British Journal of Addiction (1992) 87, 493-498
USA TRANSITIONS REPORT
AIDS and the transition to illicit drug injection—results of a randomized trial prevention program DON C. DES JARLAIS', CATHY CASRIEL^ SAMUEL R. FRIEDMAN^ & ANDREW ROSENBLUM2 ' Beth Israel Medical Center, Chemical Dependency Institute, 1st Avenue & E. 16th Street, New York, New York 10003, USA & ^ Narcotic and Drug Research, Inc., 11 Beach Street, New York, New York 10013, USA
Abstract Illicit drug injection is a major component of the AIDS epidemic in the United States, Europe and some developing countries. Prevention of illicit drug injection would not only reduce HIV transmission but would also reduce the other health, psychological and social problems associated with illicit drug injection. One hundred and four subjects who were using heroin intranasally ('sniffing') were recruited for a study of the transition to drug injection. Eligibility criteria included sniffing as the most frequent route of administration and no more than 60 injections in the past 2 years. All subjects received thorough basic information about AIDS, including HIV antibody test counseling. Subjects were then randomly assigned to a four-session social learning based AIDS/drug injection prevention program or a control condition. Eighty-three subjects were successfully followed at a mean time of 8.9 months. Twenty (24%) of the followed subjects reported injecting illicit drugs during the follow-up period. Drug injection during follow-up was associated with being in the control group, intensity of non-injected drug use, prior injection, and having close personal relationships with current intravenous (IV) drug users.
Introduction
drug injection would, however, be a highly desirable
Despite the urgency of preventing the transmission of HIV among intravenous (IV) drug users, many prevention programs have been opposed by those who believe that making drug injection 'safer' will increase the numbers of people who inject illicit drugs. This opposition has occurred despite the fact that all studies that have examined the question have failed to find any link between safer injection programs and an increase in illicit drug injection (Turner, Miller & Moses, 1989). Preventing illicit
method for preventing the transmission of HIV. It would also prevent the many other health, social and psychological problems associated with illicit drug injection, and would not involve the difficulties in achieving abstinence once addiction has developed in drug injectors or the difficulties in changing the sexual behavior of drug injectors (Des Jarlais & Friedman, 1988). The present study was undertaken to develop more knowledge of the transition to illicit drug injection in the context of an AIDS epidemic, and to
Correspondence and reprint requests to: Don C. Des Jarlais, c/o NDRI, 11 Beach Street, New York, New York 10013, USA.
«St a prevention program based On SOCial learning principles. Subjects for the Study were persons who
493
494
Don C. Des Jarlais et al.
were currently using heroin intranasally ('sniffing') and who had injected no more than 60 times in the previous 2 years. These subjects were selected because they were believed to be at very high risk of injecting illicit drugs in the future.
Methods The study was conducted in New York City from 1986 through 1988. New York had an estimated 200000 IV drug users during this period, with enough persons starting to inject to replace those who either stopped injecting or died (State of New York, 1986). New York City had over 5900 cases of AIDS among IV drug users by the end of 1988, and both the IV drug using community (Friedman et al, 1987) and the general population were aware that AIDS was associated with illicit drug injection. Eligibility criteria for participation in the study included intranasal as the primary method of heroin use for the previous 6 months, no more than 60 drug injections in the previous 2 years, and HIV antibody or hepatitis B antibody negative status. The rationale for negative serostatus was so that group sessions would focus on prevention of HIV infection rather than coping with seropositivity. The recruitment procedures are described in detail elsewhere (Casriel et al., 1990), and will only be outlined here. A variety of methods were used to recruit potential subjects, including advertisements in local newspapers, referral by heroin sniffers who had entered drug abuse treatment programs, and referral from persons applying to the study ('snowball' sampling). Overcoming an initial distrust of public authorities was a major problem in recruiting subjects. Initial participation was based on curiosity and the desire for the $20 payment that accompanied completion of the intake data collection. The intake data collection involved an informed consent, a questionnaire administered by a trained interviewer that covered drug use history, sexual behavior history, knowledge of AIDS, and a urine sample for drug testing. Subjects were also given HIV antibody pre-test counseling and asked to give a blood sample for HIV antibody testing. Post-test counseling was provided for the subjects who elected to participate in HIV antibody testing. The great majority of the subjects—90/104 (87%)—decided to take the HIV antibody test. Subjects who decided against HIV antibody testing were required to give a blood sample for hepatitis B testing, to be used as a surrogate measure for HIV antibody.
