National data      State level data       Metropolitan and other subState area data

Substance Abuse in States and Metropolitan Areas:
Model Based Estimates from the 1991-1993 National Household Surveys on Drug Abuse

Chapter 1

Executive Summary

1. Introduction

1.1 Drug Use Measures

1.2 Need for Small Area Estimates

1.3 Quality/Usefulness of the Estimates

Executive Summary

This report presents estimates of substance abuse for 26 States and 25 metropolitan statistical areas (MSAs). These estimates were developed from data collected in the National Household Survey on Drug Abuse (NHSDA) combined with local area indicators from a variety of sources. They were produced by the Substance Abuse and Mental Health Services Administration (SAMHSA) to provide State and local area policy makers with information on the prevalence of substance abuse behaviors and problems in their local area. These estimates are an inexpensive alternative to the direct survey approach for describing substance abuse in State and local areas. They are based on a consistent methodology across areas and are constructed so that they sum to national estimates produced by the NHSDA.

These State and MSA estimates are the result of a comprehensive small area estimation (SAE) project that included the development of an innovative methodology based on the methods used by other federal agencies to meet needs for small area data. The methodology employs logistic regression models that combine NHSDA data with local area indicators, such as drug-related arrests, alcohol-related death rates, and block group level characteristics from the 1990 Census, that were found to be associated with substance abuse.

An important feature of these models is that they produce estimates that are a weighted average of an indirect synthetic regression estimate and a direct survey estimate. Thus, the models require at least some NHSDA sample data for each small area under consideration. A total of 26 States met basic sample size criteria for estimation. All had at least 300 interviews and most had over 1,000. Each of the 25 metropolitan areas for which estimates were generated had at least 300 interviews.

Estimates are presented for eleven measures of substance abuse behaviors and problems including numbers and percent of people who use cigarettes, alcohol, any illicit substances, illicit substances other than marijuana, or cocaine during a month; numbers and percent of people who are dependent on alcohol or an illicit substance during a year; number and percent of people who need treatment for illicit drug use and receive treatment of illicit drug use in a year; and number and percent who need and receive treatment for alcohol abuse during a year. Yearly estimates of the numbers and percent of people arrested are also included. Estimates are based on data from the 1991 through 1993 NHSDA and represent the average across that time span.

Work was also carried out to evaluate the model used to produce these estimates. Considering a variety of evidence, it was concluded that:

• The SAE model produced estimates of all measures that were much better than States could achieve by simply applying NHSDA national prevalence rates for demographic subgroups to the population distribution in their States.

• A preponderance of evidence indicates that the estimates for alcohol, cigarette, and any illicit drug use are good in that they adequately reflect both levels of use and differences across States and MSAs in the level of use.

• For past month use of any illicit but marijuana, past month use of cocaine, past year dependency on illicit drugs and dependency on alcohol, and need for treatment, the limited evidence that we have indicates that these estimates are also reasonably good.

• For arrests, past year treatment for illicit drug use, and past year treatment for alcohol, the quality assessments resulted in mixed findings, and these small area data may not be good indicators of either differences between States or between MSAs, except in broad terms, or of the actual levels.

Finally, the estimates are subject to all of the limitations that the NHSDA national estimates are subject to, namely, reliance on the validity of self-reports of drug use and other behaviors and the exclusion of people who are homeless (not in shelters) or living in institutions such as jails and nursing homes at the time of the survey.

    

1. Introduction

This report presents estimates of substance abuse for 26 States and 25 metropolitan statistical areas (MSAs). These estimates were developed from data collected in the National Household Survey on Drug Abuse (NHSDA) combined with local area indicators from a variety of sources. They were produced by the Substance Abuse and Mental Health Services Administration (SAMHSA) to provide State and local area policymakers with information on prevalence of substance abuse behaviors and problems in their local area.

The National Household Survey on Drug Abuse (NHSDA), which is conducted by SAMHSA, provides a rich source of data on substance abuse at the national level. It is the federal government’s chief source of information on the magnitude of substance use and abuse in the United States and has been used by policymakers to monitor trends in drug use and identify problem areas. In addition to the reports on prevalence and trends in drug use that are produced by SAMHSA, the agency uses the data to investigate many special topics that are important for understanding the nature of substance abuse, and numerous researchers from around the nation conduct special analyses using the public use files that are produced from the survey. In addition, the value of the NHSDA is enhanced by its use of a rigorous sampling and data collection methodology that has been developed and tested over a number of years.

