Characteristics and Patterns of Black & Mild Use Among African American Smokers

Abstract

Introduction:

Increased consumption of little cigars and cigarillos (LCCs) among late adolescent and young adult African American smokers is an emerging public health concern. Despite a growing body of epidemiological evidence, there is scant research on systematic variations in LCC use among young adult African Americans. This study aimed to disaggregate African American Black & Mild (B&M) smokers and to identify subgroups of cigarillo smokers’ behaviors and associated characteristics.

Methods:

Using telephone screening data for a NIH-funded clinical laboratory study of toxicant exposure associated with B&M use, latent class analysis (LCA) was used to classify 331 African American B&M smokers’ based on daily use, average daily consumption, preference for flavors, preference for product design tips, and product modification behaviors.

Results:

Results showed five classes, three of which (Daily-Hypers, Daily-Flavored, and Heavy-Daily-Hypers) reflected daily use and average daily consumption rate ranging between 2.7 and 8.9 B&M cigarillos per day. Non-Daily-Hypers and Non-Daily-Flavored classes represented non-daily use patterns and averaged less than 1.0 B&M cigarillo per day. Both sets of classes defined by daily users and non-daily users included smokers who preferred flavored tobacco and who practiced product modification techniques involving hyping and blunting.

Conclusions:

Latent class analysis is a useful method to detect subtle differences in B&M product preferences and smoking behaviors among African Americans. Study findings highlight the importance of developing tailored interventions that consider within group differences in order to reduce the prevalence of cigarillo smoking among those with the greatest burden.

Implications:

The current study is the first to identify unique subgroups of African American B&M smokers based on cigarillo use behaviors and associated characteristics. Latent class analyses may prove useful for understanding other subgroups of tobacco users. Current findings concerning patterns of LCC use illustrate how future tobacco cessation and prevention interventions may be tailored for African American smokers.

Introduction

Public health research on little cigars and cigarillos (LCCs) is gaining momentum, as a growing number of studies reveal an increase in LCC use among adolescents, ^{1 , 2} ethnic/racial minorities (eg, African Americans ^{1 , 2} ), and males. ^1–3 National estimates show that 9% of African American high school students report past month large cigar/LCC use compared to 3%–8% among other ethnic/racial groups. ² LCC prevalence rates appear to increase in young adulthood (ie, ages 18–25; 14.2% ³ ), with use of cigarillos reaching epidemic proportions in some African American young adult populations (ie, 63% ⁴ ). Cigarillos are distinguished from other cigar products based on the weight of their tobacco content ⁵ and include common brands such as Black & Mild (B&M) or Swisher Sweets. Historically, rates of LCC use have been underestimated ⁶ substantially for African Americans and other vulnerable populations. ⁷

Like cigarette smoking, cigarillo smoke exposes users to several harmful toxicants including the dependence-inducing drug nicotine and cardiovascular risk agent carbon monoxide. ^8–12 Moreover, cigar/LCC products can contain flavoring additives ¹³ like menthol which is highly preferred among African American cigarette smokers. ¹⁴ If cigar/LCC flavoring additives, like menthol in cigarettes, ^{15 , 16} increase the likelihood for initiation and regular use ^{15 , 16} there may be greater propensity for dependence and poor health consequences for flavored cigar/LCC smokers. Thus, the effects of flavored cigar/LCC use may be greater African Americans who suffer a disproportionate burden of tobacco-related disease. ^{17 , 18}

A general review of research on cigarillo smoking gives the impression of a somewhat homogenous group of cigarillo users: African American adolescent males. However, like cigarette smokers, there may be considerable diversity in user attitudes, beliefs, and behaviors including frequency and intensity of smoking, within this group. These behavioral variations have both physiological and subjective impacts. ^{19 , 20} Variations in patterns of LCC smoking may differentially impact users’ nicotine dependence and disease susceptibility.

