National Household Survey on Drug Abuse/Age of First Use

Abstract

The National Household Survey on Drug Abuse/Age of First Use (NHSDA-AFU) scale is a standardized, epidemiological self-report battery designed to ascertain the chronological age at which an individual first initiates the consumption of licit and illicit substances. Developed under the auspices of the Substance Abuse and Mental Health Services Administration (SAMHSA) and the Office of Applied Studies (OAS) in conjunction with the Research Triangle Institute (RTI), the instrument constitutes a foundational core module of the United States federal drug surveillance apparatus, later institutionalized within the National Survey on Drug Use and Health (NSDUH) and the Center for Substance Abuse Prevention (CSAP) Core Measures Initiative. Comprising seven distinct substance categories—tobacco cigarettes, alcohol, marijuana or hashish, cocaine in any form, heroin, hallucinogens (LSD, PCP), and volatile inhalants—the scale utilizes a dual-format response mechanism for each item: a discrete fill-in-the-blank continuous metric capturing age in chronological years alongside a dichotomous lifetime non-use check option (“I have never used [substance] in my life”). Psychometrically, the instrument operates as a survival-analytic baseline and risk-indexing measure rather than a classical reflective sum-scale. Extensive validation literature reveals robust test-retest reliability (intraclass correlation coefficients ranging from .68 to .89 across substance classes), acceptable longitudinal stability mitigated by forward telescoping corrections, and profound predictive validity regarding secondary psychopathology, neurocognitive deficits, substance use disorder (DSM criteria), and poly-substance dependence trajectories.

Keywords

Age of First Use, NHSDA, NSDUH, SAMHSA, Substance Initiation, Epidemiology, Gateway Hypothesis, Survival Analysis, Longitudinal Risk, Psychometrics

Authors

The NHSDA-AFU was formulated and standardized by methodological teams at the Substance Abuse and Mental Health Services Administration (SAMHSA), an operating division of the U.S. Department of Health and Human Services (DHHS), in direct partnership with the Office of Applied Studies (OAS) (subsequently reorganized into the Center for Behavioral Health Statistics and Quality, CBHSQ) and the Research Triangle Institute (RTI International) located in Research Triangle Park, North Carolina. In addition, the instrument was curated and disseminated as a recommended standardized core assessment tool by the Center for Substance Abuse Prevention (CSAP) under the Core Measures Initiative Phase I Recommendations.

Principal institutional contact and administrative repositories include:

• Agency: Substance Abuse and Mental Health Services Administration (SAMHSA) / Office of Applied Studies, Rockville, MD, United States.
• Contracting Research Institution: Research Triangle Institute (RTI International), P.O. Box 12194, Research Triangle Park, NC 27709-2194.
• Program Initiative: Center for Substance Abuse Prevention (CSAP) Core Measures Initiative, State Incentive Grants (SIG) Program.

Purpose

The primary clinical, epidemiological, and theoretical purpose of the NHSDA-AFU is to systematically document the temporal onset of chemical experimentation across an individual’s developmental lifespan. Substance initiation timing has long been identified as one of the most robust single prognostic indicators of future behavioral morbidity, psychiatric comorbidity, chronic dependence, and social-occupational dysfunction. Capturing the precise chronological juncture at which an individual transitions from non-exposure to initial exposure provides vital baseline data for developmental psychopathology, public health monitoring, etiology modeling, and targeted prevention science.

From an epidemiological standpoint, calculating population-level incidence and the hazard rates of initiation across adolescence and young adulthood allows health agencies to evaluate the macroeconomic and sociological impacts of drug policy changes, public health prevention campaigns, taxation adjustments, and legislative access restrictions. For example, delaying the mean age of onset for alcohol or tobacco consumption by even 1.5 to 2 years yields measurable macro-level reductions in adult dependence prevalence, accidental vehicular deaths, and emergency department admissions.

