2)
Panama 1.53 (0.83-2.23) 12.77 118 (39.6) 393 (50.7) 600 (77.4) 130 (43.6)
Philippines 10.49 (9.26-11.73) 11.48 1176 (62.3) 1259 (54.6) 1831 (79.4) 1124 (59.6)
Poland 12.27 (11.04-13.51) 17.84 1141 (87.2) 1006 (70.6) 992 (69.6) 817 (62.4)
Qatar 2.71 (2.08-3.33) 14.64 554 (69.3) 635 (65.0) 601 (61.5) 298 (37.3)
Romania 13.74 (11.87-15.61) 17.04 567 (86.3) 466 (65.9) 547 (77.4) 471 (71.7)
Russia 23.8 (22.01-25.58) 17.77 2892 (83.0) 2727 (72.0) 3334 (88.1) 2357 (67.6)
Thailand 12.19 (10.98-13.39) 12.24 2267 (64.3) 2535 (65.1) 3303 (84.8) 2093 (59.4)
Turkey 8.05 (7.00-9.11) 20.36 1398 (87.8) 1034 (58.0) 1143 (64.1) 666 (41.8)
Uganda 0.98 (0.59, 1.38) 6.95 287 (59.9) 192 (48.0) 346 (72.2) 278 (69.5)
Ukraine 16.65 (15.25-18.06) 17.25 1622 (86.9) 1343 (66.0) 1552 (76.3) 1231 (66.0)
Uruguay 4.37 (3.24-5.50) 17.11 500 (73.9) 465 (57.3) 550 (67.7) 285 (42.1)
Viet Nam 9.88 (8.84-10.92) 14.92 1268 (70.8) 1037 (47.8) 1550 (71.4) 1167 (65.1)
Women (≥15 years):
Argentina 1.59 (0.78-2.39) 11.39 296 (58.6) 378 (54.2) 515 (73.8) 145 (28.7)
Bangladesh 3.82 (3.16-4.49) 8.57 21 (31.8) 49 (64.5) 60 (78.9) 33 (50.0)
Brazil 2.33 (2.08-2.59) 12.39 1474 (59.6) 1619 (54.6) 1272 (78.6) 1399 (54.6)
Cameroon 0.09 (-0.001, 0.19) 5.35 4 (23.5) 9 (34.6) 19 (73.1) 9 (50.0)
China 0.12 (0.04-0.19) 12.64 135 (68.5) 49 (59.8) 196 (82.4) 105 (53.3)
Egypt 0.05 (0.0027-0.09) 7.98 17 (32.1) 34 (54.8) 46 (74.2) 22 (41.5)
Greece 11.39 (9.57-13.21) 15 471 (82.2) 501 (83.1) 516 (85.6) 367 (64.0)
India 0.31 (0.21-0.40) 7.27 275 (26.0) 939 (69.7) 1086 (80.6) 625 (59.0)
Indonesia 0.22 (0.0591-0.37) 8.01 29 (32.2) 86 (63.7) 121 (89.6) 25 (27.8)
Kazakhstan 0.38 (0.06, 0.7) 10.49 30 (49.2) 75 (70.1) 85 (79.4) 34 (44.2)
Kenya 0.013 (-0.01, 0.04) 5.8 2 (25.0) 9 (34.6) 10 (38.5) 14 (70.0)
Malaysia 0.03 (0.012-0.07) 9.84 15 (60) 19 (55.9) 28 (82.4) 7 (28)
Mexico 0.22 (0.051-0.38) 8.35 56 (27.7) 192 (45.4) 252 (59.6) 62 (30.7)
Nigeria 0.042 (0.017-0.10) 18.83 5 (41.7) 8 (53.3) 12 (80.0) 7 (58.3)
Pakistan 0.2 (0.02, 2.5) 7.45 23 (30.3) 63 (74.1) 74 (87.1) 23 (29.1)
Panama 0.2 (0.047-0.36) 7.8 19 (27.5) 112 (50.7) 183 (82.8) 25 (36.2)
Philippines 0.74 (0.44-1.03) 6.88 116 (32.2) 222 (48.1) 349 (75.5) 163 (45.3)
Poland 6.78 (5.91-7.66) 15.49 742 (85.3) 676 (68.2) 654 (66.0) 502 (57.7)
Qatar 0.64 (0.20-1.08) 11.65 26 (50.0) 60 (58.8) 69 (67.6) 26 (50.0)
Romania 4.6 (3.47-5.72) 13.52 226 (75.1) 215 (61.6) 263 (75.4) 188 (62.5)
Russia 3.99 (3.14-4.85) 11.43 532 (67.7) 677 (66.3) 855 (83.7) 369 (46.9)
Thailand 0.45 (0.30-0.60) 8.61 140 (42.7) 253 (63.6) 320 (80.4) 163 (49.7)
Turkey 1.98 (1.53-2.42) 15.12 361 (69.4) 350 (55.6) 382 (60.6) 200 (38.5)
Uganda 0.06 (-0.013, 0.14) 3.47 4 (5.6) 52 (57.8) 71 (78.9) 46 (64.8)
Ukraine 2.39 (1.69-3.08) 11.3 189 (65.4) 222 (61.2) 260 (71.6) 150 (51.9)
Uruguay 2.3 (1.53-3.06) 12.96 312 (64.2) 317 (54.4) 384 (65.9) 158 (32.4)
Viet Nam 0.31 (0.14-0.49) 10.25 30 (41.1) 55 (61.8) 67 (75.3) 40 (54.8)