The purpose of this study was to use latent class analysis (LCA) to identify subgroups of African American B&M smokers using data on frequency and intensity of use, preference for flavors, preference for product design tips, and product modification. We used LCA to identify latent or unobservable subgroups within a population of B&M users to examine associations of observed characteristics. LCA can yield important information about individual differences in tobacco use and help public health officials target high-risk groups. Considering the high prevalence of cigarillo use, in particular the B&M brand among African American cigar smokers, ^{21 , 22} LCA can be a useful tool to inform tobacco prevention and cessation efforts specific to this group.

Methods

Population and Sample

As part of a larger study examining toxicant exposure associated with cigarillo use, B&M cigarillo smokers were recruited from the community via advertisements (eg, Craigslist, local newspapers, and flyers) and word-of-mouth from November 2012 to October 2013. Advertisements prompted individuals to telephone the laboratory to discuss the purpose and requirements for the study. Interested B&M smokers were asked to complete a telephone screening questionnaire that included demographics, health status, and drug use history. Study eligibility included being 18–35 years old, smoking at least 7 B&M cigarillos per week, and no more than 5 cigarettes per day for the past year. Tobacco use was verified via an expired air carbon monoxide level more than 7 ppm and a urine cotinine reading more than 2. Exclusion criteria included self-reported chronic health or psychiatric conditions, regular use of medications (except vitamins or birth control), and pregnancy (verified via urinalysis) or breastfeeding.

We completed a secondary analysis of screening data collected from individuals interested in participating in the laboratory study described above. ²³ Data for this analysis included: demographics, tobacco use history (cigarettes, cigars, and other tobacco products), cigarillo flavor and tip preferences, and cigarillo modification behaviors. A total of 454 individuals provided screening data. Fifteen participants who did not report past 30-day B&M use were excluded from the current analysis. Of the remaining participants, 302 identified as black/African American and 29 identified as “other” (black/African American and another race, mixed, bi-racial, mulatto, Caribbean, and Cape Verdean). The total sample analyzed was 331 participants. Individuals were not required to meet eligibility for the laboratory study to be included in the current analysis.

Measures

Latent Class Indicators

Dependent variables included B&M use frequency, quantity, flavor and tip preference, hyping, and rolling blunts. To assess “frequency” of use, participants were classified as daily or non-daily users (“Do you smoke Black & Milds every day?”). To assess “quantity,” the participants were asked: “Can you estimate how many Black & Milds you smoke per day?” “Flavor preference” was assessed via the following item with three options (original; flavored; both): “Do you prefer original Black & Mild or Black & Mild with flavors?” An important point to note is that original Black & Milds are flavored with Cavendish tobacco, Burley tobacco, and Virginia golden pipe tobacco. These tobaccos are highly flavored with sweet flavors including vanilla and are highly aromatic. ²⁴ “Tip preference” was assessed using four options (wood; plastic; both; none): “Do you prefer Black & Mild with a wood tip or plastic tip?” “Hyping” (“Do you ever remove the inner paper of the Black & Mild before smoking?”) was coded as a Yes/No, while “rolling blunts” (“Have you ever hollowed out a cigar like a Black & Mild by taking out the tobacco and replacing it with marijuana?”) was coded with five response options (no/I don’t anymore; hardly ever; sometimes; most of the time; all of the time).

Covariates

Demographic variables included as covariates were age, gender, high school graduate (yes; no), and monthly income after taxes (less than $500; between $500 and $1000; between $1000 and $1500; more than $1500). Assessment of “other tobacco use” items included cigarettes, cigars, waterpipe tobacco, and smokeless tobacco. For each item, participants were asked whether they smoked the specific tobacco product within the past year. Those who responded “yes” were asked “When was the most recent occasion you smoked [the tobacco product]?”, and given four options: within the past 30 days; 2 to 3 months ago; 4 to 6 months ago; more than 6 months ago. Responses to these questions were combined to produce one of two possible outcomes: “no” [(1) reported no past-year use of any other tobacco product and/or (2) past-year use but not the past 30 days)] and “yes” (reported past year and past 30-day use of any other tobacco product).

