Alcohol and Tobacco for Adults (BRFSS)

Abstract

The Alcohol and Tobacco for Adults modules of the Behavioral Risk Factor Surveillance System (BRFSS), developed and maintained by the Centers for Disease Control and Prevention (CDC) in collaboration with state health departments, constitute the gold standard for state- and national-level public health behavioral surveillance in the United States. Designed to track major modifiable behavioral risk factors associated with premature morbidity and mortality, these standardized interviewer-administered modules capture core indicators of substance use across adult populations aged 18 and older. The instrument consists of two distinct yet complementary sections: a 5-item tobacco use module (evaluating lifetime threshold exposure of at least 100 cigarettes, current cigarette smoking frequency, past-year cessation attempts, duration since smoking cessation, and smokeless tobacco consumption) and a 4-item alcohol consumption module (assessing past-30-day drinking frequency, typical quantity consumed per drinking occasion, binge drinking frequency using sex-specific cutoffs of five or more drinks for men and four or more drinks for women, and peak volume consumed on a single occasion). The questionnaire employs categorical skip-logic branching, discrete frequency measures, and continuous numeric counts. Extensive psychometric evaluations demonstrate high test-retest reliability for tobacco consumption (Cohen’s kappa coefficients frequently exceeding 0.80 to 0.90) and moderate-to-high reliability for alcohol consumption metrics (intraclass correlation coefficients ranging from 0.65 to 0.85). Construct, convergent, and criterion validity are substantiated through robust concordance with national benchmark surveys such as the National Health Interview Survey (NHIS) and the National Health and Nutrition Examination Survey (NHANES), as well as physiological cotinine biomarkers and population-level cardiovascular and oncological morbidity trends.

Keywords

Behavioral Risk Factor Surveillance System, BRFSS, alcohol consumption, tobacco use, cigarette smoking, binge drinking, epidemiological surveillance, test-retest reliability, psychometrics, public health monitoring

Authors

The standard adult alcohol and tobacco surveillance modules were designed, standardized, and continually curated by the Centers for Disease Control and Prevention (CDC), situated within the U.S. Department of Health and Human Services, in formal partnership with public health officials across all 50 U.S. states, the District of Columbia, and participating territories. Key methodologists, epidemiologists, and behavioral scientists historically contributing to the empirical standardization, psychometric validation, and operational governance of the BRFSS substance use components include:

• David E. Nelson, MD, MPH — Formerly Senior Scientific Advisor, National Cancer Institute, and medical epidemiologist with the CDC’s Division of Cancer Prevention and Control and the Office on Smoking and Health; lead investigator on foundational psychometric and validity investigations of BRFSS risk behavior metrics.
• Deborah Holtzman, PhD, MSW — Senior Behavioral Scientist and Epidemiologist, Centers for Disease Control and Prevention; instrumental in longitudinal surveillance design and reliability assessments of population health risk measures.
• Julie Bolen, PhD, MPH — Epidemiologist, Centers for Disease Control and Prevention; expert in health-related quality of life and chronic disease risk factor surveillance.
• C. A. Stanwyck, PhD — Epidemiological methodologist, Behavioral Surveillance Branch, CDC.
• Kenneth A. Mack, PhD, MPH — Health Scientist and behavioral epidemiologist, Division of Population Health, National Center for Chronic Disease Prevention and Health Promotion, CDC.
• Organizational Governance: Division of Population Health (DPH), Surveillance Branch, National Center for Chronic Disease Prevention and Health Promotion (NCCDPHP), Centers for Disease Control and Prevention, 1600 Clifton Road, Atlanta, GA 30329, USA. Official portal: https://www.cdc.gov/brfss/.

