Trajectories of hookah use among young adult college students in Texas, 2014–2019

Abstract

The purpose of the study was to examine trajectories in hookah use behaviors across young adulthood and socio-demographic correlates of the trajectories. Data were drawn from Project M-PACT for the years 2014–2019. Participants were 5,220 college students aged 18 to 25 years at baseline. Growth curve models with an accelerated design were fit to examine four hookah use trajectories across ages 18 to 30, adjusted for covariates. Prevalence of current hookah use decreased between 2014 and 2019, from 17% to 7% and the frequency of use decreased by almost one day a month. Trajectories of hookah use alone and concurrent with cigarettes or with e-cigarettes also decreased as students increased in age. Current hookah users were likely to be young, male, racial/ethnic minority, and attending a four-year college. Observed disparities among hookah users have implications, in particular regarding the role of social context, race, and ethnicity, in tailoring smoking interventions.

Introduction

While hookah has been used in India, the Middle East, and North Africa for centuries, it has grown in popularity in the United States over the last few decades. The rise can be traced back to the late 20th and early 21st centuries, when hookah use gained significant traction among young adults and college students.^1-4 Data on young adults (aged 18–25 years) indicate that current (past 30 days) hookah use increased from 7.8% in 2010 to 18.2% in 2014, whereas ever hookah use increased from approximately 28% in 2010 to 44% in 2014.^1,2

Studies suggest that hookah smokers inhale larger amounts of toxic agents than cigarette smokers during a single smoking session.^5,6 The relatively long duration of a hookah smoking session (30–90 min/session) results in greater carbon monoxide (CO) and effective nicotine exposure. Notably, peak hookah carboxyhemoglobin (COHb) levels in hookah smokers were three times higher compared to cigarette smokers.⁶ Elevated COHb levels are linked to acute and long-term health risks associated with hookah smoking, mirroring many of the health risks of cigarette use. The high intake of CO and other harmful substances during hookah use contributes to an increased risk of infections, cardiovascular and respiratory diseases, and cancers of the lung, esophagus, and bladder.⁷ Furthermore, among college students, hookah use is associated with cigarette, alcohol, and marijuana use.⁸

Among middle and high school students, correlates of hookah use include male sex, white race, younger age, and current cigarette smoking.⁹ Hookah use increases during the transition to college¹⁰ and risk of use also changes during this transition. Recent research among college students suggests that younger, ethnic minorities and males are at increased risk for hookah use.¹¹ Despite the high prevalence and adverse health consequences of hookah use, little is known about how patterns of hookah use behaviors change over time among college students. The current study seeks to examine trajectories in four hookah use behaviors (current use, concurrent use with cigarette, concurrent use with e-cigarette, and frequency of use) across young adulthood, ages 18–30 years, and to determine socio-demographic correlates of these behavioral trajectories, among an ethnically diverse sample of college students in Texas.

Method

Participants and study design

Data were drawn from the Marketing and Promotions Across Colleges in Texas Study (Project M-PACT), a multi-wave prospective study designed to capture the changes in tobacco use behaviors among Texas college students. Participants were part- and full-time college students, between 18 and 30 years at wave 1, who were recruited from one of 24 Texas colleges from Austin, Dallas/Fort Worth, Houston, and San Antonio in fall 2014 and spring 2015. Eligible students provided informed consent before joining the study. Participants completed an online survey, with six months between each of the first seven waves and one year between the last two waves. Details of recruitment and retention procedures have been published.¹² The current analysis includes data from all nine waves. Participants over the age of 25 years at baseline were excluded, resulting in a sample of 5,220 of the 5,482 participants enrolled at baseline. Project MPACT was approved by the University of Texas at Austin Institutional Review Board.

Main outcome measures

Four hookah-related outcomes were assessed at all nine waves. Current hookah use was assessed by, “During the past 30 days, have you smoked a hookah as intended (i.e., with tobacco)?” Participants who used hookah in the past 30 days were current hookah users (coded “1”), all others were non-current hookah users (coded “0”). Frequency of hookah use was assessed by “On how many of the past 30 days have you smoked a hookah as intended (i.e., with tobacco)?” with response range: 0–30 days. Current cigarette use was assessed by “On how many of the past 30 days did you smoke cigarettes?” and current e-cigarette use was assessed by “During the past 30 days, have you used any ENDS product, (i.e., an e-cigarette, vape pen, or e-hookah), even one or two puffs, as intended (i.e., with nicotine cartridges and/or e-liquid/e-juice)?” For both cigarettes and e-cigarettes, those who indicated any use in the past 30 days were considered current users (coded “1”). Participants who were coded as “1” for both current cigarette and current hookah use were labeled as concurrent hookah and cigarette users. Participants who were coded as “1” for both current e-cigarette and current hookah use were labeled as concurrent hookah and e-cigarette users.

Covariates

Time-invariant covariates, included to adjust for confounding influences, were age at baseline, sex coded, “0” (female)/“1” (male), race/ethnicity, dummy-coded (for Hispanic/Latino, non-Hispanic Black, non-Hispanic Asian, and other; non-Hispanic White was the reference group), and type of college attended, coded “0” (2-year)/“1” (4-year).