For “marijuana use,” callers were first asked whether they had used marijuana in their lifetime (“Have you ever, in your lifetime, smoked marijuana or hashish?”). Those who responded “yes” were then asked “How many days out of the last 30 have you smoked marijuana?” Never users and individuals who had not used in the past 30 days were coded as “no,” while those who reported use of marijuana on at least 1 day out of the past 30 were coded as “yes.” Perception of harm reduction was assessed among callers who confirmed that they remove the inner paper of their B&M prior to smoking. Specifically, this subgroup of callers was asked, “Compared to a Black & Mild with the inner paper removed, do you think smoking a Black & Mild without removing the inner paper is: less harmful, as harmful, or more harmful.”

Data Analysis

We used Latent Gold 5.0 to classify black/African American smokers based on the six characteristics and patterns of their B&M use described above (Statistical Innovations, 2005). The analyses involved a successive-class model-building strategy where the number of classes was increased until a solution with adequate fit was identified. Both the Bayesian Information Criterion and the Akaike Information Criterion were used as global fit indices to select the appropriate number of latent classes in the model. While parsimony was a critical factor in the determination of a true model, we also considered the likelihood of under- and over-fitting models and their relative implications to false negatives (ie, not identifying potentially vulnerable subpopulations among B&M smokers) and false positives. The bivariate residuals (BVRs) for all probable models were inspected to determine any violations related to the assumption of local independence. Missing value analysis revealed that approximately 10% of the 331 black/African American smokers had incomplete responses to B&M product items. There is no widely agreed upon missing data threshold (5% ²⁵ ; 10% ²⁶ ); however due to this study’s sample size and the significance of the research aim relative to this population, missing values were retained in the models using the Full Information Likelihood (FIML) estimation procedure available in Latent Gold. Importantly, interpretability of the latent class solutions was the final consideration for model selection. In the resultant latent class solution, the modal class assignments of members were treated as nominal variables and then analyzed using multinomial regression. IBM SPSS 22 (IBM Corporation, 2013) was used to examine class differences on demographic characteristics, cigar harm perceptions related to product modification, and other substance behaviors. These effects were estimated separately and then simultaneously in the regression model.

Results

Descriptive Characteristics

The sample was composed primarily of young adult males. A description of the participant characteristics examined (except for B&M flavor and tip preference) are displayed in Table 1 . Three quarters of participants had graduated high school and more than half had a monthly income below $500 per month. Most participants reported typically smoking B&Ms every day (70.7%). Most participants preferred original flavored B&Ms (72%) and plastic tips (56%). Aside from B&Ms, cigarettes were the most commonly used tobacco product in the past 30 days with nearly all of the cigarette smokers preferring menthol flavored cigarettes. Other tobacco product use prevalence in the past 30 days ranged from 1.5 to 14.5%. Nearly half of the sample (44.1%) reported past 30-day marijuana use. Modification behaviors were popular, with more than half reporting ever replacing tobacco from a B&M with marijuana (rolling blunts; 44.1%) and removal of the inner paper from a B&M (hyping; 72.2%). Chi-square tests of association were used to examine if examined characteristics differed by gender. Aside from females being more likely to have graduated from high school (χ ² (1) = 6.93, P < .01), males and females did not differ on the included variables.

Table 1.

Overall Participant Characteristics

Characteristic a	n	%
Age ( M , SD )	331	25.9, 5.1
Gender
Female	87	26.3%
Male	244	73.7%
High school graduate	251	75.8%
Monthly income after taxes
<$500	191	57.7%
$500–$1000	82	24.8%
$1000–$1500	41	12.4%
>$1500	17	5.1%
Cigarette use in past 30 days	221	66.8%
Daily user of Black & Milds	234	70.7%
Number of Black & Milds typically smoked per day
1 or less	131	39.6%
2–3	120	36.3%
4–5	60	18.1%
>5	20	6.0%
Other cigar use in past 30 days	48	14.5%
Waterpipe tobacco use in past 30 days	40	12.1%
Smokeless tobacco use in past 30 days	5	1.5%
Alcohol use in past 30 days	249	75.2%
Marijuana use in past 30 days	146	44.1%
Ever blunted a Black & Mild (replaced tobacco with marijuana)	145	43.8%
Frequency of blunting Black & Milds (among those who reported ever blunting)
I don’t anymore	55	37.9%
Hardly	23	15.8%
Sometimes	21	14.4%
Most of the time	12	8.2%
All the time	34	23.3%
Ever “hyped” or “freaked” Black & Milds (remove the inner paper before smoking)	228	68.9%

