Psychological Dysfunction and HIV/AIDS Risk Behavior

Summary: The relationship between psychological problems and human immunodeficiency virus HIV/AIDS risk-taking behaviors was examined among 834 daily opioid users entering methadone treatment programs. A composite measure of psychological dysfunction was created using depression, anxiety, and hostility scales. This measure was found to be significantly related to needle risk in terms of injecting with used equipment, sharing of drug paraphernalia, and sharing with strangers. Psychological dysfunction was also related to sexual risk taking in terms of number of partners, unprotected sex with other injection drug users, and trading sex. Use of cocaine was significantly related to all measures of injection and sex-related risk taking; use of speedball (heroin and cocaine) was significantly related to use of dirty equipment and sharing of paraphernalia. The implications of study findings for AIDS prevention programming are discussed.

Injection drug users (IDUs) comprise ≈30% of AIDS cases in the United States ⁽¹⁾. For many, drug abuse treatment has become the prevention strategy of choice because it is seen as increasing insulation from the disease for those retained in treatment ^(2,3) and as providing opportunities for human immunodeficiency virus (HIV) prevention counseling for all who enter treatment. With 3-month dropout from drug abuse treatment ranging from 32% for methadone maintenance to 63% for outpatient drug free ⁽⁴⁾, the importance of HIV prevention counseling for those entering treatment is apparent ⁽⁵⁾. Moreover, because early treatment dropout has been associated with psychological dysfunction ⁽⁶⁾, clients with psychological problems are less likely to receive the long-term protection from HIV infection afforded by retention in drug abuse treatment. It therefore becomes important to understand that group's level of risk taking and to clarify the nature of interventions that may be appropriate to their needs.

Previous studies of out-of-treatment IDUs in AIDS outreach programs found a relationship between sexual risk taking and psychological dysfunction ⁽⁷⁾, as well as between needle risk taking and psychological dysfunction ⁽⁸⁾. Studies of opioid users entering methadone treatment have found a relationship between needle sharing and measures of psychological distress ⁽⁹⁾, but only limited association between needle sharing and specific psychological disorders ⁽¹⁰⁾. Among male treatment admissions who were primarily cocaine users, drug abuse behaviors were associated with measures of depression and anxiety ⁽¹¹⁾. Cocaine use is of particular concern in this regard because the more frequent injection of cocaine has been seen as increasing opportunity for high-risk behaviors ^(12-15). Studies show that speedball injection and gender are also associated with elevated risk for HIV infection ^(14,16). However, the relative contributions of psychological problems, cocaine use, speedball use, and gender require greater assessment among treatment admissions. This is particularly important because speedball users and women also have an elevated incidence of psychological problems compared with nonspeedball users and men ^(8,17). The primary focus of the current investigation is to clarify the association between psychological symptoms and HIV risk behavior by considering additional risk factors such as speedball injection and cocaine use.

The relation between psychological dysfunction and HIV/AIDS risk-taking behaviors was studied among a sample of daily opiate users admitted to drug abuse treatment. It was hypothesized that psychological dysfunction-assessed by a composite measure of depression, anxiety, and hostility-would be associated with elevated levels of needle and sexual risk taking independent of other factors. That is, symptoms of psychological disorder were expected to be associated with high levels of both shared and of uncleaned needles, and to high levels of unprotected sex with multiple and risky partners. It was also hypothesized that cocaine use among opiate clients would be associated with high levels of needle and sexual risk taking independent of psychological functioning.

METHODS

Procedures

Between May 1990 and April 1993, a total of 910 opioid addicts were admitted to three methadone treatment programs participating in the Drug Abuse Treatment for AIDS-Risk Reduction (DATAR) project ⁽¹⁸⁾. A sample of 834 clients who had been daily opioid users and completed all DATAR intake forms were selected for this study (Table 1). They were drawn from programs in Corpus Christi (n = 392), Dallas (n = 221), and Houston (n = 221); each was a part of a larger multimodality clinic. Overall, 69% of the sample was male; the mean age was 37 years; 21% were African American, 39% were Hispanic (primarily Mexican American), and 36% were white; 39% were high-school graduates, and 55% reported at least some employment. There were wide variations by program site for all client characteristics other than employment. Similarly, although substantial majorities reported use of cocaine (55%) and speedball (60%), rates of use varied markedly between sites; however, crack use was <20% at every site.