Hepatitis B testing was also conducted for the subjects who agreed to HIV testing. All questions the subjects had about AIDS were answered during the intake process, leading to some lengthy discussions not only of actual means of transmission, but also of potential casual contact and mosquito-borne transmission. Subjects were randomly assigned to participate in a four-session AIDS prevention program or a control group. (The control group participated only in data collection activities, but as noted above, these provided for fairly intensive education about AIDS.) A detailed description of the theoretical basis of the experimental program and the clinical issues that arose during the sessions has been presented elsewhere (Casriel et al, 1990), and only a brief overview will be presented here. The sessions lasted 1 to 1^ hours each, and occurred over a 2week time period. The sessions included an introduction to AIDS, drug use and drug injection, sexual behavior and AIDS, and seeking entry into drug abuse treatment programs. The sessions involved didactic materials, group discussion and role playing critical situations such as refusing an offer of injection or seeking entry into a treatment program when one's non-injected drug use became too heavy. The explicit purpose of the experimental program was to prevent any drug injection among those who had never injected and any additional drug injection among those who had injected at some time in the past. Instructions on 'safer' injection procedures, such as using bleach to decontaminate injection equipment, were provided, although these were not well-received by the participants. (The subjects thought of themselves as sniffers and did not anticipate becoming injectors.) Reduction in noninjected use of illicit drugs was an additional goal of the program, but the trainers were clear not to take a condemning/punitive attitude towards persons who were using non-injected illicit drugs. The emphasis was on recognizing and admitting problems with illicit drug use and then seeking treatment to reduce/eliminate the illicit drug use. The two trainers for the group successfully encouraged a therapeutic atmosphere, in which participants felt free to discuss personal problem situations and seek help from the trainers and from their peers. Participation in the experimental groups was predicted to be associated with a reduced likelihood of injecting drugs in the follow-up period. Follow-up data collection occurred at a mean of 8.9 months (range from 5 to 21 months) after completion of each prevention cycle. Follow-up
USA transitions report efforts were begun approximately 5 months after the completion of each cycle and intensive efforts to locate all subjects were continued throughout the entire research project. The follow-up data included an interview covering drug and sexual behavior since the intake period, attitudes towards AIDS, and a second blood sample for HIV antibody and/or hepatitis B testing. All subjects who were reinterviewed also gave a second blood sample; there was no difference in the follow-up rate by whether the subjects agreed to HIV testing or participated only in the hepatitis B testing.
Results One hundred and four subjects who met the eligibility criteria were recruited. Follow-up interviews were conducted with 83 (80%) of the subjects, a relatively high percentage for studies of illicit drug use behavior. There were no statistically significant differences at intake between those assigned to the control versus experimental group, or between those who were successfully followed and those lost to follow-up. Table 1 presents demographic characteristics at intake of the 104 subjects. They were primarily young adults with substantial variation in ethnicity, education and sexual orientation. Data on their drug use histories is presented in Table 2; these subjects were heavily involved in illicit drug use and heroin use in particular. Slightly less than half (45%) had injected illicit drugs at some time, and 12% had some injection in the 6 months prior to study entry, but intranasal use was the most frequent method of administration for all subjects at entry into the study, and none had injected more than 60 times in the previous 2 years. The frequency of intranasal heroin use reported for the previous 6 months was generally high, but with substantial variation: 21% less than weekly, 8% weekly, 31% several times per week, 17% daily, and 22% several times per day. There is no reference population of persons intranasally using heroin for comparison purposes, but compared to estimates of the entire population of heroin users in New York, these subjects are younger and more likely to be white (State of New York, 1986). Seventy-eight (75%) of the subjects reported that they were heterosexual, 11 (11%) homosexual, and 14 (13%) bisexual. Fifty-eight (56%) were involved in an ongoing sexual relationship at the time of entry into the study, and 54 (52%) reported having multiple sexual partners in the prior 6 months.
495
Ninety-one (88%) reported at intake that they were practicing some form of 'safer sex', with 27 (26%) reporting some use of condoms as a method for safer sex. The subjects had good basic information about AIDS transmission at entry into the study. In response to open-ended questions about how AIDS is transmitted, 92 (88%) mentioned drug injection and 85 (82%) mentioned sexual activity. They saw little relationship between AIDS and their drug use. When asked if concerns about AIDS had affected their drug use, 103 (99%) said that it had not. The other subject stated that anxiety about AIDS had led him to increase his drug use. When asked about circumstances under which they might inject drugs, only 14 (14%) mentioned availability of clean needles as a possible factor for their injecting drugs, and 9 of these had some prior drug injection experience. (At this time sterile injection equipment was available on the illicit market in New York, at a price of $3 to $5 for a needle and syringe, compared to $1 to $2 for a used needle and syringe.) Twenty (24%) of the 83 subjects who were reinterviewed reported having injected drugs during the follow-up time period. Drug injection during follow-up was not associated with length of time to follow-up interview, though such an association might have been found over a longer follow-up period. None of the subjects seroconverted for HIV during the follow-up period; only one seroconverted for hepatitis B. The low numbers of HIV and HBV seroconversions did not permit using these variables as dependent measures in the analyses. When asked why they injected during the followup period, seeking a better high (reported by 11 subjects) and avoiding withdrawal sickness (reported by 7 subjects) were the most frequently noted reasons. None reported the availability of clean injection equipment as a reason. Univariate and multivariate analyses were conducted to determine what variables predicted drug injection during the follow-up period. Table 3 presents the six predictors of any drug injection during the follow-up period that were statistically significant in the final stepwise multivariate logistic regression analysis. (Data on univariate predictors that were not significant in the multivariate analysis are available from the senior author upon request.) Other than the experimental-control group factor, the predictors represent influences from two different time periods—the 6 months preceding entry into the study and lifetime up to the 6 months prior to study entry.
