Substance Abuse by Youth and Young Adults in Rural America

Authors


  • This study was funded under a Cooperative Agreement with the federal Office of Rural Health Policy, Health Resources and Services Administration, DHHS (CA#U1CRH03716). The conclusions and opinions expressed in the paper are the authors' and no endorsement by the University of Southern Maine, the Substance Abuse and Mental Health Services Administration (SAMHSA), or the sponsor is intended or should be inferred. We are very grateful for the generous assistance of SAMHSA and the Research Triangle Institute for being able to use the data from the National Survey on Drug Use and Health. The authors would like to thank Sarah Duffy, PhD, who was a senior economist at SAMHSA at the time this study was conducted, Beth Han, MD, PhD, MPH, who works in the Office of Applied Studies at SAMHSA, and Michael Penne, MPH, who is a Research Statistician at the Research Triangle Institute, for their assistance in the file construction and programming of the data used in this study. For further information, contact: David Lambert, PhD, Maine Rural Health Research Center, Muskie School for Public Service, P.O. Box 9300, 509 Forest Avenue, Portland, ME 04104-9300; e-mail davidl@usm.maine.edu.

Abstract

ABSTRACT: Purpose:Addressing substance abuse in rural America requires extending our understanding beyond urban-rural comparisons to how substance abuse varies across rural communities of different sizes. We address this gap by examining substance abuse prevalence across 4 geographic levels, focusing on youth (age 12-17 years) and young adults (age 18-25 years). Methods: The analysis is based on 3 years (2002-2004) of pooled data from the National Survey on Drug Use and Health. We measure rurality using a four-tier consolidation of the 2003 Rural-Urban Continuum Codes: urban, rural-adjacent, rural-large, and rural-small and medium. Findings: Rural youth have higher alcohol use and methamphetamine use than urban youth and the more rural the area, the higher the use. Rural young adults living in rural-large areas have higher rates of substance abuse than their urban peers; those living in the most rural areas have nearly twice the rate of methamphetamine use as urban young adults. Rural youth are more likely than urban youth to have engaged in the high-risk behavior of driving under the influence of alcohol or other illicit drugs. Conclusions: Higher prevalence rates, coupled with high-risk behavior, place rural youth and young adults at risk of continued substance use and problems associated with this use. Rural community infrastructure should be enhanced to support substance abuse prevention and intervention for these populations.

The increased use and impact of methamphetamine (meth) over the past decade have led county law officials to declare it America's top drug problem.1 Meth use started in California several decades ago and has spread steadily eastward, reaching high levels in a number of states that are largely rural. As a result, meth is often described as a “rural problem.”2–4 Depicting meth as a rural issue has helped to focus attention on an important problem. This should also renew interest in other substances abused in rural America, including alcohol, marijuana, and cocaine. These substances are more prevalent than meth, have significant clinical and social impact, and are associated with the use of other drugs, particularly among youth and younger adults.5

Although there is more research today on rural substance abuse than 5 or 10 years ago, there is a major limitation to much of this work: it compares all rural areas to all urban areas or to different size urban areas. Because substance abuse is a public health problem that affects certain high-risk populations more than others, a population health approach, focusing on differences among sub-populations, is appropriate.6 The literature is clear that youth and young adults are at higher risk of substance abuse than older age groups.7,8 A more limited literature suggests that rural youth and young adults may be at higher risk of abuse of some substances than urban youth and young adults and that this risk varies by size of rural community.9 If we are to develop effective substance abuse prevention and intervention programs for rural youth and young adults, we must first better understand where (which type of rural areas) the need for these programs is the highest.

This study seeks to take a first, but important, step toward understanding variations in substance abuse by size of rural community by examining substance abuse prevalence across 4 geographic levels—urban, rural-adjacent, rural-large, and rural small and medium—based on 3 years (2002-2004) of pooled data from the National Survey on Drug Use and Health (NSDUH). We examine substance use by youth (age 12-17 years), young adults (age 18-25 years), and for all persons 12 and older. We focus on youth and young adults because they have the highest rates of substance use and are the population for whom development of effective community intervention programs (prevention and treatment) may be most urgent.

