Reexamining associations between mania, depression, anxiety and substance use disorders: results from a prospective national cohort.

Separate inheritance of mania and depression together with high rates of clinical overlap of mania with anxiety and substance use disorders provide a basis for re-examining the specificity of the prospective association of manic and depression episodes that is a hallmark of bipolar disorder. We analyzed information from 34 653 adults in Waves 1 and 2 of the National Epidemiologic Survey on Alcohol and Related Conditions, a longitudinal nationally representative survey of US adults interviewed 3 years apart. Psychiatric disorders were assessed by a structured interview. We used logistic regression analyses to estimate the strength of associations between Wave 1 manic episodes and Wave 2 depression, anxiety and substance use disorders controlling for background characteristics and lifetime Wave 1 disorders. Corresponding analyses examined associations between Wave 1 major depressive episode with manic episodes and other psychiatric disorders. In multivariable models, Wave 1 manic episodes significantly increased the odds of Wave 2 major depressive episodes (adjusted odds ratio (AOR): 1.7; 95% confidence interval: 1.3-2.2) and any anxiety disorder (AOR: 1.8; 1.4-2.2), although not of substance use disorders (AOR: 1.2; 0.9-1.5). Conversely, Wave 1 major depressive episodes significantly increased risk of Wave 2 manic episodes (AOR: 2.2; 1.7-2.9) and anxiety disorders (AOR: 1.7; 1.5-2.0), although not substance use disorders (AOR: 1.0; 0.9-1.2). Adults with manic episodes have an approximately equivalent relative risk of developing depression episodes and anxiety disorders. Greater research and clinical focus is warranted on connections between manic episodes and anxiety disorders.

Oral Health in the Population Assessment of Tobacco and Health Study.

Tobacco use is a well-established risk factor for multiple adverse oral conditions. Few nationally representative oral health data sets encompass the current diversity of tobacco and nicotine products. This investigation examines the validity of oral health measures in the Population Assessment of Tobacco and Health (PATH) Study to assess relationships between tobacco use and oral health. Cross-sectional data from PATH Study wave 4 (N = 33,643 US adults, collected 2016-2018) were used to obtain estimates for 6 self-reported oral conditions (e.g., bone loss around teeth, tooth extractions) and compared with analogous estimates from the National Health and Nutrition Examination Survey (NHANES) cycle 2017-2018 (N = 5,856). Within the PATH Study, associations were calculated between tobacco use status and lifetime and past 12-mo experience of adverse oral conditions using survey-weighted multivariable logistic regression. Nationally representative estimates of oral conditions between the PATH Study and NHANES were similar (e.g., ever-experience of bone loss around teeth: PATH Study 15.2%, 95% CI, 14.4%-15.9%; NHANES 16.6%, 95% CI, 14.9%-18.4%). In the PATH Study, combustible tobacco smoking was consistently associated with lifetime and past 12-mo experience of adverse oral health (e.g., exclusive cigarette smoking vs. never tobacco use, adjusted odds ratio [AOR] for loose teeth in past 12 mo: 2.02; 95% CI, 1.52-2.69). Exclusive smokeless tobacco use was associated with greater odds of loose teeth (AOR, 1.93; 95% CI, 1.15-3.26) and lifetime precancerous lesions (AOR, 3.85; 95% CI, 1.73-8.57). Use of other noncigarette products (e.g., pipes) was inconsistently associated with oral health outcomes. PATH Study oral health measures closely align with self-reported measures from NHANES and are internally concurrent. Observed associations with tobacco use and the ability to examine emerging tobacco products support application of PATH Study data in dental research, particularly to examine potential oral health effects of novel tobacco products and longitudinal changes in tobacco use behaviors.

Sociodemographic and psychopathologic predictors of first incidence of DSM-IV substance use, mood and anxiety disorders: results from the Wave 2 National Epidemiologic Survey on Alcohol and Related Conditions.

