- Email: [email protected]
Factors affecting client response to HIV outreach efforts
Journal of Substance Abuse 13 (2001) 201 – 214
Factors affecting client response to HIV outreach efforts Penny Dorsey Tinsmana,*, Stephanie Bullmana, Xiaowu Chena, Kenneth Burgdorf a, James M. Herrellb a Caliber Associates, Suite 400, 10530 Rosehaven Street, Fairfax, VA 22030, USA Center for Substance Abuse Treatment1, Suite 740, 5600 Fishers Lane, Rockwall II, Rockville, MD 20857, USA
b
Abstract Purpose: This article describes 12 HIV Outreach Demonstration Projects funded by the Center for Substance Abuse Treatment in 1995, and the clients these projects served. The article also summarizes the findings of multivariate statistical analyses aimed at identifying important project and client characteristics that influenced project success in achieving two key outcomes: persuading at-risk clients to obtain HIV tests, and facilitating entry by substance-abusing clients into structured substance abuse treatment (SAT). Methods: Hierarchical linear modeling (HLM) analysis was used to conduct the analyses. Results: The findings support the hypothesis that HIV Outreach, as an integrated approach to addressing the multiple problems clients have due to substance abuse and related problems, can be an effective model for reaching clients who have not been reached through traditional means. Implications: By implementing a complement of comprehensive HIV Outreach interventions, many of the HIV Outreach projects were successful in addressing the various needs of their clients. However, as detailed in this article, some services and service delivery procedures, were more effective than others in achieving project objectives. D 2001 Elsevier Science Inc. All rights reserved. Keywords: HIV Outreach; Injection drug user; HIV testing; Substance abuse treatment; Hard-to-reach populations
1. Introduction The human immune deficiency virus (HIV)/acquired immune deficiency syndrome (AIDS) epidemic continues to pose a serious health threat in the US. Since the epidemic began, * Corresponding author. Tel.: +1-703-385-3200; fax: +1-703-385-3206. E-mail address: [email protected] (P.D. Tinsman). 1 Tel.: + 1-301-443-5052. 0899-3289/01/$ – see front matter D 2001 Elsevier Science Inc. All rights reserved. PII: S 0 8 9 9 - 3 2 8 9 ( 0 1 ) 0 0 0 6 6 - 9
202
P.D. Tinsman et al. / Journal of Substance Abuse 13 (2001) 201–214
injection drug use (IDU) has directly or indirectly accounted for more than one third (36%) of AIDS cases in the US (Centers for Disease Control and Prevention (CDC), 1998). Of the 48,269 new cases of AIDS reported in 1998, 14,024 (31%) were associated with IDU (CDC, 1998). IDU-associated AIDS accounts for a larger proportion of cases among women than men. Since the epidemic began, 59% of all AIDS cases among women have been attributed to IDU or sex with partners who inject drugs, compared to 31% of cases among men (CDC, 1998). IDU-associated AIDS varies across racial and ethnic populations in the US (Estrado, 1998; Freeman, Williams, & Saunders, 1999). In 1998, injection drug users (IDUs) accounted for 36% of all AIDS cases among both African American and Hispanic adults and adolescents, compared with 22% of all cases among white adults and adolescents (CDC, 1998). The majority of AIDS cases reported among females in the US has been among women of color since the onset of the epidemic in the US (Health Resource and Services Administration (HRSA), 1999). One successful method of preventing HIV transmission is substance abuse treatment (SAT). SAT has been shown to reduce risk behaviors associated with HIV transmission, such as sharing needles contaminated with blood and engaging in unsafe sexual practices — unprotected sex, trading sex for drugs or for money to buy drugs, etc. A number of studies have reported significant reduction in HIV-related drug use and sexual risk behaviors among IDUs in treatment compared to those not in treatment (Ball, Lange, Myers, & Freidman, 1988; Booth, Kwiatkowski, & Stephens, 1998; Hubbard, Marsden, Cavanaugh, Rachal, & Ginzburg, 1988; Kang & de Leon, 1993; National Institute on Drug Abuse (NIDA), 1999; Needle & Coyle, 1998). As part of its efforts to address the serious health threat presented by HIV/AIDS, the Substance Abuse and Mental Health Services Administration’s (SAMHSA) Center for Substance Abuse Treatment (CSAT) of the U.S. Department of Health and Human Services initiated a multisite, multiyear demonstration program in 1995, designed to test the effectiveness of integrating referral to SAT with street outreach models for HIV prevention to reach populations that are at highest risk for exposure to HIV, yet are historically underserved by the medical community. Specifically, the Community-Based Comprehensive HIV/STD/TB Outreach Services for High-Risk Substance Abusers Demonstration Program (HIV/AIDS Outreach Program) was designed to provide outreach services to IDUs, other chronic, hard-to-reach substance abusers, and their sex and/or needle-sharing partners. The goal of the demonstration was to reduce HIV transmission among these high-risk populations and provide greater access to SAT. Twelve projects were funded in this demonstration. Each provided four core services:
Community-based outreach services to encourage entry and facilitate access to SAT; HIV/AIDS risk-reduction interventions; Medical diagnostic testing and screening for HIV, sexually transmitted diseases (STDs), tuberculosis (TB), and pregnancy; and Linkages to primary medical care, mental health care, social services, and other ‘prophylactic means’ to effect behavioral changes most likely to reduce the risk of acquiring or transmitting HIV, STDs, TB, and related diseases.
P.D. Tinsman et al. / Journal of Substance Abuse 13 (2001) 201–214
203
Each of the 12 demonstration projects was required to conduct a comprehensive evaluation to document the effectiveness of the approaches and services being tested at each site. Additionally, CSAT funded a central, cross-site evaluation of the HIV Outreach program as a whole, based on core sets of project-level and client-level data provided by the 12 individual projects (Rowden et al., 1999). This paper, based on data assembled in the cross-site study, has two main purposes. First, it provides an overview summary of CSAT’s 12 HIV Outreach projects and of the clients they served. Second, it summarizes the findings of multivariate statistical analyses aimed at identifying important project and client characteristics that influenced project success in achieving two important outcomes: persuading at-risk clients to obtain HIV tests, and facilitating entry by substance-abusing clients into structured SAT.
2. HIV Outreach projects and their clients 2.1. Description of projects The HIV Outreach Demonstration Program was based on the hypothesis that by creating extensive linkages among service agencies, and by providing comprehensive services to clients with a multitude of problems, projects would be able to reduce risk behaviors associated with HIV transmission, even among high-risk population groups. By providing comprehensive services, such as pretreatment services, transportation to treatment, and referral to social services, HIV Outreach projects should also increase facilitated access to SAT for target populations. All projects that participated in the cross-site evaluation were located in urban areas in the US or US territories. One of the projects also served a target area that included a rural location. One project was housed in a correctional facility in New York City. Two of the projects were affiliated with local universities; these projects emphasized research and data collection. Two other projects were affiliated with government programs or with government agencies, specifically, a department of corrections and a public health commission. Eight projects were entities of nonprofit organizations whose main goal was SAT or who provided social services to indigent clients. 2.2. Target populations and project services CSAT identified the target population for the demonstration program as IDUs and their sex and needle-sharing partners, an underserved group at extremely high risk for HIV exposure/ infection. Within those parameters, projects could target populations within their communities that they considered to be most in need. Populations targeted for outreach included crack/ cocaine users, homeless individuals, prostitutes, Hispanics, African American women, and IDUs. Overall, the target populations lived in areas marked by widespread drug use, violence, poverty, and health problems.
