1. Introduction - Youth Substance Use: State Estimates from the 1999 National Household Survey on Drug Abuse

This report contains 1999 State estimates of the prevalence of use of alcohol, cigarettes, and marijuana among youths aged 12 to 17. It also presents information about risk and protective factors for youth substance use at the State level. These estimates are from the National Household Survey on Drug Abuse (NHSDA), an ongoing survey of the civilian, noninstitutionalized population of the United States, 12 years of age or older.

The Summary of Findings from the 1999 NHSDA (Substance Abuse and Mental Health Services Administration [SAMHSA], 2000b) presented national estimates of substance use and, for the first time, State estimates, for seven priority variables: past month use of any illicit drug, past month use of marijuana, past month use of any illicit drug except marijuana, past month cigarette use, past month binge use of alcohol, past year dependence on any illicit drug, and past year dependence on any illicit drug or alcohol for all persons aged 12 or older and three age groups (12 to 17, 18 to 25, and 26 or older).

This is the second presentation of State-level estimates based on the 1999 NHSDA. This report presents newly available model-based estimates for six measures associated with substance use among youths. These variables include past month use of alcohol, past month use of tobacco, average annual incidence of marijuana, perceived risk of binge drinking, perceived risk of using marijuana once a month, and perceived risk of smoking one or more packs of cigarettes a day. In addition, this report provides standard sample-weighted estimates of a number of variables related to substance use, such as average age at first use and various risk and protective factor scale scores. Also included in this report for comparison purposes, though not discussed, are model-based estimates for the same six measures for the other age groups: ages 18 to 25, ages 26 or older, and all persons aged 12 or older.

Based on the 1999 NHSDA data and the modeling procedure, State estimates for five additional substance use measures (for ages 12 or older, 12 to 17, 18 to 25, and 26 or older) have been estimated and are available on the SAMHSA website (see acknowledgments page). The estimates produced include both the estimate for each State and the 95 percent prediction interval. The five substances include past year use of cocaine, past year receipt of treatment for illicit drugs, past year receipt of treatment for illicit drugs or alcohol, past year need of treatment for illicit drugs, and past year treatment for illicit drugs or alcohol. Together with the items estimated in the first and second reports, these additional five measures represent the complete set of State model-based estimates calculated for 1999.

The NHSDA is the primary source of statistical information on the use of illicit drugs by the U.S. civilian population aged 12 or older. Conducted by the Federal Government since 1971, the survey collects data by administering questionnaires to a representative sample of the population through face-to-face interviews at their place of residence. The survey is sponsored by SAMHSA, and data collection is carried out by Research Triangle Institute (RTI), under the direction of SAMHSA's Office of Applied Studies (OAS). This section contains a brief description of the methodology. A more complete description is provided in Appendix E.

The survey covers residents of households, noninstitutional group quarters (e.g., shelters, rooming houses, dormitories), and civilians living on military bases. Persons excluded from the survey include homeless people who do not use shelters, active military personnel, and residents of institutional group quarters, such as jails and hospitals. Appendix H describes surveys that include populations that are not part of the NHSDA sampling frame.

Prior to 1999, the NHSDA was administered in about an hour and used paper-and-pencil interviewing (PAPI) methods. The NHSDA PAPI instrumentation consisted of a questionnaire booklet that was completed by the interviewer and a set of individual answer sheets that were completed by the respondent. All substance use questions and other sensitive questions appeared on the answer sheets so that the interviewer was not aware of the respondent's answers. Less sensitive questions, such as demographics, occupational status, household size, and composition, were asked aloud by the interviewer and recorded in the questionnaire booklet.

The 1999 NHSDA marked the first survey year in which the national sample was interviewed using a computer-assisted interviewing (CAI) method. The survey used a combination of computer-assisted personal interviewing (CAPI) conducted by the interviewer and audio computer-assisted self-interviewing (ACASI). For the most part, questions previously administered by the interviewer are now administered by the interviewer using CAPI. Questions previously administered using answer sheets are now administered using ACASI. CAI has many advantages over PAPI, including more efficient collection and processing of the data and improved data quality. Use of ACASI is designed to provide the respondent with a highly private and confidential means of responding to questions and should increase the level of honest reporting of illicit drug use and other sensitive behaviors. For further details on the development of the CAI procedures for the 1999 NHSDA, see SAMHSA (2001).