Background

Prevalence of Rural Substance Abuse Our understanding of substance abuse in Rural America has changed over time. For many years an assumption persisted that rural communities were more nurturing than urban areas and offered more protection against substance abuse.10 Over the last decade, several major studies and reports have found that rather than being a safe haven, rural areas experience significant rates of substance abuse, although somewhat lower than in urban areas.7,10,11 Considering the population as a whole (age 12 years and older), urban areas have higher rates than rural areas of alcohol use and abuse and illicit drug use. Rural youth (age 12-17 years), however, have higher rates of alcohol use and binge-drinking and are more likely to use other illicit drugs, except marijuana, than urban youth. Urban young adults (age 18-25 years) have higher rates of alcohol and marijuana use than rural young adults. Rural-urban differences in illicit drug use by young adults vary by type of drug.8 The literature is consistent in its findings, but is based on broad rural-urban comparisons. There are suggestions within this literature of intra-rural variations, but these variations have not been systematically examined.8,9

Geographical differences in substance abuse, both across regions of the country and within individual states, have been found using the NSDUH survey data.9,12–14 The 2004 NSDUH survey found that past month alcohol use among persons 12-20 years was highest in the Northeast (32.3%), and lower in the Midwest (31.4%), West (27.3%), and South (26.2%). The substance abuse literature suggests significant variation by race and ethnicity, but this variation is generally not broken down further by rurality (Note 1).12

The Rural Context of Substance Abuse Substance use in rural areas is influenced by both location (population and proximity to urban areas of influence) and place (cultural, demographic, and economic factors).15 Adjacency of rural to urban areas is important because of the influence that urban areas may have on rural areas in terms of increasing the supply of drugs, providing educational and economic opportunities, and promoting cultural influences. (For some types of drugs, trafficking of drugs may be bi-directional between smaller and larger rural areas.) The interplay among these factors may place residents of adjacent rural areas at relatively more or less risk of substance abuse than other rural or urban residents.

The economic downturn in much of rural America (starting with the Farm Crisis of the 1980s) has resulted in economic disadvantage and stress that are strongly associated with substance abuse and other social problems (eg, low school and educational attainment, unemployment).5,15 The similarity in prevalence of substance abuse and associated problems in very economically depressed rural areas to economically depressed inner cities led Davidson16 to coin the phrase “America's Rural Ghetto” (Note 2). Constrained economic opportunities have contributed to an aging of rural America, with significant out-migration of younger persons.17 This may result in younger adults remaining in rural areas who are less educated and less likely to be working or earning a living wage than their peers who moved away, placing the “stay-at-home” group at higher risk for substance abuse.15

Data and Methods

The data used in this study come from the 2002, 2003, and 2004 National Survey of Drug Use and Health, conducted annually by the US Substance Abuse and Mental Health Services Administration, Office of Applied Studies. Slightly fewer than 70,000 randomly selected individuals, age 12 years and older, are surveyed every year, approximately 17% of whom live in rural areas. This survey measures the prevalence and correlates of alcohol and other drug use in the United States. The sample design supports the development of national, state and regional estimates. The design of the NSDUH allowed us to pool multiple consecutive years, thereby creating a sufficiently large sample size to compare substance use across the rural-urban continuum. Pooling the data in this manner to examine substance abuse across different size rural areas has been done only on a very limited and occasional basis by the Office of Applied Studies at SAMHSA.8,9

Through an arrangement with the Office of Applied Studies, we were able to access the NSDUH analytic files rather than the more limited public use files available through the Substance Abuse and Mental Health Data Archive (Note 3 and Note 4). The analytic files contain the full nine-tiered Rural-Urban Continuum Codes upon which our analysis is based, whereas the public use files contain a more limited classification structure (ie, large-metropolitan, small-metropolitan, and non-metropolitan statistical counties).