204
P.D. Tinsman et al. / Journal of Substance Abuse 13 (2001) 201–214
Three of the projects focused solely on women. One project targeted homeless individuals. Seven projects primarily served African American clients, three focused on Hispanic/Latino or Mexican American clients, and two projects did not target a specific group outside of IDUs and their sexual partners. All of the projects followed the traditional or modified Indigenous Leader Outreach Model (ILOM) developed as part of the NIDA-funded National AIDS Demonstration Research Project (NADAR). The core elements of this approach are to provide information about HIV to IDUs and their sexual partners, conduct personalized risk assessments, and provide support for behavior change during street-based interactions, from mobile vans or at drop-in centers located within the community. Programs targeted individual users and their social networks using trained indigenous outreach staff. In the ILOM, the staff are seen as peer leaders and role models among targeted social networks due to their own successful recovery from drug addiction. Using the ILOM, outreach workers are able to provide rapid entry into communities, establish rapport with community leaders, and develop a higher level of trust and acceptance of the outreach project, than would otherwise be possible. Outreach interventions included the following:
Street outreach focused on harm reduction including distributing pamphlets, safe sex kits, and bleach kits; locating IDUs within copping areas and shooting galleries, and providing HIV and SAT information to local area businesses. Eleven of 12 projects engaged in street outreach. Mobile units or vans provided a movable center for outreach services such as counseling and testing as well as transport for outreach workers to target neighborhoods. Advantages to mobile units included high community recognition of the outreach project, safer place for worker/contact interaction, and relative privacy for outreach contacts and service provision. They also served as a focal point to initiate conversations. Outreach workers often distributed materials in front of the van, or offered food or refreshments for individuals who came inside the van to be tested or watch a safe sex demonstration. Four of the 12 projects had vans, two of which were fully equipped to provide medical testing services. Drop-in centers were located within the center of the target community. They offered a safe place for indigent clients as well as a base for outreach services. Services included testing for HIV and STDs, counseling, support groups, referrals for SAT, and other services, and a repository for clothing and food to give away to needy clients. Half of the 12 projects had a drop-in center.
2.3. Description of client recruitment The HIV Outreach projects used multiple strategies to engage their clients. Eleven of the 12 projects focused on neighborhoods known to frequented by the target population. These locations included ‘‘street locations’’ (e.g., copping areas/outdoor drug markets, shooting galleries, commercial sex worker ‘‘strolls,’’ crack houses, etc.); parks where drug trafficking occurred; social locations (e.g., gay and/or straight bars, churches, and community centers);
P.D. Tinsman et al. / Journal of Substance Abuse 13 (2001) 201–214
205
social service/institution/community agencies, including hospitals, clinics, emergency rooms, jails, prison release programs, and food distribution centers. Incentives were used to attract a clientele. Projects reported using hygiene kits (shampoo, razor, toothbrush, shower vouchers, etc.) and clothing. Food was also offered as an incentive. Sandwiches, snacks, and sodas were often offered from the project’s mobile unit; donuts and coffee would be available at the drop-in center; and in other cases, food vouchers would be available for community locations. Further, projects used multiple outreach strategies to attract the ‘‘hard-to-reach’’ population. The projects would combine regular street outreach with a mobile unit in the community that could offer services ‘‘on the street.’’ Drop-in centers, which were also successful in recruiting clients, afforded a safe location receive services or just to get off the street. Projects referred clients to medical services such as HIV, STD, TB, and pregnancy testing or were able to provide the service(s) directly to clients. Fifty-four percent of all clients received a referral for HIV testing. Ten percent of clients received a referral for TB testing. Eleven percent of clients were referred for primary health care. Twelve percent of clients received a referral to social services. Social services included housing assistance, assistance in obtaining documentation, assistance in obtaining Medicaid, etc. Fifty-three percent of all clients were tested for HIV, and 15% of all clients received a TB test. Projects provided services to facilitate their clients’ access and entry into SAT. Services included preparing clients for treatment through assessment of their readiness to enter treatment, support groups and ancillary services for clients who were on waiting lists for treatment. Thirty-one percent of clients received an HIV Education Workshop. Forty-five percent of clients attended a Risk Reduction/Education Workshop. Fifty-seven percent of clients received pretest for HIV counseling. Sixty-four percent of all clients received riskreduction counseling. Forty-one percent of the HIV Outreach clients who had reported using a substance within the past 30 days were referred to SAT. Of those who were referred, 68% entered SAT. 2.4. Description of clients The cross-site study received usable demographic data for 9296 HIV Outreach clients. Fifty-five percent of these clients were males and 45% were females. Given the approximate equality in the gender proportion, it must be noted that several projects targeted specific populations such as at-risk women (three projects), one project served only men, and the remaining projects served relatively evenly distributed proportions of male and females. Fifty-five percent of the HIV Outreach clients indicated black/African American as their race/ethnic group. Twenty percent of the clients identified themselves as Hispanic. Eighteen percent identified themselves as white. A small percentage (5%) of clients reported their race/ ethnicity as Asian/Pacific Islander. The remaining 2% of clients reported themselves as Native American (0.4%), Alaskan Native (0.1%), and other (1%). The mean age of the HIV Outreach clients was 35 years old. Forty-one percent of the clients were between the ages of 31 and 40. Twenty-five percent of the HIV Outreach clients were in the following age groups: 21 through 30 years old, and 41 through 55 years old. Only
206
P.D. Tinsman et al. / Journal of Substance Abuse 13 (2001) 201–214
6% of the clients were less than 21 years old, and 3% of the HIV Outreach clients were over the age of 55. Of those clients reporting marital status, 46 reported that they had never married and were not living with a significant other. Nearly one-quarter of the HIV Outreach clients reported that they were separated, divorced, or widowed. Fourteen percent of the clients reported that they were currently married, and 7% reported that they were living with a significant other. Additionally, one third reported that they had children. Seven percent of the female clients were pregnant at entry into the HIV Outreach program. Forty-seven percent of the HIV Outreach clients had achieved a high school diploma. Thirteen percent reported post-high school education and over one third (38%) reported that their highest level of education completed was between the seventh and eleventh grades. Half of the HIV Outreach clients reported living during the 30 days prior to entry into the program in a house, condo, or apartment. Of the other clients, 4% indicated that they had stayed in public housing 30 days prior to intake in the HIV Outreach project. Thirtytwo percent had stayed in a homeless shelter or on the street, 3% reported being in jail or prison 30 days before entering the HIV Outreach project, and 10% reported that they had stayed in another type of shelter (i.e., recovery homes, halfway houses, etc.), or other group living situations. Not all clients entering the HIV Outreach Demonstration Program were substance abusers. Many clients were the sexual partners of drug users and/or IDUs, or were engaged in sexual activities such as prostitution that put them at risk for exposure to the HIV virus. Fifty-three percent of the HIV Outreach clients reported that they had engaged in sexually risky behaviors in their lifetimes. These risky sexual behaviors include having unsafe/unprotected sex in the past 30 days, ever having unsafe sex with an injection drug user, ever using sex to obtain drugs or money, and ever having sex with someone who is or might be infected with the HIV virus. Because of the danger posed by injection behaviors, HIV Outreach projects specifically targeted IDUs. Thirty-five percent of the clients reported that they had injected drugs in their lifetime, and 31% reported that they had injected drugs in the 30 days prior to entry into the HIV Outreach project. Over three-fifths (61%) of the HIV clients reported that they had used substances in the 30 days prior to entering the HIV Outreach project. Forty percent of the clients reported using a substance other than alcohol. Table 1 presents the reports of substance abuse in the 30 days prior to entry into the HIV Outreach project.