The objective of this study was to present nationally representative findings on sociodemographic and psychopathologic predictors of first incidence of Diagnostic and Statistical Manual of Mental Disorders, 4th edn (DSM-IV) substance, mood and anxiety disorders using the Wave 2 National Epidemiologic Survey on Alcohol and Related Conditions. One-year incidence rates of DSM-IV substance, mood and anxiety disorders were highest for alcohol abuse (1.02), alcohol dependence (1.70), major depressive disorder (MDD; 1.51) and generalized anxiety disorder (GAD; 1.12). Incidence rates were significantly greater (P<0.01) among men for substance use disorders and greater among women for mood and anxiety disorders except bipolar disorders and social phobia. Age was inversely related to all disorders. Black individuals were at decreased risk of incident alcohol abuse and Hispanic individuals were at decreased risk of GAD. Anxiety disorders at baseline more often predicted incidence of other anxiety disorders than mood disorders. Reciprocal temporal relationships were found between alcohol abuse and dependence, MDD and GAD, and GAD and panic disorder. Borderline and schizotypal personality disorders predicted most incident disorders. Incidence rates of substance, mood and anxiety disorders were comparable to or greater than rates of lung cancer, stroke and cardiovascular disease. The greater incidence of all disorders in the youngest cohort underscores the need for increased vigilance in identifying and treating these disorders among young adults. Strong common factors and unique factors appear to underlie associations between alcohol abuse and dependence, MDD and GAD, and GAD and panic disorder. The major results of this study are discussed with regard to prevention and treatment implications.

Posttraumatic stress disorder among substance users from the general population.

OBJECTIVE: The purpose of the study was to evaluate the prevalence of posttraumatic stress disorder (PTSD) among substance users in the general population. METHOD: The St. Louis Epidemiologic Catchment Area study, a survey of psychiatric illness in the general population, collected data on PTSD and substance use with the Diagnostic Interview Schedule. Among the 2,663 respondents, 430 reported a traumatic event that could qualify for PTSD; however, the rate of PTSD was low, 1.35% overall. To evaluate the relationship between PTSD and substance use, respondents were hierarchically classified into one of four substance use categories ranging from polydrug use to alcohol use only. Substance users from each category as well as substance users in general were compared with persons who did not meet the substance use threshold (comparison subjects). RESULTS: Findings indicate that cocaine/opiate users are over three times as likely as comparison subjects to report a traumatic event, report more symptoms and events, and are more likely to meet diagnostic criteria for PTSD. Physical attack, but not combat-related events, was the most prevalent event reported among cocaine/opiate users. Onset of substance use preceded onset of posttraumatic symptoms, suggesting that substance use predisposes the individual to exposure to traumatic events. When other variables--including antisocial behavior--were controlled, female gender and use of cocaine/opiates predicted PTSD. CONCLUSIONS: These analyses of the co-occurrence of substance abuse and PTSD warrant further study and suggest that PTSD is much more common among substance abusers than was previously known.

New methods in cross-cultural psychiatry: psychiatric illness in Taiwan and the United States.

OBJECTIVE: Cross-cultural psychiatric research has suffered from many methodological shortcomings. To answer some of these shortcomings, the present study compared rates of psychiatric disorders in Taiwan and the United States by combining data from both countries into a single data set. METHOD: Results from large, community-based surveys in the United States and Taiwan, the National Institute of Mental Health (NIMH) Epidemiologic Catchment Area survey and the Taiwan Psychiatric Epidemiological Project, were combined into a single data set. This integration of the data sets was possible because both surveys used the NIMH Diagnostic Interview Schedule to ascertain cases. The integrated data sets were then analyzed with identical algorithms to generate lifetime prevalence rates of psychiatric disorders according to DSM-III criteria for both the United States and Taiwan. RESULTS: Lifetime prevalence rates of psychiatric illness in Taiwan were generally lower than U.S. rates. The rates of any disorder were 21.56% in Taiwan and 35.55% in the United States (Z = 22.34, p less than 10(-109]. The rates of most specific disorders were lower in Taiwan, and none of the rates was higher in Taiwan. CONCLUSIONS: While a culturally determined response bias may have lowered the rates in Taiwan somewhat, the results appear to be valid. Implications for the future use of structured diagnostic interviews in cross-cultural research are discussed.

Alcohol abuse and dependence: consistency in reporting of symptoms over ten years.

The present study assesses the stability of the diagnosis of alcoholism among a sample of St Louis, USA Epidemiological Catchment Area Study participants. The Diagnostic Interview Schedule was used at Time 1 and the Composite International Diagnostic Interview--Substance Abuse Module was used at 10-year follow-up. Alcohol abuse and dependence were diagnosed using DSM-III criteria. Kappa values and Yule's statistics for agreement in diagnoses, criterion groups and individual symptoms are reported. Fair to good measurement of agreement was shown for any diagnosis of alcohol abuse or dependence versus no diagnosis. The criterion group for "impairment in social or occupational functioning" showed the highest agreement of the three criterion groups studied. Agreement was lower for individual symptoms of alcoholism. Incident cases accounted for some of the inconsistency in responses over 10 years. Other types of inconsistency found were changing response from yes to no, changing response from no to yes but not meeting the definition of an incident case, and misstating age of onset of symptoms. Mean per cent inconsistency across all symptoms was 11.4%. Further research regarding reasons for inconsistencies is needed.