From a clinical and developmental perspective, the purpose of obtaining accurate age-of-first-use metrics is rooted in the differential neurobiological vulnerability of the nascent central nervous system. Brain regions governing executive functioning, impulse regulation, and risk evaluation—most notably the prefrontal cortex—undergo active synaptic pruning and myelination throughout adolescence and well into the third decade of life. Chemical insults occurring during early developmental windows (e.g., ages 11–14) disrupt frontostriatal reward circuits, alter endocannabinoid and dopaminergic signaling cascades, and significantly lower the threshold for compulsive re-administration compared to identical exposure initiated in fully mature neurological phenotypes (e.g., age 22 or older). The NHSDA-AFU provides researchers with the precise chronometric anchor needed to model these developmental windows of vulnerability.

Psychological Construct

The core psychological construct measured by the NHSDA-AFU is Substance Initiation Timing (chronological age at initial experimental encounter), operating as a primary behavioral proxy for risk propensity, developmental deviance, and neurobehavioral disinhibition. Unlike continuous psychological traits measured via multi-item reflective Likert batteries (e.g., neuroticism or generalized anxiety), substance onset is an irreversible, right-censored event-history construct embedded in developmental psychopathology.

The construct encompasses several distinct structural and behavioral dimensions:

• Licit/Gateway Ingress Onset: Measured via Item 1 (cigarettes) and Item 2 (alcohol). These substances represent socially pervasive, readily accessible legal drugs that traditionally constitute the initial tier of pharmacological experimentation. The construct reflects early environmental exposure, familial socialization, peer modeling, and low trait inhibitory control during late childhood or early adolescence.
• Cannabinoid Ingress Onset: Measured via Item 3 (marijuana or hashish). This dimension bridges legal substances and heavily stigmatized illicit compounds. Early initiation of cannabis serves as a psychometric marker for peer-group deviance, sensation-seeking personality structures, and susceptibility to altered states of consciousness prior to neurodevelopmental stabilization.
• Hard Illicit / Severe Risk Ingress Onset: Measured via Item 4 (cocaine), Item 5 (heroin), and Item 6 (hallucinogens). Initiation within these classes represents severe behavioral divergence from normative developmental trajectories. The psychological construct here reflects high behavioral undercontrol, severe externalizing pathology, extreme novelty seeking, and frequently adverse childhood experiences (ACEs) or self-medication phenotypes addressing severe trauma.
• Opportunistic Household/Solvent Onset: Measured via Item 7 (inhalants). Inhalant initiation reflects an opportunistic, high-risk exploratory construct typically observed among younger adolescents (ages 10–14) who lack institutional or economic access to conventional substances. Inhalant experimentation is strongly correlated with executive dysfunction, conduct disorder, and heightened vulnerability to neurotoxic pathology.

Additionally, each dimension incorporates a crucial categorical counter-state: lifetime abstinence. The presence of the right-censored option (“I have never used [substance] in my life”) ensures that the construct captures both the timing of initiation and complete behavioral avoidance, preserving empirical fidelity in longitudinal survival analyses.

Theoretical Framework

The design and interpretation of the NHSDA-AFU are rooted in three major theoretical frameworks within developmental psychopathology, behavioral sociology, and addiction neuroscience:

1. The Stage Theory of Drug Involvement (Gateway Hypothesis)

Formulated extensively by Denise Kandel (1975, 2002), the Stage Theory of Drug Involvement posits that substance consumption follows a regular, progressive, and developmental sequence. According to this model, individuals rarely initiate illicit substance use without prior experience with legal or culturally normalized psychoactive agents. The typical sequence originates with beer or wine, transitions to distilled spirits and tobacco, advances to cannabis, and subsequently branches into harder illicit substances (e.g., cocaine, hallucinogens, prescription opioids, and heroin). The NHSDA-AFU operationalizes this framework by presenting substances in an arrangement that tracks the epidemiological probability of exposure, allowing researchers to examine transitional probabilities, non-normative sequencing, and accelerations along the developmental trajectory.

2. Problem Behavior Theory (PBT)

Pioneered by Richard Jessor and Shirley Jessor (1977), Problem Behavior Theory suggests that adolescent risk behaviors—including early alcohol consumption, smoking, drug experimentation, delinquency, and precocious sexual activity—do not occur in isolation. Instead, they reflect a shared underlying latent syndrome of “unconventionality” and purposive departure from mainstream societal norms. In PBT, early age of first use signifies a pronounced skew in the balance between instigation factors (e.g., peer approval of deviance, value on autonomy) and personal controls (e.g., religiosity, parental academic expectations). The earlier the age of first use, the greater the saturation of the “problem behavior syndrome.”