HCS – hardcore smoking, CI – confidence interval



Figure 1.  Proportions of hardcore smokers (HCS) among daily smoking men and women in 27 GATS countries.
joghr-02-e2018019-f1


Overall (both sexes), proportion of daily smokers who were defined as HCS was higher in Greece, Cameroon, Russia, Ukraine, Romania and Poland (36.2-56.2%), in other countries the proportion was between 30% and 10%, except Mexico (8.3%). The proportion of male HCS was over 55% in Greece, and in Russia, Ukraine, Poland and Romania it was nearly 40%, whereas in most other countries it ranged between 30-15%. In all countries except Qatar, Bangladesh and Nigeria, the proportion of HCS was higher among males where prevalence of current daily smoking was also low. In both sexes, proportion of daily smokers who were defined as HCS tended to be higher among countries with higher prevalence of daily smoking ( Figure 2 ).

Figure 2.  Scatter diagram for relationship of percentage of hardcore smokers (HCS) and prevalence of daily smoker among men and women in 27 GATS countries. Comparison of prevalence of daily smoker with percentage of HCS by each country can also be found in Table S1 in Online Supplementary Document(Online Supplementary Document) .
joghr-02-e2018019-f2


Overall in 27 GATS countries there were an estimated 111 million HCS (101.6 million men and 12.4 million women, Table 3 ). Among the GATS countries, four countries together accounted for nearly 70% of all HCS: China (30 million), India (21 million), Russia (16 million) and Indonesia (11 million). China and India had largest burden of HCS in terms of absolute number of HCS due to their large population sizes. India and China had relatively lower prevalence of current daily smoking than Greece and eastern European countries of GATS which had very high current daily smoking prevalence. Age-standardization led to similar estimated prevalence for both daily smoking and HCS, compared to crude estimates (see Table S1 in Online Supplementary Document(Online Supplementary Document) ).

