ORIGINAL ARTICLE
Prevalence, correlates, and risk perception of HIV
infection among heroin users in Central Taiwan
Tsang-Yaw Lin
a, Vincent C.H. Chen
b,c, Chao-Hui Lee
d, Chung-Ying Chen
a,
Wen-Chuan Shao
a, Sheng-Huang Chang
a, Jen-Yu Chou
e, Te-Jen Lai
b,f,
Cleusa P. Ferri
g, Michael Gossop
h, Charles T.C. Lee
i,*
a
Tsaotun Psychiatric Center, Department of Health, Nan-Tou, Taiwan
b
Department of Psychiatry, Chung San Medical University Hospital, Taichung, Taiwan
c
Department of Psychiatry, Chung Shan Medical University, Taichung, Taiwan
d
Department of Clinical Psychology, Taipei City Hospital, Song-De Branch, Taipei,Taiwan
e
Graduate Institute for Humanities in Medicine, Taipei Medical University, Taipei, Taiwan
f
Institute of Medicine, Chung Shan Medical University, Taichung, Taiwan
g
Department of Psychobiology, Federal University of Sao Paulo, Sao Paulo, Brazil
hKing’s College London, Institute of Psychiatry, London, United Kingdom i
Department of Public Health, Kaohsiung Medical University, Kaohsiung, Taiwan
Received 7 September 2012; accepted 25 April 2013 Available online 12 August 2013
KEYWORDS Degree of risk perception; Heroin users; HIV infection; Methadone maintenance treatment program
Abstract We investigated the prevalence and correlated factors of human immunodeficiency virus (HIV) among heroin users attending methadone maintenance treatment (MMT) programs in Central Taiwan, and explored the degree of risk perception of HIV infection among the par-ticipants. Our study participants were 781 heroin users seeking treatment at the MMT program at Tsaotun Psychiatric Center in Taiwan. The presence of HIV antibodies was identified by microparticle enzyme immunoassay and confirmed by western blot. Multivariate logistic regression was used to identify the independent correlates of HIV infection. The mean age of the sample was 36.1 years [standard deviation (SD)Z 7.6]; of the patients, 710 (90.9%) were men. The prevalence of HIV infection among our study population was 20.7%. Multivariate lo-gistic regression analysis revealed that HIV infection was independently associated with the age of the patients of initial heroin use, heroin injection use, nondrug-related criminal convic-tions, needle-sharing behaviors, and sharing injection paraphernalia. A strong agreement ex-isted between self-reported HIV serostatus and the results of laboratory analyses, with 88.8% of patients reporting their condition correctly. We found a high rate of HIV infection among patients in the MMT program. Factors associated with HIV infection were mostly related
* Corresponding author. Department of Public Health, Kaohsiung Medical University, No. 100, Shih-Chuan 1st Road, Sanmin District, Kaohsiung City 80708, Taiwan.
E-mail address:[email protected](C.T.C. Lee).
1607-551X/$36 Copyrightª 2013, Kaohsiung Medical University. Published by Elsevier Taiwan LLC. All rights reserved. http://dx.doi.org/10.1016/j.kjms.2013.04.003
Available online atwww.sciencedirect.com
to drug-use behaviors. These findings stress the importance of education regarding drug-risk behaviors.
Copyrightª 2013, Kaohsiung Medical University. Published by Elsevier Taiwan LLC. All rights reserved.
Introduction
Human immunodeficiency virus (HIV) infection is the most serious global health problem of the 20th century. The estimated number of adults and children newly infected with HIV in 2008 was 2.7 million. In the same year, 2 million AIDS-related deaths were reported [1]. Compared with European countries, certain Asian countries have signifi-cantly higher HIV seroprevalence rates[2]. Approximately 4.7 million people are estimated to have HIV infection in Asia[1], 75% of whom are from China, India, and Thailand
[3].
Compared with the global adult HIV prevalence rate of 0.8% in 2008 [1], the prevalence rate is relatively low in Taiwan (0.082% in 2009) [4]. Most of the HIV-infected in-dividuals in Taiwan are men, aged 20e49 years[4]. Prior to 2003, less than 2% of the newly diagnosed cases each year were injecting drug users (IDUs)[5]. However, in 2006, 60% (nZ 1778) of the newly diagnosed patients contracted HIV through intravenous drug use [6]. Injection drug use has become the main risk factor of HIV infection in the country, with more than one-third (38%) of IDUs in Taiwan becoming HIV/AIDS patients[7].