3. Neurodevelopmental Dual-Systems and Frontostriatal Maturation Models

Modern cognitive neuroscience (e.g., Steinberg, 2008; Volkow et al., 2016) conceptualizes substance initiation through the lens of a developmental mismatch between the socioemotional/dopaminergic reward system (which matures rapidly around puberty) and the cognitive control system centered in the prefrontal cortex (which matures gradually into mid-adulthood). Adolescents characterized by early pubertal timing experience heightened reward sensitivity and sensation seeking before executive inhibitory pathways are fully operationalized. Retrospective measurement of the age of first use enables researchers to locate the precise window of neurological vulnerability at which an external chemical reinforcer hijacked developing brain circuitry.

Validity

The psychometric validity of the NHSDA-AFU has been rigorously evaluated across decades of federal surveillance studies, independent academic re-analyses, and cross-national epidemiological research.

Construct and Criterion-Related Validity

Construct validity is evidenced by the consistent, ubiquitous inverse relationship documented between the reported age of first use and lifetime severity of substance use disorders. In landmark epidemiological analyses of the NHSDA and subsequent NSDUH datasets (Grant & Dawson, 1997; Chen et al., 2009), individuals initiating alcohol consumption prior to age 14 exhibited a four- to fivefold increased risk of developing lifetime alcohol dependence compared to those who delayed initiation until age 21 or older (approx. 40% vs. 10%). Analogous findings are documented for cannabis: early initiation (prior to age 15) significantly predicts higher rates of cannabis use disorder, secondary mood disorders, cognitive deficits, and elevated likelihood of transitioning to stimulant or opioid abuse.

Convergent and Discriminant Validity

Convergent validity has been established by cross-referencing NHSDA-AFU findings with contemporaneous school-based and longitudinal surveillance systems, such as the Monitoring the Future (MTF) study conducted by the University of Michigan and the Youth Risk Behavior Surveillance System (YRBSS) administered by the Centers for Disease Control and Prevention (CDC). Across these comparative frameworks, cohort initiation curves, median age-of-onset milestones, and demographic risk stratifications demonstrate correlations exceeding .85.

Discriminant validity is supported by the measure’s sensitivity to distinct substance-specific pharmacological classes. While licit substances (cigarettes, alcohol) exhibit broad, unimodal onset curves peaking between ages 13 and 16, hard illicit compounds (cocaine, heroin) yield distinct, positively skewed distributions with substantially later median onsets (ages 18 to 22) and high lifetime abstinence rates. This differentiation confirms that the instrument does not suffer from halo-effect response sets or generalized acquiescence.

Methodological Caveat: Recall Bias and Forward Telescoping

A primary psychometric challenge in retrospective age-of-onset measurement is the phenomenon of forward telescoping, wherein adult respondents systematically recall distant adolescent events as having occurred more recently (i.e., at an older age) than they actually did. Methodological validation studies (e.g., Johnson & Mott, 1993; Engels et al., 1997) demonstrate that when identical cohorts are followed longitudinally from adolescence into middle adulthood, the reported mean age of first use exhibits a gradual upward drift of 1.0 to 2.5 years. Despite this systematic recall artifact, the rank-order stability of participants remains highly robust, preserving predictive validity in relative risk models.

Reliability

Because the NHSDA-AFU is an event-history questionnaire gathering factual chronological markers rather than a reflective multi-item scale, standard indices of internal consistency (such as Cronbach’s alpha) are structurally inapplicable. Instead, psychometric reliability is established through test-retest reliability, inter-rater consistency across proxy reports, and longitudinal concordant stability.