The 1999 NHSDA sample employed a 50-State design with an independent, multistage area probability sample for each of the 50 States and the District of Columbia. The eight States with the largest population (which together account for 48 percent of the total U.S. population aged 12 or older) were designated as large sample States (California, Florida, Illinois, Michigan, New York, Ohio, Pennsylvania, and Texas). For these States, the design provided a sample large enough to support direct State estimates. For the remaining 42 States and the District of Columbia, smaller, but adequate, samples were selected to support State estimates using small area estimation (SAE) techniques (described in Appendix G). The design also oversampled youths and young adults, so that each State's sample was approximately equally distributed among three major age groups: 12 to 17 years, 18 to 25 years, and 26 years or older.

Nationally, 169,166 addresses were screened and 66,706 persons were interviewed within the screened addresses. The survey was conducted from January through December 1999. The weighted response rate for household screening was 89.6 percent. The weighted interview response rate for youths aged 12 to 17 was 78.1 percent. Unweighted response rates for individual States for youths aged 12 to 17 ranged from 69 to 90 percent.

Estimates in this report have been adjusted to reflect the probability of selection, record nonresponse, poststratification to known benchmarks, item imputation, and other aspects of the estimation process.

Subsequent to the publication of Summary of Findings from the 1999 NHSDA (Substance Abuse and Mental Health Services Administration [SAMHSA], 2000b), an error was discovered in the imputation procedure that affected some of the national and state prevalence estimates. There is an expanded discussion of the impact of that error in the new Summary of Findings from the 2000 National Household Survey on Drug Abuse (SAMHSA, in press). Revised 1999 national estimates are available in the 2000 Summary of Findings report and in the associated summary tables available on the SAMHSA website (see acknowledgments). Revised 1999 model-based state estimates for 12 measures will also be available on the SAMHSA website. These measures are as follows: past month any illicit drug use, past month marijuana use, past month any illicit drug other than marijuana use, past year cocaine use, past month alcohol use, past month "binge" use, past month any tobacco use, past month cigarette use, annual average marijuana incidence, and perceptions of great risk of marijuana, alcohol, and cigarettes. A report on State estimates from the 2000 NHSDA, to be released in the fall of 2001, will also include revised 1999 State estimates. All estimates presented in this report are based on the corrected imputation procedure, and the general procedure for imputation is described in Section E.2 of Appendix E.

The findings presented in this report are divided into five main chapters: alcohol use, tobacco use, marijuana use, risk and protective factors for substance use, and sequences in the initiation of cigarettes, alcohol, and marijuana. The discussions of alcohol, cigarettes, and marijuana in Chapters 2 through 4 are further divided into two sections within each chapter. The first includes estimates referred to in this report as model-based estimates. These are produced by combining the prevalence rate based on the State sample data and the prevalence rate based on a national regression model applied to local-area county and Census block group/tract-level estimates from the State. The methodology, validation, and discussion of data limitations for these estimates are discussed further in Appendix G.

The State model-based estimates are portrayed in the maps showing all 50 States and the District of Columbia (Appendix A), in tables that include all 50 States and the District of Columbia by four age categories (Appendix B), and in individual State tables arranged to display all of the estimates discussed in this report by the four age categories for a given State (Appendix D). The color of each State on the maps in Appendix A indicates how the State ranks relative to other States for each indicator. States could fall into one of five groups according to their ranking by quintiles. Because there are 51 areas to be ranked, the middle quintile was assigned 11 areas and the remaining groups 10 each. In some cases, a group may have contained more or fewer areas than this number because the estimates were the same as for other States in the group. Those States with the highest rates for a given variable are in red; those with the lowest estimates are in white. Although the tables show estimates rounded to one decimal place, the rankings for the maps are based on estimates calculated to two decimal places. For the model-based estimates, when two or more States fell on the border between adjoining quintiles and had identical estimates to two decimal places, those States were assigned to the lower quintile.

At the top of each table in Appendix B is a "national" total that represents the (weighted) sum of the estimates from the 50 States and the District of Columbia. Those totals are generally slightly different from the corresponding national estimates calculated by summing the sample-weighted records across the entire sample. The latter estimates are the preferred unbiased estimates for the Nation and are used in the text for comparison with the State-level estimates.

The second sections in Chapters 2 through 4 include related information based on direct sample-weighted estimates. Chapter 5 on risk and protective factors includes scale scores that are also based on sample-weighted estimates. Chapter 6 analyzes national and State data with respect to the progression of substance use, particularly the relationship between the use of cigarettes and/or alcohol and the use of marijuana. The State sample-based estimates are provided in Appendix C.

The estimates based only on the State sample data have been largely restricted to averages and other scales that are more precisely estimated than estimates of proportions, such as the past month prevalence rate for marijuana. The State sample mean estimates tend to vary in a relatively narrow range around the national mean, with a number of States sharing the same estimate, even to three or four decimals. As a result, in the discussion of the quintiles based on sample estimates, as many decimals as necessary were used to break ties so as not to distort the distribution of 10 (or 11 for the middle group) areas per group.