Table 1 Substance use patterns Substance
Percentage of clients reporting use in the last 30 days (%)
Cocaine Crack Marijuana/hashish Heroin Alcohol
18 28 17 14 46
P.D. Tinsman et al. / Journal of Substance Abuse 13 (2001) 201–214
207
3. Methods and results 3.1. Analytic approach This paper examines factors associated with project success in achieving one or another of two important outreach objectives: (1) persuading high-risk clients to obtain HIV tests, and (2) facilitating entry into SAT of hard-to-reach, high-risk clients with substance abuse problems. The cross-site database contains two broad categories of independent variables that are potentially useful in explaining such client outcomes: variables describing projectlevel differences in outreach approaches, components, procedures, etc., and variables describing client-level differences in demographic characteristics, HIV risk factors, and exposure to various outreach services. The cross-site database is hierarchical in structure: clients are nested within projects. The multivariate analytic approach that is most appropriate for use with this kind of data structure is hierarchical linear modeling (HLM) analysis. This approach permits simultaneous identification of: (1) project-level effects, controlling for between-project differences in client composition, (2) client-level effects that are consistent across projects, and (3) interaction effects among variables, both within and between levels. For these analyses, we used the SAS macro GLIMMIX. We began by entering selected client characteristics into the HLM model. We then entered project-level variables into the model one at a time, retaining those that made statistically significant improvements in goodness of fit, either as a main effect or in interaction with other variables (Littell, Milliken, Stroup, & Wolfnger, 1996). 3.2. HIV testing Arguably, it is desirable that all of the high-risk clients served by HIV Outreach projects be screened for the presence of the HIV virus. Unfortunately, the HIV status of the clients is not available. Of the 12 CSAT-funded HIV Outreach projects, 8 provided the cross-site study with information about whether or not their clients actually received any HIV tests. These eight projects collectively served a total of 5851 high-risk clients, 63% of who were reported to have received one or more HIV tests during the course of their contact with the projects. In seeking to account for the differences between clients who received HIV testing and those who did not, our HLM analysis examined the influences of the following independent variables. 3.2.1. Client-level variables Gender; age; race; use/nonuse in the past 30 days of cocaine, crack, marijuana, heroin, alcohol; whether or not any IDU in the past 30 days; whether or not client had unprotected sex in the past 30 days; whether or not client is a sexual partner of a drug user; whether or not client received HIV risk education services at the Outreach project; and whether or not client received any transportation services through the Outreach project (bus tokens, rides to drop-in centers in project-operated vehicles, rides to SAT or other service locations, etc.).
208
P.D. Tinsman et al. / Journal of Substance Abuse 13 (2001) 201–214
3.2.2. Project-level variables Whether or not project was gender-focused (two projects in the analysis set primarily targeted women clients, the rest were not gender-focused); whether or not project provided on-site HIV testing (six projects provided testing at Outreach project facilities, the rest provided testing through referral to other facilities); and whether or not the project had a mobile unit in which HIV tests could be performed (two projects had such units, the rest did not). Projects with mobile units for HIV testing represent a subset of those with capabilities for on-site HIV testing. The other on-site testing projects did HIV testing primarily at drop-in facilities. Table 2 summarizes the results of the HLM analysis of the relationships between these variables and whether or not an at-risk client actually received HIV testing. Only variables that produced statistically significant (at P < .05) associations with HIV testing are shown. Relationships are expressed as odds ratios (ORs), which indicate how much more or less likely it was that a client with the indicated characteristic was tested for HIV, as compared to a client who did not have the indicated characteristic. ORs above 1.0 indicate that clients with a given characteristic were more likely than others to obtain an HIV test; conversely, an OR less than 1.0 indicates that a client with a given characteristic was less likely than others to receive an HIV test. For example, one important finding in Table 1 is that, consistently across projects, clients who received an initial HIV risk education program were more than four Table 2 Results of HLM analysis of factors associated with whether or not HIV Outreach clients received HIV testinga 95% Confidence interval Variable Client level Used crack in last 30 days Used marijuana in last 30 days Had unprotected sex in last 30 days Sexual partner of drug user Received HIV risk education Project level On-site HIV testing Mobile unit
Odds ratio
Upper
P
0.53 0.58 0.43 0.27 4.64
0.43 0.45 0.35 0.22 3.73
0.65 0.76 0.54 0.35 5.77
< .0001 < .0001 < .0001 < .0001 < .0001
21.22 86.06
1.14 6.40
396.68 1156.40
.0409 .0008
3.57
38.15
< .0001
0.37
11.31
.0834
Interaction of client- and project-level variables 11.67 Projects without on-site HIV testing: Clients who received transportation services vs. clients who did not receive transportation services 2.10 Projects with on-site HIV testing: Clients who received transportation services vs. clients who did not receive transportation services a
Lower
Observations are relative to clients/projects with characteristics.
P.D. Tinsman et al. / Journal of Substance Abuse 13 (2001) 201–214
209
times as likely to obtain an HIV test as clients who did not receive such a program (OR = 4.64, P < .0001). Several client-level variables were not significantly associated with HIV testing. This includes all three client demographic characteristics (gender, age, race) and several major HIV risk factors, including IDU and recent use of heroin or cocaine. Several other HIV risk factors were negatively associated with HIV testing, as indicated by ORs below 1.0. For example, recent crack or marijuana users were only about half as likely as nonusers of these substances to obtain HIV testing (OR = 0.53 and 0.58, respectively, with P < .0001 for both). Clients who reported having had unprotected sex in the past 30 days were also about half as likely to receive HIV testing as clients without this risk factor (OR = 0.43, P < .0001). Clients who were sexual partners of drug users were only about one-fourth as likely as other clients to obtain HIV testing (OR = 0.27, P < .0001). These risk factor associations are somewhat difficult to interpret, however, considering that all clients in the outreach database were identified as being at high risk of HIV in one way or another. Thus, clients who did not have any particular risk factor typically did have one or more other risk factors. The strongest and most important associations in Table 2 concern project characteristics, particularly the mechanisms projects used to deliver HIV testing services. The projects that used mobile testing units to reach their clients were vastly more successful than other projects in their HIV testing efforts: all else being equal, outreach clients at projects with mobile units were 86 times more likely to obtain an HIV test than those at other outreach projects. More generally, clients who were served by projects that had any capability for on-site HIV testing were substantially (21 times) more likely to receive HIV testing than those served by projects that used referrals to other agencies as their testing method. However, it is possible that not all clients tested by outside testing facilities are counted; projects relied on both client self-report and linkage partners to provide the information. Table 2 also shows an interesting and understandable interaction effect between whether or not a project used on-site HIV testing methods and whether or not the client received projectprovided transportation services. For projects that did not have on-site testing facilities, it appears that it was critically important to provide clients with transportation support, presumably to help them get to the project’s testing location: among clients at projects without on-site testing, those who received transportation support were 11.67 times more likely to receive HIV testing than those without transportation help ( P < .0001). For clients at projects that provided on-site HIV testing, presence or absence of transportation support was not significantly associated with HIV testing outcomes. 3.3. Entering SAT Perhaps the single most important objective of the CSAT-funded HIV Outreach demonstration projects was to find effective methods of facilitating entry into structured SAT programs for IDUs and other groups of substance abusers who are at high risk for HIV/AIDS. It is important to note that project staff played an integral role in whether a client entered SAT; because of the possible availability of treatment slots within the organization and additional linkage agreements, staff rarely referred a client to SAT without first securing an available
210
P.D. Tinsman et al. / Journal of Substance Abuse 13 (2001) 201–214
bed. Waiting lists would not be a viable option for this population. Of the 12 demonstration projects, 9 were able to provide outcome information to the cross-site study about whether or not individual clients received any form of structured SAT. These nine projects served 6155 clients, 4083 of whom reported past-month IDU or other past-month use of alcohol or other drugs. Among these SAT-eligible outreach clients, 1686 (41%) actually entered a structured SAT program. It was disappointing to note, in our initial analysis of the raw data, that the rate of SAT entry was considerably lower among IDU clients (30%) than among other substanceusing clients (51%). Our HLM analysis of factors associated with entering SAT included the same set of client demographic, risk factor, and enabling services variables that was used in the previous analysis of HIV testing. The analysis also examined two project-level variables: whether or not the project was gender-focused (i.e., whether it primarily targeted women clients or was Table 3 Results of HLM analysis of factors associated with whether or not HIV Outreach clients with substance abuse problems enter substance abuse treatmenta 95% Confidence interval Variable
Odds ratio
Lower
Client level Used crack in past 30 days Used marijuana in past 30 days Had unprotected sex in past 30 days
2.47 1.28 0.56
2.03 1.05 0.47
3.00 1.55 0.68
< .0001 .0123 < .0001
Project level Internal SAT Gender-focused
8.03 0.53
1.06 0.07
60.68 3.86
.0436 .5340
1.27
2.43
.0007
5.98
1586.55
< .0001
1.09
2.66
.0184
86.28
2022.35
< .0001
Interaction of client- and project-level variables 1.76 Projects with external SAT: Clients who received transportation services vs. clients who did not receive transportation services 16.76 Projects with internal SAT: Clients who received transportation services vs. clients who did not receive transportation services Interaction of client- and project-level variables 1.71 Mixed gender-focused projects: Clients who received HIV/risk education vs. clients who did not receive HIV/risk education 188.44 Women-targeted projects : Clients who received HIV/risk education vs. clients who did not receive HIV/risk education a
Observations are relative to clients/projects with characteristics.