Table 1

Measures

A comprehensive set of intake, during-treatment, and follow-up data collection instruments are used in the DATAR project to record client, counselor, and treatment process and outcome information ⁽¹⁹⁾. Selected data from that system are described below.

Predictor Measures

As part of the admission process, each client completed the Self-Rating Form (SRF), an 88-item self-administered instrument assessing 11 areas of psychosocial functioning. Three scales were used in this study. The Depression Scale consists of six items (α = 0.77) and correlates highly with the Symptoms Check List (SCL-90) Depression Scale (r = 0.81) ⁽²⁰⁾, and the Beck Depression Scale (r = 0.75) ⁽⁷⁾. Seven items comprise the Anxiety Scale (α = 0.82), and it correlates highly with the SCL-90 Anxiety Scale (r = 0.74) ⁽²⁰⁾. Finally, the Hostility Scale consists of eight items (α = 0.83), and is highly correlated with the SCL-90 Hostility Scale (r = 0.61) ⁽²⁰⁾.

Item responses on the three SRF scales ranged from 0 to 4, corresponding to never, rarely, sometimes, often, and almost always. Principal components factor analysis results showed that a one-factor solution was a good fit for the three scales. Therefore, the Psychological Dysfunction measure consisted of the average of the three scales (α = 0.78); high scores reflect a high level of psychological symptoms. After the completion of the SRF, a structured intake interview lasting ≈1 h was conducted by a treatment counselor. The assessment provided detailed information regarding client sociodemographic background, family functioning, peer relations, criminal history, health and psychological status, and drug abuse history.

Criterion Measures

Items from the intake interview were selected to measure HIV/AIDS risk behavior in the 6 months before intake. Four items measured injection-related risk in terms of (a) use of dirty works (needles, cooker, or cotton used by others and not cleaned with bleach); (b) the number of persons with whom dirty works were shared; (c) times shared dirty works with strangers (did not know well), and with (d) sex traders-people who trade sex for money, drugs, or gifts. Four items measured sex-related risks in terms of (a) number of sex partners, and the frequency of unprotected sex with (b) an IDU, (c) while high, and (d) while trading sex.

RESULTS

Table 2 presents needle and sex risk behaviors for all clients sampled. About 34% used dirty works daily in the 6 months preceding treatment entry, and 38% shared with more than one person. Nearly 19 and 16% shared dirty works at least once with strangers or sex traders, respectively. Most of the sample reported having only one or no sex partners (65%). However, rates of involvement in specific risky sexual activities ranged from 10 to 75%. The wide variation in individual risk behavior is indicated by the large standard deviations associated with each risk measure.

Table 2

Due to the highly skewed nature of these data, multiple logistic regression analyses were employed. Thus, in most cases any self-reported risk was recoded as 1 and the absence of risk was left at 0. The predictor variables of gender, site, race-ethnicity, and completion of high school were dummy coded with 1s and 0s. Scores for Psychological Dysfunction remained intact (ranged from 0 to 4), and cocaine and speedball use scores ranged from 0 to 2 corresponsing to never, less than weekly, and more than weekly.

Table 3 shows results of the set of multiple logistic regression analyses for each HIV risk measure tested using Psychological Dysfunction, site, age, gender, race-ethnicity, education, cocaine use, and speedball use as predictors (incidence of crack use was low and therefore not included). These analyses show that the Psychological Dysfunction measure was significantly related to three of the four injection-related measures (dirty works, people with whom shared dirty works, and times shared dirty works with strangers; b = 0.31-0.69, p < 0.05), and three of the four sex-related measures (number of sex partners, unprotected sex with IDU and while trading sex; b = 0.28-0.50, p < 0.05).

