2004 State Estimates of Substance Use
and Mental Health
This report presents State estimates for 22 measures of substance use or mental health problems based on the 2003 and 2004 National Surveys on Drug Use and Health (NSDUHs).1 Sponsored by the Substance Abuse and Mental Health Services Administration (SAMHSA), NSDUH is an ongoing survey of the civilian, noninstitutionalized population of the United States aged 12 years or older. Interview data from approximately 135,500 persons were collected in 2003-2004. State estimates presented in this report have been developed using a small area estimation (SAE) procedure in which State-level NSDUH data are combined with local-area county and census block group/tract-level data from the State. This model-based methodology provides more precise estimates of substance use at the State level than those based solely on the sample, particularly for smaller States.
Starting in 1999, the NSDUH sample was expanded to produce State-level estimates. The samples in each State were selected to represent proportionately the geography and demography of that State. The first report with State estimates was published in 2000 (Office of Applied Studies [OAS], 2000). It utilized the 1999 survey data and the SAE procedure. Because the SAE procedure requires significant preparatory steps for the modeling and extensive computation to generate results, the number of variables estimated has been limited to ones with high policy value. The first report included only seven measures. Subsequent State reports have been published annually, gradually extending the capabilities of the SAE procedure and increasing the number of measures estimated (Wright, 2002a, 2002b, 2003a, 2003b, 2004; Wright & Sathe, 2005). The current practice is to base annual estimates on a 2-year moving average of NSDUH data in order to enhance the precision for States with smaller samples.
Recently, State estimates have been produced for additional measures by combining 3 (or more) years of NSDUH data and using sampling weights and direct estimation. The advantage of this approach is that it can be used on any variable in the dataset; however, the estimates typically are not as accurate as the SAE measures. These estimates have been included in some reports and in tables on the SAMHSA website.
NSDUH 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 planned and managed by SAMHSA's OAS, and the data are collected and processed by RTI International.2 This section briefly describes the national survey methodology. 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 prisons and long-term hospitals.
The 1999 survey marked the first 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 an interviewer and audio computer-assisted self-interviewing (ACASI). 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 OAS (2001).
The 1999 through 2001 NHSDAs and the 2002 through 2004 NSDUHs 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 accounted 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). Collectively, the sample allocated to these States ensured adequate precision at the national level while providing individual State samples large enough to support both model-based (SAE) and design-based estimates. For the remaining 42 States and the District of Columbia, smaller, but adequate, samples were selected to support State estimates using SAE techniques (described in Appendix E of the 2001 NHSDA State report, Wright, 2003b). 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.
In 2002, several changes were introduced to the survey. Incentive payments of $30 were given to respondents for the first time in order to address concerns about the national and State response rates. Other changes included a change in the survey name, new data collection quality control procedures, and a shift from the 1990 decennial census to the 2000 census as a basis for population count totals and to calculate any census-related predictor variables that are used in the estimation.
An unanticipated result of these changes was that the prevalence rates for 2002 were in general substantially higher than those for 2001—substantially higher than could be attributable to the usual year-to-year trend—and thus are not comparable with estimates for 2001 and prior years.3 Therefore, the 2002 NSDUH was established as a new baseline for the national, as well as the State, estimates. Given the varying effects of the incentive and other changes on the States, not only are the estimates for 2002 and later years not comparable with prior years, but also the relative rankings of States may have been affected. Therefore, the rankings of States for 2002-2003 or later should not be compared with those for prior years.
By combining data across 2 years, the precision of the small area estimates for the small States, and thus their rankings, have been improved significantly. In addition, by combining 2 years of data, the impact of the national model on those States has been reduced significantly relative to estimates based on a single year's data.4
Nationally, approximately 261,000 addresses were screened and about 135,500 persons responded within the screened addresses. The survey is conducted from January through December each year. The screening response rate for 2003-2004 combined averaged 90.8 percent, and the interviewing response rate averaged 77.2 percent, for an overall response rate of 70.1 percent. The State overall response rates for 2003-2004 ranged from 60.6 percent in New York to 81.5 percent in Utah (Table A.9).
Estimates in this report have been adjusted to reflect the probability of selection, unit nonresponse, poststratification to known benchmarks, item imputation, and other aspects of the estimation process. These procedures are described in the NSDUH Methodological Resource Books (MRBs) for each survey year (see http://www.oas.samhsa.gov/nhsda/methods.cfm).
