RYO - Smokefree Coalition

Nicotine & Tobacco Nicotine & Tobacco ResearchResearch Advance Access published September 16, 2010

Original Investigation

Prevalence, Correlates of, and Reasons for Using Roll-Your-Own Tobacco in a High RYO Use Country: Findings from the ITC New Zealand Survey David Young, Ph.D.,1 Nick Wilson, F.N.Z.C.P.H.M., M.P.H.,2 Ron Borland, Ph.D.,1 Richard Edwards, M.D.,2 & Deepa Weerasekera, Ph.D.2 1 2

Tobacco Control Unit, Cancer Council Victoria, Carlton, Victoria, Australia Department of Public Health, University of Otago, Wellington, New Zealand

Received April 28, 2010; accepted August 20, 2010

Abstract Aim: To describe the prevalence, correlates of, and reasons for use of roll-your-own (RYO) tobacco in a high RYO use and ethnically diverse country: New Zealand (NZ). Methods: The NZ arm of the International Tobacco Control Policy Evaluation Survey (ITC Project) is sampled from the New Zealand Health Survey, with boosted sampling of Māori, Pacific peoples, and Asian New Zealanders. We surveyed 1,376 current adult smokers using standard ITC project procedures in 2007–2008. Results: Prevalence of regularly smoking RYOs was 53% (with 38% of all smokers being exclusive RYO smokers). RYO use was higher among disadvantaged smokers, heavier smokers, those with a relatively low intention of quitting, and those with more friends who smoke. RYO use increased more in the youngest age groups as disadvantage increased. “Lower price” dominated the reasons smokers’ cited for smoking RYOs (at 83%). About one fifth cited “less health concerns” as a reason. Conclusions: RYO smoking is particularly associated with individual deprivation and high levels of dependence. Its capacity to blunt price signals provided by tobacco taxes is accompanied by misperceptions that it is less hazardous to health and it is particularly prevalent among vulnerable disadvantaged populations (including Māori, young people, and those with mental health problems). Governments should reconsider removing any tax advantages given to RYO tobacco, ensure RYO smokers are properly informed of health risks, and supported to quit as strongly as other smokers. However, governments should also examine a broader range of options including a higher differential tax on RYO tobacco, removing flavors, and controlling all tobacco marketing.

Introduction Roll-your-own (RYO) cigarette use is an important component of the tobacco market in many countries. The International Tobacco Control Policy Evaluation South East Asia survey (ITC-SEA) reports that 58% of smokers make regular use of RYOs in Thailand and 17% do so in Malaysia (Young, Yong, et al., 2008 ). In the ITC four-country survey (ITC-4), there were large differences between countries in the prevalence of RYO use, with 28% of U.K. smokers, 24% of Australian smokers, 17% of Canadian smokers, but only 7% of U.S. smokers making regular use of RYO cigarettes (Young et al., 2006 ). A range of data sources suggest that RYO smoking in New Zealand (NZ) is far higher than in the four countries in the aforementioned ITC Project and is increasing. For example, in the 2006 New Zealand Tobacco Use Survey, among a nationally representative sample of smokers aged 15–64 years, 49% of European/other smokers, 60% of Māori smokers, 24% of Pacific smokers, and 13% of Asian smokers predominantly smoked RYOs (for daily or nondaily smokers; Ministry of Health, 2007). Returns from manufacturers to the Ministry of Health show that in 2006, 32% of tobacco sold for use in cigarettes was in the form of RYO tobacco (Laugesen, 2006). There was a 38% increase in weight of RYO tobacco sold from 1990 to 2006 (12% increase from 1999 to 2006). This compared with a decrease of 46% in tobacco weight sold in the form of FM cigarettes during the same period (22% decrease from 1999 to 2006). The significant attributes of RYO smokers have also been explored in the two ITC studies. In the four-country study, RYO use was associated with lower income, male sex, greater nicotine addiction, less intention to quit, and being more likely to believe that RYOs are less harmful (Young et al., 2006). In ITC-SEA, where most RYO tobacco is sourced from the informal economy, RYO smoking was more common among rural and older smokers and, especially, among female smokers with this profile (Young, Yong,

doi: 10.1093/ntr/ntq155 © The Author 2010. Published by Oxford University Press on behalf of the Society for Research on Nicotine and Tobacco. All rights reserved. For permissions, please e-mail: [email protected]