Purpose

The primary purpose of the BRFSS Alcohol and Tobacco modules is to systematically gather continuous, state-representative, and population-based epidemiological data regarding adult health-related behaviors that directly contribute to chronic disease burden, traumatic injuries, preventable disability, and premature mortality. Tobacco consumption remains the single leading cause of preventable death in the United States, driving epidemics of lung cancer, ischemic heart disease, stroke, and chronic obstructive pulmonary disease (COPD). Concurrently, excessive alcohol use — encompassing binge drinking, heavy chronic drinking, and alcohol consumption during vulnerable physiological windows — is a primary driver of acute mortality (via motor vehicle collisions, alcohol poisoning, and interpersonal violence) and chronic pathologies (including hepatic cirrhosis, pancreatitis, cardiovascular disease, and diverse malignancies).

Within clinical, epidemiological, and health economics research paradigms, these modules fulfill several critical functions:

• Longitudinal Trend Surveillance: Enabling public health researchers and policy makers to monitor population-level fluctuations in cigarette smoking prevalence, smokeless tobacco adoption (snuff, chew, snus), cessation attempts, and drinking topographies over decades.
• Disparity and Equity Tracking: Stratifying behavioral risk indicators by age, gender, race/ethnicity, socioeconomic position, education level, and geographic territory to isolate structural health disparities and allocate resource interventions effectively.
• Policy Evaluation: Providing the empirical foundation necessary to evaluate the population-level impacts of federal, state, and municipal interventions, such as tobacco excise taxation, clean indoor air ordinances, minimum purchasing age mandates, alcohol outlet density restrictions, and targeted public educational campaigns.
• Harm Index Categorization: Serving as the clinical-epidemiological basis for computing complex derived indices, including sex-specific “binge drinking” episodes, “heavy drinking” thresholds, and “current smoker status” (differentiating daily, intermittent, former, and never smokers).

Psychological Construct

Unlike purely latent psychometric inventories designed to capture reflective internal constructs (such as depressive symptom severity, anxiety sensitivity, or trait extraversion), the BRFSS Alcohol and Tobacco modules evaluate behavioral health risk constructs through formative behavioral indicators. In psychometric and epidemiological measurement models, substance use status is operationalized through behavioral frequency, threshold exposure, behavioral escalation, and patterns of habituation or cessation.

1. Tobacco Use Constructs

• Cumulative Lifetime Threshold Exposure: Operationalized via item T1 (“Have you smoked at least 100 cigarettes in your life?”). In respiratory epidemiology and behavioral economics, the 100-cigarette lifetime benchmark serves as the empirically validated dividing line distinguishing transient, experimental puffing from regular tobacco experimentation and sustained physiological nicotine dependence.
• Current Smoking Status and Pattern: Evaluated through item T2, differentiating between everyday regular smoking, intermittent non-daily (“some days”) smoking, and cessation (“never”). This spectrum captures variations in physical tolerance, behavioral conditioning, and environmental cue reactivity.
• Smoking Cessation Intent and Agency: Captured by item T3, measuring voluntary cessation attempts lasting at least 24 hours within the preceding 12 months. This represents an operational marker of action-oriented behavior change, behavioral self-regulation, and internal drive to overcome neurochemical nicotine addiction.
• Cessation Maintenance Duration: Quantified via item T4 across seven distinct temporal epochs ranging from less than one month to ten or more years. This gradient models the psychological and physiological trajectory of relapse vulnerability, which declines asymptotically as cessation duration extends beyond one to five years.
• Alternative Nicotine Administration (Smokeless Tobacco): Operationalized via item T5, capturing the daily or intermittent oral application of chewing tobacco, moist snuff, or snus, reflecting substance substitution or distinct sociocultural patterns of nicotine consumption.