Upper
P
P.D. Tinsman et al. / Journal of Substance Abuse 13 (2001) 201–214
211
structured to serve a mixed-gender clientele), and whether the project had ‘‘internal SAT’’ capabilities, in the sense of being able to offer SAT at treatment facilities operated by its parent organization, or whether it sought to facilitate client entry into SAT programs operated by other agencies with which it had established referral linkages. Table 3 presents the results of this second HLM analysis. Again, only variables that were found significantly associated with treatment entry are shown in the table. As in the previous analysis of HIV testing outcomes, client demographic measures and several other client-level variables were not statistically associated with SAT entry. Most conspicuously absent from the list of client variables associated with SAT entry is the classification of whether or not the client is an injection drug user. In contrast to the earlier raw data comparison that showed a much lower rate of treatment entry for IDUs than for non-IDUs, this indicates that, after adjusting for the effects of other client and project variables, IDUs were not less likely than other substance-abusing clients to enter SAT. Other variables that were not associated with treatment entry were recent client use of heroin, cocaine, or alcohol, and whether or not the client was a sexual partner of a drug user. Three client-level variables were statistically associated with entry into SAT: clients who reported recent crack use were about 2 1/2 times more likely to enter SAT than other clients (OR = 2.47, P < .0001); clients who reported recent marijuana use were 28% more likely to enter SAT than other clients (OR = 1.28, P = .01); and clients who reported having unprotected sex in the past month were about half as likely to enter SAT as other clients, all else being equal (OR = 0.56, P < .0001). Again, as in the HIV testing analysis, project-level variables had the strongest statistical associations with client entry into SAT, as main effects and/or in interaction with other variables. All else being equal, SAT-eligible clients served by projects with internal SAT facilities were eight times more likely to receive SAT than similar clients served by projects
Fig. 1. Interaction between client transportation and internal/external SAT.
212
P.D. Tinsman et al. / Journal of Substance Abuse 13 (2001) 201–214
Fig. 2. Interaction between HIV education and gender focus.
that sought to facilitate client entry through referral to external SAT programs. Among treatment-eligible clients served by projects with internal SAT facilities, those who received transportation assistance were much more likely to receive SAT than those who did not receive transportation help (OR = 16.76, P < .0001). Transportation assistance was much less of a factor among clients at projects using external rather than internal SAT facilities (Fig. 1). Another interesting project-level variable is whether or not the project has a genderspecific focus in its programming. Gender focus did not emerge as having a significant overall effect on project success in getting eligible clients into SAT. However, it did have a significant interaction with whether or not an SAT-eligible client received one or more sessions of HIV risk education. Receiving a pretreatment intervention of this kind was strongly linked to subsequent entry into formal SAT at women-focused projects (OR = 188.44, P < .0001), much more than at mixed-gender projects (OR = 1.71, P < .02). This interesting interaction is shown graphically in Fig. 2.
4. Discussion The national cross-site evaluation of the Community-Based Comprehensive HIV/STD/TB Outreach Services for High-Risk Substance Abusers Demonstration Program findings indicate that the provision of a variety of services to hard-to-reach, underserved target populations can increase the rate of medical testing for HIV and associated diseases, and can increase access to SAT. The findings from this cross-site study support the hypothesis that HIV Outreach, as an integrated approach to addressing the multiple problems clients have due to substance abuse and related problems, is an effective model for reaching clients who have
P.D. Tinsman et al. / Journal of Substance Abuse 13 (2001) 201–214
213
not been reached through traditional means. By implementing a complement of comprehensive HIV Outreach interventions, the projects were successful in addressing the various needs of their clients. This study’s findings offer some clear direction for similar projects. To promote HIV and related (STDs, TB, etc.) testing to the hard-to-reach populations, projects should employ a mobile unit to provide testing services on the street, or provide on-site HIV testing. In this study, projects that used mobile units were 86 times more likely to succeed in having their clients tested for HIV, and for those projects that provided on-site HIV testing, their clientele were 21 times more likely to be tested for HIV. Further, project-provided transportation services and HIV risk education also had strong impacts on the testing of clients. Clients who had received these types of services were much more likely to be tested for HIV. To promote treatment entry, it appears highly desirable that HIV Outreach projects should be housed within agencies that provide SAT. Projects that were housed within institutions/ agencies that had internal SAT facilities were eight times more likely to have clients enter SAT, as compared to projects that could not provide such a direct entry into treatment. Another important finding was that clients who received transportation services were 16 times more likely to enter SAT in projects that had internal SAT. Preparatory services such HIV risk education were also associated with increased rates of entry into SAT. Such ancillary services are important in preparing clients for the possibility of treatment entry.
Acknowledgments This article was produced under Contract 270-97-7030 funded by the Center for Substance Abuse Treatment, Substance Abuse and Mental Health Services Administration, U.S. Department of Health and Human Services. Contents solely represent the responsibility of the authors and do not necessarily represent the official views of the agency.
References Ball, J. C., Lange, W. R., Myers, C. P., & Freidman, S. R. (1988). Reducing the risk of AIDS through methadone maintenance treatment. Journal of Health and Social Behavior, 29, 214 – 226. Booth, R. E., Kwiatkowski, C. F., & Stephens, R. C. (1998). Effectiveness of HIV/AIDS interventions on drug use and needle risk behaviors for out-of-treatment injection drug users. Journal of Psychoactive Drugs, 8 (3), 269 – 278 (July – September). Centers for Disease Control and Prevention (CDC). (1998). U.S. HIV and AIDS cases reported through December 1998. HIV/AIDS Surveillance Report: Year-End Edition, 11 (2), 5. Estrado, A. L. (1998). Drug use and HIV risks among African-American, Mexican-American, and Puerto Rican drug injectors. Journal of Psychoactive Drugs, 30 (3), 247 – 254. Freeman, R. C., Williams, M. L., & Saunders, L. A. (1999). Drug use, AIDS knowledge, and HIV risk behavior of Cuban-, Mexican-, and Puerto Rican-born drug injectors who are recent entrants into the Unites States. Substance Use and Misuse, 34 (13), 1765 – 1793. Health Resources and Services Administration (HRSA). (1999). HIV disease in women of color. HRSA Care Action. U.S. Department of Health and Human Services, HIV/AIDS Bureau (May): Rockville, MD.
214
P.D. Tinsman et al. / Journal of Substance Abuse 13 (2001) 201–214
Hubbard, R. L., Marsden, M. E., Cavanaugh, E., Rachal, J. V., & Ginzburg, H. M. (1988). Role of drug-abuse treatment in limiting the spread of AIDS. Review of Infectious Diseases, 10, 377 – 384. Kang, S. Y., & de Leon, G. (1993). Correlates of drug injection behaviors among methadone outpatients. American Journal of Drug and Alcohol Abuse, 19, 107 – 118. Littell, R. C., Milliken, G. A., Stroup, W. W., & Wolfnger, R. D. (1996). SAS system for mixed models (p. 432). Cary, NC: SAS Institute. National Institute on Drug Abuse (NIDA). (1999). 25 Years of discovery to advance the health of the public. Rockville, MD: U.S. Department of Health and Human Services, National Institutes of Health. Needle, R. H., & Coyle, S. (1998). Community-based outreach risk-reduction strategy to prevent HIV risk behaviors in out-of-treatment injection drug users. In: NIH Consensus Development Conference on Interventions to Prevent HIV Risk Behaviors. National Institutes of Health: Bethesda, MD, USA. Rowden, D.W, Dorsey, P. E., Bullman, S., Lestina, R. P., Han, C., & Herrell, J. M. (1999). HIV Outreach for hardto-reach populations: a cross-site perspective. Journal of Evaluation and Program Planning, 22 (3), 251 – 258.