1

Downloaded from ntr.oxfordjournals.org at University of Otago on September 16, 2010

Corresponding Author: David Young, Ph.D., Tobacco Control Unit, Cancer Council Victoria, 100 Drummond Street, Carlton, Victoria 3053, Australia. Tel: +61-3-96355123; Fax: +61-3-96355440; E-mail: [email protected]

Prevalence, correlates of, and reasons for using roll-your-own tobacco et al., 2008). The proportion of RYO smokers was also related to socioeconomic status (SES) with higher prevalence among low income, less educated, and unemployed smokers. RYO smokers also smoked fewer cigarettes and had a lower intention to quit (Malaysia only).

Research suggests that RYO cigarettes are at least as harmful or possibly more harmful than FM cigarettes (e.g., studies reviewed elsewhere [O’Connor et al., 2007 ]). For example, a stronger association has been noted with the use of hand-rolled tobacco, compared with manufactured cigarette use, and the increased risk of esophageal cancer (Tuyns & Esteve, 1983), and with risk of cancer of the mouth, pharynx, and larynx (De Stefani, Oreggia, Rivero, & Fierro, 1992). A prospective study of 26,000 smokers in Norway reported increased risk of lung cancer in hand-rolled tobacco smokers compared with those smoking manufactured cigarettes (Engeland, Haldorsen, Andersen, & Tretli, 1996). Given that the prevalence of RYO smoking is increasing internationally (Connolly & Alpert, 2008), including in some countries where RYO use has traditionally been relatively uncommon (e.g., Ireland), it has become important to identify the reasons smokers cite for smoking RYO cigarettes in order to inform both population-level tobacco control strategies and cessation programs targeted at individuals. This is particularly true for young smokers, where initiation most frequently occurs, and for those disadvantaged groups, where tobacco smoking is increasingly concentrated. Taking all this into consideration, the study being reported investigated the prevalence, attributes, and reasons for smoking RYOs in a high RYO use developed country, with an ethnically diverse population, and in more detail than has previously been the case.

Methods The ITC Project The International Tobacco Control Policy Evaluation Survey (the ITC Project) is a multicountry study on tobacco use epidemiology and tobacco control policy evaluation. It has expanded to over 17 participating countries, including NZ. A full description of the ITC Project conceptual framework and methods have been published elsewhere (Fong et al., 2006 ; Thompson et al., 2006 ). The NZ arm of the ITC Project survey differs somewhat from the other ITC Project countries in that the smokers in-

2

Participants From the NZHS sample, we had an additional sampling frame of adult smokers who were 18+ years and expressed willingness to participate in further research (85.2% of adult smokers in the NZHS). Of 2,438 potential respondents who met these criteria, a total of 1,376 completed a telephone questionnaire giving a response rate of 56.4% (see an online Methods Report [Wilson, 2009] for more detail).

Procedures Surveying of these participants was carried out using a computerassisted telephone survey (subcontracted to Roy Morgan Research). The first wave of participants were interviewed between March 2007 and February 2008, usually 3–4 months after their NZHS interview. The study protocol was cleared by the MultiRegion Ethics Committee in New Zealand (MEC/06/07/071) and by the Office of Research Ethics, University of Waterloo, Waterloo, Canada (ORE #13547).

