Appendix 1: Methodology

How were the sites selected? (See map in the Introduction) A total of 20 sites were studied for this issue of Pulse Check. We selected sites using Census Bureau regions and divisions with a goal of achieving geographic and demographic diversity. In addition, we made an effort to select sites in areas with special drug abuse problems of national concern. More specifically, we applied the following methodology in selecting sites.

We purposely selected the most populous States in the four census regions: New York in Region I (Northeast Region); Texas in Region II (South Region); Illinois in Region III (Midwest Region); and California in Region IV (West Region). In three of these States, we selected the most populous metropolitan areas: New York City, Chicago, and Los Angeles. In Texas, however, we selected El Paso—a known high trafficking area with particularly high levels of unemployment, population growth, and poverty—because of its proximity to the United States border with Mexico.

We included four rural States, one per census region. (Rural States are defined by the Census Bureau as those in which 50 percent or more of the State's population reside in census-designated rural areas.) The four rural sites selected are as follows:

• Region I (Northeast): Portland, Maine—Of the three rural States in the Northeast Region (including New Hampshire and Vermont), Maine has the only Atlantic coastline and shares the longest border with Canada. It also includes an ONDCP-designated High Intensity Drug Trafficking Area (HIDTA). Portland is Maine's most populous metropolitan area.
• Region II (South): Columbia, South Carolina—The three other rural States in the South census region are Kentucky, Mississippi, and West Virginia. However, South Carolina's location along a major drug trafficking corridor makes that State a strategic choice. Recent cocaine seizures in Columbia further highlight its strategic importance.
• Region III (Midwest): Sioux Falls, South Dakota—Sioux Falls is the most populous metropolitan area within the Midwest Region's two rural States (North Dakota and South Dakota).
• Region IV (West): Billings, Montana—Montana is the only census-designated rural State in the West Region, and Billings is its most populous metropolitan area.

The remaining 12 sites were selected to ensure that the entire list included at least 2 sites from each of the 9 Census Bureau divisions (East North Central, Mountain, Middle Atlantic, New England, Pacific, South Atlantic, South East Central, South West Central, and West North Central). Additional selection criteria included population density, representation of racial/ethnic minorities, and emphasis on high drug trafficking areas. Applying these criteria resulted in the final selection of the following 20 Pulse Check sites:

How do the 20 sites vary demographically? Appendix 2 highlights the demographic diversity of these 20 sites. For example, their population density per square kilometer ranges from a sparse 18.4 in Billings, Montana, to a crowded 2,897.4 in New York City (1997 estimates). Their unemployment rates range from a 2.7 low in Columbia, South Carolina, to an 11.2 high in El Paso, Texas. The racial/ethnic breakdowns in the 20 sites further exemplify their diversity: white representation ranges from 31.4 percent in Honolulu, Hawaii, to 96.7 percent in Sioux Falls, South Dakota; black representation ranges from 0.5 percent in Billings, Montana, to 42.1 percent in Memphis, Tennessee.

What other data are available at the 20 selected sites? Information from other national-level data sources will be useful for framing, comparing, corroborating, enhancing, or explaining the information obtained for Pulse Check. The following data sources are available in nearly every site: ONDCP's past Pulse Check reports; the National Institute on Drug Abuse (NIDA) Community Epidemiology Work Group (CEWG); the Substance Abuse and Mental Health Services Administration (SAMHSA) Drug Abuse Warning Network (DAWN); and the National Institute of Justice (NIJ) Arrestee Drug Abuse Monitoring (ADAM) program.

Who are the Pulse Check sources, and how were they selected? Consistent with previous issues, the information sources for Pulse Check were telephone discussions with 4 knowledgeable individuals in each of the 20 sites: 1 ethnographer or epidemiologist, 1 law enforcement official, and 2 treatment providers. Ethnographers and epidemiologists were recruited based on several criteria: past participation in the Pulse Check program; membership in NIDA's CEWG; research activities in local universities; or service in local community programs. We recruited law enforcement officials by contacting local police department narcotic units, Drug Enforcement Administration (DEA) local offices, and HIDTA directors.

To identify treatment sources, we randomly selected providers from the 1998 Uniform Facility Data Set (UFDS), a listing of Federal, State, local, and private facilities that offer drug abuse and alcoholism treatment services. For this purpose, we excluded facilities that reported more than 50 percent of their clientele as having a primary alcohol abuse problem, served a caseload of fewer than 100 clients, or provided only prevention or detox services. We then divided the remaining facilities into two groups— methadone and non-methadone treatment facilities—in order to capture two client populations whose demographic characteristics and use patterns often differ widely. We selected one from each of these two categories of programs for each of the 20 selected sites. Because Billings, Montana, has no UFDS-listed methadone treatment facilities, we selected two non-methadone facilities in this site. Despite several attempts, in a few sites we were unable to recruit a second treatment provider, so we solicited a referral from the one treatment provider successfully recruited from the UFDS data file.

Of the 80 sources identified and recruited, we successfully obtained information for this Pulse Check issue from 74: a response rate of 93 percent. The nonresponding individuals included a law enforcement official in Detroit, the methadone treatment providers in Denver, Miami, and Honolulu, and the non-methadone treatment providers in Philadelphia and St. Louis.

What kind of data were collected, and how? For each of the 74 sources, we conducted a single telephone discussion, lasting about 1 hour. We asked sources to explore with us their perceptions of the change in the drug abuse situation between 1999 and 2000, reflecting back on the first 6 months of 1999 compared with the same 6 months in 2000. In general, we discussed the following topic areas with Pulse Check sources, noting source opinions about the drug use problem in 2000 compared with the previous year:

• An overall snapshot of the community's drug use situation, including the perceived seriousness of the problem and the relative concern about specific major drugs of abuse—heroin, cocaine, marijuana, methamphetamine, club drugs, and any emerging drug abuse problems
• Population groups in the site most likely to use major drugs of abuse, user characteristics (such as age, gender, race/ethnicity, socioeconomic status, and residence), and patterns of use (such as route of administration, concomitance, setting, and context)
• Availability of major drugs of use (such as drug forms, supply levels, quantities, prices, and purity)
• Trafficking patterns (such as how drugs are manufactured or grown, where drugs originate, where they are transshipped, how they enter local areas, and where they ultimately end up)
• Seller characteristics (such as gang or organized crime affiliation, age, gender, race/ethnicity, seller drug use patterns, and drug-associated violence)
• Sales practices (such as where drugs are sold, settings, street names, packaging, adulterants, marketing strategies)
• Other activities associated with drug sales or use (such as gang activity, prostitution, violence)
• Community context issues that might have impacted on the drug use situation during the study periods (such as treatment availability, medical consequences, large Seizures, law enforcement policy initiatives, policing and sentencing practices, legislation, task forces, media campaigns, or major news events)

Not surprisingly, ethnographic and epidemiologic sources seemed to be very knowledgeable about users and patterns of use; they were somewhat knowledgeable about drug availability; and they were less informed about sellers, distribution, and trafficking patterns. Treatment providers had a similar range of knowledge, but they generally focused on the specific populations targeted by their programs. Some providers, however, were able to provide a broader perspective about the communities extending beyond their individual programs. Among the three Pulse Check source types, law enforcement officials appeared to be most knowledgeable about drug availability, trafficking patterns, seller characteristics, sales practices, and other associated activities; they were, understandably, less knowledgeable about user groups and characteristics.