Factors affecting client response to HIV outreach efforts Penny Dorsey Tinsmana,*, Stephanie Bullmana, Xiaowu Chena, Kenneth Burgdorf a, James M. Herrellb a Caliber Associates, Suite 400, 10530 Rosehaven Street, Fairfax, VA 22030, USA Center for Substance Abuse Treatment1, Suite 740, 5600 Fishers Lane, Rockwall II, Rockville, MD 20857, USA
b
Abstract Purpose: This article describes 12 HIV Outreach Demonstration Projects funded by the Center for Substance Abuse Treatment in 1995, and the clients these projects served. The article also summarizes the findings of multivariate statistical analyses aimed at identifying important project and client characteristics that influenced project success in achieving two key outcomes: persuading at-risk clients to obtain HIV tests, and facilitating entry by substance-abusing clients into structured substance abuse treatment (SAT). Methods: Hierarchical linear modeling (HLM) analysis was used to conduct the analyses. Results: The findings support the hypothesis that HIV Outreach, as an integrated approach to addressing the multiple problems clients have due to substance abuse and related problems, can be an effective model for reaching clients who have not been reached through traditional means. Implications: By implementing a complement of comprehensive HIV Outreach interventions, many of the HIV Outreach projects were successful in addressing the various needs of their clients. However, as detailed in this article, some services and service delivery procedures, were more effective than others in achieving project objectives. D 2001 Elsevier Science Inc. All rights reserved. Keywords: HIV Outreach; Injection drug user; HIV testing; Substance abuse treatment; Hard-to-reach populations
1. Introduction The human immune deficiency virus (HIV)/acquired immune deficiency syndrome (AIDS) epidemic continues to pose a serious health threat in the US. Since the epidemic began, * Corresponding author. Tel.: +1-703-385-3200; fax: +1-703-385-3206. E-mail address: [email protected] (P.D. Tinsman). 1 Tel.: + 1-301-443-5052. 0899-3289/01/$ – see front matter D 2001 Elsevier Science Inc. All rights reserved. PII: S 0 8 9 9 - 3 2 8 9 ( 0 1 ) 0 0 0 6 6 - 9
202
P.D. Tinsman et al. / Journal of Substance Abuse 13 (2001) 201–214
injection drug use (IDU) has directly or indirectly accounted for more than one third (36%) of AIDS cases in the US (Centers for Disease Control and Prevention (CDC), 1998). Of the 48,269 new cases of AIDS reported in 1998, 14,024 (31%) were associated with IDU (CDC, 1998). IDU-associated AIDS accounts for a larger proportion of cases among women than men. Since the epidemic began, 59% of all AIDS cases among women have been attributed to IDU or sex with partners who inject drugs, compared to 31% of cases among men (CDC, 1998). IDU-associated AIDS varies across racial and ethnic populations in the US (Estrado, 1998; Freeman, Williams, & Saunders, 1999). In 1998, injection drug users (IDUs) accounted for 36% of all AIDS cases among both African American and Hispanic adults and adolescents, compared with 22% of all cases among white adults and adolescents (CDC, 1998). The majority of AIDS cases reported among females in the US has been among women of color since the onset of the epidemic in the US (Health Resource and Services Administration (HRSA), 1999). One successful method of preventing HIV transmission is substance abuse treatment (SAT). SAT has been shown to reduce risk behaviors associated with HIV transmission, such as sharing needles contaminated with blood and engaging in unsafe sexual practices — unprotected sex, trading sex for drugs or for money to buy drugs, etc. A number of studies have reported significant reduction in HIV-related drug use and sexual risk behaviors among IDUs in treatment compared to those not in treatment (Ball, Lange, Myers, & Freidman, 1988; Booth, Kwiatkowski, & Stephens, 1998; Hubbard, Marsden, Cavanaugh, Rachal, & Ginzburg, 1988; Kang & de Leon, 1993; National Institute on Drug Abuse (NIDA), 1999; Needle & Coyle, 1998). As part of its efforts to address the serious health threat presented by HIV/AIDS, the Substance Abuse and Mental Health Services Administration’s (SAMHSA) Center for Substance Abuse Treatment (CSAT) of the U.S. Department of Health and Human Services initiated a multisite, multiyear demonstration program in 1995, designed to test the effectiveness of integrating referral to SAT with street outreach models for HIV prevention to reach populations that are at highest risk for exposure to HIV, yet are historically underserved by the medical community. Specifically, the Community-Based Comprehensive HIV/STD/TB Outreach Services for High-Risk Substance Abusers Demonstration Program (HIV/AIDS Outreach Program) was designed to provide outreach services to IDUs, other chronic, hard-to-reach substance abusers, and their sex and/or needle-sharing partners. The goal of the demonstration was to reduce HIV transmission among these high-risk populations and provide greater access to SAT. Twelve projects were funded in this demonstration. Each provided four core services:
Community-based outreach services to encourage entry and facilitate access to SAT; HIV/AIDS risk-reduction interventions; Medical diagnostic testing and screening for HIV, sexually transmitted diseases (STDs), tuberculosis (TB), and pregnancy; and Linkages to primary medical care, mental health care, social services, and other ‘prophylactic means’ to effect behavioral changes most likely to reduce the risk of acquiring or transmitting HIV, STDs, TB, and related diseases.
P.D. Tinsman et al. / Journal of Substance Abuse 13 (2001) 201–214
203
Each of the 12 demonstration projects was required to conduct a comprehensive evaluation to document the effectiveness of the approaches and services being tested at each site. Additionally, CSAT funded a central, cross-site evaluation of the HIV Outreach program as a whole, based on core sets of project-level and client-level data provided by the 12 individual projects (Rowden et al., 1999). This paper, based on data assembled in the cross-site study, has two main purposes. First, it provides an overview summary of CSAT’s 12 HIV Outreach projects and of the clients they served. Second, it summarizes the findings of multivariate statistical analyses aimed at identifying important project and client characteristics that influenced project success in achieving two important outcomes: persuading at-risk clients to obtain HIV tests, and facilitating entry by substance-abusing clients into structured SAT.
2. HIV Outreach projects and their clients 2.1. Description of projects The HIV Outreach Demonstration Program was based on the hypothesis that by creating extensive linkages among service agencies, and by providing comprehensive services to clients with a multitude of problems, projects would be able to reduce risk behaviors associated with HIV transmission, even among high-risk population groups. By providing comprehensive services, such as pretreatment services, transportation to treatment, and referral to social services, HIV Outreach projects should also increase facilitated access to SAT for target populations. All projects that participated in the cross-site evaluation were located in urban areas in the US or US territories. One of the projects also served a target area that included a rural location. One project was housed in a correctional facility in New York City. Two of the projects were affiliated with local universities; these projects emphasized research and data collection. Two other projects were affiliated with government programs or with government agencies, specifically, a department of corrections and a public health commission. Eight projects were entities of nonprofit organizations whose main goal was SAT or who provided social services to indigent clients. 2.2. Target populations and project services CSAT identified the target population for the demonstration program as IDUs and their sex and needle-sharing partners, an underserved group at extremely high risk for HIV exposure/ infection. Within those parameters, projects could target populations within their communities that they considered to be most in need. Populations targeted for outreach included crack/ cocaine users, homeless individuals, prostitutes, Hispanics, African American women, and IDUs. Overall, the target populations lived in areas marked by widespread drug use, violence, poverty, and health problems.