496
Don C. Des Jarlais et al. Table 1. Subject demographics Baseline « = 104
Follow-up n = 83 n %
M
%
Sex Male Females
73 31
70 30
57 26
69 31
Race White Black Hispanic
53 27 24
51 26 23
42 19 22
51 23 26
63 8 11
76 10 13
Age Mean = 27.3 yrs., SD = 6.9 yrs.. Range 16-48 Follow-up: Mean = 27.5, SD = 7.0, Range 16-48 Education Highest grade completed (mean) 12.8 yrs. SD = 2.1, Range 7-16 Follow-up: Mean = 12.7, SD = 2.2, Range 7-16
Sexual orientation* Heterosexual Homosexual Bi-sexual
78 11 14
75 11 14
* One subject did not report sexual orientation. Table 2. Subject drug use histories = 104 IV use ever IV use in past 6 months Sniffed heroin in past 6 months Sniffed cocaine in past 6 months Sniffed heroin and cocaine in combination in past 6 months Alcohol use in past 6 months Marijuana use in past 6 months History of drug abuse treatment
The eligibility criteria for entry into the study permitted no more than 60 injections in the previous two years and required intranasal use as the most frequent route of drug administration at the time of entry. Even within this highly restricted range, prior experience with drug injection—for both the lifetime and prior six month time periods—was associated with follow-up period drug injection. Social relationships with IV drug users also had an independent effect on follow-up period drug injection. As shown in Table 3, having a close personal relationship to a current IV drug user was a significant predictor of follow-up drug injection. ('Close personal relationships' included relatives and close friends.) Believing that one might try
47 12 104 72 32 81 74 42
45% 12% 100% 69% 31% 78% 71% 40%
Follow-up
35 9 83 60 23 64 60 36
42% 11% 100% 72% 28% 77% 72% 43%
injecting because of the influence of friends also was significantly associated with follow-up drug injection in the univariate analyses, but not in the multivariate analysis. A number of variables measuring intensity of prior drug use, particularly non-injected drug use, were associated with follow-up drug injection. The two variables significant in the final multivariate logistic equation were sniffing a combination of heroin and cocaine ('speedball') in the 6 months prior to intake, and having been in drug abuse treatment. On both methodological and conceptual grounds we interpret these results in terms of a relationship between intensity of prior drug use to follow-up period drug injection rather than these
USA transitions report
497
Table 3. Predictors offollow-up drug injection Variable Experimental groupf Close relationship to IV drug user Previous drug treatment Sniffed heroin and cocaine in combination in prior 6 months IV use in prior 6 months IV use before prior 6 months
Regression coefficient
Standard error Odds ratio
Confidence interval for odds ratio
- 0.67* 1.34*** 0.91*
0.40 0.43 0.43
0.512 3.82 2.48
0.266 to 0.986 1.64 to 8.87 1.07 to 5.75
1.03** 1.71*** 0.81*
0.43 0.55 0.41
2.80 5.53 2.25
1.21 to 6.51 1.88 to 16.2 1.01 to 5.02
t One-tailed testing. * /) < 0.05, **/) < 0.01, *** p < 0.005.
two specific variables. The eligibility criteria for entry into the study undoubtedly attenuated variation in some of the prior drug use variables, which could have altered the intercorrelation matrix. Slightly different eligibility criteria might have produced differences in the correlation matrix of the prior drug use variables. Because the outcomes of multivariate regression analyses can be particularly sensitive to small differences in the correlations among the predictor variables, it does not seem appropriate to give special primacy to these two indicators of intensity of prior drug use. The relationship between intensity of non-injected drug use prior to intake and drug injection during followup is consistent with 'avoiding' withdrawal sickness, the second most common self-reported reason for injecting during follow-up. Because injection is a very efficient method of using drugs, users who are physically dependent upon a drug and who have limited financial resources for purchasing the drug, would be more likely to use by injection. Being in the experimental group for this study was associated with a lower probability of injecting during the follow-up period. While the predicted effect of the experimental program was observed, we also want to note that the experimental program clearly did not prevent all drug injection. Fifteeen per cent (6/40) of the persons assigned to the experimental group injected during the follow-up period compared to 33% (14/43) of those assigned to the control group {x^ = 3.5,/) < 0.05, one-tailed testing). Seventy-five (90%) of the subjects practised some form of 'safer sex' during the follow-up period, with an increase to 41 (49%) using condoms at least some of the time. The increase in condom use was not associated with participation in the experimental group. In a multivariate logistic analy-
sis, using condoms during the 6 months prior to intake and a lower frequency of intranasal use of heroin during the 6 months prior to intake were each significantly associated with condom use during follow-up (both/> = < 0.01, data available from the senior author).