Table 3.  Estimated number of HCS and their 95CIs in 27 GATS countries by both sexes and sex-wise
Country Total number HCS (both sexes) Number of male HCS Number of female HCS
Estimates 95% CIs Estimates 95% CIs Estimates 95% CIs
Argentina 600,767 426,451 775,083 347,841 231,894 463,788 255,510 125,344 384,068
Bangladesh 3,596,984 3,154,423 4,048,961 3,471,427 2,889,753 4,053,100 1,819,390 1,505,045 2,138,498
Brazil 3,388,870 3,012,329 3,750,929 3,270,889 3,004,735 3,537,043 1,742,445 1,555,488 1,936,881
Cameroon 102,667 59,033 145,017 99,426 55,165 143,045 5,777 -64 12,196
China 29,676,736 25,786,527 33,566,945 28,801,771 24,751,522 32,908,274 658,744 225,071 1,043,012
Egypt 1,630,698 1,481,549 1,779,847 1,620,481 1,448,305 1,792,657 11,222 666 21,713
Greece 1,980,614 1,821,070 2,140,158 1,416,617 1,273,168 1,560,066 553,263 464,858 641,669
India 20,541,066 18,696,562 22,469,411 19,129,782 17,285,431 21,017,025 1,269,425 859,933 1,637,968
Indonesia 11,224,127 10,243,707 12,221,452 11,039,063 9,234,358 12,852,202 186,353 50,061 313,412
Kazakhstan 393,981 328,093 461,213 351,459 291,825 411,093 14,205 1,420 177,561
Kenya 252,969 138,220 365,110 255,229 138,627 370,535 1,706 919 525
Malaysia 863,312 705,414 1,021,210 859,365 681,165 1,038,619 2,989 1,195 7,272
Mexico 526,817 383,140 678,476 441,100 299,181 583,019 91,221 21,147 157,564
Nigeria 351,291 171,561 522,851 331,155 155,357 502,865 17,141 6,938 40,812
Pakistan 3,831,873 3,191,048 4,485,777 3,720,310 3,089,066 4,351,554 17,039 1,704 212,993
Panama 20,982 11,817 29,906 18,435 10,001 26,869 2,462 567 4,345
Philippines 3,449,610 3,092,329 3,806,891 3,227,957 2,849,464 3,609,527 228,130 135,645 317,533
Poland 3,039,755 2,803,689 3,279,055 1,894,431 1,704,525 2,085,882 1,145,704 998,688 1,294,408
Qatar 29,551 23,401 35,358 35,948 27,591 44,173 2,442 763 4,122
Romania 1,523,880 1,341,014 1,708,438 1,119,847 967,437 1,272,257 403,960 304,726 502,315
Russia 15,649,954 14,684,652 16,603,190 13,031,482 12,051,383 14,006,105 2,629,798 2,069,565 3,196,621
Thailand 3,277,696 3,011,216 3,544,175 3,150,351 2,837,642 3,460,475 123,535 82,356 164,713
Turkey 2,798,306 2,477,374 3,113,608 2,258,669 1,964,060 2,556,084 559,270 432,163 680,727
Uganda 88,860 53,316 124,403 86,213 51,904 121,403 5,385 -1,167 12,564
Ukraine 3,485,824 3,195,011 3,776,637 2,959,780 2,710,910 3,210,428 514,390 363,732 662,895
Uruguay 84,635 68,895 100,376 53,530 39,688 67,372 31,174 20,738 41,475
Viet Nam 3,216,745 2,879,506 3,553,985 3,118,793 2,790,499 3,447,086 103,190 46,602 163,106
Total 115,628,570 103,241,349 128,108,462 10,6111,349 92,834,654 119,492,544 12,395,870 9,274,104 15,770,969

HCS – hardcore smoking, CI – confidence interval



DISCUSSION

The prevalence of HCS and the proportion of daily smokers defined as ‘hardcore’ varied widely between 27 GATS countries and sexes. In general, we found that proportion of daily smokers defined as “hardcore” was higher among countries with a higher prevalence of current daily smoking. In terms of absolute numbers China, India, Russia and Indonesia have very large numbers of HCS, whereas Russia has highest proportion of daily smokers defined as “hardcore”. In all GATS countries, women ranked lower than men in prevalence of HCS and number of HCS. Both prevalence of HCS and the proportion of HCS among current daily smokers were higher in high-income and upper-middle-income GATS countries, Qatar being an exception to this.

A strength of our analysis was a robust survey design of GATS, which generates nationally representative samples of men and women. The standardised methodology and survey instrument of GATS enabled us to construct a definition of HCS for cross-country comparison ( 23 ). Nevertheless, estimates reported in this paper should be interpreted with caution against the possible limitations inherent in GATS design. Self-reported smoking behaviour is known to lead to under-estimates due to smoking-related stigma ( 24 ). Furthermore, social desirability bias related to quit attempts and intentions to quit may have resulted in underestimation of HCS ( 25 ). Both duration of smoking as well as smoking intensity (ie, daily number of cigarettes smoked) indicate nicotine addiction but duration of smoking could not be computed in many low-income and lower-middle-income GATS countries because of missing data on age of smoking initiation. The surveys we analysed were implemented at different time periods, smoking patterns change over time and are sensitive to country-specific tobacco control policies ( 2 ); hence our results should be interpreted in the context of the local tobacco control environment in the GATS countries.

The prevalence of HCS among high-income countries has been well reported and ranged from 16% in the UK ( 11 ) to 6% in Norway ( 13 ). Our estimates varied widely and ranged from 21% in Greece to 0.4% in Nigeria. However, our estimates of HCS are not comparable to those from high-income countries due to heterogeneity in the constructs these studies used to define a HCS ( 15 ). Given the relationship, we identified between daily rates of smoking and HCS, our estimates may also differ from high-income countries because the prevalence of daily smoking differs as well. Two prior GATS-based studies used different definitions for HCS and reported a population prevalence of HCS of 3.1% (India), 3.8% (Bangladesh) and 6.0% (Thailand) ( 19 ) whereas 10.0% in Poland ( 18 ) which were nearly same as our estimates.