^a For items with missing data, available n for analysis are listed: Daily user of Black & Milds, 317; Other cigar use in the past 30 days, 89; Waterpipe tobacco use in past 30 days, 92; Smokeless tobacco use in past 30 days, 16; Alcohol use in the past 30 days, 322; Ever blunted a Black & Mild, 262; Frequency of blunting Black & Milds, 145; Ever “hyped” or “freaked” Black & Milds, 316.

Latent Class Model

We estimated a series of latent class models to identify underlying groups within a heterogeneous sample of B&M smokers. Based on the Bayesian Information Criterion, the overall fit of the model improved substantially from the 1-class model to 5-class model (3,816-3,223), but deteriorated when estimating subsequent models. Comparatively, the Akaike Information Criterion revealed improved model fit from the 1-class model to a 7-class model (3,771-2,921), with model fit weakening for the 8-class (2,965) and 9-class (2,984) models ( Table 2 ). As a consequence, a model with 5, 6 and 7 latent classes was probed further using additional diagnostics to determine best fit. The model with 5 through 7 latent classes was examined using the conditional bootstrap LR results to compare k -class versus k -1class models, observing classification errors between successive classes, and considering the interpretability of classes.

Table 2.

Baseline Latent Class Model Comparisons

No. of classes	AIC	BIC	Classification errors	Entropy R2
1	3771.3	3816.9	0.00	1.00
2	3380.5	3464.2	0.00	0.99
3	3184.3	3306.0	0.01	0.98
4	3081.8	3241.5	0.01	0.98
5	3025.7	3223.4	0.01	0.96
6	3012.8	3248.5	0.09	0.87
7	2921.8	3195.6	0.02	0.96
8	2965.6	3277.4	0.08	0.89
9	2984.3	3334.1	0.08	0.88

AIC = Akaike’s Information Criterion; BIC = Bayesian Information Criterion. Selected model is noted by boldface type.

There was strong evidence for the more parsimonious 5-class model based on the above-referenced diagnostics. Additional support for the 5-class model was revealed by the high degree of similarity between two classes within the 7-class model and a larger class within the 5-class model. Specifically, cluster proportions, means and interpretability across model indicators for two classes within the 7-class model approximated the larger class within the 5-class model. The residual correlations for the 5-class model were low (<2.0), thereby confirming the assumption of independence between indicators. Classification error for the 5-class model also was low (<1.0%) and the entropy R² was 0.96.

Importantly, an inspection of the 5-class model revealed a marginally significant parameter estimate for rolling blunts ( P ≥ .10). We reestimated the model to include a restriction on this effect across the five latent classes. This procedure produced no noticeable improvements in classification estimates and in the probabilistic interpretation of the re-scaled parameters. For these reasons and reasons related to the significance of study aims about B&M product modification behaviors, we did not implement any parameter restrictions on the 5-class model.

Three of the five latent classes (Daily-Hypers; Daily-Flavored; Heavy-Daily-Hypers) reflected daily use characteristics and an average daily consumption rate ranging between 2.7 and 8.9 B&M cigarillos per day. Two other classes, Non-Daily-Hypers and Non-Daily-Flavored, averaged less than 1.0 B&M cigarillo on a daily basis. Both sets of classes included smokers who preferred flavored tobacco and who practiced product modification techniques involving hyping and rolling blunts.