Measuring Rural and Urban Areas We measure rurality using a four-tier consolidation of the 2003 Rural-Urban Continuum Codes (RUCC). The RUCC subdivide the Office of Management and Budget classification of metro and non-metro counties into 3 metro and 6 non-metro groupings (Note 5). This scheme distinguishes metropolitan counties by population size of their metro areas and non-metropolitan counties by the size of their largest cities or towns and their proximity to a metro area(s).18 In this study we combine the 3 metro groupings into a single category, referred to as urban, and the 6 non-metro groupings into 3 categories, referred to as rural, resulting in the following four-tier classification: Urban consists of metropolitan counties in metropolitan areas of any population size; Rural–Adjacent consists of non-metro counties with any mix of urban and rural populations adjacent to a metro area; Rural–Large consists of non-metro counties with urban population of 20,000 or more, not adjacent to a metro area; and Rural–Medium/Small consists of non-metro counties with urban population of 2,500 to 19,999, not adjacent to a metro area (medium), as well as non-metro counties that are completely rural or have fewer than 2,500 urban population, not adjacent to a metro area (small).

We consolidated the RUCC in this manner because our primary interest was to examine variations in the prevalence of substance use across different types of rural areas and compare each rural category with the urban population (Note 6). We wanted to be able to examine the rural adjacent areas, separate from non-adjacent rural areas, because adjacent areas typically have a higher degree of social and economic integration with the adjoining urban areas19 and have lower levels of poverty, higher economic performance, and greater in-migration patterns than rural non-adjacent counties.20–22 These factors are likely to influence the level and patterns of substance abuse.5 We combined medium and small rural county categories into one category given the relatively small sample size and relatively low prevalence rates for some substances in small rural counties (Note 7).

Age was examined in terms of 3 categories: youth (age 12-17 years), young adults (age 18-25 years), and all respondents (age 12 years and older). We focused our analysis on youth and young adults because the literature suggests it is these groups that tend to be the higher users of many substances and they, therefore, may be at risk for impaired life chances and prospects from substance abuse.

Measures of Substance Use We examined the use of 6 commonly abused substances: alcohol, marijuana, cocaine, inhalants, methamphetamine, and OxyContin. The first 4 substances have long histories of use and abuse and significantly impact the lives and well-being of those using them. Methamphetamine and OxyContin have emerged more recently and have been widely reported to pose significant problems for rural youth and young adults. These 6 substances are reported in terms of past-year use. Although past month use is a preferred measure of recent substance use, the typically lower past month prevalence rates limit the sample size available to conduct our analyses. We used past-year rather than past-month use to ensure sufficient cell size to make comparisons across geographic categories for all 6 substances. Two additional measures of alcohol use are reported: binge-drinking (5 drinks on a single occasion one or more times in the past month) and heavy drinking (participated in binge-drinking 5 or more times in the past month). Finally, we report the prevalence of driving under the influence in the past year, in terms of alcohol only and any illicit drug or alcohol (illicit drugs include marijuana, cocaine, inhalants, hallucinogens [including LSD, PCP, or Ecstasy], heroin, or non-medical use of psychotherapeutics, which include stimulants, sedatives, tranquilizers, and pain relievers).

Study SampleTable 1 presents the number of respondents to the NSDUH for 2002-2004, by geographic area and by age category.

Table 1.  Sample Sizes and Percentages (Weighted) by Urban-Rural Classification of Geographic Area and Age, 2002-2004, National Survey on Drug Use and Health
Geographic ClassificationAge 12-17 N (%)Age 18-25 N (%)Age 26 + N (%)All Ages N (%)
  1. Data Source: SAMHSA, Office of Applied Studies, National Survey on Drug Use and Health, 2002, 2003, and 2004.