The findings in this report are presented in six chapters, including this introductory chapter, along with U.S. maps of estimates for States at the ends of Chapters 2 through 6 and data tables in Appendices B and C at the end of the report. In each chapter except Chapter 6, there are separate estimates for three age groups (12 to 17, 18 to 25, and 26 or older) and all ages 12 or older combined.
Chapter 2 presents State estimates of the prevalence of any illicit drug use, marijuana use, the prevalence of perceived risk of marijuana use, incidence of marijuana use, any illicit drug use other than marijuana, cocaine use, and nonmedical use of pain relievers. Chapter 3 discusses analogous estimates of alcohol use, binge alcohol use, and the perceived risk of binge alcohol use. Chapter 3 also includes for the first time estimates of underage (ages 12 to 20) alcohol use and binge alcohol use. Chapter 4 presents estimates for tobacco use, cigarette use, and the perceived risk of heavy cigarette use. Chapter 5 discusses the substance treatment–related measures (i.e., dependence on and abuse of alcohol or illicit drugs and needing but not receiving treatment). Chapter 6 presents estimates of serious psychological distress (SPD), formerly referred to as serious mental illness (SMI) for persons aged 18 or older.
At the ends of Chapters 2 through 6, State model-based estimates are portrayed in U.S. maps showing all 50 States and the District of Columbia. The maps reflect the ranking of States into fifths from lowest to highest for each measure to simplify the discussion in the chapters. Appendix A gives a brief description of the SAE methodology for 2003-2004 and discusses minor refinements in that methodology for these analyses relative to prior years. For a more detailed discussion of the SAE methodology, see Appendix E of the 2001 State report (Wright, 2003b). Also included in Appendix A are the State sample sizes and response rates for 2002, 2003, 2004, 2002-2003 combined, and 2003-2004 combined (Table s A.1 to A.12). Tables of model-based estimates for each substance use or mental health measure are included in Appendix B. The quintile rankings can be determined from these tables that include all 50 States and the District of Columbia, listed in alphabetical order, by four age categories. Estimates of change between 2002-2003 and 2003-2004 are presented in Appendix C. Tables for individual States are available on the SAMHSA website and display all of the estimates discussed in this report by the appropriate age categories. Also available on the SAMHSA website are tables of the total number of persons associated with each measure corresponding to the estimated percentages or rates for each substance use or mental health measure in Appendix B (see http://www.oas.samhsa.gov/2k4State/lot.htm).
The color of each State on the U.S. maps indicates how the State ranks relative to other States for each measure. 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 "quintile" could have more or fewer States than desired because two (or more) States have the same estimate (to two decimal places). When this occurs at the "boundary" between two "quintiles," all States with the same estimate were assigned to the lower quintile. Those States with the highest rates for a given outcome are in red, with the exception of the perceptions of risk measures, for which the lowest perceptions of great risk are in red. Those States with the lowest estimates are in white, with the exception of the perceptions of risk measures, for which the highest perceptions of great risk are in white.
At the top of each table in Appendix B is a national total that represents the (population-weighted) mean of the estimates from the 50 States and the District of Columbia. These totals have been benchmarked in order to agree with the corresponding national estimates calculated as sample-weighted averages or proportions across the entire sample. (For more details, refer to Appendix A, Section A.4.) Associated with each State estimate is a 95 percent prediction interval (PI). These intervals indicate the precision of the estimate. For example, the State with the highest estimated past month alcohol rate for youths aged 12 to 17 (a model-based estimate) was Wisconsin, with a rate of 24.4 percent (Table B.9). The 95 percent PI on that estimate is from 21.5 to 27.7 percent. Therefore, the probability is 0.95 that the true prevalence for Wisconsin for persons aged 12 to 17 will fall between 21.5 to 27.7 percent. The PI indicates the uncertainty due to both sampling variability and model bias.
In this report, State rankings are discussed in terms of the range and the national average because the latter provide a useful context for the discussion. However, the differences between the highest (or lowest) rate and the next-to-highest (or next-to-lowest) rate are typically very small and not statistically significant. For example, although Alaska (11.8 percent) had the highest rate of past month use of an illicit drug among persons aged 12 or older for 2003-2004, the estimate for New Mexico (11.3 percent) was only half of a percentage point lower and statistically no different than the Alaska estimate (Table B.1). Therefore, it is important to consider the PI when comparing States. For Alaska, one can say that 95 percent of the time the true value would fall in the range of approximately 10.2 to 13.7 percent. Clearly, the estimate for New Mexico falls into this range, but Mississippi's estimate (5.8 percent) does not.