Inhalant use among urban American Indian youth.

AIMS: To assess the prevalence of inhalant use among urban American Indian youth and to examine differences between inhalant users and non-users. DESIGN: Baseline (T1) self-report questionnaires completed in 5th-6th grade and at seven annual follow-up assessments (T2-T8). SETTINGS: Seattle metropolitan area. PARTICIPANTS: Two hundred and twenty-four Indian youth. MEASUREMENTS: Youth-completed measures of substance use, ethnic self-identity, involvement in traditional Indian activities, family conflict, family history of alcoholism, peer and sibling deviance, self-esteem, delinquency, aggression, anxiety, depression, sensation seeking, conduct disorder and alcohol dependence. FINDINGS: Lifetime inhalant use was reported by 12.3% of adolescents. At T1, inhalant users had significantly lower perceived self-worth and average annual household incomes and significantly greater density of familial alcoholism and expression of aggressive and delinquent conduct than non-users. Aggressive behavior was the most important T1 predictor of inhalant use. Lifetime conduct and alcohol dependence disorders were 3.3 and 2.6 times more prevalent among inhalant users than non-users at T5. Inhalant users had more extensive deviant peer networks, were more sensation-seeking, and evidenced lower perceived self-worth than non-users at T8. CONCLUSIONS: Inhalant use was less prevalent in this particular sample of urban Indian adolescents than in most studies of reservation Indian youth. As with other studies of inhalant abuse, aggressive and delinquent males of low SES and low-perceived self-worth with family histories of alcohol dependence, were at highest risk for inhalant use.

Cocaine use and HIV risk in out of treatment drug abusers.

During the initial phases of the HIV epidemic, epidemiological and prevention initiatives included a necessary focus on injection drug use which targeted, for the most part, heroin injectors. More recently, cocaine users have become a central focus for many HIV prevention interventions. This illustrates how researchers and public health practitioners have adapted to the changing and intersecting epidemics of HIV and drug abuse in order to more efficiently prevent the spread of HIV among drug abusers. In this special issue of Drug and Alcohol Dependence, the contributing authors focus attention on the HIV risk behaviors of cocaine abusers using data from the National Institute on Drug Abuse (NIDA) Cooperative Agreement for AIDS Community-Based Outreach/Intervention Research Program. The Cooperative Agreement consists of 23 sites, with 22 sites located in the US (including Puerto Rico) and one site in Brazil, and has included over 28,000 subjects recruited using street outreach. All Cooperative Agreement subjects were administered the NIDA risk behavior assessment (RBA) which provides detailed information about injection drug use behaviors, crack cocaine use and sexual behaviors. The RBA serves as the source of basic data for all papers in this issue which includes work describing local studies in St. Louis, North Carolina and Brazil as well as analysis of the National Database from researchers in Denver, Washington and Kentucky. In general, the studies focus on the etiology of HIV risk behaviors by empirically identifying the antecedents of individual HIV risk behaviors among cocaine users. These predictors of risk and protective behavior include the severity of crack cocaine, comorbid antisocial personality disorder and depression, alcohol consumption, and typologies of crack and injection drug use. Overall, the studies show that NIDA's prevention intervention is feasible and effective in many settings. The success of the studies which are included in this special issue is a testament to the feasibility of NIDA's science-based prevention interventions.

Results of the NIDA treatment demonstration grants' cocaine workgroup: characteristics of cocaine users and HIV risk behaviors.