Discussion It is not possible to assess quantitatively the 'representativeness' of this recruited sample of persons at high risk for injecting illicit drugs. The high follow-up rate achieved (80%), particularly for drug users not associated with treatment programs, however, does permit some insight into the transition process from non-injection use to injection of illicit drugs. First, the transition should not be considered a single event in which a first injection leads to exclusive use of injection as a method of drug administration. Once a person experiments with injection of heroin or cocaine, he or she may return to non-injected use of these drugs. While slightly less than half of the subjects had some previous injection, only 12% had injected in the 6 months prior to intake, and for all subjects sniffing heroin was the predominant method of administration at intake. The transition process to being a confirmed/ exclusive injector may take months or years, and many persons who experiment with injection undoubtedly do not become confirmed injectors. Previous injection experience does, however, increase the likelihood of later injection. The relatively high rates of persons who injected during the follow-up period (16% in the experimental group, 33% in the control group) raise concerns about preventing drug injection for persons with high levels of non-injection use of heroin and
498
Don C. Des Jarlais et al.
cocaine. The intake process, that included answering all questions the subjects had about AIDS and HIV antibody test counseling, was in itself a relatively intensive intervention for education about AIDS compared to mass media campaigns. That one third of the control group injected in the follow-up period despite participating in this process is a strong indication of the weaknesses of providing information about AIDS and HIV antibody test counseling in deterring drug injection among persons at high risk. Studies of drug education programs have consistently found that 'information only' or 'fear arousal' programs are not effective in the prevention of licit or illicit drug use (Schaps et al., 1981). A recent study of a fear arousal AIDS prevention program for British drug users was also found to have no effect on behavior (Sherr, 1990). That the extensive education about AIDS provided to the control group in this study was not sufficient to deter illicit drug injection in many subjects is thus consistent with the relevant prior studies. Determining whether AIDS education programs that provide only information about negative consequences and modes of transmission are going to be effective will require much additional research (Miller, Turner & Moses, 1990), but the present literature suggests that information only programs are not likely to prevent high risk sexual and drug use behaviors. The theoretical basis for developing the experimental program was that persons sniffing heroin needed to develop skills to manage social pressures to inject drugs and resources to cope with/reduce/ eliminate their non-injection drug use. The followup data appear to support this formulation of the transition from non-injecting to injecting drug use. Having a close personal relationship to a current IV drug user and frequency of non-injected drug use were both strongly associated with follow-up drug injection, and the program had a modest effect in reducing the number of persons injecting during
follow-up. The conceptual basis of designing drug injection/AIDS prevention programs around social pressure/modelling issues and coping with noninjected drug use problems was confirmed by the results of this study. It is our opinion, however, based on the follow-up data presented here and on our clinical impressions from working with the subjects, that many persons at risk for injecting illicit drugs will need more intensive and extensive services than can be provided in a four session AIDS prevention program. Additional research to develop stronger prevention programs and to assess the effectiveness of the different components of such programs is clearly needed.
References CASRIEL, C , DES JARLAIS, D . C , RODRIGUEZ, R. et al.
(1990) Working with heroin sniffers: primary prevention of IV drug use related AIDS, Journal of Substance Abuse Treatment, 7, pp. 1-10. DBS JARLAIS, D . C , FRIEDMAN, S. R. (1988) HIV infection
among persons who inject illicit drugs: problems and prospects, Journal of Acquired Immune Deficiency Syndromes, 1, pp. 267-273. FRIEDMAN, S. R., DES JARLAIS, D . C , SOTHERAN, J. L . et
al. (1987) AIDS and self-organization among intravenous drug users. International Journal of the Addictions, 22, pp. 201-220. MILLER, H . G., TURNER, C . S. & MOSES, L . E. (Eds)
(1990) AIDS: the second decade (Washington, National Academy Press). SCRAPS, E., D I BARTOLO, R., MOSKOWITZ, J., PALLEY, C . &
CHURGIN, S. (1981) A review of 127 drug abuse prevention program evaluations, Jouma/ of Drug Abuse, 11, pp. 17-43. SHERR, L. (1990) Fear arousal and AIDS: do shock tactics work? AIDS, 4, pp. 361-364. STATE OP NEW YORK (1986) Third Annual Update to the Division of Substance Abuse Services'Statewide Comprehensive Five-Year Plan for 1984-85 through 1988-89. (Albany, New York, Division of Substance Abuse Services). TtTRNER, C. F., MILLER, H . G. & MOSES, L . E. (Eds)
(1989) AIDS Sexual Behavior and Intravenous Drug Use (Washington, National Academy Press).