Table 3 presents the assigned labels and conditional probabilities for all model indicators. Overall, about 71% B&M smokers belonged to one of three daily user classes, whereas 29% belonged to the remaining two, non-daily user classes. Original flavored cigarillos were preferred by more than two-thirds of the total sample of smokers (68%), representing about 75% of daily users of B&Ms within Daily-Hypers and Heavy-Daily-Hypers. Of note, essentially none of the B&M user classes had a preference for both original flavored and specialty flavored products. In terms of product modification behaviors, about three-fourths of B&M daily users (Daily-Hypers and Heavy-Daily-Hypers) and more than half of non-daily users (Non-Daily-Hypers and Non-Daily-Flavored) endorsed removing the inner paper from the cigarillo product before smoking (ie, hyping). The majority of B&M users within each class (62%–85%) reported not rolling blunts with cigar products. Interestingly, only the B&M user classes defined by a preference for original flavored cigarillos (Daily-Hypers, Non-Daily-Hypers, and Heavy-Daily-Hypers) reported a preference for hyping cigarillos or rolling blunts.

Table 3.

Conditional Probabilities for 5-Class Latent Class Model

	DH	DF	NDH	NDF	HDH
Cluster size indicators	0.49	0.18	0.16	0.13	0.04
Black & Mild frequency and quantity
Frequency
Non-daily user	0.02	0.01	0.99	0.99	0.01
Daily user	0.98	0.99	0.01	0.01	0.99
Quantity
Average daily consumption ( M )	3.00	2.76	0.51	0.53	8.98
Black & Mild product design features
Flavor preference
Original	0.99	0.00	0.99	0.01	0.85
Flavored	0.01	0.97	0.01	0.99	0.14
Both	0.00	0.03	0.00	0.00	0.01
Tip preference
Plastic tip	0.65	0.51	0.69	0.18	0.73
Wood tip	0.22	0.30	0.21	0.78	0.26
Both	0.13	0.19	0.10	0.04	0.01
Black & Mild product modification behaviors
Ever “hyped” or “freaked”
No	0.01	0.99	0.01	0.99	0.14
Yes	0.99	0.01	0.99	0.01	0.86
Frequency of blunting
No/I don’t anymore	0.70	0.85	0.71	0.73	0.62
Hardly	0.07	0.06	0.07	0.07	0.07
Sometimes	0.07	0.04	0.07	0.06	0.08
Most of the time	0.04	0.02	0.04	0.04	0.05
All of the time	0.12	0.03	0.11	0.10	0.18
M	1.82	1.31	1.74	1.69	2.08
Model covariates ( I )
Age ( M )	25.6	26.5	26.3	24.6	27.5
Gender
Female	0.29	0.22	0.15	0.32	0.26
Male	0.71	0.78	0.85	0.68	0.74
High school graduate
No	0.27	0.22	0.20	0.16	0.32
Yes	0.73	0.78	0.80	0.84	0.68
Monthly income after taxes
<$500	0.57	0.52	0.64	0.58	0.59
$500–$1000	0.26	0.24	0.22	0.23	0.29
$1000–$1500	0.10	0.20	0.10	0.14	0.11
>$1500	0.07	0.04	0.04	0.05	0.01
Other tobacco use in past 30 days
No	0.25	0.40	0.17	0.30	0.31
Yes	0.75	0.60	0.83	0.70	0.69
Marijuana use in past 30 days
No	0.52	0.69	0.51	0.56	0.62
Yes	0.48	0.31	0.48	0.44	0.38
Removing inner paper from Black & Mild
Less harmful vs. regular	0.28	0.24	0.24	0.28	0.27
Equal or more harmful vs. regular	0.72	0.76	0.76	0.72	0.73

DH = Daily-Hypers; DF = Daily-Flavored; HDH = Heavy-Daily-Hypers; NDH = Non-Daily-Hypers; NDF = Non-Daily-Flavored.

Table 4 presents the probability mean estimates (ie, an index for overall misclassification) for the five latent classes. B&M daily users compared to non-daily users had a substantially higher probability of reporting an average daily consumption rate of about 1 (daily user total response probability = 0.77) to 5 or more cigarillos (daily user total response probability = 1.00). Specifically, the Daily-Hypers class had the highest probabilities of consuming one or more cigarillos per day (0.54–0.74). Daily users also had a strong preference for original flavored B&M cigarillos, especially Daily-Hypers members (0.71). Flavored cigarillos preferred by daily users (daily user total response probability = 0.68). Daily-Flavored members had a high probability of using flavored cigarillos (0.65) and both products (1.00). Daily-Hypers members had a preference for the plastic tip (0.57) or both tip types (0.55). Further, B&M smokers who endorsed hyping had a high probability of being Daily-Hypers members (0.71) and those who did not were more likely to be classified as Daily-Flavored members (0.56). Smokers who reported rolling blunts most or all of the time (0.67 and 0.58, respectively) were more likely to Daily-Hypers members compared to other classes.