Heroin is the most commonly injected illicit drug in Taiwan[7]. According to the National Bureau of Controlled Drugs, in 2008, 52,708 people were incarcerated because of drug-related crimes, 77% of them being heroin users [8]. Half of the drug users in Taiwan have injected heroin at least once[9].
Heroin injection is strongly associated with the rapidly developing HIV epidemic[10]and is specifically associated with the sharing of injection paraphernalia [7,11]. The prevalence of HIV is significantly higher among intravenous heroin users than among noninjection heroin users[12]. A study conducted in Spain found that approximately one-third of heroin injectors were HIV positive[13]. Although injecting drug use is strongly associated with the HIV epidemic in most Asian countries, including China, Indonesia, Thailand, and Hong Kong[1], each country has a different set of independent risk factors. In Hong Kong, older men were more likely to be involved in HIV-related risk behaviors [14]. Indonesia, where more than half of IDUs (52%) had contracted HIV, was the only country in Asia that females had a higher HIV prevalence rate than male injectors[1].
Other studies have shown that the main risk factors for HIV infection also include young age[11], male sex [15], lower education[16], sex behaviors[17], criminal convic-tions[18], incarceration [19], depression [20], social sup-port[20], impulsivity, and severity of dependence[21].
Understanding risk perceptions is important for designing appropriate strategies for HIV/AIDS prevention because people’s behavior is guided more by the perceived
risk of infection than by their actual, but unknown, HIV status [22]. In Taiwan, the degree of risk perception con-cerning HIV infection among heroin users is rarely reported or remains unreported, although the scale of HIV infection is low in this population.
Studies on HIV infection among heroin users in Taiwan included incarcerated injection heroin users [7,23]; how-ever, studies investigating the prevalence of HIV infection and its related factors among opiate-dependent patients in methadone maintenance treatment (MMT) programs are rare. Our study investigates the prevalence and correlates of HIV infection among heroin users seeking treatment in the Taiwanese MMT program and explores the degree of participant risk perception.
Materials and methods
Participants and procedure
Participants in our study were heroin-dependent patients, aged 18 years, recruited from 2006 to 2010 at an MMT program at Tsaotun Psychiatric Center in Nantou County, Central Taiwan. Nantou City comprises approximately 540,000 inhabitants. All attendees were interviewed by a senior psychiatrist to confirm the heroin-dependence diagnosis according to the Diagnostic and Statistical Manual of Mental Disorders-Fourth Edition (DSM-IV) criteria. They also completed a set of self-reported questionnaires administered by a trained psychiatric nurse on the first day of the MMT program. Exclusion criteria included illiteracy and a diagnosis of mental retardation. The assessment included demographic data, history of heroin substance abuse and other substance use, criminal history, and psy-chosocial information prior to MMT.
Measurements
The heroin-related behaviors of patients were assessed prior to MMT. Questions included participant ages for initial substance use and first heroin use, frequency of heroin detoxification for more than 2 weeks (lifetime), number of criminal convictions (substance related and nonsubstance related), main method of heroin administration (injection or smoking, lifetime), sharing needles with others (yes/no, lifetime), and sharing paraphernalia, with the exception of needles (yes/no, lifetime). The method of heroin use was determined from the answers of the following question: “What has been your main method for using heroin in the past 6 months?” Heroin in Taiwan is sold using the tradi-tional measuring unit of a half qian (1 qianZ 3.75 g). Pa-tients were asked the amount of heroin (in half qian/d) they used in the 2 weeks prior to the interview.
Blood samples were taken from the patients on entry to treatment. The presence of HIV antibodies was tested by a microparticle enzyme immunoassay (AxSYM HIV 1/2 gO; Abbott Laboratories, Diagnostics Division, Abbott Park, IL) and confirmed by western blot (NEW LAV BLOT 1; Bio-Rad, France). The application of chemiluminescent immunoas-says with the ARCHITECT i2000 system was used to detect anti-hepatitis C virus (anti-HCV) and hepatitis B surface an-tigen (HBsAg). These tests were performed at the Depart-ment of Laboratory Medicine, DepartDepart-ment of Health, Tai-Chung Hospital, Taiwan.
We used the following self-administered questionnaires for this study: the Chinese version of the Centre for Epidemiologic Studies Depression Scale (CESD)[24,25]and the Chinese version of the Severity of Dependence Scale (SDS[Ch]) [26]. A positive correlation was found between SDS[Ch] scores and DSM-IV criteria for heroin dependence
[27]; the family adaptation, partnership, growth, affection, resolve scale (APGAR[Ch])[28,29]; and the Chinese version of the Barratt Impulsiveness Scale (BIS-11-CH)[30,31]. All scales were validated in Taiwan, and total scores were used for statistical analysis. Higher scores indicated more severe conditions for all instruments. Risk perception regarding HIV infection was assessed by the question “Do you have HIV?” (the answer options: “yes”, “no”, or “not sure”).
