Section A: Overview and Findings

A.1. Introduction

This report presents estimates of substance use for substate areas defined by each State for the purpose of allocating the substance abuse treatment block grant funds from the Substance Abuse and Mental Health Services Administration (SAMHSA). Estimates are provided for each of the 331 substate areas representing collectively the 50 States and the District of Columbia. These estimates are based on combined data from the 1999-2001 National Surveys on Drug Use and Health (NSDUHs), formerly called the National Household Survey on Drug Abuse (NHSDA). The survey, sponsored by SAMHSA, collects information from residents of households, residents of noninstitutionalized group quarters (e.g., shelters, rooming houses, dormitories), and civilians living on military bases. In 1999-2001, NSDUH collected data from approximately 207,000 respondents aged 12 or older. The survey is stratified to obtain representative samples from all 50 States and the District of Columbia.

This report provides a more nuanced perspective to the variations in substance use rates both within and across States than was possible with previous State reports (Wright & Sathe, 2005). The 12 measures presented are marijuana use in past month, average annual rate of first use of marijuana, perceptions of great risk of smoking marijuana once a month, any illicit drug use in past month, any illicit drug use other than marijuana in past month, cocaine use in past year, alcohol use in past month, binge alcohol use in past month, perceptions of great risk of having five or more drinks of an alcoholic beverage once or twice a week, cigarette use in past month, any tobacco product use in past month, and perceptions of great risk of smoking one or more packs of cigarettes per day.

A.2. Format of the Report

Section A provides background on the survey, how substate areas were formed, the general methodological approach, and a brief discussion of the findings. Section B gives further information on the small area estimation (SAE) methodology. Section C includes the estimates for each of the 12 measures above and the corresponding prediction intervals (PIs). It also contains maps that show the prevalence of each outcome for each substate region on a national map. The tables have been ordered alphabetically by State, and the substate areas have been ordered alphabetically within each State. There are four separate tables, each having three related outcomes. Section D contains definitions of State planning areas. Section E includes the population counts for persons aged 12 or older for each substate area and the combined 1999, 2000 and 2001 NSDUH sample sizes. Users may find these counts helpful in calculating the weighted average prevalence estimate for any combination of substate areas. Section F lists the references. Refer to Appendix E (Tables E.18 to E.20) of the 2001 State report (Wright, 2003) for the 1999, 2000, and 2001 NSDUH response rates.

A.3. Overview: Substate Areas, Ranking Areas, and Small Area Estimation Methods

The substate areas for each State were developed in a series of communications between SAMHSA staff and State treatment representatives in the summer and fall of 2004 and in early 2005. The goal of the project was to provide substate-level estimates showing the geographic distribution of substance prevalence for areas that States would find useful for treatment planning purposes.¹ The final substate region boundaries were based on the individual recommendations of the State staff assuming the NSDUH sample sizes provided adequate precision. Most States defined areas in terms of counties or groups of counties. A few States defined the areas in terms of census tracts.

Maps and tables (612 maps and 204 tables) showing the substate estimates for each State separately will be available on the SAMHSA website at http://oas.samhsa.gov/metro.htm. National tables (12 total) showing estimates by age group (12 to 17, 18 to 25, and 26 or older) also will be available at this same Web location for those substate areas having sufficient precision.

For each measure in this report, the 331 substate areas have been ranked from lowest to highest and divided into seven categories designed to represent distributions that are somewhat symmetric, like a normal distribution–but flatter. Colors were assigned to all areas such that the third having the lowest prevalence are in blue, the middle third are in white, and the third with the highest prevalence are in red (the only exception being the three perception of risk outcomes, which have the highest estimates represented in blue and the lowest represented in red). To further discriminate among the areas that display relatively higher prevalence, the "highest" third has been further divided into three categories: dark red for the 15 areas with the highest estimates, medium red for the 30 areas with the next highest, and light red for the 65 areas with the next highest. The "lowest" third is categorized in a similar way, using three distinct shades of blue. Due to tied values of prevalence, the number of substate regions in each category can vary.

Estimates in this report are based on hierarchical Bayes estimation methods that combine survey data with a national model. Applying this methodology to the State substance use measures has been shown to result in more precise estimates than using the sample-based results alone (Wright, 2002). The methodology used to produce estimates in this report is the same as that used to produce State estimates from the NSDUH data since 1999. Sample data have been combined across 3 years (1999-2001) to improve the precision of substate area estimates. The estimate for each area is accompanied by a 95 percent PI (for more details, see Section B, Substate Region Estimation Methodology).