USA TRANSITIONS REPORT
AIDS and the transition to illicit drug injection—results of a randomized trial prevention program DON C. DES JARLAIS', CATHY CASRIEL^ SAMUEL R. FRIEDMAN^ & ANDREW ROSENBLUM2 ' Beth Israel Medical Center, Chemical Dependency Institute, 1st Avenue & E. 16th Street, New York, New York 10003, USA & ^ Narcotic and Drug Research, Inc., 11 Beach Street, New York, New York 10013, USA
Abstract Illicit drug injection is a major component of the AIDS epidemic in the United States, Europe and some developing countries. Prevention of illicit drug injection would not only reduce HIV transmission but would also reduce the other health, psychological and social problems associated with illicit drug injection. One hundred and four subjects who were using heroin intranasally ('sniffing') were recruited for a study of the transition to drug injection. Eligibility criteria included sniffing as the most frequent route of administration and no more than 60 injections in the past 2 years. All subjects received thorough basic information about AIDS, including HIV antibody test counseling. Subjects were then randomly assigned to a four-session social learning based AIDS/drug injection prevention program or a control condition. Eighty-three subjects were successfully followed at a mean time of 8.9 months. Twenty (24%) of the followed subjects reported injecting illicit drugs during the follow-up period. Drug injection during follow-up was associated with being in the control group, intensity of non-injected drug use, prior injection, and having close personal relationships with current intravenous (IV) drug users.
Introduction
drug injection would, however, be a highly desirable
Despite the urgency of preventing the transmission of HIV among intravenous (IV) drug users, many prevention programs have been opposed by those who believe that making drug injection 'safer' will increase the numbers of people who inject illicit drugs. This opposition has occurred despite the fact that all studies that have examined the question have failed to find any link between safer injection programs and an increase in illicit drug injection (Turner, Miller & Moses, 1989). Preventing illicit
method for preventing the transmission of HIV. It would also prevent the many other health, social and psychological problems associated with illicit drug injection, and would not involve the difficulties in achieving abstinence once addiction has developed in drug injectors or the difficulties in changing the sexual behavior of drug injectors (Des Jarlais & Friedman, 1988). The present study was undertaken to develop more knowledge of the transition to illicit drug injection in the context of an AIDS epidemic, and to
Correspondence and reprint requests to: Don C. Des Jarlais, c/o NDRI, 11 Beach Street, New York, New York 10013, USA.
«St a prevention program based On SOCial learning principles. Subjects for the Study were persons who
493
494
Don C. Des Jarlais et al.
were currently using heroin intranasally ('sniffing') and who had injected no more than 60 times in the previous 2 years. These subjects were selected because they were believed to be at very high risk of injecting illicit drugs in the future.
Methods The study was conducted in New York City from 1986 through 1988. New York had an estimated 200000 IV drug users during this period, with enough persons starting to inject to replace those who either stopped injecting or died (State of New York, 1986). New York City had over 5900 cases of AIDS among IV drug users by the end of 1988, and both the IV drug using community (Friedman et al, 1987) and the general population were aware that AIDS was associated with illicit drug injection. Eligibility criteria for participation in the study included intranasal as the primary method of heroin use for the previous 6 months, no more than 60 drug injections in the previous 2 years, and HIV antibody or hepatitis B antibody negative status. The rationale for negative serostatus was so that group sessions would focus on prevention of HIV infection rather than coping with seropositivity. The recruitment procedures are described in detail elsewhere (Casriel et al., 1990), and will only be outlined here. A variety of methods were used to recruit potential subjects, including advertisements in local newspapers, referral by heroin sniffers who had entered drug abuse treatment programs, and referral from persons applying to the study ('snowball' sampling). Overcoming an initial distrust of public authorities was a major problem in recruiting subjects. Initial participation was based on curiosity and the desire for the $20 payment that accompanied completion of the intake data collection. The intake data collection involved an informed consent, a questionnaire administered by a trained interviewer that covered drug use history, sexual behavior history, knowledge of AIDS, and a urine sample for drug testing. Subjects were also given HIV antibody pre-test counseling and asked to give a blood sample for HIV antibody testing. Post-test counseling was provided for the subjects who elected to participate in HIV antibody testing. The great majority of the subjects—90/104 (87%)—decided to take the HIV antibody test. Subjects who decided against HIV antibody testing were required to give a blood sample for hepatitis B testing, to be used as a surrogate measure for HIV antibody.