This journal issue includes seven articles (six plus this introduction) from the "cocaine workgroup" of the National Institute on Drug Abuse (NIDA) treatment demonstration grants. In this introduction, results of the first attempts to compare data from seven disparate demonstration grant sites are summarized: Overall, rates of recent cocaine use were high in all locations, injection drug use was common, age of first drug use was between 14 and 17 years with age of first cocaine use between 20 and 25 years, arrests were common at all sites especially among cocaine injectors, and polydrug use was the norm. Interestingly, both gender and ethnic status were significantly associated with polydrug use and marijuana use among the cocaine users. These results indicated that it is possible to define variables precisely for analysis across sites and laid the groundwork for the next set of analyses in which the common theme of human immunodeficiency virus (HIV) risk behaviors among cocaine abusers was agreed upon. This next set of analyses are included in the following six papers. Overall, these reports confirm recent data about the association of cocaine use with HIV risk behaviors. They extend considerably the literature on the association of cocaine with HIV risk behaviors, and the report from New York in which therapeutic community treatment was shown to be feasible and possibly useful to methadone clients represents an interesting and new finding. In conclusion, cross-site collaborations can take different forms and this collection of papers represents one successful approach.

Agreement between DSM-III and III-R substance use disorders.

With proposed criteria for DSM-IV substance dependence imminent, an evaluation of the impact of changes from DSM-III to DSM-III-R would be informative. Recent admissions to St. Louis drug treatment centers were interviewed with the DIS-III-R, which covers criteria from both systems. Kappa values for system agreement, diagnostic overlap and percent positive agreement are reported by substance. The DSM-III-R system cast a wider net for dependence than DSM-III for alcohol, tobacco and amphetamines. Neither system predominated for cannabis, opioids and barbiturates/sedatives/hypnotics. Reasons for differences and implications of findings are discussed.

Subjective reports of withdrawal among cocaine users: recommendations for DSM-IV.

The proposed DSM-IV criteria for substance use disorders have included, as an option, a subtyping for physiologic dependence, characterized by either tolerance or withdrawal. Even if this option is not chosen at this stage of system revision, this weighting scheme justifies wider surveillance of these symptoms, especially for the more newly described cocaine dependence disorder. Wider surveillance of withdrawal is possible with the CIDI Substance Abuse Module (SAM), a WHO/ADAMHA diagnostic interview which covers criteria of substance use disorders according to the DSM-III, III-R, ICD-10 and proposed DSM-IV systems. To aid in this effort, we used the SAM, which includes a master list of all symptoms (n = 16) in the DSM manuals related to withdrawal from any substance to assess withdrawal symptoms from all substances. In this study, we hypothesized that the persons who used opiates with cocaine might misattribute their symptoms to cocaine; thus, we compared the responses of persons who used cocaine and opiates (opiate users) with the responses of persons who used cocaine without opiates (non-opiate users). Data from two St. Louis studies were combined for these analyses, users not in treatment or newly enrolled to drug-free or methadone treatment from a NIDA demonstration project and users selected for the St. Louis DSM-IV Field Trial. Of the 196 persons included from the field trial, 80% reported lifetime cocaine use compared with 91% of the 412 persons from the demonstration project. The symptoms mentioned in the diagnostic manuals were among the most frequently endorsed by both cocaine use groups.(ABSTRACT TRUNCATED AT 250 WORDS)

Is antisocial personality disorder associated with increased HIV risk behaviors in cocaine users?

Previous reports have shown antisocial personality disorder (ASPD) to be strongly associated with injection equipment sharing and increased rates of human immunodeficiency virus (HIV) infection in a sample of heroin injectors. Another report has shown ASPD to be associated with injection drug use, needle sharing, sexual promiscuity, and prostitution in a sample of methadone maintenance clients. The current study extends this work by examining the relationship of ASPD and human immunodeficiency virus (HIV) risk behaviors in a sample of cocaine users (48% out of treatment and 52% just entering treatment). Associations were tested for sexually risky behaviors in addition to injection behaviors. The principle finding of this study is that ASPD was shown to be associated with increased rates of injection drug use and sharing syringes, with earlier age of onset of injection drug use, with certain venereal diseases, and with a variety of HIV risk sexual behaviors. When men and women were tested separately, the pattern of association of risky behaviors with ASPD varied considerably. Overall, this work confirms that psychiatric status, especially the presence of ASPD, may have to be considered in evaluating the results of HIV risk-reduction interventions.

Diagnostic concordance of DSM-III-R, DSM-IV, and ICD-10 inhalant use disorders.