Table 4.

Probability Means for 5-Class Latent Class Model

	DH	DF	NDH	NDF	HDH
Cluster size indicators	0.49	0.18	0.16	0.13	0.04
Black & Mild frequency and quantity
Frequency
Non-daily user	0.03	0.00	0.53	0.44	0.00
Daily user	0.69	0.25	0.00	0.00	0.06
Quantity
0.03–0.71	0.00	0.00	0.65	0.35	0.00
0.86–1.71	0.54	0.23	0.13	0.10	0.00
2.00–2.50	0.69	0.31	0.00	0.00	0.00
3.00–4.00	0.74	0.26	0.00	0.00	0.00
5.00 or more	0.61	0.12	0.00	0.00	0.27
Missing	0.20	0.04	0.12	0.62	0.02
Black & Mild product design features
Flavor preference
Original	0.71	0.00	0.23	0.00	0.05
Flavored	0.00	0.65	0.00	0.32	0.03
Both	0.00	1.00	0.00	0.00	0.00
Missing	0.00	0.06	0.00	0.94	0.00
Tip preference
Plastic tip	0.57	0.16	0.19	0.02	0.06
Wood tip	0.40	0.19	0.12	0.25	0.04
Both	0.55	0.29	0.13	0.03	0.00
Missing	0.05	0.06	0.06	0.83	0.00
Black & Mild product modification behaviors
Ever “hyped” or “freaked”
No	0.00	0.56	0.00	0.42	0.02
Yes	0.71	0.00	0.23	0.00	0.06
Frequency of blunting
No/I don’t anymore	0.47	0.21	0.15	0.13	0.04
Hardly	0.52	0.04	0.31	0.09	0.04
Sometimes	0.45	0.14	0.13	0.24	0.04
Most of the time	0.67	0.08	0.08	0.17	0.00
All of the time	0.58	0.06	0.18	0.09	0.09

DH = Daily-Hypers; DF = Daily-Flavored; HDH = Heavy-Daily-Hypers; NDH = Non-Daily-Hypers; NDF = Non-Daily-Flavored.

Latent Class Model With Covariates

We tested whether demographic characteristics (ie, age, gender, education level and monthly income), cigar harm perceptions related to product modification, and other substance behaviors (ie, other tobacco use and marijuana use) predicted class membership for B&M smokers. Daily-Hypers was specified as the referent class. Age was the only predictor of class membership ( P = .02). Specifically, members in Heavy-Daily-Hypers were significantly older than their Daily-Hypers counterparts, OR = 1.18 (1.07, 1.28). Of particular note is that Heavy-Daily-Hypers members reported consuming, on average, more than 8 B&M cigarillos daily. There were no significant effects between members of Daily-Hypers and other classes for any of the other demographic characteristics, cigar harm perceptions, or other tobacco behavior predictors.

Discussion

Recent work suggests that cigar use is most prevalent among African Americans. ^27–30 Still, few studies that report on the demographic profiles of cigar users have differentiated between product types (large cigars, little cigars, and cigarillos). Little is known about the consumers and use characteristics of individual cigar types such as cigarillos. Thus, the purpose of this study was to characterize African American smokers of B&M cigarillos based on their patterns of use, preference for product types, and product modification behaviors. The LCA analysis identified five latent classes that were distinguished by frequency of B&M use, flavor preference, and hyping behaviors with the majority of smokers being classified into a class of daily users who prefer original flavored cigarillos. The findings show that B&M users are not a monolithic population of tobacco users, and suggest that prevention and cessation interventions may be differentially effective based on patterns of daily and unorthodox use.