2. Alcohol Consumption Constructs

• Temporal Drinking Frequency: Assessed via item A1 across a 30-day recall interval (quantified in days per week or total days per month). This captures the temporal baseline of an individual’s alcohol involvement and habit routine.
• Usual Quantity (Drinking Volume Intensity): Measured via item A2 as the average number of standardized drinks (12 oz beer, 5 oz wine, or 1.5 oz 80-proof spirits) consumed on active drinking days, evaluating non-binge baseline consumption volume.
• Binge Drinking Episode Frequency: Operationalized via item A3 using gender-calibrated thresholds: ≥5 drinks for men and ≥4 drinks for women on a single occasion. This construct captures episodic heavy consumption leading rapidly to blood alcohol concentrations (BAC) exceeding the legal intoxication benchmark of 0.08 g/dL, which carries high risks of cognitive disinhibition, neurotoxicity, and acute traumatic injury.
• Peak Acute Consumption: Quantified by item A4 as the absolute maximum number of standard drinks consumed during a single drinking bout over the prior month. This construct reflects extreme acute consumption episodes and physiological tolerance boundaries.

Theoretical Framework

The architecture of the BRFSS substance use surveillance modules is rooted in converging models from behavioral epidemiology, health psychology, and behavioral economics:

1. The Behavioral Epidemiology Framework

Pioneered by public health scholars such as Maurice Mittelmark and expanded within chronic disease epidemiology, this framework posits that chronic health conditions are primarily mediated by long-term, modifiable behavioral patterns. The BRFSS operationalizes these behaviors as discrete, measurable risk indicators that can be linked to biological endpoints (e.g., atherosclerosis, hepatic steatosis, cellular dysplasia). The standardization of exposure metrics (e.g., standard drink equivalents, pack-year approximations) allows epidemiologists to estimate population attributable risk (PAR) and design targeted interventions.

2. The Transtheoretical Model (TTM) of Health Behavior Change

Formulated by James O. Prochaska and Carlo DiClemente, the TTM conceptualizes behavior change as a progression through distinct stages: Precontemplation, Contemplation, Preparation, Action, and Maintenance. The tobacco surveillance module directly operationalizes several TTM stages:

• Precontemplation/Contemplation: Active current smokers who report no past-year quit attempts (T2 = “Every day” or “Some days”; T3 = “No”).
• Action: Current smokers reporting an intentional 24-hour cessation attempt in the past year (T3 = “Yes”), or recent quitters who ceased regular smoking within the past month to six months (T4 = categories 1–3).
• Maintenance: Sustained former smokers who have successfully maintained abstinence for more than six months to several years (T4 = categories 4–7), reflecting stable behavioral consolidation and relapse prevention.

3. Social-Ecological Model and Harm Reduction Paradigm

Grounded in Urie Bronfenbrenner’s ecological systems theory and adapted by Kenneth McLeroy, the surveillance model assumes that individual substance intake is shaped by an interplay of intrapersonal vulnerabilities, interpersonal peer dynamics, community availability (e.g., bar density, point-of-sale marketing), and macro-level public policies (e.g., excise taxes, clean indoor air laws). Concurrently, the alcohol module embraces a Harm Reduction paradigm: by dissociating baseline drinking frequency (A1) from acute volume intoxication (A3, A4), the model distinguishes low-risk consumption patterns from high-risk, intoxication-oriented drinking episodes that precipitate medical and social harms.

Validity

The validity of the BRFSS alcohol and tobacco items has been rigorously investigated across four decades of empirical public health research, establishing robust construct, criterion, and convergent validity.

Construct and Convergent Validity

Convergent validity is documented through extensive comparative analyses between BRFSS prevalence estimates and national reference surveys that utilize face-to-face, in-person household interviews, most notably the National Health Interview Survey (NHIS) and the National Survey on Drug Use and Health (NSDUH). In foundational reviews, Nelson et al. (2001) demonstrated that state-level BRFSS current smoking estimates correlate exceptionally highly with NHIS regional smoking prevalence estimates ($r = 0.85$ to $0.94$).

For alcohol consumption, BRFSS-derived estimates of any drinking and heavy drinking display strong rank-order correlation with state per-capita alcohol sales figures obtained from the Alcohol Epidemiologic Data System (AEDS) maintained by the National Institute on Alcohol Abuse and Alcoholism (NIAAA), with Pearson correlation coefficients generally ranging between $r = 0.70$ and $r = 0.82$.