The hardening hypothesis has implications for tobacco control, since when the prevalence of smoking decreases it would make sense to target HCS to achieve further reductions ( 5 ). Studies on HCS have primarily been based on cross-sectional data ( 12 , 14 , 18 , 19 ). Additional studies that have examined sequential cross-sectional surveys have provided some evidence against the ‘hardening theory’ ( 13 , 17 , 26 ); however, one study reported that hardening may be occurring in the UK ( 5 ). Nevertheless, we are only able to provide cross-sectional estimates of HCS for 27 GATS countries. Some GATS countries have a huge burden of estimated population who are HCS, mainly among men in China, India, Russia and Indonesia. These estimate numbers have implications for tobacco control since smoking cessation can help to avoid a substantial tobacco-attributable mortality, particularly in low-and-middle-income countries (LMIC) currently experiencing epidemic levels of smoking ( 27 ). Moreover, these numbers have implications for healthcare services in LMICs which typically neither provided smoking cessation services ( 28 ) nor implemented tobacco-related training programs healthcare professionals ( 29 ) as per the recommendations of FCTC Article 14 or the FCTC Article 14 guidelines. LMICs with higher burden of HCS and low quit rates are likely to require larger healthcare costs to implement smoking cessation services.

Tobacco control strategies should consider smoking intensity, duration of smoking, time to smoke first cigarette, as these factors are known to be associated with tobacco dependence ( 30 ). More intensive and universal measures should be undertaken to achieve higher quit rates among all daily smokers ( 5 ). Reaching out to the hardcore smokers who lack motivation (don't want to quit) and have severe nicotine addiction (can’t quit) ( 17 ) requires a more holistic approach. Motivational interviewing maybe one of the approaches to help smokers decide to quit ( 31 ), as it is well known that smokers who are in contemplation stages are more likely to remain tobacco-free ( 32 ). HCS constructs in most studies including ours do not include socio-economic deprivation and mental health which are associated with “hardening” ( 14 ). As smoking is also known to be common among those with socio-economic deprivation ( 33 ) wider community level interventions, such as health promotion and socio-economic development may not only help cessation but could also potentially help to decrease smoking initiation ( 34 ).

In addition to conflicting reports on the ‘hardening hypotheses’, there is also a lack of consensus on uniform definitions of hardening across the published literature. Hence, there is a need for further studies defining HCS based on all the four domains of constructs described by Edwards et al ( 17 ). Current conflicting literature, mainly from developed countries, has tested the ‘hardening hypothesis’ with sequential cross-sectional surveys ( 10 , 11 , 13 , 16 , 17 , 26 ); cohort studies on daily smokers who cannot quit would confirm the hardening hypothesis and examine the changes in proportion of HCS over time.

CONCLUSIONS

HCS constitute a fifth to third of daily smokers in some GATS countries, which presents challenges to tobacco control efforts. As the proportion of HCS tends to increase with the prevalence of daily smoking, efforts should be made to counter smoking initiation. Further HCS also poses a challenge to health services in LMIC where smoking cessation services are sub optimal. Future studies should include all four domains of HCS to better understand about “hardening hypothesis” in LMIC.

Acknowledgements

The authors thank the Centre for Disease Control for making the datasets of Global Adult Tobacco Survey that enables us prepare this manuscript.

Ethics approval: All the Global Adult Tobacco Surveys were approved by ethical boards of surveys countries and CDC, Atlanta. The data used for this manuscript are available in the public domain are de-identified. Therefore, a separate ethical approval was not required for this manuscript preparation.

Notes

[1] Funding: This study did not receive any funding.

[2] Competing interests: The authors have completed the Unified Competing Interest form at www.icmje.org/coi_disclosure.pdf (available on request from the corresponding author) and declare no competing interests.