Consistent with other work examining characteristics of cigar smokers nearly three-fourths of our sample identified as male. ^{3 , 31 , 32} The majority were also daily B&M smokers, with a range of 2.7 to 8.9 B&M cigarillos per day. This level of daily use is higher than previously reported (eg, 13.3% ²⁹ ; <10% ³³ ), though most surveys have assessed use over longer periods of time (“lifetime,” “past year,” and/or “past 30-day”). ^{3 , 27 , 29 , 34 , 35} With regard to product characteristics, most smokers (>70%) in this study preferred B&M cigarillos of original flavor, consistent with published work among youth (65%) ³⁶ and adult (nearly 60% ³¹ ) smokers of various cigar product types. Still, that at least a third of smokers across samples report use of a flavored cigar product is notable. Smokers also preferred B&M cigarillos with plastic tips; no published work has reported on tip preferences among cigarillo smokers for comparison. Understanding the importance of these product features for user acceptability and their effects on smoking behaviors can ultimately be used for product regulatory actions.

The most popular other tobacco product used was cigarettes. Work conducted by our group ³⁷ and others ^38–40 has shown high rates of dual cigarette-cigar use, relative to dual use of cigarettes and other tobacco products (eg, smokeless tobacco, hookah). Likewise, up to 60% of sampled cigar smokers also reported cigarette use. ^{3 , 27 , 29 , 34} Also popular among our sample of cigar smokers was the use of marijuana. Nearly half of respondents reported concurrent cigar and marijuana use, consistent with a rate of 45.9% among surveyed college students aged 18–25 years. ³⁴ Still, these prevalence rates are higher than those found in other studies (>25% ³ ). Also consistent with previous data is the observation that many cigar users use these products to smoke blunts (current study = 55.3%; >33% ³ ). There has been discussion over whether blunting constitutes concurrent cigar and marijuana use or whether it signifies marijuana use alone. ^41–43 At the least, this cigar modification practice has important implications for understanding the health effects of cigars and the impact of point of sale regulations. It is unclear what proportion of cigar users in the current study used cigars for blunting versus cigar-only smoking.

Three (Daily-Hypers, Daily-Flavored, and Heavy-Daily-Hypers) of the five classes consisted of daily B&M smokers. Daily-Hypers and Heavy-Daily-Hypers classes of daily smokers were similar in that both generally preferred original flavored, plastic tipped B&M cigarillos and endorsed hyping. Heavy-Daily-Hypers members, however, were older and smoked a higher number of B&Ms per day compared to Daily-Hypers members. Heavy-Daily-Hypers members also had a higher rate of blunt use versus Daily-Hypers members (eg, 23% vs. 16% endorsed “most or all of the time”). While the percentage of Heavy-Daily-Hypers members reporting marijuana use was lower than that for Daily-Hypers members, these findings may suggest that Heavy-Daily-Hypers members who smoke marijuana do so primarily via cigar modification. Alternatively, Heavy-Daily-Hypers members may use their B&M cigarillos as a “blunt chaser,” or a tobacco product smoked after blunt use. ⁴⁴ Disentangling the use of cigarillos for cigar versus marijuana use is an important area for future work as it also may impact prevalence estimates of cigar smoking. ^{43 , 45}

Daily-Flavored class members were also daily B&M smokers, though they preferred flavored cigarillos, had more variability in tip preference versus other groups, and did not endorse hyping. The other two classes (Non-Daily-Hypers and Non-Daily-Flavored) were nondaily B&M smokers. While Non-Daily-Hypers members preferred original flavored products, plastic tips, and hyping, Non-Daily-Flavored members preferred flavored, wood tips, and no hyping. Studies have reported that specialty-flavored, rather than original flavored, cigars may be used for blunting in effort to mask the smell of marijuana. ⁴⁶ However, in our sample, blunting was no more likely among classes that endorsed flavored B&M use, relative to those that endorsed original flavored. Of course, participants in our study may use brands other than B&M cigarillos for blunt use, and differences in blunt use for those brands may differ based on product flavorings. While the potential benefits of tip type—plastic, wood, or none—on cigarillo smoking are unknown, those classes that preferred original flavored cigarillos and plastic tips, either daily or nondaily users, were those that also hyped B&Ms. Preferences for plastic tips may be due to the logistics of hyping in that wood tipped B&Ms are more difficult to pry apart to remove the paper lining. Granted, banning both tip types may not decrease the practice of hyping, as untipped cigarillos are likely able to be modified for this same purpose. However, respondents in this study were not given the option of “none.”