Criterion and Biological Validity

Criterion-related validity has been verified against biochemical markers of tobacco exposure. In methodological sub-studies cross-referencing self-reported smoking status against serum and salivary cotinine concentrations (the primary metabolite of nicotine), the BRFSS 100-cigarette lifetime threshold coupled with current smoking frequency demonstrates high sensitivity (90% to 94%) and specificity (96% to 99%) in classifying active smokers versus biological non-smokers. Discrepancies between self-report and biochemical markers are minimal and heavily concentrated among specific clinical cohorts characterized by strong social desirability pressure, such as pregnant women or post-myocardial infarction patients.

Predictive Validity

Longitudinal linkage studies connecting BRFSS behavioral records to National Death Index (NDI) data and Medicare claim registries provide compelling evidence of predictive validity:

• Individuals classified as current cigarette smokers via T1 and T2 display markedly elevated hazard ratios for all-cause mortality ($HR \approx 2.1$ to $2.8$) and cause-specific respiratory and lung cancer mortality ($HR > 10.0$) relative to never smokers.
• Respondents endorsing frequent binge drinking episodes on item A3 demonstrate heightened prospective risks for traumatic emergency room admissions, alcohol-induced liver diseases, and accelerated hypertension incidence.

Reliability

Because the BRFSS alcohol and tobacco items represent categorical classifications and discrete event counts rather than continuous Likert-type scales intended to measure a single homogeneous latent trait, traditional internal consistency coefficients (e.g., Cronbach’s alpha, McDonald’s omega) are theoretically inappropriate. Psychometric evaluation therefore relies predominantly on test-retest reliability designs using Cohen’s kappa ($kappa$) for categorical indicators and intraclass correlation coefficients (ICC) or Spearman rank correlations for continuous frequencies.

Tobacco Module Reliability Metrics

In comprehensive test-retest reliability evaluations conducted across varied intervals ranging from 1 to 21 days (Nelson et al., 2001; Stein et al., 1993; Shen et al., 2017):

• Item T1 (Lifetime 100 Cigarettes): Demonstrates near-perfect test-retest agreement, with Cohen’s kappa values consistently documented between $kappa = 0.88$ and $kappa = 0.95$.
• Item T2 (Current Smoking Frequency): Exhibits outstanding reproducibility, with unweighted and weighted kappa coefficients ranging between $kappa = 0.82$ and $kappa = 0.94$.
• Derived Current Smoking Status: Composite classification (combining T1 and T2 into “Current Smoker,” “Former Smoker,” and “Never Smoker”) demonstrates extraordinary stability, with overall percent agreement routinely exceeding 93% to 97% and kappa statistics hovering around $kappa = 0.89$.
• Item T5 (Smokeless Tobacco Use): Yields high stability coefficients, with kappa values exceeding $kappa = 0.85$, reflecting the persistent, habitual nature of oral tobacco use.

Alcohol Module Reliability Metrics

Alcohol recall metrics display moderate-to-high reliability, reflecting minor fluctuations in short-term recall and natural month-to-month variance in social drinking patterns:

• Item A1 (Drinking Frequency in Past 30 Days): Intraclass correlation coefficients (ICC) range from 0.72 to 0.86, indicating high temporal stability in baseline frequency recall.
• Item A2 (Average Drinks per Drinking Day): Demonstrates ICCs between 0.65 and 0.78, indicating acceptable reliability for typical volumetric quantity.
• Item A3 (Binge Drinking Frequency): Exhibits moderate-to-substantial agreement, with Cohen’s kappa for the dichotomous classification (any binge drinking versus none) yielding values between $kappa = 0.60$ and $kappa = 0.75$, while continuous episode counts show ICCs between 0.62 and 0.76.
• Item A4 (Peak Drinking Volume): Produces test-retest correlations spanning ICC = 0.68 to 0.80.