204
P.D. Tinsman et al. / Journal of Substance Abuse 13 (2001) 201–214
Three of the projects focused solely on women. One project targeted homeless individuals. Seven projects primarily served African American clients, three focused on Hispanic/Latino or Mexican American clients, and two projects did not target a specific group outside of IDUs and their sexual partners. All of the projects followed the traditional or modified Indigenous Leader Outreach Model (ILOM) developed as part of the NIDA-funded National AIDS Demonstration Research Project (NADAR). The core elements of this approach are to provide information about HIV to IDUs and their sexual partners, conduct personalized risk assessments, and provide support for behavior change during street-based interactions, from mobile vans or at drop-in centers located within the community. Programs targeted individual users and their social networks using trained indigenous outreach staff. In the ILOM, the staff are seen as peer leaders and role models among targeted social networks due to their own successful recovery from drug addiction. Using the ILOM, outreach workers are able to provide rapid entry into communities, establish rapport with community leaders, and develop a higher level of trust and acceptance of the outreach project, than would otherwise be possible. Outreach interventions included the following:
Street outreach focused on harm reduction including distributing pamphlets, safe sex kits, and bleach kits; locating IDUs within copping areas and shooting galleries, and providing HIV and SAT information to local area businesses. Eleven of 12 projects engaged in street outreach. Mobile units or vans provided a movable center for outreach services such as counseling and testing as well as transport for outreach workers to target neighborhoods. Advantages to mobile units included high community recognition of the outreach project, safer place for worker/contact interaction, and relative privacy for outreach contacts and service provision. They also served as a focal point to initiate conversations. Outreach workers often distributed materials in front of the van, or offered food or refreshments for individuals who came inside the van to be tested or watch a safe sex demonstration. Four of the 12 projects had vans, two of which were fully equipped to provide medical testing services. Drop-in centers were located within the center of the target community. They offered a safe place for indigent clients as well as a base for outreach services. Services included testing for HIV and STDs, counseling, support groups, referrals for SAT, and other services, and a repository for clothing and food to give away to needy clients. Half of the 12 projects had a drop-in center.
2.3. Description of client recruitment The HIV Outreach projects used multiple strategies to engage their clients. Eleven of the 12 projects focused on neighborhoods known to frequented by the target population. These locations included ‘‘street locations’’ (e.g., copping areas/outdoor drug markets, shooting galleries, commercial sex worker ‘‘strolls,’’ crack houses, etc.); parks where drug trafficking occurred; social locations (e.g., gay and/or straight bars, churches, and community centers);
P.D. Tinsman et al. / Journal of Substance Abuse 13 (2001) 201–214
205
social service/institution/community agencies, including hospitals, clinics, emergency rooms, jails, prison release programs, and food distribution centers. Incentives were used to attract a clientele. Projects reported using hygiene kits (shampoo, razor, toothbrush, shower vouchers, etc.) and clothing. Food was also offered as an incentive. Sandwiches, snacks, and sodas were often offered from the project’s mobile unit; donuts and coffee would be available at the drop-in center; and in other cases, food vouchers would be available for community locations. Further, projects used multiple outreach strategies to attract the ‘‘hard-to-reach’’ population. The projects would combine regular street outreach with a mobile unit in the community that could offer services ‘‘on the street.’’ Drop-in centers, which were also successful in recruiting clients, afforded a safe location receive services or just to get off the street. Projects referred clients to medical services such as HIV, STD, TB, and pregnancy testing or were able to provide the service(s) directly to clients. Fifty-four percent of all clients received a referral for HIV testing. Ten percent of clients received a referral for TB testing. Eleven percent of clients were referred for primary health care. Twelve percent of clients received a referral to social services. Social services included housing assistance, assistance in obtaining documentation, assistance in obtaining Medicaid, etc. Fifty-three percent of all clients were tested for HIV, and 15% of all clients received a TB test. Projects provided services to facilitate their clients’ access and entry into SAT. Services included preparing clients for treatment through assessment of their readiness to enter treatment, support groups and ancillary services for clients who were on waiting lists for treatment. Thirty-one percent of clients received an HIV Education Workshop. Forty-five percent of clients attended a Risk Reduction/Education Workshop. Fifty-seven percent of clients received pretest for HIV counseling. Sixty-four percent of all clients received riskreduction counseling. Forty-one percent of the HIV Outreach clients who had reported using a substance within the past 30 days were referred to SAT. Of those who were referred, 68% entered SAT. 2.4. Description of clients The cross-site study received usable demographic data for 9296 HIV Outreach clients. Fifty-five percent of these clients were males and 45% were females. Given the approximate equality in the gender proportion, it must be noted that several projects targeted specific populations such as at-risk women (three projects), one project served only men, and the remaining projects served relatively evenly distributed proportions of male and females. Fifty-five percent of the HIV Outreach clients indicated black/African American as their race/ethnic group. Twenty percent of the clients identified themselves as Hispanic. Eighteen percent identified themselves as white. A small percentage (5%) of clients reported their race/ ethnicity as Asian/Pacific Islander. The remaining 2% of clients reported themselves as Native American (0.4%), Alaskan Native (0.1%), and other (1%). The mean age of the HIV Outreach clients was 35 years old. Forty-one percent of the clients were between the ages of 31 and 40. Twenty-five percent of the HIV Outreach clients were in the following age groups: 21 through 30 years old, and 41 through 55 years old. Only
206
P.D. Tinsman et al. / Journal of Substance Abuse 13 (2001) 201–214
6% of the clients were less than 21 years old, and 3% of the HIV Outreach clients were over the age of 55. Of those clients reporting marital status, 46 reported that they had never married and were not living with a significant other. Nearly one-quarter of the HIV Outreach clients reported that they were separated, divorced, or widowed. Fourteen percent of the clients reported that they were currently married, and 7% reported that they were living with a significant other. Additionally, one third reported that they had children. Seven percent of the female clients were pregnant at entry into the HIV Outreach program. Forty-seven percent of the HIV Outreach clients had achieved a high school diploma. Thirteen percent reported post-high school education and over one third (38%) reported that their highest level of education completed was between the seventh and eleventh grades. Half of the HIV Outreach clients reported living during the 30 days prior to entry into the program in a house, condo, or apartment. Of the other clients, 4% indicated that they had stayed in public housing 30 days prior to intake in the HIV Outreach project. Thirtytwo percent had stayed in a homeless shelter or on the street, 3% reported being in jail or prison 30 days before entering the HIV Outreach project, and 10% reported that they had stayed in another type of shelter (i.e., recovery homes, halfway houses, etc.), or other group living situations. Not all clients entering the HIV Outreach Demonstration Program were substance abusers. Many clients were the sexual partners of drug users and/or IDUs, or were engaged in sexual activities such as prostitution that put them at risk for exposure to the HIV virus. Fifty-three percent of the HIV Outreach clients reported that they had engaged in sexually risky behaviors in their lifetimes. These risky sexual behaviors include having unsafe/unprotected sex in the past 30 days, ever having unsafe sex with an injection drug user, ever using sex to obtain drugs or money, and ever having sex with someone who is or might be infected with the HIV virus. Because of the danger posed by injection behaviors, HIV Outreach projects specifically targeted IDUs. Thirty-five percent of the clients reported that they had injected drugs in their lifetime, and 31% reported that they had injected drugs in the 30 days prior to entry into the HIV Outreach project. Over three-fifths (61%) of the HIV clients reported that they had used substances in the 30 days prior to entering the HIV Outreach project. Forty percent of the clients reported using a substance other than alcohol. Table 1 presents the reports of substance abuse in the 30 days prior to entry into the HIV Outreach project.