In spite of the wealth of information that can be gleaned from the NHSDA, SAMHSA recognized that there was considerable need for information on drug use and abuse for States and other smaller geographic areas. In order to help States meet their needs for data on substance abuse, SAMHSA undertook a comprehensive small area estimation (SAE) project that included the development of a new methodology for producing estimates for small areas and an evaluation of that methodology. This report summarizes the results of this project. A more detailed report of the results and methodology is available including additional tabulations and cross-tabulations. ^(footnote#1)

    

1.1 Drug Use Measures

Estimates for eleven measures of substance abuse behaviors and problems are presented in this report. They include:

Use of legal (licit) substances:

1. Cigarette Use In Past Month - Smoked cigarettes at least once within the past month.

2. Alcohol Use In Past Month - Had at least one drink of an alcoholic beverage, that is, beer, wine or liquor or a mixed alcohol drink within the past month.

Use of illicit substances:

Three measures of illicit drug use are presented including:

3. Any Illicit Drug Use In Past Month - Use within the past month of hallucinogens, heroin, marijuana, cocaine, inhalants, or the nonmedical use of sedatives, tranquilizers, stimulants or analgesics.

4. Any Illicit Drug Use Other Than Marijuana In Past Month - Past month use of any illicit drug excluding those whose only illicit drug use was marijuana.

5. Cocaine Use In Past Month - Use within the past month of cocaine in any form, including crack.

Current use of any illicit drug provides a broad measure of illicit drug use; however it is dominated by marijuana use. Therefore estimates of the use of illicit drugs other than marijuana and of cocaine are presented since these latter two measures reflect different types of drug abuse.

Drug or alcohol dependence:

Substance dependence measures the most severely affected substance abusers who are likely to require some kind of treatment or intervention. The dependence measure developed for the NHSDA is based on an algorithm that approximates the DSM-III-R criteria. ^(footnote#2) Two dependence measures are presented:

6. Dependent On Illicit Drugs In Past Year - Dependent on marijuana, inhalants, cocaine, hallucinogens, heroin, opiates or nonmedical use of sedatives, tranquilizers, analgesics, or stimulants. Those who are dependent on both alcohol and another illicit substance are included, but those who are dependent on alcohol only are not.

7. Dependent On Alcohol and Not Illicit Drugs In Past Year - Dependent on alcohol and not dependent on any illicit drugs.

Treatment for drug and alcohol problems:

Three measures related to drug and alcohol treatment are presented. Two are based on items in the NHSDA that asked about treatment use. A measure of drug abuse treatment need based on an algorithm developed by SAMHSA’s Office of Applied Studies is also included. Comparing the estimated number of people receiving treatment to the estimated number needing treatment can demonstrate gaps in the provision of services for substance abusers:

8. Received Treatment For Illicit Drugs In Past Year - Received treatment in the past 12 months at any location (including hospitals, clinics, self-help groups, doctors) for any illicit drug. These estimates include those who received treatment in the past 12 months for both drinking and illicit drugs.

9. Received Treatment For Alcohol Use, But Not Illicit Drugs, In Past Year - Received treatment in the past 12 months for drinking (including hospitals, clinics, self-help groups, doctors). These estimates exclude those who received treatment in the past 12 months for both drinking and illicit drugs.

10. Needed Treatment For Illicit Drug Use In Past Year - Persons who either: were dependent on illicit drugs in the past year; were a past year heroin user; received treatment in the past 12 months for any illicit drugs; were a needle user of heroin, stimulants or cocaine in the past 12 months; were a daily marijuana user during the past 12 months; or in the past 12 months were weekly users of hallucinogens, cocaine, inhalants or had weekly nonmedical use of stimulants, sedatives, tranquilizers or analgesics.