Table 3

The use of cocaine was consistently related to HIV risk; all the injection and sex-related measures tested were significant (dirty works, people with whom shared dirty works, times shared dirty works with strangers and sex traders, number of sex partners, unprotected sex with IDU, while high, and while trading sex; b = 0.30-0.59, p < 0.01). Speedball injection was significantly associated with higher use of dirty works and more sharing partners (b = 0.31 and 0.31, p < 0.01). Gender differences were evident in that males were more likely to share injection equipment with strangers (b = 0.47, p < 0.05) and females were more likely to have sex with an IDU and engage in sex trading (b = -0.71 and -1.35, p < 0.001). Results also showed that clients from Corpus Christi were consistently at higher risk for HIV than were clients from other sites with all four injection-related measures statistically significant (b = 1.10-1.91, p < 0.001). It was hypothesized that IDUs in Corpus Christi took more risks because of less exposure to the AIDS epidemic. This was supported by findings that 22% of the clients in Corpus Christi reported knowing someone with AIDS compared with 30% in Dallas and 53% in Houston.

It was also of interest to test whether gender, cocaine, or speedball use moderated the relationship between Psychological Dysfunction and risk behavior, but χ² analyses did not support this contention. Finally, analyses were performed to see whether any particular component of the Psychological Dysfunction measure was differentially associated with risk behaviors. Results showed that correlations between each risk measure and depression, anxiety, and hostility were similar.

DISCUSSION

The results of this study extend earlier findings of a relationship between pscyological symptoms and HIV/AIDS risk behavior ^(7-11). Injecting with paraphernalia previously used by others, number of people with whom share paraphernalia, and sharing with strangers were found to be related to higher levels of psychological symptoms as reflected by a composite measure of depression, anxiety, and hostility. With regard to sex risk, more sex partners, unprotected sex with IDUs, and sex trading were found to be related to higher levels of psychological symptoms. These findings clarify the independent effect of psychological symptoms relative to other client characteristics such as gender, cocaine and speedball use, which have also been found to be associated with risk behaviors.

The data support earlier suggestions that cocaine injectors may be more at risk for HIV infection than individuals injecting heroin alone ^{(13,15,16,21)} and conflicts with Booth's ⁽²²⁾ report of safer needle use by cocaine users. Compared to heroin injectors, those who also injected cocaine were more likely to use dirty paraphernalia and to share paraphernalia with a larger number of injectors, strangers, and sex traders. They also had more sex partners, were more likely to have unprotected sex with other IDUs, and were more likely to trade sex for drugs. The greater sharing of injection equipment likely combines with more frequent injecting by cocaine users ^(13,23), heightening the risk of HIV infection to that population. Results also showed that Corpus Christi clients were consistently engaging in needle-related risks and were least likely to know someone with AIDS. Exposure to the AIDS epidemic is much lower in Corpus Christi because only 1% of AIDS cases in Texas come from that region whereas 23 and 36% of cases come from Dallas and Houston, respectively ⁽²⁴⁾.

In general, the findings highlight both the potential and special responsibilities of drug abuse treatment programs in relation to HIV prevention. Several studies point to high rates of both psychological problems and cocaine use among treatment admissions. Significant levels of psychological disturbance among IDUs were first noted in the mid 1970s ⁽²⁵⁾. During the 1980s these reports were supplemented by findings of significant levels of drug abuse among mental health clients ^(26,27). Admissions for cocaine use doubled between 1985 and 1987 ⁽²⁸⁾, and cocaine is now the most frequently reported drug among treatment admissions (S. G. Craddock et al., unpublished observations).

The association between risk behaviors and both psychological problems and cocaine use, and the potential for these to be associated with negative treatment outcomes ^(29-31), has implications for both drug abuse treatment and HIV prevention counseling. With regard to treatment, the data suggest the importance of developing early retention strategies designed to extend the period of treatment protection for these at-risk populations. With regard to HIV prevention counseling, the data suggest the urgency of providing education/prevention interventions early in treatment while also educating the client regarding community AIDS prevention resources available.

Acknowledgment: This work was supported by the National Institute on Drug Abuse (NIDA) (grant no. DA06162). The interpretations and conclusions, however, do not necessarily represent the position of NIDA or the Department of Health and Human Services.