In addition to the substate area estimates, comparable estimates are provided for the 50 States and the District of Columbia using the same methodology. Because these estimates are based on 3 years' data, they are not directly comparable with the State estimates in earlier reports that are based on only 2 consecutive years. However, if the national map based on the substate areas is compared with one of the national maps based on 2000-2001 data, it can be seen that most of the States with significant areas in dark red correspond to States that were in the highest fifth in 2000-2001. For example, the 10 States with the highest estimates of past month use of marijuana among persons aged 12 or older were Alaska, Colorado, Connecticut, Delaware, Maine, Massachusetts, New Hampshire, Oregon, Rhode Island, and Vermont. Based on the 1999-2001 data, each of those States displays some areas in red, indicating a high prevalence level.

Because the SAE methods used here tend to borrow strength from both the national model and the State-level random effects, substate estimates with sample sizes that were closer to the minimum (250) tend to be shrunk more toward the corresponding State prevalence estimate than substate areas with large sample sizes. This results generally in exaggerating the clustering of substate estimates around their State means. Relatively high estimates for small substate areas tend to be pulled down, while relatively low estimates tend to be pulled up. In most cases, the larger the fraction of a State's population contained in a substate area, the more the substate area estimate will tend toward the State mean.

A.4. Discussion of Findings

The substate-level estimates provide for the first time a picture of the dispersion in substance use within States and across the United States. At the State level, planners can use the information to tailor programs to the range of prevalence. However, it is important to note that the range within each State is greatly influenced by the total number of sample observations for the State and the number of substate areas that can support estimation. Some of the large States such as Ohio specified 20 or more areas, while others such as New York only specified two. Because of this, States like Ohio and Pennsylvania often can display more of the diversity of prevalence levels.

Where States share a border, a smoother transition was expected; however, the boundaries at borders often appear to be abrupt. For example, the northernmost area in Idaho (Region 1) was in the lowest group of past year cocaine use, while the neighboring areas in Washington and Montana displayed higher rates. However, many of these apparent differences tended to fall within sampling error. This was not the case for Colorado (Regions 1 and 4) and Nebraska (Regions 1 and 2) or Oklahoma (Northwest and Southwest) and Texas (Region 1), where the apparent differences exceeded sampling error.

Generally, related measures of substance use had relatively similar distributions across substate areas. This was true for related measures, such as past month alcohol use and past month binge alcohol use; past month marijuana use and average annual rates of first use of marijuana (marijuana incidence); past month use of any illicit drug and past month marijuana use (marijuana being the most common illicit drug); and past month use of a substance (cigarettes, binge use of alcohol, and marijuana) and perceptions of great risk of using that substance. However, the patterns for the use of marijuana, cocaine, alcohol, and cigarettes were quite different. The maps for past month use of alcohol and binge use of alcohol show the heavier alcohol levels more typical of areas in the Northern United States. Higher rates of past month cigarette use were more clustered in the South and Midwest. Past month marijuana use was relatively high in the Northeast and in the West and low in the Midwest and the South. However, both Idaho and Utah reported low levels of marijuana use compared with other Western States. Substate regions in five States (Colorado, Delaware, District of Columbia, Massachusetts, and New Mexico) were in the highest group for past year use of cocaine, whereas substate regions in eight States (California, Colorado, the District of Columbia, Hawaii, Massachusetts, Oregon, Rhode Island, and Vermont) had the highest levels of past month marijuana use. By this measure, cocaine use was more clustered than past month marijuana use.

The distribution of substance use in the United States is likely to be a function of many factors, including attitudes toward substance use, various socioeconomic factors, supply of and demand for the substances, extent of law enforcement activities, and the effectiveness of various prevention, treatment, and mental health programs. Data from these other sources need to be analyzed in conjunction with the different substance use measures in order to better understand the underlying dynamics of the process.

In 2006, SAMHSA will produce a second set of substate estimates for the above treatment planning areas that also will reflect any State geographic changes in the interim. Estimates will be made for approximately 20 substance use measures, including the ones in this report, plus measures of alcohol and illicit drug dependence and abuse, and serious mental illness (SMI). The estimates will be based on data from the combined 2002-2004 NSDUHs. To the extent that States maintain their current definitions of substate treatment planning areas, the newer estimates can be compared with the earlier ones. Given that a number of methodological changes occurred in the NSDUH procedures between 2001 and 2002, including a $30 incentive for completing the survey, the estimates for 2002-2004 will be higher than those for 1999-2001. The differences in prevalence between those two periods, therefore, cannot be interpreted as a true trend for that period. Nevertheless, by comparing the substate differences with comparable changes in the national prevalence estimates for those two periods, analysts should be able to draw some conclusions for substate areas that show significant differences, assuming that the methodological effect is reasonably constant across all States.