All metropolitan52,990 (82.9)54,338 (84.4)51,816 (82.9)159,144 (83.1)
Rural–adjacent 8,546 (10.9)7,872 (9.8) 8,042 (11.2) 24,460 (11.0)
Rural–large 2,368 (2.0) 2,811 (2.5) 2,271 (1.9)   7,450 (2.0) 
Rural–small and medium 4,707 (4.2) 3,612 (3.3) 4,297 (4.0)  12,616 (3.9) 
Total68,611 (10.5)68,633 (13.3)66,426 (76.2) 203,670 (100.0)

Analysis To protect the confidentiality of NSDUH respondents (given the need for the full geographic identifier), we developed an arrangement with SAMHSA that gave us access to the results of our analyses without having direct access to the full analytic data files. The data runs for the analyses reported in this article were conducted by the Research Triangle Institute (Research Triangle Park, NC), which is under contract to the Substance Abuse and Mental Health Services Administration to maintain the NSDUH database using standard suppression criteria developed for analysis of the NSDUH. The authors developed the analysis plan. File construction and programming were done jointly by the authors and the Research Triangle Institute. Significance tests are conducted, using SUDAAN (RTI International, Research Triangle Park, NC) to adjust for the weighted sample design, in 2 ways: (1) chi-square tests for differences across the four-tier classification of rural-urban; and (2) t tests comparing each of the 3 rural categories to the urban category.

Findings

Prevalence Overall, all persons aged 12 years and older living in urban areas report past-year use of alcohol, cocaine, and marijuana slightly more often than persons living in rural areas (Table 2). However, different use patterns emerge when the data are analyzed for individual age groups. Rural youth (age 12-17 years) in all 3 rural categories have higher rates of past year use of alcohol, cocaine, methamphetamine, and inhalants than urban youth. That rural youth use alcohol and illicit drugs other than marijuana at higher rates than urban youth has been previously established.10,11Table 2 confirms and extends these findings: youth in all 3 categories of rural areas use these substances (with the exception of marijuana and OxyContin in small/medium rural areas) at greater rates than youth in urban areas. Methamphetamine and alcohol use are higher in rural than in urban areas, and the more rural the area, the higher the use. The use of marijuana and cocaine among youth is greatest in large rural and rural–adjacent areas. Although comparisons were not statistically significant for inhalant and OxyContin use (likely due to small sample size and prevalence rates), they show patterns of higher use in rural than in urban areas.

Table 2.  Past Year Substance Use by Urban and Rural Youth and Young Adults, Average of 2002-2004 (Weighted Percent)
SubstanceAgeOverall P ValueUrbanRural–Adjacent§Rural–Large§Rural–Small/Medium§
  1. Data Source: SAMHSA, Office of Applied Studies, National Survey on Drug Use and Health, 2002, 2003, and 2004.

  2. Based on 2004 survey only.

  3. ‡Overall comparisons by rural-urban classification: *P < .05; **P < .01; ***P < .001.

  4. §t test comparing each rural category to urban: *P < .05; **P < .01; ***P < .001 (t tests were not run when the overall comparison was not significant).

Alcohol12-17***33.736.3***37.0*37.7***
18-25***78.276.5*  81.8*74.2** 
All ages***66.758.9***64.1*56.5***
Cocaine12-17* 1.7 2.2*   2.6* 2.1   
18-25* 6.7 6.4   7.2 5.1** 
All ages*** 2.6 2.0** 2.1 1.7***
Inhalants12-17ns 4.4 4.9   4.7 4.8   
18-25ns 2.2 2.1   1.7 1.5   
All ages* 0.9 0.8   1.0 0.7*  
Marijuana12-17ns15.016.0   15.8 13.6   
18-25***29.226.1***30.6 23.0***
All ages***11.18.7***11.8  7.1***
Methamphetamine12-17* 0.7 0.9     1.2*  1.2**
18-25*** 1.5 2.2**   2.7*** 2.9***
All ages*** 0.6 0.8   1.3 0.8   
Oxycontin12-17ns 0.7 1.2   1.2 0.5   
18-25ns 1.7 1.8   2.8 2.8   
All agesns 0.5 0.6   0.8 0.6   