The diagnostic concordance of DSM-III-R, DSM-IV, and ICD-10 inhalant use disorders was assessed using the sample of lifetime inhalant users (n=76) participating in the DSM-IV Field Trial for Substance Use Disorders. Substantially smaller proportions of lifetime inhalant users met DSM-IV inhalant abuse or dependence criteria than met comparable DSM-III-R or ICD-10 criteria. DSM-III-R and ICD-10 performed similarly, although DSM-III-R tended to be more inclusive vis-à-vis diagnoses of inhalant dependence. Kappa coefficients indicated a moderate degree of concordance between the three nosologies for inhalant use disorder diagnosis rates. Inclusion of possible withdrawal symptomatology criteria (that are not normally included) in the DSM-IV and ICD-10 criteria sets for inhalant dependence exerted little effect on diagnosis rates.

Reliability of self-reported antisocial personality disorder symptoms among substance abusers.

It is estimated that from 20 to 60% of substance abusers meet criteria for Antisocial Personality Disorder (APD). An accurate and reliable diagnosis is important because persons meeting criteria for APD, by the nature of their disorder, are less likely to change behaviors and more likely to relapse to both substance abuse and high risk behaviors. To understand more about the reliability of the disorder and symptoms of APD, the Diagnostic Interview Schedule Version III-R (DIS) was administered to 453 substance abusers ascertained from treatment programs and from the general population (St Louis Epidemiological Catchment Area (ECA) follow-up study). Estimates of the 1 week, test-retest reliability for the childhood conduct disorder criterion, the adult antisocial behavior criterion, and APD diagnosis fell in the good agreement range, as measured by kappa. The internal consistency of these DIS symptoms was adequate to acceptable. Individual DIS criteria designed to measure childhood conduct disorder ranged from fair to good for most items; reliability was slightly higher for the adult antisocial behavior symptom items. Finally, self-reported 'liars' were no more unreliable in their reports of their behaviors than 'non-liars'.

Cocaine users with antisocial personality improve HIV risk behaviors as much as those without antisocial personality.

Previous work has documented that antisocial personality disorder (APD) is associated with increased rates of HIV risk behaviors and with worse substance abuse treatment outcomes. The question addressed by this paper is whether cocaine users with APD respond to an HIV risk-reduction intervention as well as cocaine users without the disorder. The study subjects were 333 cocaine users followed up at 18 months as part of a NIDA-funded treatment demonstration project. The total sample improved across a wide range of HIV risk behaviors. Improving significantly (P < 0.05) from baseline to the 18-month follow-up were several drug-related behaviors: cocaine use; current cocaine dependence; use of drugs other than cocaine drug injection; injection equipment sharing; and use of syringes that were not cleaned. Several sex-related HIV risk behaviors also improved significantly: having multiple sex partners; being intoxicated during sex; giving drugs for sex; receiving money for sex; and receiving drugs for sex. When the sample was stratified by APD status, very similar improvement was seen in respondents with and without APD. To examine further the relationship of APD to change in HIV risk behaviors, separate logistic regression models of improving and worsening HIV risk behaviors were tested. What the authors found was no association of APD with improvement in HIV risk behaviors but a significant association of APD with worsening HIV risk behaviors. It appears that cocaine users with APD improve their HIV risk behaviors just as much as those without APD but may be at higher HIV risk for the development of such behaviors.

Self-report stability for substance use over 10 years: data from the St. Louis Epidemiologic Catchment Study.

In substance abuse research many study protocols require the recall and report of behavior from the distant past that may affect reliability. This study addresses the stability of substance use reports over a 10-year follow-up period. We reinterviewed 223 ECA subjects who reported baseline drug use. Reports from baseline to follow-up were categorized as either incident cases, agreements, or discrepancies. Different patterns were found for each substance group. The authors found that cannabis, cocaine, sedatives, and opiates had high agreement rates ranging from 82-86% and amphetamines had the lowest (73%). Implications for future drug research are discussed.

The effects of psychiatric comorbidity on response to an HIV prevention intervention.

Drug abusers with psychiatric comorbidity are at high risk for becoming exposed to HIV. To address this compelling public health issue, our randomized HIV prevention study compares the effectiveness of the NIDA standard HIV testing and counseling protocol to a four session, peer-delivered, educational intervention for out-of-treatment cocaine users with and without antisocial personality disorder (ASPD) and major depression. Among the 966 out-of-treatment cocaine users who have completed the 3 month follow-up, all groups, regardless of assignment to standard vs. peer-delivered intervention or psychiatric status, improved significantly in: crack cocaine use, injection drug use, number of IDU sex partners and overall number of sex partners, but not in condom use. Nevertheless, when stratified by psychiatric status, ASPD was associated with significantly less improvement in crack cocaine use (P = 0.04) and with a trend for less improvement in having multiple sex partners and having IDU sex partners (P = 0.06 and 0.08, respectively). ASPD status was not associated with change in injection drug use or condom use. Depression was associated with a trend (P = 0.07) for greater improvement in crack cocaine use but not in any of the other behaviors. When examining the standard and peer intervention groups separately, no consistent differences in the association of psychiatric comorbidity with outcome were discerned between the two groups. We conclude that persons with ASPD and depression respond well to standard HIV prevention interventions, but these psychiatric disorders respectively attenuate and enhance response somewhat. Behavioral interventions tailored for persons with these conditions may be indicated if long-term change in HIV risk behaviors is to be achieved.