Test-Retest Stability

Short-term test-retest studies (intervals spanning 2 to 6 weeks) conducted by the Research Triangle Institute utilizing cognitive interviewing and re-interview protocols have yielded exceptionally high reliability coefficients. For continuous age responses, intraclass correlation coefficients (ICCs) regularly fall between .75 and .92:

• Cigarette Smoking (even one or two puffs): ICC = .82 to .88
• Alcohol (first full drink): ICC = .78 to .85
• Marijuana / Hashish: ICC = .84 to .91
• Cocaine / Crack: ICC = .76 to .86
• Inhalants / Hallucinogens / Heroin: ICC = .68 to .81 (partially suppressed by lower base-rate sample sizes)

When evaluated categorically (i.e., concordance in identifying whether initiation occurred prior to age 15 versus at or after age 15), Cohen’s kappa (κ) values consistently range between .70 and .86, reflecting substantial to near-perfect agreement.

Longitudinal Consistency and Reliability Enhancement

In extensive multi-wave reliability checks across annual administrations, inconsistencies (e.g., an individual reporting age 14 at Time 1 and age 16 at Time 2) are statistically modeled using computer-assisted self-interviewing (CASI) logic checks. The implementation of audio computer-assisted self-interviewing (ACASI) in later iterations of the NHSDA substantially attenuated socially desirable underreporting, yielding superior inter-temporal reliability relative to traditional paper-and-pencil formats.

Factor Analysis

Because age-of-first-use data are continuous, non-normally distributed, and strictly bounded by the respondent’s current age (with extensive right-censoring for individuals reporting lifetime non-use), conventional Exploratory Factor Analysis (EFA) and Confirmatory Factor Analysis (CFA) based on Pearson correlation matrices are mathematically inappropriate. Instead, psychometric structural evaluations utilize Latent Class Analysis (LCA), Item Response Theory (IRT) for unfolding thresholds, and Cox Proportional Hazards Latent Modeling.

Latent Class and Growth Mixture Modeling

Structural modeling of the NHSDA-AFU items routinely reveals a distinct three- or four-class latent taxonomy representing qualitative patterns of developmental initiation:

• Class 1: Early-Onset Poly-Substance Experimenters (approx. 8–12% of epidemiological samples): Characterized by precocious initiation of tobacco and alcohol (ages < 12), rapid transition to cannabis (ages 12–14), and high conditional probabilities of initiating inhalants, cocaine, or hallucinogens prior to age 18.
• Class 2: Normative / Adolescent-Limited Licit Experimenters (approx. 45–55% of samples): Characterized by mid-to-late adolescent initiation of alcohol and cigarettes (ages 16–18), low-to-moderate cannabis uptake around legal majority, and complete right-censoring (zero probability) on heroin and cocaine.
• Class 3: Late-Emerging / College-Age Initiators (approx. 15–20% of samples): Characterized by delayed alcohol and occasional cannabis exposure occurring exclusively post-high school (ages 18–22), with near-zero penetration of illicit drugs.
• Class 4: Abstainers / Minimal Initiators (approx. 20–25% of samples): Characterized by complete lifetime non-use across all items, or isolated late experimentation with alcohol only.

Guttman Scalability

When examined through deterministic or probabilistic Guttman scale analysis, the seven items display a high Coefficient of Reproducibility (CR > .90) and a Coefficient of Scalability (CS > .65). The hierarchy strictly follows the canonical sequence: Tobacco/Alcohol → Marijuana → Inhalants/Hallucinogens → Cocaine → Heroin. Deviations from this sequence (e.g., initiating heroin prior to ever sampling alcohol or tobacco) represent statistically rare aberrant response vectors indicative of severe pathology or reporting deception.