Hepatitis B testing was also conducted for the subjects who agreed to HIV testing. All questions the subjects had about AIDS were answered during the intake process, leading to some lengthy discussions not only of actual means of transmission, but also of potential casual contact and mosquito-borne transmission. Subjects were randomly assigned to participate in a four-session AIDS prevention program or a control group. (The control group participated only in data collection activities, but as noted above, these provided for fairly intensive education about AIDS.) A detailed description of the theoretical basis of the experimental program and the clinical issues that arose during the sessions has been presented elsewhere (Casriel et al, 1990), and only a brief overview will be presented here. The sessions lasted 1 to 1^ hours each, and occurred over a 2week time period. The sessions included an introduction to AIDS, drug use and drug injection, sexual behavior and AIDS, and seeking entry into drug abuse treatment programs. The sessions involved didactic materials, group discussion and role playing critical situations such as refusing an offer of injection or seeking entry into a treatment program when one's non-injected drug use became too heavy. The explicit purpose of the experimental program was to prevent any drug injection among those who had never injected and any additional drug injection among those who had injected at some time in the past. Instructions on 'safer' injection procedures, such as using bleach to decontaminate injection equipment, were provided, although these were not well-received by the participants. (The subjects thought of themselves as sniffers and did not anticipate becoming injectors.) Reduction in noninjected use of illicit drugs was an additional goal of the program, but the trainers were clear not to take a condemning/punitive attitude towards persons who were using non-injected illicit drugs. The emphasis was on recognizing and admitting problems with illicit drug use and then seeking treatment to reduce/eliminate the illicit drug use. The two trainers for the group successfully encouraged a therapeutic atmosphere, in which participants felt free to discuss personal problem situations and seek help from the trainers and from their peers. Participation in the experimental groups was predicted to be associated with a reduced likelihood of injecting drugs in the follow-up period. Follow-up data collection occurred at a mean of 8.9 months (range from 5 to 21 months) after completion of each prevention cycle. Follow-up
USA transitions report efforts were begun approximately 5 months after the completion of each cycle and intensive efforts to locate all subjects were continued throughout the entire research project. The follow-up data included an interview covering drug and sexual behavior since the intake period, attitudes towards AIDS, and a second blood sample for HIV antibody and/or hepatitis B testing. All subjects who were reinterviewed also gave a second blood sample; there was no difference in the follow-up rate by whether the subjects agreed to HIV testing or participated only in the hepatitis B testing.
Results One hundred and four subjects who met the eligibility criteria were recruited. Follow-up interviews were conducted with 83 (80%) of the subjects, a relatively high percentage for studies of illicit drug use behavior. There were no statistically significant differences at intake between those assigned to the control versus experimental group, or between those who were successfully followed and those lost to follow-up. Table 1 presents demographic characteristics at intake of the 104 subjects. They were primarily young adults with substantial variation in ethnicity, education and sexual orientation. Data on their drug use histories is presented in Table 2; these subjects were heavily involved in illicit drug use and heroin use in particular. Slightly less than half (45%) had injected illicit drugs at some time, and 12% had some injection in the 6 months prior to study entry, but intranasal use was the most frequent method of administration for all subjects at entry into the study, and none had injected more than 60 times in the previous 2 years. The frequency of intranasal heroin use reported for the previous 6 months was generally high, but with substantial variation: 21% less than weekly, 8% weekly, 31% several times per week, 17% daily, and 22% several times per day. There is no reference population of persons intranasally using heroin for comparison purposes, but compared to estimates of the entire population of heroin users in New York, these subjects are younger and more likely to be white (State of New York, 1986). Seventy-eight (75%) of the subjects reported that they were heterosexual, 11 (11%) homosexual, and 14 (13%) bisexual. Fifty-eight (56%) were involved in an ongoing sexual relationship at the time of entry into the study, and 54 (52%) reported having multiple sexual partners in the prior 6 months.
495
Ninety-one (88%) reported at intake that they were practicing some form of 'safer sex', with 27 (26%) reporting some use of condoms as a method for safer sex. The subjects had good basic information about AIDS transmission at entry into the study. In response to open-ended questions about how AIDS is transmitted, 92 (88%) mentioned drug injection and 85 (82%) mentioned sexual activity. They saw little relationship between AIDS and their drug use. When asked if concerns about AIDS had affected their drug use, 103 (99%) said that it had not. The other subject stated that anxiety about AIDS had led him to increase his drug use. When asked about circumstances under which they might inject drugs, only 14 (14%) mentioned availability of clean needles as a possible factor for their injecting drugs, and 9 of these had some prior drug injection experience. (At this time sterile injection equipment was available on the illicit market in New York, at a price of $3 to $5 for a needle and syringe, compared to $1 to $2 for a used needle and syringe.) Twenty (24%) of the 83 subjects who were reinterviewed reported having injected drugs during the follow-up time period. Drug injection during follow-up was not associated with length of time to follow-up interview, though such an association might have been found over a longer follow-up period. None of the subjects seroconverted for HIV during the follow-up period; only one seroconverted for hepatitis B. The low numbers of HIV and HBV seroconversions did not permit using these variables as dependent measures in the analyses. When asked why they injected during the followup period, seeking a better high (reported by 11 subjects) and avoiding withdrawal sickness (reported by 7 subjects) were the most frequently noted reasons. None reported the availability of clean injection equipment as a reason. Univariate and multivariate analyses were conducted to determine what variables predicted drug injection during the follow-up period. Table 3 presents the six predictors of any drug injection during the follow-up period that were statistically significant in the final stepwise multivariate logistic regression analysis. (Data on univariate predictors that were not significant in the multivariate analysis are available from the senior author upon request.) Other than the experimental-control group factor, the predictors represent influences from two different time periods—the 6 months preceding entry into the study and lifetime up to the 6 months prior to study entry.