Measures The particular questions relating to the use and beliefs around RYO cigarettes were identical to those from the standard ITC survey questionnaire used in other countries. To determine RYO use, smokers were asked, “do you smoke factory-made cigarettes, roll-your-own cigarettes or both?” Smokers were categorized as exclusive RYO, exclusive FM, or “mixed” users. Some sociodemographic questions were asked in the NZHS but most of the smoking behavior and smoking-related belief questions were from the Wave 4 of the four-country ITC survey. We used some of the indices used elsewhere in our ITC Project analyses (Borland et al., 2004 ; Young et al., 2007 ). These are more precisely described in an online Methods Report (Wilson, 2009). Where indices were developed, we calculated scores for assessing internal consistency (Cronbach’s alpha), and these indices were only used if the scores were at least 0.5. Other measures included SES, which was assessed using a small area-based SES score developed for New Zealand (NZDep). In particular, NZDep2006 measures the level of socioeconomic deprivation for each neighborhood (Meshblock) according to a combination of 2006 Census variables (Salmond, Crampton, & Atkinson, 2007). This index and its predecessors have been extensively validated (White, Gunston, Salmond,


There are less detailed data available about the characteristics of RYO smokers in NZ. The data that are available suggest that RYO smoking is commoner among younger, Māori, and lower SES smokers. In the 2006 Tobacco Use Survey, use of RYO was significantly greater among younger age groups, among European/other ethnicities, and to a lesser extent among Māori (Ministry of Health, 2007). For example, the proportion of 15- to 19-year-old smokers who smoked any RYO cigarettes was 69% among Māori and 58% among non-Māori, whereas for 60- to 64-year-olds, it was 50% among Māori and 30% among non-Māori. Males had higher levels of RYO use among Europeans/others, but there was no consistent pattern by gender among Māori (Ministry of Health, 2007).

volved are respondents from the New Zealand Health Survey (NZHS; with this survey being conducted in 2006–2007). Methods of the NZHS are detailed more fully in the report on the key results (Ministry of Health, 2008b) and a detailed methodology report (Ministry of Health, 2008a). Respondents were selected by a complex sample design, which included systematic boostedsampling of the Māori, Pacific peoples, and Asian populations (with Pacific peoples in NZ being in order of decreasing population size: Samoan, Cook Islands Māori, Tongan, Niuean, Fijian, Tokelauan, and Tuvaluan). Interviews were conducted faceto-face in respondents’ homes by trained interviewers (on contract to the Ministry of Health) and resulted in a total of 11,924 interviews with respondents aged 18 years and over. The overall response rate was 67.9%. Other issues around the NZHS response rate as it relates to the ITC project are detailed in an online Methods Report (Wilson, 2009).

Nicotine & Tobacco Research Atkinson, & Crampton, 2008). We also used an individual-level deprivation score created for the New Zealand setting (NZiDep; (Table 1) and two measures of financial stress (see Table 1 and table footnotes). All these variables can be collectively included in the multivariate model without destabilizing the model with intercorrelation.

To put the NZ results concerning “reasons for RYO use” into a wider context, we analyzed comparable data from Australian smokers in Wave 5 (2006–2007) of the four-country ITC study (for methods, see elsewhere [Borland et al., 2009 ; Thompson et al., 2006]). Australia is a similar Anglophone country, albeit with a lower prevalence of RYO usage.

Table 1. Use of RYO Tobacco and FM Cigarettes by Sociodemographic Variables (weighted sample with adjustment for complex sample design)

Variable

Exclusive RYO (%)

Mixed FM + RYO (%)

Any RYO (%)

37.4 39.7 48.4 53.3 59.3

44.3 42.6 36.4 36.1 31.2

18.3 17.7 15.2 10.6 9.5

62.6 60.3 51.6 46.7 40.6

1.00 (referent) 0.91 (0.50–1.63) 0.64 (0.36–1.12) 0.52 (0.29–0.95) 0.41 (0.23–0.74)

1.00 (Referent) 0.99 (0.43–2.28) 0.98 (0.44–2.21) 1.41 (0.58–3.41) 1.35 (0.52–3.53)

46.7 48.2

37.8 38.6

15.5 13.2

53.3 51.8

1.00 (referent) 0.94 (0.70–1.27)

1.00 (referent) 1.20 (0.75–1.92)