Statistical analysis
We conducted a descriptive analysis to describe the socio-demographic characteristics, heroin-related problems, and HIV prevalence in the sample. Descriptive statistics, such as means, standard deviations (SDs), and proportions, were used when appropriate. The Chi-square test and t test were used to assess the association of all variables, and all sig-nificant factors were entered into a multivariate logistic regression model (backward elimination) to identify signifi-cant variables related to HIV infection after adjusting for the potential effects of other variables. Kappa was used to represent the agreement between self-reported HIV and laboratory findings. A similar statistical procedure was per-formed to establish significant variables related to needle sharing. The variance inflation factor (VIF) was used to detect multicollinearity in the regression model. Explana-tory variables with a VIF of5 indicates a multicollinearity problem [32]. Differences were considered significant if p< 0.05. An analysis was conducted using SPSS version 15.0 (SPSS Inc., Chicago, IL, USA).
Ethics
Our study was approved by the institutional review board of the Tsaotun Psychiatric Center. All patients provided writ-ten informed consent prior to participation.
Results
The total study sample comprised 781 heroin users with a mean age of 36.1 years (SDZ 7.6); 710 (90.9%) were men and 59.4% of all participants had received<9 years of ed-ucation (mean 9.9, SDZ 2.0). Six patients were excluded because of illiteracy or mental retardation. The mean age
for first heroin use was 26.2 years (SDZ 12.0). The most frequently reported method of heroin administration (used by 81.4% of the patients) was injection. Of the patients, 92.0% had criminal convictions related to substance-use offences (possession/selling; nZ 2.3, SD Z 1.5) and 66.3% related to nonsubstance use (nZ 1.2, SD Z 1.3).
Table 1shows the sample characteristics related to HIV serostatus. The prevalence of HIV infection among the study population was 20.7%. The prevalence of HCV and hepatitis B virus (HBV) was 90.1% and 66.2%, respectively. In total, 154 men and eight women were diagnosed as HIV seropositive. In the univariate analysis, HIV-positive pa-tients were younger in age (34.2 6.9 years vs. 36.5 7.9 years; pZ 0.001) and younger in age for their first illegal drug (22.1 6.2 years vs. 24.3 7.6 years; p < 0.001), had more drug-related (2.7 1.6 vs. 2.2 1.5; p < 0.001) and nondrug-related (1.6 1.5 vs. 1.1 1.2; p < 0.001) crim-inal convictions, and were more likely to initiate heroin use at a younger age (23.9 5.7 years vs. 26.8 13.1 years; p Z 0.006) than HIV-negative patients. HIV-infected pa-tients also tended to be more depressed (pZ 0.009). The prevalence of HIV-positive users among the injecting heroin users in this study group was 23.1%. Those who injected heroin were more likely to be HIV positive (pZ 0.001). HIV-positive patients were also more likely to test HIV-positive for HCV infection (p< 0.001). Compared to HIV-negative her-oin users, HIV-positive herher-oin users were more likely to be men (p Z 0.039). HIV-positive patients reported more sharing of needles (p< 0.001) and injection paraphernalia (p< 0.001). The HIV subgroups did not differ significantly regarding HBV. Factors such as family support, impulsivity, education, detoxification frequency, and severity of dependence were not significant in our study.
With respect to the degree of risk perception, 779 heroin users responded regarding the risk perception of their HIV status. Among them, 150 (19.3%) reported HIV infection, 553 (71.0%) reported HIV noninfection, and 76 (9.7%) were unsure of their HIV status. A strong agreement existed be-tween self-reported HIV serostatus and the results of the laboratory analyses, with 88.8% of responders reporting their condition correctly. Among the 150 HIV-infected her-oin users identified by blood test, 91.9% reported their condition correctly. We categorized the first answer as infection (self-report) and the last two answers as no infection (self-report) to calculate the kappa value with laboratory results. The kappa was 0.94 [95% confidence interval (CI), 0.91e0.97; p Z 0.005], showing a high level of agreement [33]. We treated HIV serostatus as the gold standard; the positive and negative predictive values were 0.987 and 0.979, respectively, in this study. The sensitivity and specificity were 0.919 and 0.997, respectively. These measurements also showed high levels of agreement be-tween the self-reports and the HIV serostatus.