Table 1 Substance use patterns Substance
Percentage of clients reporting use in the last 30 days (%)
Cocaine Crack Marijuana/hashish Heroin Alcohol
18 28 17 14 46
P.D. Tinsman et al. / Journal of Substance Abuse 13 (2001) 201–214
207
3. Methods and results 3.1. Analytic approach This paper examines factors associated with project success in achieving one or another of two important outreach objectives: (1) persuading high-risk clients to obtain HIV tests, and (2) facilitating entry into SAT of hard-to-reach, high-risk clients with substance abuse problems. The cross-site database contains two broad categories of independent variables that are potentially useful in explaining such client outcomes: variables describing projectlevel differences in outreach approaches, components, procedures, etc., and variables describing client-level differences in demographic characteristics, HIV risk factors, and exposure to various outreach services. The cross-site database is hierarchical in structure: clients are nested within projects. The multivariate analytic approach that is most appropriate for use with this kind of data structure is hierarchical linear modeling (HLM) analysis. This approach permits simultaneous identification of: (1) project-level effects, controlling for between-project differences in client composition, (2) client-level effects that are consistent across projects, and (3) interaction effects among variables, both within and between levels. For these analyses, we used the SAS macro GLIMMIX. We began by entering selected client characteristics into the HLM model. We then entered project-level variables into the model one at a time, retaining those that made statistically significant improvements in goodness of fit, either as a main effect or in interaction with other variables (Littell, Milliken, Stroup, & Wolfnger, 1996). 3.2. HIV testing Arguably, it is desirable that all of the high-risk clients served by HIV Outreach projects be screened for the presence of the HIV virus. Unfortunately, the HIV status of the clients is not available. Of the 12 CSAT-funded HIV Outreach projects, 8 provided the cross-site study with information about whether or not their clients actually received any HIV tests. These eight projects collectively served a total of 5851 high-risk clients, 63% of who were reported to have received one or more HIV tests during the course of their contact with the projects. In seeking to account for the differences between clients who received HIV testing and those who did not, our HLM analysis examined the influences of the following independent variables. 3.2.1. Client-level variables Gender; age; race; use/nonuse in the past 30 days of cocaine, crack, marijuana, heroin, alcohol; whether or not any IDU in the past 30 days; whether or not client had unprotected sex in the past 30 days; whether or not client is a sexual partner of a drug user; whether or not client received HIV risk education services at the Outreach project; and whether or not client received any transportation services through the Outreach project (bus tokens, rides to drop-in centers in project-operated vehicles, rides to SAT or other service locations, etc.).
208
P.D. Tinsman et al. / Journal of Substance Abuse 13 (2001) 201–214
3.2.2. Project-level variables Whether or not project was gender-focused (two projects in the analysis set primarily targeted women clients, the rest were not gender-focused); whether or not project provided on-site HIV testing (six projects provided testing at Outreach project facilities, the rest provided testing through referral to other facilities); and whether or not the project had a mobile unit in which HIV tests could be performed (two projects had such units, the rest did not). Projects with mobile units for HIV testing represent a subset of those with capabilities for on-site HIV testing. The other on-site testing projects did HIV testing primarily at drop-in facilities. Table 2 summarizes the results of the HLM analysis of the relationships between these variables and whether or not an at-risk client actually received HIV testing. Only variables that produced statistically significant (at P < .05) associations with HIV testing are shown. Relationships are expressed as odds ratios (ORs), which indicate how much more or less likely it was that a client with the indicated characteristic was tested for HIV, as compared to a client who did not have the indicated characteristic. ORs above 1.0 indicate that clients with a given characteristic were more likely than others to obtain an HIV test; conversely, an OR less than 1.0 indicates that a client with a given characteristic was less likely than others to receive an HIV test. For example, one important finding in Table 1 is that, consistently across projects, clients who received an initial HIV risk education program were more than four Table 2 Results of HLM analysis of factors associated with whether or not HIV Outreach clients received HIV testinga 95% Confidence interval Variable Client level Used crack in last 30 days Used marijuana in last 30 days Had unprotected sex in last 30 days Sexual partner of drug user Received HIV risk education Project level On-site HIV testing Mobile unit
Odds ratio
Upper
P
0.53 0.58 0.43 0.27 4.64
0.43 0.45 0.35 0.22 3.73
0.65 0.76 0.54 0.35 5.77
< .0001 < .0001 < .0001 < .0001 < .0001
21.22 86.06
1.14 6.40
396.68 1156.40
.0409 .0008
3.57
38.15
< .0001
0.37
11.31
.0834
Interaction of client- and project-level variables 11.67 Projects without on-site HIV testing: Clients who received transportation services vs. clients who did not receive transportation services 2.10 Projects with on-site HIV testing: Clients who received transportation services vs. clients who did not receive transportation services a
Lower
Observations are relative to clients/projects with characteristics.
P.D. Tinsman et al. / Journal of Substance Abuse 13 (2001) 201–214
209
times as likely to obtain an HIV test as clients who did not receive such a program (OR = 4.64, P < .0001). Several client-level variables were not significantly associated with HIV testing. This includes all three client demographic characteristics (gender, age, race) and several major HIV risk factors, including IDU and recent use of heroin or cocaine. Several other HIV risk factors were negatively associated with HIV testing, as indicated by ORs below 1.0. For example, recent crack or marijuana users were only about half as likely as nonusers of these substances to obtain HIV testing (OR = 0.53 and 0.58, respectively, with P < .0001 for both). Clients who reported having had unprotected sex in the past 30 days were also about half as likely to receive HIV testing as clients without this risk factor (OR = 0.43, P < .0001). Clients who were sexual partners of drug users were only about one-fourth as likely as other clients to obtain HIV testing (OR = 0.27, P < .0001). These risk factor associations are somewhat difficult to interpret, however, considering that all clients in the outreach database were identified as being at high risk of HIV in one way or another. Thus, clients who did not have any particular risk factor typically did have one or more other risk factors. The strongest and most important associations in Table 2 concern project characteristics, particularly the mechanisms projects used to deliver HIV testing services. The projects that used mobile testing units to reach their clients were vastly more successful than other projects in their HIV testing efforts: all else being equal, outreach clients at projects with mobile units were 86 times more likely to obtain an HIV test than those at other outreach projects. More generally, clients who were served by projects that had any capability for on-site HIV testing were substantially (21 times) more likely to receive HIV testing than those served by projects that used referrals to other agencies as their testing method. However, it is possible that not all clients tested by outside testing facilities are counted; projects relied on both client self-report and linkage partners to provide the information. Table 2 also shows an interesting and understandable interaction effect between whether or not a project used on-site HIV testing methods and whether or not the client received projectprovided transportation services. For projects that did not have on-site testing facilities, it appears that it was critically important to provide clients with transportation support, presumably to help them get to the project’s testing location: among clients at projects without on-site testing, those who received transportation support were 11.67 times more likely to receive HIV testing than those without transportation help ( P < .0001). For clients at projects that provided on-site HIV testing, presence or absence of transportation support was not significantly associated with HIV testing outcomes. 3.3. Entering SAT Perhaps the single most important objective of the CSAT-funded HIV Outreach demonstration projects was to find effective methods of facilitating entry into structured SAT programs for IDUs and other groups of substance abusers who are at high risk for HIV/AIDS. It is important to note that project staff played an integral role in whether a client entered SAT; because of the possible availability of treatment slots within the organization and additional linkage agreements, staff rarely referred a client to SAT without first securing an available
210
P.D. Tinsman et al. / Journal of Substance Abuse 13 (2001) 201–214
bed. Waiting lists would not be a viable option for this population. Of the 12 demonstration projects, 9 were able to provide outcome information to the cross-site study about whether or not individual clients received any form of structured SAT. These nine projects served 6155 clients, 4083 of whom reported past-month IDU or other past-month use of alcohol or other drugs. Among these SAT-eligible outreach clients, 1686 (41%) actually entered a structured SAT program. It was disappointing to note, in our initial analysis of the raw data, that the rate of SAT entry was considerably lower among IDU clients (30%) than among other substanceusing clients (51%). Our HLM analysis of factors associated with entering SAT included the same set of client demographic, risk factor, and enabling services variables that was used in the previous analysis of HIV testing. The analysis also examined two project-level variables: whether or not the project was gender-focused (i.e., whether it primarily targeted women clients or was Table 3 Results of HLM analysis of factors associated with whether or not HIV Outreach clients with substance abuse problems enter substance abuse treatmenta 95% Confidence interval Variable
Odds ratio
Lower
Client level Used crack in past 30 days Used marijuana in past 30 days Had unprotected sex in past 30 days
2.47 1.28 0.56
2.03 1.05 0.47
3.00 1.55 0.68
< .0001 .0123 < .0001
Project level Internal SAT Gender-focused
8.03 0.53
1.06 0.07
60.68 3.86
.0436 .5340
1.27
2.43
.0007
5.98
1586.55
< .0001
1.09
2.66
.0184
86.28
2022.35
< .0001
Interaction of client- and project-level variables 1.76 Projects with external SAT: Clients who received transportation services vs. clients who did not receive transportation services 16.76 Projects with internal SAT: Clients who received transportation services vs. clients who did not receive transportation services Interaction of client- and project-level variables 1.71 Mixed gender-focused projects: Clients who received HIV/risk education vs. clients who did not receive HIV/risk education 188.44 Women-targeted projects : Clients who received HIV/risk education vs. clients who did not receive HIV/risk education a
Observations are relative to clients/projects with characteristics.