Analytic methods

To examine trajectories in (1) current hookah use, (2) concurrent hookah use with cigarettes, (3) concurrent hookah use with e-cigarette use, and (4) frequency in hookah use across young adulthood, multi-level growth curve models (GCM) were fit using the STATA MELOGIT and MIXED procedure. All models included random intercepts for participants and the college that they attended to allow for nesting of repeated measures within participants and participants within their colleges. An accelerated longitudinal design was applied, so that participant age, not wave, was used to model the trajectories and each participant contributed up to 4.5 years of data. Participants were 18–25 at wave 1 and 23–30 at wave 9. Thus, each trajectory reflects the fitted trend for the entire age range in the sample, 18–30 years, based on the contribution from participants overlapping age ranges.

First, unconditional growth models were fit (i.e., a model without covariates) with time-varying age as the only predictor to identify the optimal function form of each of the four trajectories across increasing age.¹³ Then, the three unconditional growth models: linear age, quadratic age, and log-transformed age were compared using Akaike information criterion (AIC). The linear age effect in the unconditional growth model for current hookah use and concurrent hookah use with cigarette outcomes yielded the lowest AIC. The log-transformed age effect for concurrent hookah use with e-cigarettes outcome yielded the lowest AIC, while the quadratic age effect yielded the lowest AIC for frequency.

A sensitivity analysis showed no substantive differences in the interpretation of the linear age effect and either the log-transformed age effect or the quadratic age effect. Thus, linear models were estimated for all four trajectories (current use, concurrent use with cigarette, concurrent use with e-cigarette, and frequency of use) to be consistent. Each of the conditional growth models included the following covariates: age at baseline, participant sex, race/ethnicity, and type of college attended. Models are based on maximum likelihood estimated using adaptive Gaussian quadrature, which allows all time points with complete data to be used; thus, time points are included from participants who did not provide data at every time point.

Results

At baseline, 58% of participants were women, 95% attended a 4-year college; and 34% were Hispanic, 33% non-Hispanic White, 17% non-Hispanic Asian, 9% non-Hispanic Other, and 7% were non-Hispanic African American. There were 887 participants who reported current hookah use at baseline; 45% reported concurrent hookah use with cigarettes and 50% reported concurrent use with e-cigarettes. Current and concurrent hookah use declined from 2014 to 2019. In 2014, 17% of students were current hookah users compared with 7% in 2019 (data not shown).

Results from the unconditional model (data not shown) revealed that trajectories of current hookah use, concurrent use with cigarettes and/or e-cigarettes, and frequency of use decreased from ages 18–30 years. Estimates from the conditional models corroborated the decreases (see Figure 1 and Table 1) and revealed that males were more likely than females to use hookah, with a cigarette and/or e-cigarette. Four-year college students were more likely to use hookah and concurrently use hookah with e-cigarettes than were two-year students. Compared to non-Hispanic White students, college students of all other racial/ethnic groups were more likely to report current hookah use and to use hookah on more days. Hispanic and non-Hispanic Other college students were more likely to use hookah concurrently with cigarettes and/or e-cigarettes compared to non-Hispanic White students.

Figure 1.

Mean number of days of hookah use across young adulthood, Texas, 2014–2019. Note: Models adjusted for participant baseline age, sex, race/ethnicity, and college type (2-y vs 4-y).

Table 1.

Parameters estimates from conditional growth curve models for current hookah, concurrent use with cigarettes, concurrent use with electronic cigarettes and frequency of hookah use, 2014–2019.

	Current use OR (95% CI) (n = 5,218)	Frequency of use B (95% CI) (n = 1,807)	Concurrent with cigarettes OR (95% CI) (n = 5,218)	Concurrent with E-cigarettes OR (95% CI) (n = 5,218)
Intercept	0.05 (0.02, 0.13)	3.94 (1.60, 6.29)	0.004 (0.001, 0.017)	0.05 (0.02, 0.18)
Current age	0.71 (0.69, 0.74)	−0.33 (−0.43, −0.23)	0.69 (0.66, 0.73)	0.68 (0.65, 0.72)
Baseline age	1.35 (1.28, 1.43)	0.32 (0.18, 0.48)	1.46 (1.34, 1.59)	1.32 (1.23, 1.41)
Sex (male)	1.49 (1.26, 1.77)	0.32 (−0.06,0.72)	2.41 (1.88, 3.10)	1.58 (1.28, 1.94)
4-Year college	1.71 (1.10, 2.66)	−0.63 (−1.50, 0.24)	1.38 (0.73, 2.59)	1.79 (1.02, 3.14)
NH-White	ref	ref	Ref	Ref
Hispanic	1.47 (1.19, 1.82)	−0.15 (−0.64, 0.34)	1.44 (1.05, 1.95)	1.39 (1.07, 1.80)
NH-AA	2.03 (1.46, 2.80)	0.32 (−0.39, 1.03)	0.59 (0.32, 1.07)	1.34 (0.89, 2.02)
NH-Asian	1.64 (1.28, 2.09)	0.93 (0.36, 1.49)	1.17 (0.80, −1.70)	1.38 (1.02, 1.88)
NH-Other	2.22 (1.61, 3.05)	0.50 (−0.22, 1.21)	1.58 (0.98, 2.54)	1.84 (1.24, 2.72)

NH = Non-Hispanic; AA = African American; CI = confidence interval; OR = odds ratio. Note: Models adjusted for participant baseline age, sex, race/ethnicity, and college type (2-year vs 4-year).