 

Past year arrest:

Estimates of past year arrest are also presented.^(footnote#3)  This attribute was included because it is correlated with substance abuse:

11. Arrested For Any Crime In Past Year - Arrested and booked at least once for breaking a law in the past 12 months.

    

1.2 Need for Small Area Estimates

The State and Metropolitan Statistical Area (MSA) estimates in this report were developed to assist people who need information about substance abuse in their areas. The impetus for developing these local area statistics was based on the increasing demand for such estimates by State and Federal substance abuse planners. While there are existing data systems that provide small area indicators of drug abuse for selected MSAs, like the emergency room drug-related episode reports from the Drug Abuse Warning Network (DAWN) and the voluntary urine test reports among arrestees (DUF), these data systems do not provide estimates of the desired total population statistics on use, dependence, treatment and the need for treatment. Furthermore, national surveys like the NHSDA do not provide large enough samples for direct State level estimation, a limitation that persists even after pooling results over three years of NHSDA data. Local area surveys designed to provide reliable direct estimates tend to be costly and generally do not provide comparable estimates between areas due to different sampling and data collection methodologies. Thus, the model based approach to deriving the small area estimates has been used increasingly by government agencies as the demand for local area data has increased and the required statistical and computational tools have improved.

The estimates in this report will be useful to those States and local areas that need information on substance abuse behaviors in their areas. These estimates:

• Take into account local area characteristics

• Provide an inexpensive alternative to the direct survey approach for describing substance abuse in States or local areas

• Use a consistent methodology across States and are constructed so that they sum to national estimates produced by the NHSDA.

Chapter 4 presents a summary of the work done to evaluate the model used to produce these estimates and other models which were constructed for comparative purposes. The evaluation revealed that: for most of the estimates presented in this report, we have evidence that they adequately reflect the prevalence of substance abuse characteristics for States and MSAs. Because there was no independently existing "gold-standard" to which to compare these estimates, a variety of methods were used in an attempt to evaluate their quality. These evaluations included some statistical tests that were appropriate for the methodology used and some comparisons to external data which were available for a few of the measures (alcohol and cigarette use, arrest, and drug treatment). Taken together, the preponderance of evidence indicates that:

• The estimates appear to be good for the drug use characteristics - particularly, alcohol, cigarettes, and any illicit drug use in that they adequately reflect both level of use and differences across States and MSAs.

The pictures for the other measures is less clear.

• For cocaine use, past year dependence on an illicit drug or alcohol, and need for treatment, we had only the statistical evaluation to use to consider the quality of the estimates. These statistical evaluations indicated that these estimates were of reasonable quality; however, we did not have any external figures to compare them with, and these are rare characteristics affecting less than about 3 percent of the population nationally.
• For arrests, past year treatment for illicit drug use, and past year treatment for alcohol, the quality assessments resulted in mixed findings. Internal statistical evaluations showed that the estimates are reasonably good; however, correlations between these estimates and administrative data on treatment and arrest were low. This conflicting evidence is difficult to interpret because the administrative data that were available for comparison were collected using different definitions and methods than those that were used in the NHSDA. Also, there may be differences across states in these definitions and methods.
• Finally, all the evaluations showed that the models used to produce these estimates are better than States could achieve by simply applying previously published NHSDA prevalence rates to the population distribution in their States.

In spite of these findings supporting the validity of these small area estimates, they should be used with caution. Chapter 3, which presents the estimates, includes a chart which shows the 95 percent confidence intervals for the estimates. ^(footnote#4)    Examining these, it can be seen that the estimates for some States are more precise than those for some others. In general:

Finally, the estimates are subject to all of the limitations that the NHSDA national estimates are subject to, namely, reliance on the validity of self-reports of drug use and other behaviors and the exclusion of people who are homeless (not in shelters) or living in institutions such as jails and nursing homes at the time of the survey. In particular, estimates of the number of people receiving treatment for drug abuse and the number of people arrested are believed to be substantially underestimated, based on comparisons with administrative data. Some heavy drug use measures, such as the number of people dependent on drugs or needing treatment, may also be underestimated. ^(footnote#5)  Estimates of current cigarette use shown in this report are also conservative, as they are based on interviewer-administered questions, which may be less conducive to complete reporting (particularly for youths) than the more private self-administered answer sheets used in the NHSDA beginning in 1994.