Upper
P
P.D. Tinsman et al. / Journal of Substance Abuse 13 (2001) 201–214
211
structured to serve a mixed-gender clientele), and whether the project had ‘‘internal SAT’’ capabilities, in the sense of being able to offer SAT at treatment facilities operated by its parent organization, or whether it sought to facilitate client entry into SAT programs operated by other agencies with which it had established referral linkages. Table 3 presents the results of this second HLM analysis. Again, only variables that were found significantly associated with treatment entry are shown in the table. As in the previous analysis of HIV testing outcomes, client demographic measures and several other client-level variables were not statistically associated with SAT entry. Most conspicuously absent from the list of client variables associated with SAT entry is the classification of whether or not the client is an injection drug user. In contrast to the earlier raw data comparison that showed a much lower rate of treatment entry for IDUs than for non-IDUs, this indicates that, after adjusting for the effects of other client and project variables, IDUs were not less likely than other substance-abusing clients to enter SAT. Other variables that were not associated with treatment entry were recent client use of heroin, cocaine, or alcohol, and whether or not the client was a sexual partner of a drug user. Three client-level variables were statistically associated with entry into SAT: clients who reported recent crack use were about 2 1/2 times more likely to enter SAT than other clients (OR = 2.47, P < .0001); clients who reported recent marijuana use were 28% more likely to enter SAT than other clients (OR = 1.28, P = .01); and clients who reported having unprotected sex in the past month were about half as likely to enter SAT as other clients, all else being equal (OR = 0.56, P < .0001). Again, as in the HIV testing analysis, project-level variables had the strongest statistical associations with client entry into SAT, as main effects and/or in interaction with other variables. All else being equal, SAT-eligible clients served by projects with internal SAT facilities were eight times more likely to receive SAT than similar clients served by projects
Fig. 1. Interaction between client transportation and internal/external SAT.
212
P.D. Tinsman et al. / Journal of Substance Abuse 13 (2001) 201–214
Fig. 2. Interaction between HIV education and gender focus.
that sought to facilitate client entry through referral to external SAT programs. Among treatment-eligible clients served by projects with internal SAT facilities, those who received transportation assistance were much more likely to receive SAT than those who did not receive transportation help (OR = 16.76, P < .0001). Transportation assistance was much less of a factor among clients at projects using external rather than internal SAT facilities (Fig. 1). Another interesting project-level variable is whether or not the project has a genderspecific focus in its programming. Gender focus did not emerge as having a significant overall effect on project success in getting eligible clients into SAT. However, it did have a significant interaction with whether or not an SAT-eligible client received one or more sessions of HIV risk education. Receiving a pretreatment intervention of this kind was strongly linked to subsequent entry into formal SAT at women-focused projects (OR = 188.44, P < .0001), much more than at mixed-gender projects (OR = 1.71, P < .02). This interesting interaction is shown graphically in Fig. 2.
4. Discussion The national cross-site evaluation of the Community-Based Comprehensive HIV/STD/TB Outreach Services for High-Risk Substance Abusers Demonstration Program findings indicate that the provision of a variety of services to hard-to-reach, underserved target populations can increase the rate of medical testing for HIV and associated diseases, and can increase access to SAT. The findings from this cross-site study support the hypothesis that HIV Outreach, as an integrated approach to addressing the multiple problems clients have due to substance abuse and related problems, is an effective model for reaching clients who have
P.D. Tinsman et al. / Journal of Substance Abuse 13 (2001) 201–214
213
not been reached through traditional means. By implementing a complement of comprehensive HIV Outreach interventions, the projects were successful in addressing the various needs of their clients. This study’s findings offer some clear direction for similar projects. To promote HIV and related (STDs, TB, etc.) testing to the hard-to-reach populations, projects should employ a mobile unit to provide testing services on the street, or provide on-site HIV testing. In this study, projects that used mobile units were 86 times more likely to succeed in having their clients tested for HIV, and for those projects that provided on-site HIV testing, their clientele were 21 times more likely to be tested for HIV. Further, project-provided transportation services and HIV risk education also had strong impacts on the testing of clients. Clients who had received these types of services were much more likely to be tested for HIV. To promote treatment entry, it appears highly desirable that HIV Outreach projects should be housed within agencies that provide SAT. Projects that were housed within institutions/ agencies that had internal SAT facilities were eight times more likely to have clients enter SAT, as compared to projects that could not provide such a direct entry into treatment. Another important finding was that clients who received transportation services were 16 times more likely to enter SAT in projects that had internal SAT. Preparatory services such HIV risk education were also associated with increased rates of entry into SAT. Such ancillary services are important in preparing clients for the possibility of treatment entry.
Acknowledgments This article was produced under Contract 270-97-7030 funded by the Center for Substance Abuse Treatment, Substance Abuse and Mental Health Services Administration, U.S. Department of Health and Human Services. Contents solely represent the responsibility of the authors and do not necessarily represent the official views of the agency.
References Ball, J. C., Lange, W. R., Myers, C. P., & Freidman, S. R. (1988). Reducing the risk of AIDS through methadone maintenance treatment. Journal of Health and Social Behavior, 29, 214 – 226. Booth, R. E., Kwiatkowski, C. F., & Stephens, R. C. (1998). Effectiveness of HIV/AIDS interventions on drug use and needle risk behaviors for out-of-treatment injection drug users. Journal of Psychoactive Drugs, 8 (3), 269 – 278 (July – September). Centers for Disease Control and Prevention (CDC). (1998). U.S. HIV and AIDS cases reported through December 1998. HIV/AIDS Surveillance Report: Year-End Edition, 11 (2), 5. Estrado, A. L. (1998). Drug use and HIV risks among African-American, Mexican-American, and Puerto Rican drug injectors. Journal of Psychoactive Drugs, 30 (3), 247 – 254. Freeman, R. C., Williams, M. L., & Saunders, L. A. (1999). Drug use, AIDS knowledge, and HIV risk behavior of Cuban-, Mexican-, and Puerto Rican-born drug injectors who are recent entrants into the Unites States. Substance Use and Misuse, 34 (13), 1765 – 1793. Health Resources and Services Administration (HRSA). (1999). HIV disease in women of color. HRSA Care Action. U.S. Department of Health and Human Services, HIV/AIDS Bureau (May): Rockville, MD.
214
P.D. Tinsman et al. / Journal of Substance Abuse 13 (2001) 201–214
Hubbard, R. L., Marsden, M. E., Cavanaugh, E., Rachal, J. V., & Ginzburg, H. M. (1988). Role of drug-abuse treatment in limiting the spread of AIDS. Review of Infectious Diseases, 10, 377 – 384. Kang, S. Y., & de Leon, G. (1993). Correlates of drug injection behaviors among methadone outpatients. American Journal of Drug and Alcohol Abuse, 19, 107 – 118. Littell, R. C., Milliken, G. A., Stroup, W. W., & Wolfnger, R. D. (1996). SAS system for mixed models (p. 432). Cary, NC: SAS Institute. National Institute on Drug Abuse (NIDA). (1999). 25 Years of discovery to advance the health of the public. Rockville, MD: U.S. Department of Health and Human Services, National Institutes of Health. Needle, R. H., & Coyle, S. (1998). Community-based outreach risk-reduction strategy to prevent HIV risk behaviors in out-of-treatment injection drug users. In: NIH Consensus Development Conference on Interventions to Prevent HIV Risk Behaviors. National Institutes of Health: Bethesda, MD, USA. Rowden, D.W, Dorsey, P. E., Bullman, S., Lestina, R. P., Han, C., & Herrell, J. M. (1999). HIV Outreach for hardto-reach populations: a cross-site perspective. Journal of Evaluation and Program Planning, 22 (3), 251 – 258.