Discussion

Overall, prevalence rates for current hookah use among college students in Texas decreased between 2014 and 2019, from 17% to 7%. In contrast, national data suggest that the prevalence of hookah use among college students increased, from 5.2% in 2015 to 6.2% in 2018.14 The prevalence observed in Texas is higher than the national rate and underscores the need for targeted interventions to prevent use and increase cessation among users. Results from our study also show that the trajectories of hookah use alone or concurrent with cigarettes and/or e-cigarettes, and the frequency of hookah use decreased significantly as young adults increased in age. These findings are consistent with research showing developmental declines in all types of substance use as young adults increase in age.¹⁵ Concomitantly, declining use may reflect that hookah has become a more prominent focus of public policy and intervention programs since the amendment to the Family Smoking Prevention and Tobacco Control Act in 2016, which brings hookah tobacco products under the US Food and Drug Administration’s regulations for packaging and labeling, health warning statements, and the disclosure of product constituents.¹⁶

Of concern, some individuals remained at increased risk. Among college students in Texas, current hookah users are likely to be non-White young males who attend a four-year college. This finding is consistent with more recent research among college students in the US, regardless of where they attend college.¹² Shreya and Berg ¹² report that non-Hispanic Black students use hookah at higher rates than students of other ethnic groups. Others report a higher percentage of Hispanic and non-Hispanic Black college/Caribbean West Indian students reported current hookah use compared to non-Hispanic White and non-Hispanic Asian students.^3,4 Among a sample of college students attending an HBCU in rural Maryland, young Black males were at highest risk for hookah use.¹⁷

The popularity of hookah smoking among college students may in part be driven by the lower legal age and greater accessibility of hookah products, compared with cigarettes, as well as the attractive flavors used in the tobacco mixtures, the misperceptions that it is safer and contains less nicotine than cigarettes.¹⁸ Active socializing while smoking hookah (e.g., among friends in popular social settings/hookah lounges) and the good taste/smell of the tobacco encourage and maintain hookah use among youth and young adults.^4,19 The overall decrease in prevalence rates underscores the possibility that positive beliefs and misperceptions about hookah have shifted over time as people become more aware of the potential health consequences. However, sub-groups of college students, such as fraternity and sorority members, remain vulnerable to hookah use, in part because of their group identity, shared norms, attitudes, and beliefs that facilitate the use of the product.²⁰ Thus, these preventive interventions need to address multi-level risk factors in various domains, especially at personal, social, and environmental levels. For example, organizing smoke-free community events or anti-hookah campaigns with the involvement of local leaders, influencers, and celebrities can effectively discourage hookah smoking, especially among younger individuals who are often influenced by their peer groups.

Strengths and limitations

This is one of the first studies to prospectively assess hookah use behaviors among an ethnically diverse sample of college students. However, there are limitations. First, self-reported data were used; students may have under-reported hookah smoking. Second, the convenience sample limits the generalizability of the results; the prevalence rates are based on data from students attending colleges in the largest metropolitan areas in Texas and should be generalized with caution to other populations. Third, while these data were gathered from 2014 to 2019, the results remain relevant as they identify demographic risk factors associated with hookah use when usage was at its peak. Thus, future research can build on this work to explore changes and trends both in who uses hookah as well as the prevalence of use over time. Finally, current hookah use was assessed as “at least one day in the past 30 days,” which did not differentiate the frequency and intensity of hookah use per session or day. Future research should use better measures of patterns of hookah use.

Conclusion

Although prevalence rates for hookah use decreased in Texas from 2014 to 2019, and exclusive and concurrent use of hookah with cigarettes or e-cigarettes declined from ages 18–30 years, some individuals remain at increased risk. Among college students in Texas, consistent with other locations in the United States, certain demographic groups, namely younger males, ethnic minorities, and those attending four-year colleges, are identified as remaining at heightened risk for current hookah usage. These disparities could be due to a mix of social, cultural, normative (i.e., peer influence), and socioeconomic factors. For example, younger males might be more likely to use hookah as part of social gatherings due to peer influence. Additionally, unique stressors and cultural influences affect minority groups, potentially leading to higher usage rates. Addressing these observed disparities requires tailored public health interventions that consider the varied influences on hookah smoking behavior.²¹

Data availability statement

The data supporting this research are available from Dr. Loukas Alexandra on reasonable request.