Flavor may be a main factor that contributes to B&M use and hyping behavior. Tobacco product characteristics are often manipulated by the industry in an effort to increase their appeal. For instance, flavor additives promote product use among novices by masking the taste of tobacco and facilitating smoke inhalation. ⁴⁷ Objective data confirm that, at least among regular users of smokeless tobacco products like dip and chew, the majority report initiating use with a flavored product like mint or wintergreen. ⁴⁸ A similar trajectory may occur for users of flavored cigar products. The flavoring and hyping association may help explain why hyping-related harm reduction beliefs did not systematically vary between latent classes. It may be that hyping is specifically used to enhance the flavor of original tobacco flavored B&Ms rather than to decrease the potential harm. This premise is supported by the results of our previous qualitative investigation ²² where themes related to hyping included “making it easier to smoke” and “enhancing flavor.” These latter beliefs may be better covariates that explain hyping behavior patterns than those related to harm reduction.

Multinomial regression results showed one significant difference between Daily-Hypers (the largest class) and other latent classes for demographic characteristics, cigar harm perceptions, or other tobacco behavior predictors. Heavy-Daily-Hypers members were significantly older than Daily-Hypers members on average by about 2 years. The total sample and smaller class sizes for all B&M smokers other than Daily-Hypers may have influenced our ability to detect differences between these groups. Given that we used a convenience sampling approach, additional work is needed to determine whether these class sizes accurately represent the population of B&M cigarillo smokers.

Limitations

While most characteristics identified in our sample are consistent with those described in previous work, the extant literature rarely differentiates between cigar types (large cigars, small cigars, cigarillos). That is, cigar use has typically been measured via generic items (ie, “cigar” not operationalized ³² ) or an aggregate item (ie, “little cigars/cigarillos/bidis” ^{30 , 35} ). Consequently, observed differences between samples may be due to differences in the criteria used to define the sample and/or the measurements (including our measure of blunting) used to assess product use. Of course, differences between studies may also be a product of our convenience sampling of “B&M smokers.” Moreover, given that the large majority of participants identified as African American males, study results may not generalize to B&M smokers of other races or to those who are female. Notable, however, is survey work that shows cigar use is more prevalent among males than females, ^{37 , 38 , 49} as well as among non-Hispanic African Americans relative to other racial/ethnic groups. ^{27 , 50}

Conclusions

Despite these limitations, this study fills a gap in the literature by being the first to characterize African American B&M cigarillo smokers. More importantly, by shifting from the paradigm of African American B&M smokers as a monolithic group, it allows an examination of how subgroups within this population differ in terms of important cigar smoking behaviors and demographic characteristics that may place these smokers at an even greater risk for tobacco-related morbidity. Understanding these differences will permit public health practitioners to develop tailored prevention and treatment strategies that account for the heterogeneity within these different groups of tobacco users. For instance, interventions that are informed by a greater understanding of the characteristics of African American B&M users may be able to dispel harm reduction beliefs associated with B&M use and modification behaviors such as hyping. Additionally, interventions may help African American B&M users find cessation products that satisfy their flavor preferences such as using flavored nicotine replacement therapy products. These improvements to current tobacco cessation interventions may help improve cessation rates among African Americans. Though this study provides useful data on B&M use among African Americans, more research is needed to provide the FDA information about how best to regulate flavored cigars.

Funding

This work was supported by the National Cancer Institute at the National Institutes of Health (R21CA161317 to AN and TE). Additionally, this work was supported by the National Institute on Drug Abuse at the National Institutes of Health (P50DA036105 to AN, MCG, EKS, COC, TE) and the Center for Tobacco Products of the US Food and Drug Administration. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health or the Food and Drug Administration.

Declaration of Interests

None declared.