Associated with each State estimate based on the sample data and sampling weights is a 95 percent confidence interval. Also, associated with each State estimate based on the modeling approach is a 95 percent prediction interval. These intervals indicate the precision of the estimate. For example, the State with the highest estimated past month alcohol rate for youths (a model-based estimate) was North Dakota, with a rate of 24.7 percent (Table B.1B). The 95 percent prediction interval on that estimate is from 20.6 to 29.1 percent. Therefore, the probability is 0.95 that the true prevalence for North Dakota will fall between 20.6 and 29.1 percent. The prediction interval indicates the uncertainty due to both sampling variability and model bias. The interpretation of the 95 percent confidence interval is fairly similar; however, the estimates are assumed to be unbiased and the interval measures uncertainty due only to sampling.

Some of the estimates in this report are based on special calculations. Each set of estimates presented in this report is explained in further detail below.

Incidence rates are typically calculated as the number of new initiates of a substance during a period of time (such as in the past year) divided by the estimate of the number of person years of exposure (in thousands). The incidence measure in this report is the result of a similar definition but is based on the model-based methodology mentioned in Section 1.2 and discussed further in Appendix G. The following definition is used in this report:

"initiates" is the count of persons in 1999 who first used marijuana in the past 24 months, and "never users" is the count of persons in 1999 who had never used marijuana.

Note that this estimate uses a 2-year time period to accumulate incidence cases. By assuming further that the distribution of first use for the incidence cases is uniform across the 2-year interval, the total number of person years of exposure is 1 year on average for the incidence cases plus 2 years for all the never users at the end of the time period. This approximation to the person years of exposure permits one to recast the incidence rate as a function of two population prevalence rates, namely, the fraction of youths who first used marijuana in the past 2 years and the fraction who had never used marijuana. Both of these prevalence estimates were estimated using the survey-weighted hierarchical Bayes estimation approach. Also note that for estimates for age groups, the age is based on the age at the time of the interview, whereas the usual estimates of incidence based on direct estimation use the age at the time of first use to determine the age group.

The count of youths who first used marijuana in the past 2 years is based on a "moving" 2-year period that ranges over 3 calendar years. Youths were asked when they first used marijuana. If a youth indicated first use of marijuana between the day of the interview and 2 years prior, the youth was included in the count. Thus, it is possible for a youth interviewed in the first part of 1999 to indicate first use as early as the first part of 1997 or as late as the first part of 1999. Similarly, a youth interviewed in the last part of 1999 could indicate first use as early as the last part of 1997 or as late as the last part of 1999. Therefore, the reported period of first use ranged from early 1997 to late 1999 and was "centered" in 1998. About half of the youths reported first use in 1998, while a quarter each reported first use in 1997 and 1999. Youths who responded in 1999 that they had never used marijuana were included in the count of "never used." For further information on the general procedures for calculating incidence rates and other limitations, see Section F.4 in Appendix F.

Each survey respondent who reported having used a particular substance was asked at what age he or she first used the substance. An early age at first use has been associated with increased likelihood of problem use later in life and early use of other substances (Kandel & Yamaguchi, 1993;Kandel, Yamaguchi, & Chen, 1992). A low estimated State average age at first use may offer a partial explanation for States that have high prevalence levels among youths.

The average age at first use (AFU) of a substance was defined to be the average over all persons who initiated first use at age 25 or younger and who had initiated their first use in 1995, 1996, or 1997. Although the 1999 NHSDA obtained information about initiation for 1998, estimates of the number of initiates for such substances as alcohol and cigarettes for 1998 would be inaccurate because they would exclude a significant number of youths who began use at ages 10 and 11. Therefore, it was decided not to use the 1998 data relating to age at first use for any of the substances discussed in this report in order to have consistent definitions when comparing alcohol, cigarettes, and marijuana (see Appendix C for further discussion).

The confidence intervals based on 3 years' data are quite large despite the relatively large sample sizes. Average age at first use was calculated for alcohol, cigarettes, and marijuana. Because marijuana had the lowest prevalence of the three substances, it also had the smallest sample size. However, even the number of persons who reported first use of marijuana in the smaller States was respectable, ranging from 59 to 118 (data not shown in table). In the eight largest States, sample sizes ranged from 234 to 420 (data not shown in table). These sample sizes are usually considered adequate for the estimation of most statistics. Therefore, the underlying (large) size of the standard deviation is diagnostic of the generally wide variation in the reported age at first use among youths.

The 1999 NHSDA questionnaire included seven questions about problems due to substance use that were asked of each person who had used the substance in the past year. The respondent could answer "yes" (i.e., they had the problem) or "no." These seven questions were based on criteria in the Diagnostic and Statistical Manual of Mental Disorders (4^th ed., DSM-IV) for diagnosing whether a person is dependent on a substance (American Psychiatric Association [APA], 1994).