47.4 36.9 64.5 85.9

38.6 46.5 17.0 13.1

14.0 16.6 18.5 1.0

52.6 63.1 35.5 14.1

1.00 Referent 1.54 (1.13–2.09) 0.50 (0.26–0.94) 0.15 (0.06–0.36)

1.00 Referent 1.02 (0.65–1.62) 0.34 (0.12–0.92) 4.98 (0.58 –42.69)

61.6 45.3 52.5 47.1 40.1

28.2 44.3 33.9 38.4 41.2

10.2 10.4 13.6 14.5 18.7

38.4 54.7 47.5 52.9 59.9

1.00 Referent 1.93 (1.01–3.69) 1.45 (0.77–2.74) 1.80 (0.98–3.29) 2.39 (1.34–4.29)

1.00 Referent 1.55 (0.51–4.75) 0.91 (0.30–2.79) 0.97 (0.32–2.92) 0.80 (0.29–2.25)

57.4 43.4 40.4 29.5 28.3 37.1

32.5 38.7 45.9 50.9 43.7 44.0

10.1 17.9 13.7 19.6 28.0 18.9

42.6 56.6 59.6 70.5 71.7 62.9

1.00 Referent 1.76 (1.17–2.64) 1.99 (1.19–3.32) 3.22 (2.02–5.13) 3.41 (1.92–6.03) 2.28 (1.68–3.09)

1.00 Referent 0.67 (0.35–1.26) 1.03 (0.49–2.19) 0.80 (0.40–1.61) 0.48 (0.21–1.09) 0.72 (0.44–1.19)

36.8

39.3

23.9

63.2

1.60 (0.93–2.75)

0.59 (0.28–1.24)

43.3

39.2

17.5

56.7

1.25 (0.88–1.76)

0.79 (0.47–1.32)

OR for exclusive RYO vs. mixed RYO (95% CI)

Note. FM = Factory-made; NZ = New Zealand; OR = odds ratios; RYO = roll-your-own. a Based on NZHS data with the age data collected a few months prior to the ITC Project survey. b Ethnicity results are for prioritized ethnicity where all those with Māori or both Māori and other ethnic affiliations were classified as Māori; where all those with Pacific or both Pacific and other ethnic affiliations were classified as Pacific (unless Māori affiliation was also reported) etc. For more detail, see an online Methods Report (Wilson, 2009). The European grouping includes other (non-Māori, non-Pacific, and non-Asian) ethnic groups. c Deprivation level was based on a NZ–specific small area deprivation index (NZDep2006) and also an individual measure of deprivation (NZiDep) also designed for NZ (Wilson, 2009). d We considered two measures of financial stress which are correlated with each other (and the small area deprivation measure; Wilson, 2009) but involve significant conceptual differences (Siahpush, Yong, Borland, Reid, & Hammond, 2009). The first question was, “. . . because of a shortage of money, were you unable to pay any important bills on time, such as electricity, telephone or rent bills?”. The second question was, “In the last six months, have you spent money on cigarettes that you knew would be better spent on household essentials like food?” For more detail, see an online Methods Report (Wilson, 2009).

3


Agea 18–24 (n = 125) 25–34 (n = 307) 35–44 (n = 320) 45–54 (n = 261) 55+ (n = 223) Gender Male (n = 529) Female (n = 847) Ethnicityb European (includes other) (n = 560) Māori (n = 556) Pacific (n = 71) Asian (n = 49) Area deprivation level (quintiles)c 1 and 2 (least deprived) (n = 105) 3 and 4 (n = 183) 5 and 6 (n = 210) 7 and 8 (n = 281) 9 and 10 (most deprived) (n = 457) Individual deprivation (NZiDep scores)c 0, i.e., least deprived individuals (n = 559) 1 (n = 230) 2 (n = 152) 3–4 (n = 176) 5–8, i.e., most deprived individuals (n = 118) 1–8 (any deprivation) (n = 676) Financial stressd Unable to pay any important bills on time—“Yes” (n = 104) (referent = “No”) Not spending on household essentials—“Yes” (n = 336) (referent = “No”)

Exclusive FM (%)

OR for any smoking RYO (i.e., both and RYO only) vs. FM only (95% CI)

Prevalence, correlates of, and reasons for using roll-your-own tobacco

Results

Weighting and Statistical Analyses Weighting of the results was necessary, given the sampling design (e.g., boosted sampling of Māori, Pacific peoples, and Asian New Zealanders in the NZHS) and nonresponse for the NZHS and ITC Project survey. A full description of the weighting process is detailed in an online report (Clark, 2008).