Comparing assessments of DSM-IV substance dependence disorders using CIDI-SAM and SCAN.

The main question addressed by this paper is whether DSM-IV substance dependence diagnoses obtained from two different instruments (the semi-structured WHO Schedules for Clinical Assessment in Neuropsychiatry, SCAN and the highly structured WHO Composite International Diagnostic Interview--Substance Abuse Module, SAM) are as consistent as diagnoses obtained from a single instrument (SAM) administered twice. Such comparisons of results from the two different instruments provide some measure of validity of the lay-administered SAM and of the underlying diagnostic concepts. Chance-corrected concordance was estimated using the kappa coefficient for SAM/SCAN (test/validation) and SAM/SAM (test/retest) comparisons. Analyses of agreement between SAM and SCAN for DSM-IV dependence diagnoses indicated good agreement for alcohol and cocaine, and fair agreement for opiates and cannabis. SAM/SAM (test/retest) agreement was excellent for alcohol and opiate dependence, good for cocaine dependence, and fair for cannabis dependence. Agreement on individual dependence criteria was generally consistent with overall diagnostic agreement though more variable. Notable was the poor agreement for cannabis criteria in the SAM/SCAN protocol. This may indicate that the dependence syndrome is less applicable to cannabis, while the consistency of agreement for alcohol, opiate, and cocaine dependence criteria supports the validity of these dependence syndromes. Finally, these data indicate that both the clinical (SCAN) and non-clinical (SAM) interviews can be used effectively for a variety of substances and dependence diagnoses.

Achieving a 96.6 percent follow-up rate in a longitudinal study of drug abusers.

Longitudinal studies can be hampered by poor follow-up rates, seriously reducing generalizability of the findings. Understanding the barriers, as well as approaches to overcome and adapt to these impediments, resulted in a 96.6% 18 month follow-up rate of 479 drug abusers enrolled in an NIDA funded demonstration project aimed at reducing HIV transmission among St. Louis' most vulnerable drug-using population. In this paper, we discuss the importance of phone and systems tracking, creative team work and persistence and procedures for tracking out-of-treatment drug-users by analyzing the efforts needed to reach the study subjects. We also compared the characteristics of hard-to-reach respondents with those who were less difficult. The results revealed that employment status was the only characteristic associated with being hard-to-reach. Gender, race, age, education and psychiatric status did not discriminate recruitment difficulty in this sample. The study findings suggest that although unemployment predicted recruitment difficulty, and that the locator information provided by the respondent can be very helpful, a comprehensive tracking strategy as well as persistence and creative team work are the most important determinants of the rate of success of a follow-up investigation.

Reaching and enrolling drug users for HIV prevention: a multi-site analysis.

Since 1994, several sites have participated in a NIDA Cooperative Agreement for AIDS Community-based Outreach/Intervention Research Program to examine rates of HIV risk behaviors and evaluate HIV risk reduction interventions among out-of-treatment drug injection and crack cocaine and heroin smokers. We studied the process and outcome of community outreach for recruitment of drug users in AIDS research and education projects in three metropolitan areas: St. Louis, MO; San Antonio, TX, and Durham and Wake Counties, NC. There were two primary areas of focus: (1) the level of accuracy among community health outreach workers (CHOWs) in identifying potentially eligible persons for HIV prevention, and (2) overall effectiveness in recruiting and enrolling persons in formal assessment and intervention studies. We found cross-site and within-site differences in levels of accuracy and in recruitment and enrollment yields. Drug users who had never been in treatment and drug users who had never been tested for HIV infection were underrepresented at all sites. We discuss the factors which may have contributed to cross-site and within-site differences. The findings suggest a need for continued study, refinement, and evaluation of community outreach strategies in order to enroll a broad spectrum of vulnerable groups in HIV prevention activities.