Instrument / Measurement Tool

• Instrument Name: National Household Survey on Drug Abuse – Age of First Use Scale (NHSDA-AFU)
• Alternative Titles: NSDUH Substance Initiation Module; CSAP Core Measures Age-of-Onset Questionnaire
• Item Count: 7 core items
• Administration Format: Self-administered paper-and-pencil questionnaire, Computer-Assisted Personal Interviewing (CAPI), or Audio Computer-Assisted Self-Interviewing (ACASI)
• Target Population: Adolescents and adults (ages 12 and older)
• Estimated Completion Time: 2 to 4 minutes
• Response Structure: Dual-response option per item:
  • Continuous chronological integer: “The first time I [used substance], I was _____ years old”
  • Dichotomous lifetime abstinence check box: “I have never [used substance] in my life”
• Substance Breakdown:
  1. Cigarettes (even one or two puffs)
  2. Alcoholic beverages (full drink, excluding sips)
  3. Marijuana or hashish
  4. Cocaine in any form (including crack)
  5. Heroin
  6. LSD, PCP, or any other hallucinogen
  7. Inhalants (used for kicks or to get high)
• Scoring and Data Parameterization:
  • Direct Metric: Raw reported age in years for initiated substances.
  • Censoring Flag: Non-users are coded as right-censored at their current chronological age at the time of interview.
  • Composite Early Initiation Index: Dichotomous operationalization classifying onset into binary risk brackets (e.g., ≤ 14 years [Early/High Risk] vs. ≥ 15 years [Normative/Low Risk]).
  • Survival Analysis Framing: Formatted as time-to-event data where the survival time is calculated from birth ($T_0$) to the age of reported onset, with non-initiators censored at current age.

Permissions & Fee and Test Year

The NHSDA-AFU was developed under contracts funded by the United States Federal Government through the Substance Abuse and Mental Health Services Administration (SAMHSA) and the Office of Applied Studies (OAS), originally published in this standardized format in 1999 (incorporating instruments from the 1998 survey wave and the CSAP Core Measures Initiative Phase I Recommendations). As a work of the U.S. Federal Government, the instrument is in the public domain.

No licensing fees, copyright permissions, or royal payments are required for academic, clinical, educational, or commercial use. Researchers are free to reproduce, administer, modify, or integrate these items into digital diagnostic platforms and epidemiological batteries, provided proper academic attribution is accorded to SAMHSA and the Research Triangle Institute.

References

• Chen, C. Y., Storr, C. L., & Anthony, J. C. (2009). Early-onset drug use and risk for drug dependence problems. Addictive Behaviors, 34(3), 319–322. https://doi.org/10.1016/j.addbeh.2008.10.021
• Engels, R. C., Knibbe, R. A., & Drop, M. J. (1997). Inconsistencies in self-reported blunts, drinks and cigarettes: Longitudinal findings from a Dutch adolescent sample. Journal of Substance Abuse, 9, 75–85. https://doi.org/10.1016/S0899-3289(97)90008-7
• Grant, B. F., & Dawson, D. A. (1997). Age at onset of alcohol use and its association with DSM-IV alcohol abuse and dependence: Results from the National Longitudinal Alcohol Epidemiologic Survey. Journal of Substance Abuse, 9, 103–110. https://doi.org/10.1016/S0899-3289(97)90009-9
• Jessor, R., & Jessor, S. L. (1977). Problem behavior and psychosocial development: A longitudinal study of youth. Academic Press.
• Johnson, T. P., & Mott, J. A. (1993). Age of first alcohol use: A comparison of alternative retrospective measures. Journal of Studies on Alcohol, 54(4), 481–487. https://doi.org/10.15288/jsa.1993.54.481
• Kandel, D. B. (1975). Stages in adolescent involvement in drug use. Science, 190(4217), 912–914. https://doi.org/10.1126/science.1188374
• Kandel, D. B. (Ed.). (2002). Stages and pathways of drug involvement: Examining the gateway hypothesis. Cambridge University Press. https://doi.org/10.1017/CBO9780511499593
• Steinberg, L. (2008). A social neuroscience perspective on adolescent risk-taking. Developmental Review, 28(1), 78–106. https://doi.org/10.1016/j.dr.2007.08.002
• Substance Abuse and Mental Health Services Administration. (1999). Summary of Findings from the 1998 National Household Survey on Drug Abuse. Office of Applied Studies, DHHS Publication No. (SMA) 99-3328. Rockville, MD.
• Volkow, N. D., Koob, G. F., & McLellan, A. T. (2016). Neurobiologic advances from the brain disease model of addiction. New England Journal of Medicine, 374(4), 363–371. https://doi.org/10.1056/NEJMra1511480

13. Items of the Scale (Questionnaire)