Estimates of change between 2002-2003 and 2003-2004 are presented in Appendix C for 22 measures, by age group (see Table s C.1 to C.23). This is the first time that the estimates of change have been published for 10 of these measures. These tables show the estimates for 2002-2003 and 2003-2004 and a p value to test the hypothesis that there was "no change" over this period. The report only discusses differences if they are significant at p values of 0.05 or less (corresponding to a probability of 95 percent that the change was not 0). However, p values greater than 0.05 but less than or equal to 0.10 also have been marked to highlight other possible changes because the year-to-year changes are often small and relatively hard to detect, especially for those measures with low prevalence rates. The methodology for estimating change involves estimating one model for 2002-2003 based on the predictor variables and the sample for those years and a separate model for 2003-2004 based on the predictor variables and sample for those years. This can lead to slightly different national models (i.e., models with slightly different model coefficients for the two sets of years). The change between 2002-2003 and 2003-2004 estimates the average yearly change between 2002 and 2004. "Average yearly change" indicates the change between 2002 and 2004 divided by 2. For more details on this topic, see the section on measuring change (Section A.8) in Appendix A.
Throughout the report, there are a number of drug measures that are related, such as marijuana use and any illicit drug use. It might appear that one could draw new conclusions by subtracting one from the other (e.g., subtracting the percentage who used marijuana in the past month from the percentage who used any illicit drug in the past month to find the percentage who used an illicit drug other than marijuana in the past month). Because related measures have not been estimated jointly, but with different models, subtracting one measure from another related measure at the State level can give misleading results, perhaps even a "negative" estimate, and should not be done.
Estimates for 2003-2004 were developed for 22 measures:
The national results from the 2004 survey were released in September 2005 (OAS, 2005c). Additional methodological information on NSDUH, including the questionnaire, is available electronically on the OAS webpages at http://www.oas.samhsa.gov. Brief descriptive reports and in-depth analytic reports focusing on specific issues or population groups also are produced by OAS. Further information on access to NSDUH publications, detailed tables, and public use files is contained in "Accessing Data from the National Survey on Drug Use and Health (NSDUH)" (OAS, 2004a). A complete listing of previously published reports from NSDUH and other data sources is available from OAS. Most of these reports are available through the Internet (http://www.oas.samhsa.gov). In addition, OAS makes public use data files available to researchers through the Substance Abuse and Mental Health Data Archive (SAMHDA, 2006). Currently, data files are available from the 1979 to 2004 surveys at http://www.icpsr.umich.edu/SAMHDA/index.html.
In 2006, estimates for substate planning areas based on combined 2002-2004 NSDUH data will be available at the SAMHSA website. The substate planning area definitions for all 50 States and the District of Columbia are based on the areas for substate allocation of funds under SAMHSA's Substance Abuse Prevention and Treatment (SAPT) block grant. This will be the second time that substate data for the entire United States have been collected and estimated using comparable methods.6 Estimates will be available for each State and the District of Columbia for the 22 measures listed in Section 1.3. Along with the substate estimates will be comparable State and national estimates summarized in tables and maps that indicate the distribution of prevalence rates across the United States. The methodology used for producing substate estimates is similar to the SAE methodology used to produce the State estimates in this report.
1In 2002, the name of the survey was changed from the National Household Survey on Drug Abuse (NHSDA) to NSDUH.
2RTI International is a trade name of Research Triangle Institute.
3For a thorough discussion of the impact of these changes, see Section C.2 of Appendix C in OAS (2005c).
4Combining data across 2 years permits the estimation of change at the State level by expressing it as the difference of two consecutive 2-year SAE moving averages. Because the 2004 data are available for analysis, estimates of change between combined 2003-2004 data and the combined 2002-2003 data can be developed. This method is similar to the one used in the 2001 State report (Wright, 2003b).
6Substate data were first reported in May 2005 using data from the 1999 to 2001 surveys (OAS, 2005d) and in June 2005 using only marijuana use data from the same survey years (OAS, 2005a).
This page was last updated on December 30, 2008.