496
Don C. Des Jarlais et al. Table 1. Subject demographics Baseline « = 104
Follow-up n = 83 n %
M
%
Sex Male Females
73 31
70 30
57 26
69 31
Race White Black Hispanic
53 27 24
51 26 23
42 19 22
51 23 26
63 8 11
76 10 13
Age Mean = 27.3 yrs., SD = 6.9 yrs.. Range 16-48 Follow-up: Mean = 27.5, SD = 7.0, Range 16-48 Education Highest grade completed (mean) 12.8 yrs. SD = 2.1, Range 7-16 Follow-up: Mean = 12.7, SD = 2.2, Range 7-16
Sexual orientation* Heterosexual Homosexual Bi-sexual
78 11 14
75 11 14
* One subject did not report sexual orientation. Table 2. Subject drug use histories = 104 IV use ever IV use in past 6 months Sniffed heroin in past 6 months Sniffed cocaine in past 6 months Sniffed heroin and cocaine in combination in past 6 months Alcohol use in past 6 months Marijuana use in past 6 months History of drug abuse treatment
The eligibility criteria for entry into the study permitted no more than 60 injections in the previous two years and required intranasal use as the most frequent route of drug administration at the time of entry. Even within this highly restricted range, prior experience with drug injection—for both the lifetime and prior six month time periods—was associated with follow-up period drug injection. Social relationships with IV drug users also had an independent effect on follow-up period drug injection. As shown in Table 3, having a close personal relationship to a current IV drug user was a significant predictor of follow-up drug injection. ('Close personal relationships' included relatives and close friends.) Believing that one might try
47 12 104 72 32 81 74 42
45% 12% 100% 69% 31% 78% 71% 40%
Follow-up
35 9 83 60 23 64 60 36
42% 11% 100% 72% 28% 77% 72% 43%
injecting because of the influence of friends also was significantly associated with follow-up drug injection in the univariate analyses, but not in the multivariate analysis. A number of variables measuring intensity of prior drug use, particularly non-injected drug use, were associated with follow-up drug injection. The two variables significant in the final multivariate logistic equation were sniffing a combination of heroin and cocaine ('speedball') in the 6 months prior to intake, and having been in drug abuse treatment. On both methodological and conceptual grounds we interpret these results in terms of a relationship between intensity of prior drug use to follow-up period drug injection rather than these
USA transitions report
497
Table 3. Predictors offollow-up drug injection Variable Experimental groupf Close relationship to IV drug user Previous drug treatment Sniffed heroin and cocaine in combination in prior 6 months IV use in prior 6 months IV use before prior 6 months
Regression coefficient
Standard error Odds ratio
Confidence interval for odds ratio
- 0.67* 1.34*** 0.91*
0.40 0.43 0.43
0.512 3.82 2.48
0.266 to 0.986 1.64 to 8.87 1.07 to 5.75
1.03** 1.71*** 0.81*
0.43 0.55 0.41
2.80 5.53 2.25
1.21 to 6.51 1.88 to 16.2 1.01 to 5.02
t One-tailed testing. * /) < 0.05, **/) < 0.01, *** p < 0.005.
two specific variables. The eligibility criteria for entry into the study undoubtedly attenuated variation in some of the prior drug use variables, which could have altered the intercorrelation matrix. Slightly different eligibility criteria might have produced differences in the correlation matrix of the prior drug use variables. Because the outcomes of multivariate regression analyses can be particularly sensitive to small differences in the correlations among the predictor variables, it does not seem appropriate to give special primacy to these two indicators of intensity of prior drug use. The relationship between intensity of non-injected drug use prior to intake and drug injection during followup is consistent with 'avoiding' withdrawal sickness, the second most common self-reported reason for injecting during follow-up. Because injection is a very efficient method of using drugs, users who are physically dependent upon a drug and who have limited financial resources for purchasing the drug, would be more likely to use by injection. Being in the experimental group for this study was associated with a lower probability of injecting during the follow-up period. While the predicted effect of the experimental program was observed, we also want to note that the experimental program clearly did not prevent all drug injection. Fifteeen per cent (6/40) of the persons assigned to the experimental group injected during the follow-up period compared to 33% (14/43) of those assigned to the control group {x^ = 3.5,/) < 0.05, one-tailed testing). Seventy-five (90%) of the subjects practised some form of 'safer sex' during the follow-up period, with an increase to 41 (49%) using condoms at least some of the time. The increase in condom use was not associated with participation in the experimental group. In a multivariate logistic analy-
sis, using condoms during the 6 months prior to intake and a lower frequency of intranasal use of heroin during the 6 months prior to intake were each significantly associated with condom use during follow-up (both/> = < 0.01, data available from the senior author).