Sample sizes in most States were not sufficient to produce reliable estimates of dependence for youths. An alternative measure of "problem" substance use was therefore developed. This "dependence score" was determined for each youth in the sample based on the youth's responses to the seven DSM-IV questions. Reasoning that the likelihood of having a problem with a substance was less for those who did not use the substance in the past year than it was for those who did, a score of 0 was assigned to youths who were not past year users. Past year users with no problems (i.e., a "no" response on all seven questions) were assigned a score of 1. For each "yes" response to one of the questions, the score was incremented by 1. Therefore, a person who answered all seven questions "yes" would have a score of 8.¹

The full text of the questions on dependence is provided in Appendix I. These questions deal with the time spent on activities associated with the substance, the extent of use, the development of tolerance, interference with work, emotional and psychological health problems, and inability to reduce use. The dependence scores for alcohol, cigarettes, and marijuana were analyzed using the same seven questions for each so that the scores would be comparable across the three substances.

It should be noted that there are other methods of combining youth scores that may be useful, such as only assigning scores to youths who were past year users. Although that score would not be comparable from State to State, it would provide a State with a measure of the size of the dependence problem within the set of past year users of the substance. Other measures could include scores that are just based on the heavier users or those with scores above a specified threshold, perhaps relative to a national percentile.

The 1999 NHSDA collected information on a number of risk and protective factors for substance use. Risk and protective factors involve attitudes and behavior associated with the higher likelihood of use or nonuse of drugs. These factors are typically classified into a number of domains, such as peer/individual, school, family, and community. A detailed report on these factors based on the 1999 data is to be released later in 2001; however, a short discussion is included here of four of the constructs in the peer/individual domain because that domain typically has a large impact on whether youths use illicit drugs. For additional information, see also the report on Risk and Protective Factors for Adolescent Drug Use (Lane, Gerstein, Huang, & Wright, 2001), which presents findings from the 1997 NHSDA.

The four constructs discussed are peer antisocial behavior, favorable attitudes toward substance use, peer attitudes favorable toward substance use, and peer substance use (for the exact wording of questions, see Appendix I).

The construct for antisocial behavior was based on an average of six NHSDA items. "During the past year have you...gotten into a serious fight at school or work; taken part in a group fight of your friends against another group; carried a handgun; sold illegal drugs; stolen or tried to steal something worth more than $50; attacked someone with the intent to seriously hurt them?"

The construct to measure favorable attitudes toward substance use had three items: "How do you feel about someone your age...smoking one or more packs of cigarettes a day; trying marijuana or hashish once or twice; having one or two drinks of an alcoholic beverage nearly every day?"

The construct for peer attitudes favorable toward substance use had the same three items as above, except the question was worded as follows: "How do you think your close friends would feel about you..."

The construct on peer substance use had four items depending on the substance: "How many of your friends...smoke cigarettes; use marijuana or hashish; drink alcoholic beverages; get drunk at least once a week?"

In August 2000, the first report of the 1999 NHSDA data was released—the Summary of Findings from the 1999 NHSDA (SAMHSA, 2000b). In addition to national results, that report included estimates for the 50 States and the District of Columbia for seven model-based variables by four age groups (ages 12 or older, ages 12 to 17, ages 18 to 25, and ages 26 or older). Additional tabulations have been generated from the 1999 data and are available at the SAMHSA website. Further methodological information will be posted to the website as it becomes available. Analytic reports focusing on specific issues or population groups will continue to be produced by SAMHSA. A few of the reports in progress focus on the following topics:

A complete listing of previously published reports from the NHSDA and other data sources is available from OAS. Most of these reports are also available through the Internet (see Acknowledgments page). In addition, OAS makes public use data files available to researchers through the Substance Abuse and Mental Health Data Archive (SAMHDA, 2001). Currently, files are available from the 1979 through the 1998 NHSDAs. The 1999 public use file will be available in the fall of 2001.

¹ Respondents with missing data ("don't know" or "refused") for one or more of the questions were handled in two ways. Those who had missing data for four or more answers were eliminated from the analysis. Those who were only missing one to three answers were retained in the analysis, and their responses for missing items were imputed based on their percentage of "yes" responses to the questions that were answered.

SAMHSA, an agency in the Department of Health and Human Services, is the Federal Government's lead agency for improving the quality and availability of substance abuse prevention, addiction treatment, and mental health services in the United States.

* Adobe™ PDF and MS Office™ formatted files require software viewer programs to properly read them. Click here to download these FREE programs now