All analyses were conducted in Stata (version 10; StataCorp, College Station, TX), and all of the presented results were weighted and adjusted for the complex sample design of the NZHS to make the sample representative of all NZ smokers.

The results reveal that 47.5% (95% CI: 43.7%–51.2%) of current smokers smoked FM cigarettes exclusively, 38.2% (95% CI: 34.5%–41.8%) smoked RYO exclusively, and 14.3% (95% CI: 11.6%–17.1%) smoked both. Table 1 shows the use of RYO by sociodemographic factors. Use was most common among younger smokers, with over 60% of smokers aged less than 35 smoking RYOs and over 40% smoking RYOs exclusively. Compared with European/others, the prevalence of RYO smoking was higher among Māori but lower among Pacific and Asian smokers. Over 60% of Māori smoked some RYO, with almost half smoking RYO exclusively. RYO use was more prevalent among smokers with high individual deprivation scores, from the more deprived areas and smokers in financial stress.

Smoker-Related Behaviors and Mental Health The results in Table 2 indicate that RYO smokers are heavier smokers, have a lower intention to quit, and more friends who smoke than do smokers of FM cigarettes. They also suggest a consistent relationship between RYO use and various measures of psychological risk and mental health. It is apparent from the table that

Table 2. Differences Between RYO Versus FM Smokers for Key Smoking Behaviors, Mental Health, and Drug Usea Characteristic Smoking-related behaviors Heaviness of Smoking Index (alternate version)b Strength of intention of quitting (4-point scale)c Average number of friends who are smokers Mental health and alcohol use Ever been diagnosed with a mental health disorder (from a list of eight items in the NZHSd)—“Yes” (to any in that list; n = 247) (reference = “No”) Ever been diagnosed with an “alcohol-related disorder”—“Yes” (n = 27) Ever been diagnosed with a “drug-related disorder”—“Yes” (n = 22) AUDIT score = 0 (no drinking) (n = 159) AUDIT score = 1–7 (n = 699) AUDIT score ≥ 8 (hazardous alcohol use) (n = 378) p Value for trend (increasing AUDIT score in RYO users) Health and mental health measures Mean SF-36 score (overall) Mean SF-36 (physical functioning) Mean SF-36 (role limitation—physical) Mean SF-36 (mental health score) Psychological distress Kessler 10-item scale (K10)

RYO only and mixed (FM + RYO) smoker

FM only smoker

Score 1.45 1.03 3.18 % 62.9

Score 0.57 1.20 2.56 % 37.1

74.8 70.3 45.6 50.8 57.6 χ2 = 7.49, p = .006 Score 74.2 (72.8–75.5) 85.2 (83.0–87.4) 83.7 (81.0–86.3) 79.6 (78.2–81.1) 5.0 (4.5–5.6)

25.2 29.7 54.4 49.2 42.4 Score 76.7 (75.6–77.9) 86.8 (85.0–88.6) 87.3 (85.3–89.3) 81.1 (79.6–82.7) 3.9 (3.4–4.5)

Difference 0.88 (p < .001) −0.17 (p = .022) 0.62 (p < .001) OR (95% CI) 1.70 (1.18–2.47) 2.71 (0.90–8.19) 2.16 (0.64–7.28) 1.00 (reference) 1.23 (0.78–1.93) 1.62 (0.99–2.66) Difference −2.5 (p = .005) −1.6 (p = .283) −3.6 (p = .031) −1.5 (p = .165) 1.1 (p = .005)