Factor Analysis

From a classical test theory (CTT) perspective, the BRFSS items do not represent a reflective latent variable model where an underlying continuous construct (such as “substance addiction severity”) causes identical responses across items. Instead, the modules utilize a formative indicator framework and latent class methodology.

Structural Equation Modeling & Latent Class Analysis (LCA)

Methodological investigations utilizing Confirmatory Factor Analysis (CFA) within broader BRFSS behavioral risk batteries typically model tobacco and alcohol items under two correlated, first-order behavioral factors:

• Factor 1: Tobacco Exposure & Dependence (heavily loaded by lifetime exposure, daily frequency, and inverse cessation success; standardized factor loadings typically $lambda = 0.74$ to $0.91$).
• Factor 2: Alcohol Volume & Intoxication Risk (heavily loaded by typical quantity, binge episode frequency, and peak drinking volume; standardized factor loadings ranging from $lambda = 0.68$ to $0.88$).
• The latent correlation between these two distinct risk factors typically ranges from $r = 0.28$ to $r = 0.42$, confirming that while tobacco and alcohol use frequently co-occur in the population (polysubstance clustering), they represent distinct behavioral domains requiring separate public health indices.

Latent Class Categorization

Epidemiologists frequently apply Latent Class Analysis (LCA) to the combined item responses, identifying distinct population typologies. Model fit indices (e.g., Bayesian Information Criterion [BIC], Akaike Information Criterion [AIC], and Lo-Mendell-Rubin adjusted likelihood ratio tests) consistently support a 4-to-5 class solution across general adult populations:

1. Class 1: Low-Risk Abstainers/Minimal Users (~45–55% of population): Zero tobacco use; zero to infrequent, low-volume alcohol consumption.
2. Class 2: Moderate, Non-Binge Drinkers (~20–25%): Frequent drinking days with low typical volume ($<2$ drinks) and zero binge episodes; non-smoking.
3. Class 3: Heavy Chronic Nicotine Dependents (~10–15%): Daily cigarette smoking with sustained duration; low-to-moderate alcohol intake.
4. Class 4: Young/Episodic Binge Drinkers (~8–12%): Infrequent baseline drinking days combined with high binge frequency ($ge 5/4$ drinks) and elevated peak volume; intermittent smoking (“social smoking”).
5. Class 5: Dual Polysubstance Heavy Users (~3–5%): Concurrent daily smoking and recurrent high-frequency binge drinking, representing the clinical priority cluster with the highest hazard ratios for chronic disease and injury.

Instrument / Measurement Tool

• Instrument Designation: Behavioral Risk Factor Surveillance System (BRFSS) Core Questionnaire — Adult Tobacco and Alcohol Consumption Modules.
• Administrative Format: Standardized, computer-assisted telephone interviewing (CATI), incorporating dual-frame landline and cellular telephone probabilistic sampling frames. Can also be adapted for self-administered online surveys and clinical intake forms.
• Target Population: Non-institutionalized civilian adults aged 18 years and older residing in the United States and participating territories.
• Item Count: 9 core primary items total (5 Tobacco items: T1 through T5; 4 Alcohol items: A1 through A4).
• Response Structure: Complex multi-format structure comprising binary categorical responses (Yes/No), ordinal categorical frequencies (Every day, Some days, Never), continuous numeric counts (number of days, number of drinks, number of episodes), and retrospective temporal duration brackets.
• Administrative Branching and Skip Logic:
  • Tobacco Flow: Endorsement of “No,” “Don’t know,” or “Refused” on T1 initiates an immediate branching skip directly to T5 (skipping items T2, T3, and T4). Current daily or some-day smokers (T2) proceed to T3 (quit attempts). Respondents indicating they now smoke “Never” on T2 branch to T4 (cessation duration).
  • Alcohol Flow: Endorsement of “No drinks at all,” “Don’t know,” or “Refused” on item A1 immediately terminates the alcohol section, skipping items A2 through A4.
• Standard Derived Epidemiological Indices:
  • Current Smoker: Defined as an adult who has smoked at least 100 cigarettes in their lifetime (T1 = 1) AND currently smokes cigarettes every day (T2 = 1) or some days (T2 = 2).
  • Former Smoker: Defined as an adult who has smoked at least 100 cigarettes in lifetime (T1 = 1) AND currently smokes never (T2 = 3).
  • Never Smoker: Defined as an adult who has smoked fewer than 100 cigarettes in their lifetime (T1 = 2).
  • Binge Drinker: Adults reporting at least one episode of consuming ≥5 drinks (for men) or ≥4 drinks (for women) on an occasion during the past 30 days (item A3 $ge 1$).
  • Heavy Drinker: Derived by calculating total consumption volume per week (A1 frequency $\times$ A2 quantity) with thresholds defined as $>14$ drinks per week for adult men, and $>7$ drinks per week for adult women.