Some, but not all, rural young adults have higher alcohol, cocaine, and marijuana use than their urban peers (Table 2). Past-year use of each of these substances is highest in large rural areas, next highest in urban areas, lower in rural–adjacent areas, and lowest in rural–small/medium areas. The difference in past-year alcohol use by young adults in rural–large and in urban areas is statistically significant (P < .05). However, the differences in past-year cocaine and marijuana use by young adults living in rural-large and in urban areas are not statistically significant. In contrast to alcohol, cocaine and marijuana, the use of methamphetamine and OxyContin by young adults increases across the urban-rural continuum with the highest use in small/medium rural areas. These differences (overall and specific rural-urban comparisons) are statistically significant for methamphetamine but not for OxyContin. (The lack of significance for OxyContin use may reflect the small sample size resulting from OxyContin being included as a separate substance for the first time in the 2004 survey.) The depiction of methamphetamine as a problem for rural areas appears to be well founded: young adults living in the most rural (small/medium) areas use methamphetamine at nearly twice the rate of their urban peers (2.9% vs 1.5%, P < .001).

High-Risk Behavior The frequency of substance use is an important indicator of the potential seriousness and consequence of substance use. Table 3 presents rates of binge-drinking (having 5 drinks or more on a single occasion within the past month) and heavy-drinking (having participated in binge-drinking 5 or more times in the past month). The general pattern of use, by age group and across level of rurality, is similar to the patterns for any alcohol use in the past year shown in Table 2. With the exception of binge-drinking rates by youth in large rural areas, rates of youth binge and heavy drinking increase with degree of rurality with the highest rates found in the most rural (rural–small/medium) areas (15.0%, P < .001 and 4.1%, P < .001, respectively). Young adults living in rural-large areas have the highest rates of binge-drinking (48.0%, P < .01) and heavy drinking (20.0%, P < .05).

Table 3.  Binge-Drinking and Heavy Drinking by Urban and Rural Youth and Young Adults, Average of 2002-2004 (Weighted Percent)
Drinking Risk BehaviorAgeOverall P Value§UrbanRural–AdjacentRural–LargeRural–Small/Medium||
  1. Data Source: SAMHSA, Office of Applied Studies, National Survey on Drug Use and Health, 2002, 2003, and 2004.

  2. Binge-drinking defined as 5 drinks on a single occasion 1 or more times in the past month.

  3. Heavy drinking defined as having participated in binge-drinking 5 or more times in the past month. Source: SAMHSA, Office of Applied Studies, National Survey on Drug Use and Health, 2002, 2003, and 2004.

  4. § Overall comparisons by rural-urban classification: *P < .05; **P < .01; ***P < .001.

  5. t test comparing each rural category to urban: *P < .05; **P < .01; ***P < .001 (t tests were not run when the overall comparison was not significant).

Binge-drinking (past month)12-17***10.312.5***11.5  15.0***
18-25*41.141.6   48.0**38.6   
All ages***23.021.1***24.7  20.1***
Heavy drinking (past month)12-17*** 2.53.0*  3.1   4.1***
18-25**14.915.6   20.0* 12.8*  
All ages** 6.8 6.9    8.1*  5.8** 

Rural youth are more likely to have driven in the past year while under the influence of alcohol or illicit drugs, or alcohol only, than youth in urban areas (Table 4). The more rural the community, the more likely young persons are to have driven under the influence. Over a quarter of all young adults reported having driven under the influence of either alcohol or an illicit drug during the past year (Table 4). The intra-rural differences are not as large for young adults as for youth. The lowest rates of driving under the influence are found in rural–adjacent areas; higher rates are found in small/medium rural areas and the highest rates are found in rural-large areas.

Table 4.  Driving Under the influence of Alcohol or Any Illicit Drug or Alcohol During the Past Year, by Urban and Rural Youth and Young Adults, Average of 2002-2004 (Weighted Percent)
SubstanceAgeOverall P ValueUrbanRural–Adjacent§Rural–Large§Rural–Small/Medium§
  1. Data Source: SAMHSA, Office of Applied Studies, National Survey on Drug Use and Health, 2002, 2003, and 2004.