REFERENCES

1 E Gakidou, A Afshin, AA Abajobir, KH Abate, C Abbafati, and KM Abbas. Global, regional, and national comparative risk assessment of 84 behavioural, environmental and occupational, and metabolic risks or clusters of risks, 1990 - 2013;2016: a systematic analysis for the Global Burden of Disease Study 2016. Lancet. 2017;390:p.1345–422. DOI: 10.1016/S0140-6736(17)32366-8. [28919119]

2 GBD Tobacco Collaborators. . Tobacco in transition, 1990-2015: a systematic analysis of smoking prevalence and attributable burden from the Global Burden of Disease Study 2015. Lancet. 2017;389:p.1885–906. DOI: 10.1016/S0140-6736(17)30819-X

3 JR DiFranza and MP Guerrera. . Hard-core smokers. JAMA. 1989;261:p.2634–5. DOI: 10.1001/jama.1989.03420180058016. [2709539]

4 A Darville and EJ Hahn. . Hardcore smokers: what do we know? Addict Behav. 2014;39:p.1706–12. DOI: 10.1016/j.addbeh.2014.07.020. [25117846]

5 G Docherty and A McNeill. . The hardening hypothesis: does it matter? Tob Control. 2012;21:p.267–8. DOI: 10.1136/tobaccocontrol-2011-050382. [22345265]

6 JR Hughes. . The hardening hypothesis: is the ability to quit decreasing due to increasing nicotine dependence? A review and commentary. Drug Alcohol Depend. 2011;117:p.111–7. DOI: 10.1016/j.drugalcdep.2011.02.009. [21411244]

7 E Augustson and S Marcus. . Use of the current population survey to characterize subpopulations of continued smokers: a national perspective on the “hardcore” smoker phenomenon. Nicotine Tob Res. 2004;6:p.621–9. DOI: 10.1080/14622200410001727876. [15370158]

8 S Emery, EA Gilpin, C Ake, AJ Farkas, and JP Pierce. . Characterizing and identifying “hard-core” smokers: implications for further reducing smoking prevalence. Am J Public Health. 2000;90:p.387–94. DOI: 10.2105/AJPH.90.3.387. [10705856]

9 A Sorg, J Xu, SB Doppalapudi, S Shelton, and JK Harris. . Hardcore smokers in a challenging tobacco control environment: the case of Missouri. Tob Control. 2011;20:p.388–90. DOI: 10.1136/tc.2010.039743. [21330318]

10 R Mathews, WD Hall, and CE Gartner. . Is there evidence of ‘hardening’ among Australian smokers between 1997 and 2007? Analyses of the Australian National Surveys of Mental Health and Well-Being. Aust N Z J Psychiatry. 2010;44:p.1132–6. DOI: 10.3109/00048674.2010.520116. [21070109]

11 G Docherty, A McNeill, C Gartner, and L Szatkowski. . Did hardening occur among smokers in England from 2000 to 2010? Addiction. 2014;109:p.147–54. DOI: 10.1111/add.12359. [24103060]

12 MJ Jarvis, J Wardle, J Waller, and L Owen. . Prevalence of hardcore smoking in England, and associated attitudes and beliefs: cross sectional study. BMJ. 2003;326:p.1061 . DOI: 10.1136/bmj.326.7398.1061. [12750206]

13 M Lund, KE Lund, and E Kvaavik. . Hardcore smokers in Norway 1996-2009. Nicotine Tob Res. 2011;13:p.1132–9. DOI: 10.1093/ntr/ntr166. [21849413]

14 AK Ferketich, S Gallus, P Colombo, R Pacifici, P Zuccaro, and VC La. . Hardcore smoking among Italian men and women. Eur J Cancer Prev. 2009;18:p.100–5. DOI: 10.1097/CEJ.0b013e32830c8d98. [19337056]

15 ML Costa, JE Cohen, MO Chaiton, D Ip, P McDonald, and R Ferrence. . “Hardcore” definitions and their application to a population-based sample of smokers. Nicotine Tob Res. 2010;12:p.860–4. DOI: 10.1093/ntr/ntq103. [20601409]

16 J Bommelé, GE Nagelhout, M Kleinjan, TM Schoenmakers, MC Willemsen, and D van de Mheen. . Prevalence of hardcore smoking in the Netherlands between 2001 and 2012: a test of the hardening hypothesis. BMC Public Health. 2016;16:p.754 . DOI: 10.1186/s12889-016-3434-x. [27506600]

17 R Edwards, D Tu, R Newcombe, K Holland, and D Walton. . Achieving the tobacco endgame: evidence on the hardening hypothesis from repeated cross-sectional studies in New Zealand 2008-2014. Tob Control. 2017;26:p.399–405. DOI: 10.1136/tobaccocontrol-2015-052860. [27382047]