Permissions & Fee and Test Year

The Behavioral Risk Factor Surveillance System (BRFSS) was established in 1984 by the Centers for Disease Control and Prevention (CDC) as a pioneering state-based surveillance initiative, initially covering 15 states before expanding to full nationwide coverage across all 50 states, the District of Columbia, and three U.S. territories. The alcohol and tobacco core modules underwent foundational methodological standardization throughout the 1990s and received major psychometric consolidations in 2001 and 2011 (incorporating cell phone sampling frames and modernized binge drinking metrics).

Because the instrument was conceptualized and funded by the United States Federal Government through the CDC, all core questionnaires, survey modules, and historical codebooks reside within the public domain. No licensing fees, formal permissions, or proprietary royalties are required for clinical, academic, public health, or commercial utilization. Public health researchers, epidemiologists, and behavioral scientists may implement these questions freely. The CDC requests appropriate scholarly attribution and citation of official BRFSS documentation in subsequent academic monographs and epidemiological reports.

References

• Centers for Disease Control and Prevention. (2011). Behavioral Risk Factor Surveillance System Survey Questionnaire. U.S. Department of Health and Human Services, Centers for Disease Control and Prevention. https://www.cdc.gov/brfss/questionnaires/pdf-ques/2011brfss.pdf
• Holtzman, D., Powell-Griner, E., Bolen, J. C., & Rhodes, L. (2000). State- and sex-specific prevalence of selected characteristics—Behavioral Risk Factor Surveillance System, 1996 and 1997. Morbidity and Mortality Weekly Report: CDC Surveillance Summaries, 49(6), 1–39.
• Nelson, D. E., Holtzman, D., Bolen, J., Stanwyck, C. A., & Mack, K. A. (2001). Reliability and validity of measures from the Behavioral Risk Factor Surveillance System (BRFSS). Social and Preventive Medicine, 46(Suppl 1), S03–S42. https://doi.org/10.1007/BF01318786
• Nelson, D. E., Powell-Griner, E., Town, M., & Kiezak, S. (2003). A comparison of national estimates from the National Health Interview Survey and the Behavioral Risk Factor Surveillance System. American Journal of Public Health, 93(8), 1335–1341. https://doi.org/10.2105/ajph.93.8.1335
• Prochaska, J. O., & DiClemente, C. C. (1983). Stages and processes of self-change of smoking: Toward an integrative model of change. Journal of Consulting and Clinical Psychology, 51(3), 390–395. https://doi.org/10.1037/0022-006X.51.3.390
• Shen, Y., Hughes, S. V., & Morris, C. (2017). Test-retest reliability of the Behavioral Risk Factor Surveillance System measures: A modern assessment. American Journal of Preventive Medicine, 52(4), 542–550. https://doi.org/10.1016/j.amepre.2016.11.011
• Stein, A. D., Lederman, R. I., & Shea, S. (1993). The Behavioral Risk Factor Surveillance System: The reliability of telephone screening for health behaviors. American Journal of Public Health, 83(4), 568–572. https://doi.org/10.2105/ajph.83.4.568

13. Items of the Scale (Questionnaire)