Note. FM = Factory-made; OR = odds ratios; NZHS = New Zealand Health Survey; RYO = roll-your-own. a For more detail on the particular questions in the NZHS, see elsewhere (Ministry of Health, 2008b). b The “Heaviness of Smoking Index” (HSI) has been developed by others and we used the “alternative version” (HSI-AV) utilized by others (Borland et al., 2004). This is calculated as the square root of the daily cigarette consumption minus the natural logarithm of time to first cigarette of the day. The specific equations are detailed in an online Methods Report (Wilson, 2009). c The “strength of intention of quitting” scale has been used in other ITC Project work (Young et al., 2007). For more detail, see an online Methods Report (Wilson, 2009). d This list was depression, bipolar disorder (manic depression), anxiety disorder (including panic, phobia, posttraumatic stress disorder, obsessive compulsive disorder), eating disorder, alcohol-related disorder, drug-related disorder, schizophrenia, and “any other mental health condition (please specify).”

4


Univariate and bivariate analyses of the key socioeconomic and smoking variables associated with RYO smoking were initially conducted, followed by a multivariate logistic regression. The latter used a conceptual framework, which assumed that there would be hierarchical relationships between demographic and sociodemographic factors (Victora, Huttly, Fuchs, & Olinto, 1997), that would dominate over smoking-related behaviors and beliefs. Model 1 included age, gender, ethnicity, and key sociodemographic variables (e.g., SES and financial stress). Model 2 added mental health conditions (available from the NZHS—see elsewhere [Ministry of Health, 2008a; Wilson, 2009]) and heaviness of smoking, number of friends who smoke, and intention of quitting. Model 3 included all these variables plus significant interactions.

Prevalence of RYO Use

Nicotine & Tobacco Research RYO smokers are disproportionately represented among those who have the highest Alcohol Use Disorders Identification Test (AUDIT) scores (representing more hazardous alcohol use), those who have ever been diagnosed with a mental health disorder (selfreport), and those who have a lower overall Short Form Health Survey 36 (SF-36) score (with a significant contribution by physical role limits). In addition, they have higher “psychological distress” scores.

Multivariate Analysis The multivariate analysis of correlates of any RYO use reported in Table 3 was complemented by an analysis, which included the interaction between all other correlates and level of individual deprivation, with the only significant effect included in Model 3.

The significant interaction between deprivation and age 35– 49 is because, while RYO use was most common among the most deprived young and older smokers, use increases far more in the youngest age groups as disadvantage increases (see Figure 1).

Reasons for Smoking RYO Cigarettes

This study furthers the contribution that the ITC studies have made to our understanding of RYO smoking. The study shows RYO smoking is especially high in NZ and higher than the United Kingdom, another country that is concerned about high levels of use ( Young, Fong, Borland, Cummings, & Hammond, 2008; Young et al., 2006; Young, Yong, et al., 2008). The levels found in NZ are comparable with those found in Thailand, a developing country where RYO use is concentrated among the rural peasant class, and thus, the determinants of use may be quite different to NZ. As in other developed countries (Young et al., 2006), RYO smokers from NZ are more disadvantaged, younger, and, importantly, they appear to be more deeply trapped by their dependency and embedded within a smoking culture (i.e., more dependent, with more friends who smoke, and a lower intention to quit) than FM smokers. However, in NZ, RYO use is not dominated by men, as found in the ITC 4-country study (Young et al., 2006), being equally common among women. The main driver of RYO use in NZ appears to be individuallevel deprivation. The difference for the most deprived smokers versus the least deprived is marked (i.e., 72% vs. 43%; Table 1), and it swamped all other effects in the multivariate analysis, including ethnicity. If smoking is increasingly a sign of disadvantage, then RYO smoking is even more so. The interaction between deprivation and age shows that for some reason, the increase in RYO use with deprivation is not as marked in the middle-aged group. This group has the lowest proportion of deprived smokers, so it may be that RYO use only increases with deprivation when supported by the individual’s subculture. This interpretation could also help explain the low prevalence among Pacific peoples, even though they are a disadvantaged community, suggesting that it is not just deprivation alone that correlates with RYO use, but disadvantage in a specific cultural context. As such, this is an issue that merits further research. On the other hand, the higher level of Māori RYO use appears to be primarily due to their higher levels of individual deprivation.