Discussion It is not possible to assess quantitatively the 'representativeness' of this recruited sample of persons at high risk for injecting illicit drugs. The high follow-up rate achieved (80%), particularly for drug users not associated with treatment programs, however, does permit some insight into the transition process from non-injection use to injection of illicit drugs. First, the transition should not be considered a single event in which a first injection leads to exclusive use of injection as a method of drug administration. Once a person experiments with injection of heroin or cocaine, he or she may return to non-injected use of these drugs. While slightly less than half of the subjects had some previous injection, only 12% had injected in the 6 months prior to intake, and for all subjects sniffing heroin was the predominant method of administration at intake. The transition process to being a confirmed/ exclusive injector may take months or years, and many persons who experiment with injection undoubtedly do not become confirmed injectors. Previous injection experience does, however, increase the likelihood of later injection. The relatively high rates of persons who injected during the follow-up period (16% in the experimental group, 33% in the control group) raise concerns about preventing drug injection for persons with high levels of non-injection use of heroin and
498
Don C. Des Jarlais et al.
cocaine. The intake process, that included answering all questions the subjects had about AIDS and HIV antibody test counseling, was in itself a relatively intensive intervention for education about AIDS compared to mass media campaigns. That one third of the control group injected in the follow-up period despite participating in this process is a strong indication of the weaknesses of providing information about AIDS and HIV antibody test counseling in deterring drug injection among persons at high risk. Studies of drug education programs have consistently found that 'information only' or 'fear arousal' programs are not effective in the prevention of licit or illicit drug use (Schaps et al., 1981). A recent study of a fear arousal AIDS prevention program for British drug users was also found to have no effect on behavior (Sherr, 1990). That the extensive education about AIDS provided to the control group in this study was not sufficient to deter illicit drug injection in many subjects is thus consistent with the relevant prior studies. Determining whether AIDS education programs that provide only information about negative consequences and modes of transmission are going to be effective will require much additional research (Miller, Turner & Moses, 1990), but the present literature suggests that information only programs are not likely to prevent high risk sexual and drug use behaviors. The theoretical basis for developing the experimental program was that persons sniffing heroin needed to develop skills to manage social pressures to inject drugs and resources to cope with/reduce/ eliminate their non-injection drug use. The followup data appear to support this formulation of the transition from non-injecting to injecting drug use. Having a close personal relationship to a current IV drug user and frequency of non-injected drug use were both strongly associated with follow-up drug injection, and the program had a modest effect in reducing the number of persons injecting during
follow-up. The conceptual basis of designing drug injection/AIDS prevention programs around social pressure/modelling issues and coping with noninjected drug use problems was confirmed by the results of this study. It is our opinion, however, based on the follow-up data presented here and on our clinical impressions from working with the subjects, that many persons at risk for injecting illicit drugs will need more intensive and extensive services than can be provided in a four session AIDS prevention program. Additional research to develop stronger prevention programs and to assess the effectiveness of the different components of such programs is clearly needed.
References CASRIEL, C , DES JARLAIS, D . C , RODRIGUEZ, R. et al.
(1990) Working with heroin sniffers: primary prevention of IV drug use related AIDS, Journal of Substance Abuse Treatment, 7, pp. 1-10. DBS JARLAIS, D . C , FRIEDMAN, S. R. (1988) HIV infection
among persons who inject illicit drugs: problems and prospects, Journal of Acquired Immune Deficiency Syndromes, 1, pp. 267-273. FRIEDMAN, S. R., DES JARLAIS, D . C , SOTHERAN, J. L . et
al. (1987) AIDS and self-organization among intravenous drug users. International Journal of the Addictions, 22, pp. 201-220. MILLER, H . G., TURNER, C . S. & MOSES, L . E. (Eds)
(1990) AIDS: the second decade (Washington, National Academy Press). SCRAPS, E., D I BARTOLO, R., MOSKOWITZ, J., PALLEY, C . &
CHURGIN, S. (1981) A review of 127 drug abuse prevention program evaluations, Jouma/ of Drug Abuse, 11, pp. 17-43. SHERR, L. (1990) Fear arousal and AIDS: do shock tactics work? AIDS, 4, pp. 361-364. STATE OP NEW YORK (1986) Third Annual Update to the Division of Substance Abuse Services'Statewide Comprehensive Five-Year Plan for 1984-85 through 1988-89. (Albany, New York, Division of Substance Abuse Services). TtTRNER, C. F., MILLER, H . G. & MOSES, L . E. (Eds)
(1989) AIDS Sexual Behavior and Intravenous Drug Use (Washington, National Academy Press).