18 D Kaleta, B Usidame, E Dziankowska-Zaborszczyk, T Makowiec-Dabrowska, and M Leinsalu. . Prevalence and factors associated with hardcore smoking in Poland: findings from the Global Adult Tobacco Survey (2009-2010). BMC Public Health. 2014;14:p.583 . DOI: 10.1186/1471-2458-14-583. [24916122]

19 J Kishore, PK Jena, C Bandyopadhyay, M Swain, S Das, and I Banerjee. . Hardcore smoking in three South-East asian countries: results from the global adult tobacco survey. Asian Pac J Cancer Prev. 2013;14:p.625–30. DOI: 10.7314/APJCP.2013.14.2.625. [23621209]

20 KE Warner and DM Burns. . Hardening and the hard-core smoker: concepts, evidence, and implications. Nicotine Tob Res. 2003;5:p.37–48. DOI: 10.1080/1462220021000060428. [12745505]

21World Health Organization. Global adult tobacco survey. Geneva: World Health Organization; 2009.

22Ahmadi B, Boschi-Pinto C, Lopez A, Murray C, Lozano R, Inoue M. Age standardization of rates: a new WHO standard. Geneva: World Health Organisation; 2001.

23Kalsbeek WD, Bowling JM, Hsia J, Mirza S, Palipudi KM, Asma S. The Global Adult Tobacco Survey (GATS): sample design and related methods. 2010.

24 S Connor Gorber, S Schofield-Hurwitz, and J Hardt. . Levasseur Gv, Tremblay M. The accuracy of self-reported smoking: a systematic review of the relationship between self-reported and cotinine-assessed smoking status. Nicotine Tob Res. 2009;11:p.12–24. DOI: 10.1093/ntr/ntn010. [19246437]

25 J-F Etter, MM Bergman, J-P Humair, and TV Perneger. . Development and validation of a scale measuring self-efficacy of current and former smokers. Addiction. 2000;95:p.901–13. DOI: 10.1046/j.1360-0443.2000.9569017.x. [10946439]

26 MC Kulik and SA Glantz. . The smoking population in the USA and EU is softening not hardening. Tob Control. 2016;25:p.470–5. DOI: 10.1136/tobaccocontrol-2015-052329. [26108654]

27 P Jha and R Peto. . Global effects of smoking, of quitting, and of taxing tobacco. N Engl J Med. 2014;370:p.60–8. DOI: 10.1056/NEJMra1308383. [24382066]

28 H Piné-Abata, A McNeill, R Murray, A Bitton, N Rigotti, and M Raw. . A survey of tobacco dependence treatment services in 121 countries. Addiction. 2013;108:p.1476–84. DOI: 10.1111/add.12172. [23451932]

29 GR Kruse, NA Rigotti, M Raw, A McNeill, R Murray, and H Pine-Abata. Tobacco dependence treatment training programs: an international survey. Nicotine Tob Res. 2016;18:p.1012–8. DOI: 10.1093/ntr/ntv142. [26117835]

30 K Fagerström. . urberg H. A comparison of the Fagerström Test for Nicotine Dependence and smoking prevalence across countries. Addiction. 2008;103:p.841–5. DOI: 10.1111/j.1360-0443.2008.02190.x. [18412764]

31 CJ Heckman, BL Egleston, and MT Hofmann. . Efficacy of motivational interviewing for smoking cessation: a systematic review and meta-analysis. Tob Control. 2010;19:p.410–6. DOI: 10.1136/tc.2009.033175. [20675688]

32 CC DiClemente, JO Prochaska, SK Fairhurst, WF Velicer, MM Velasquez, and JS Rossi. . The process of smoking cessation: an analysis of precontemplation, contemplation, and preparation stages of change. J Consult Clin Psychol. 1991;59:p.295–304. DOI: 10.1037/0022-006X.59.2.295. [2030191]

33 CT Sreeramareddy, S Harper, and L Ernstsen. . Educational and wealth inequalities in tobacco use among men and women in 54 low-income and middle-income countries. Tob Control. 2018;27:p.26–34. DOI: 10.1136/tobaccocontrol-2016-053266. [27885168]

34 KM Cummings. . Community-wide interventions for tobacco control. Nicotine Tob Res. 1999;1 Suppl 1:p.S113–6. DOI: 10.1080/14622299050011701. [11072415]

Additional Material