From a list of five reasons (plus “other”), RYO smokers gave an average of 3.0 reasons for their use of RYOs. Lower cost (83%), better taste (73%), and greater satisfaction (63%) were the most frequently cited reasons (Figure 2). Around half cited reducing the amount smoked, and just over 20% cited lower health concerns (“not as bad for your health”). So far as smoking RYO because of lower cost (84% of exclusive smokers vs. 80% of mixed smokers), reducing the amount smoked (50% vs. 52%), and less health concerns (22% vs. 20%) were concerned, there was clearly little difference between mixed and exclusive RYO smokers. On the other hand, far more exclusive RYO users responded that they smoked RYOs because they taste better (82% vs. 51%) and because they provide greater satisfaction (70% vs. 43%). Given the findings above on the importance of deprivation level, we explored covariates of reporting cost as a reason. Women were more likely to do so (odds ratio [OR] = 1.78 [95% CI: 1.04–3.03]), as were those reporting not spending on household essentials (OR = 1.88 [95% CI: 1.05–3.36]) and there was a significant association with increasing level of individual deprivation (χ2 = 5.61, p = .018).

The results support anecdotal evidence for extensive use of RYO tobacco by people with mental health and/or multiple drug use problems (including hazardous alcohol intake), and deprivation also seems to be the main mediating mechanism here. Given the often parlous financial situation of these people, it is not surprising that they make disproportionate use of RYO. Those diagnosed with a mental illness are consistently identified as heavier smokers (Chaiton, Cohen, O’Loughlin, & Rehm, 2009; Lawrence, Mitrou, & Zubrick, 2009; McKee, Weinberger, Harrison, Coppola, & George, 2009; Mexal, Berger, Logel, Ross, Freedman, & Leonard, 2010), as well as financially deprived (Marsh & McKay, 1994) and younger (Marsh & McKay). If this analysis is correct, it suggests that RYO use is likely to become prevalent in any economically disadvantaged group where the cultural conditions do not inhibit its use.

Figure 2 compares the reasons for smoking RYO in NZ with those in Australia. It is noteworthy that while the results for relative cost are almost identical, NZ smokers are more likely to cite taste and satisfaction as reasons for smoking RYO, whereas Australian smokers are more likely to believe that RYO smoking carries lower health risks and that it reduces the amount smoked.

The high level of RYO use among younger smokers is an issue of concern, with 63% of 18- to 24-year-olds in NZ smoking RYO, and moreover, the data presented in Figure 1 demonstrates that it is a combination of relative youthfulness and disadvantage that is most strongly associated with RYO use. In addition, a 2006 national survey of 14- to 15-year-old children

5


The only significant interaction is between individual deprivation and age (35–49 years). In the final model, as in the earlier analyses, individual deprivation is the most important independent correlate of RYO use, with its adjusted odds ratio (AOR) maximized by the inclusion of the interaction with age (AOR = 4.7). Age, AUDIT score, history of mental illness, and psychological distress all drop out as an independent correlates, but Pacific peoples and Asian ethnicity are still significant correlates, with both groups far less likely to use RYO. The indices of smoking behavior are also still predictive, with RYO users having a higher Heaviness of Smoking Index and more friends who smoke than do FM smokers.

Discussion

Prevalence, correlates of, and reasons for using roll-your-own tobacco

Table 3. Logistic Regression Analysis Comparing Regular Smokers of any RYO Tobacco With FM-only Smokers (weighted and adjusted results) AOR (95% CI)a Model 2 (+ mental health and smoking behavior)

Model 3 (+ interaction effects)

Hosmer–Lemeshow, χ2 = 6.76, df = 8 (p = .563)b



Variables

n = 1,235

n = 1,133

n = 1,152

Demographic Age 35–49 vs. 50+

1.46 (1.01–2.12)