  2. During the past 12 months have you operated a vehicle while under the influence of…?

  3. Overall comparisons by rural-urban classification: *P < .05; **P < .01; ***P < .001.

  4. §t test comparing each rural category to urban: *P < .05; **P < .01; ***P < .001 (t tests were not run when the overall comparison was not significant).

Alcohol12-17*** 3.7 5.3*** 6.1*** 7.2***
18-25*25.924.7   30.2*  26.6   
All ages***14.311.5***16.2*  11.8***
Any illicit drug or alcohol12-17*** 4.9 6.9*** 8.3*** 8.4***
18-25ns28.927.9   32.6   29.5   
All ages***15.312.6***17.3*  12.7***

Limitations Under the data analysis arrangement, we did not have direct access to the data files, limiting our ability to conduct follow-up analyses. As a result, this study is an important, but exploratory, analysis of intra-rural variation in substance abuse. There are several limitations to our study, which arise from the need to preserve sufficient cell-size to conduct our analyses. We used a four-tier classification to measure rurality. Without the cell-size constraint, we would have explored using multiple categories of urban (eg, large metropolitan, small metropolitan) and separate categories for rural–medium and rural-small areas. Although past-month use is a preferred measure of recent substance use, we measured substance-abuse prevalence in terms of past-year use to ensure sufficient cell size to conduct our analysis. Finally, we were not able to explore regional and racial/ethnic differences in prevalence within the framework of our analysis because cell sizes would not be large enough across all comparisons to conduct the analysis.

Discussion

Until a decade ago the literature suggested, and policymakers assumed, that substance abuse was lower in rural than urban populations. Over the last decade, research has shown that among all ages substance abuse is slightly lower in rural than in urban areas and that alcohol and illicit drug use, other than marijuana, is higher among rural than urban youth. What was not known was how substance use varied across different levels of rurality and how the use of more powerful and increasingly prevalent drugs, such as OxyContin and methamphetamine, differed between rural and urban areas and across different levels of rurality.

The data in this study show that rural youth (12-17 years) and young adults (18-25 years) have high prevalence rates in the use of a number of substances. In general, substance use by youth is highest in rural–small/medium areas and highest for young adults in rural large-non-adjacent areas. While we were not able to explore these differences within the current study, it is possible to speculate on this pattern. For youth, there may be fewer alternatives for recreation and higher sources of stress in the most rural and isolated areas. Young adults may move from smaller to larger areas either for college, work, or both. Young adults living in rural large-non-adjacent areas may have “stayed at-home” and not moved to rural-adjacent or urban areas.

While some of the geographic differences reported in this study are modest, and the prevalence rates for some substances such as meth are small, these differences are significant both statistically and from a policy perspective. Overall, these prevalence rates, coupled with high-risk behavior (eg, binge-drinking, heavy drinking, and driving under the influence) place rural youth and young adults at risk of continued substance use and substance use-related injuries. Rural America's youth and young adults have a substance abuse problem that needs to be addressed.

Where do we start? The Carsey Institute's recent report “Substance Abuse in Rural and Small Town America” notes the need for alcohol prevention and treatment programs for rural youth and the need to begin to address the meth problem in rural areas.11 The latter is a daunting challenge given the scarcity of rural substance abuse prevention and treatment services. Developing intervention programs for youth alcohol abuse is a logical and important starting place. Alcohol use is, by far, the most commonly abused substance, often co-occurs with the use of illicit drugs, and is correlated with poor school performance and social problems. It is also important to address meth and other illicit drug use among rural youth and young adults, particularly given the prevalence and likely consequence of use of these substances.

Rather than craft intervention policies and programs one substance at a time, it is important to consider what can be done to enhance the infrastructure of rural communities to support substance abuse prevention and early intervention for rural youth and young adults. One way to do this is by adapting best practice and model prevention programs to rural areas. While progress has been made in identifying such practices and programs,23 we need to understand how to implement them and make them work in diverse rural communities. Environmental prevention strategies tailored to the unique cultural, economic, and geographic realities of rural communities should be developed and evaluated.24 This is particularly important for the most rural areas, which have the highest rates of youth alcohol and meth use and young adult meth use. These areas face the greatest challenge, from a law enforcement perspective, from a prevention perspective, and from a treatment perspective, due to low population density, limited resources stretched over large geographic areas (Note 8) and, in many cases, patterns of poverty, unemployment, and declining economic sectors.

Pooling three consecutive years of data with discrete geographic identifiers enabled us to conduct the intra-rural analysis reported in this article. We recommend that the NSDUH, which is conducted every year, continue to be used in this manner and that the full analytic file with geographic identifiers be made available to other external researchers, with proper safeguards for quality and confidentiality. A number of important questions should be examined. To what extent do observed intra-rural differences in prevalence vary by major region of the country? To what extent do they vary by major ethnic and cultural groups? To what extent are they explained by underlying demographic characteristics? What are the intra-rural differences in alcohol and co-occurring drug use? To what extent do they vary by major ethnic and cultural group and by region? Understanding regional, ethnic, and demographic differences is critical in developing culturally competent intervention programs.

Addressing the shortage of health and social services in rural America is a long and vexing problem.25,26 Developing a coordinated substance abuse policy in the United States has been a similarly uphill struggle.27 However, the increased attention in policy and research circles to both rural health and substance abuse suggests that we can begin to address systematically substance abuse by rural youth and young adults.

Endnotes

Note 1. The 2004 NSDUH reported that Native Americans/Alaska Natives had the highest past-month rates of illicit drug use (12.3%), followed by African Americans (8.7%), whites (8.1%), Hispanics (7.2%), and Asians (3.1%).7 There is significant variation among different Hispanic sub-groups. Most of the research on substance abuse by African Americans has focused on alcohol use (which tends to be lower among African American youth than other ethnic groups), but tends not to be specific to, or broken down by, rurality.13

Note 2. Subsequent reports and studies have lent support to Davidson's conceptualization. A 1994 U.S. News and World Report article28 reported by Conger15 described Waterloo, Iowa, as one of the rural communities that include a growing white underclass, as identified by having more than 40% of its population living below the poverty line and having high rates of crime, domestic violence, single-headed households, and inter-generational poverty.

Note 3. Our agreement with the Office of Applied Studies did not provide us with direct access to the data but to results of data runs on the full analytic file based on our specifications. Our project liaison at the Research Triangle Institute (SAMHSA's contractor on this project) ran the analyses on the full analytic file according to our specifications.

Note 4. The public use files have been modified to protect the privacy of respondents by eliminating and/or collapsing geographic and other variables that could be used to identify individual respondents and randomly eliminating approximately 12,500 records per year through a sampling sub-step that maintains the integrity of the public use data file.

Note 5. The 2003 Rural-Urban Continuum Codes are based on the 2000 Census. The 3 metropolitan (metro) county groupings are (1) counties in metro areas of 1 million population or more; (2) counties in metro areas of 250,000 to 1 million population; (3) counties in metro areas of fewer than 250,000 population. The 6 non-metropolitan (non-metro) groupings are (4) urban population of 20,000 or more, adjacent to a metro area; (5) urban population of 20,000 or more, not adjacent to a metro area; (6) urban population of 2,500 to 19,999, adjacent to a metro area; (7) urban population of 2,500 to 19,999, not adjacent to a metro area; (8) completely rural or less than 2,500 urban population, adjacent to a metro area; (9) completely rural or less than 2,500 urban population, not adjacent to a metro area.

Note 6. A similar classification was used in a study of heavy and binge-drinking using Behavioral Risk Factor Surveillance System data.29

Note 7. Exploration of the data revealed that small sample size and low prevalence triggered the NSDUH suppression criteria for many of the cells in the analysis of small/isolated as well as the medium rural counties. As a result, we were unable to compare differences across these 2 categories of rural counties, leading us to combine the small/isolated and medium rural categories to achieve sufficient cell size to report use of the substances included in this study.

Note 8. As with health and human services more generally, there are fewer substance abuse prevention and treatment services in rural than in urban areas.13

Ancillary