We performed a multivariate logistic regression analysis with backward elimination to identify factors related to HIV infection among heroin users. HIV infection was indepen-dently associated with the age of initial heroin use [odds ratio (OR), 0.95; 95% CI, 0.91e0.98; p Z 0.001], heroin injection use (OR, 1.95; 95% CI, 1.01e3.49; p Z 0.025), nondrug-related criminal convictions (OR, 1.95; 95% CI, 1.09e3.49; p Z 0.005), needle-sharing behaviors (OR, 2.11; 95% CI, 1.17e3.80; p Z 0.013), and sharing of injection
paraphernalia (OR, 5.17; 95% CI, 3.18e8.41; p < 0.001;
Table 2). The VIFs of explanatory variables of<5 indicated limited multicollinearity in the previous final regression model (Table 2). All models in the model selection process also included variables with VIFs of<5.
We performed a similar analysis to identify factors related to needle sharing among heroin users. Needle sharing was independently associated with the severity of heroin dependence (OR, 1.17; 95% CI, 1.05e1.32; pZ 0.006), family support (OR, 0.91; 95% CI, 0.85e0.98; pZ 0.010), sharing injection paraphernalia (OR, 20.59; 95% CI, 10.78e39.32; p < 0.001), and HIV infection (OR, 2.75; 95% CI, 1.47e5.15; p Z 0.002). The VIF also showed limited multicollinearity in the model selection process (Table 3).
Discussion
The results showed a high rate of HIV infection (20.7%) among the heroin users seeking treatment at the MMT program. The prevalence was much higher than that of the
Table 1 Patient demographics and clinic characteristics of heroin users in Central Taiwan, 2006e2010.
Variable Total valid sample HIV p (t test)
Negative Positive na Mean SD n Mean SD n Mean SD
Age (y) 781 36.1 7.6 619 36.5 7.9 162 34.2 6.9 0.001
Education (y) 781 9.9 2.0 619 9.9 2.1 162 9.7 1.9 0.193
Onset age of substance (y) 781 23.8 7.3 619 24.3 7.6 162 22.1 6.2 <0.001 Onset age of heroin use (y) 781 26.2 12.0 619 26.8 13.1 162 23.9 5.7 0.006 Dose of heroin (half qian/db) 780 5.0 5.1 618 4.9 5.2 162 5.3 5.0 0.459
Detoxification for heroin more than 2 wk (lifetime) 778 2.8 2.8 616 2.8 2.7 162 2.7 3.1 0.632 Severity of heroin dependence (SDS,ctotal score) 781 7.4 2.7 619 7.4 2.7 162 7.4 2.5 0.962
No. of criminal convictions (illegal substances) 779 2.3 1.5 618 2.2 1.5 161 2. 7 1.6 <0.001 No. of criminal convictions (nonsubstance causes) 777 1.2 1.3 617 1.1 1.2 160 1.6 1.5 <0.001 Impulsivity (Barrat Impulsivity Scalec) 725 72.5 8.7 569 72.2 7.8 156 73.5 11.3 0.104 Family support (APGARc) 781 5.7 4.3 619 5.6 4.3 162 6.1 4.3 0.161 Depression (CESD Scalec) 781 23.05 9.8 619 22.6 9.8 162 24.8 9.6 0.009
Variable Total valid sample size HIV p (Chi-square )
Negative Positive n % n % Sex Male 710 556 89.8 154 95.1 0.039 Female 71 63 10.2 8 4.9 Injection Yes 618 475 77.6 143 89.9 0.001 No 153 137 22.2 16 10.1
Needle sharing Yes 86 38 6.2 48 29.6 <0.001
No 693 579 93.8 114 70.4
Paraphernalia sharing Yes 148 70 11.3 78 48.1 <0.001
No 632 548 88.7 84 51.9
HCV Yes 704 542 88.0 162 100.0 <0.001
No 74 74 12.0 0 0.0
HBV Yes 517 411 66.2 106 65.4 0.777
No 262 206 33.4 56 34.6
APGAR Z adaptation, partnership, growth, affection, resolve score; CESD Z Centre for Epidemiologic Studies Depression; HCVZ hepatitis C virus; HBV Z hepatitis B virus; SD Z standard deviation; SDS Z Severity of Dependence Scale.
a Total number of observations do not sum up to 781 for all variables due to missing values. b 1 qianZ 3.75 g (number of half qians of heroin used per day).
c Chinese versions.
Table 2 Multivariate logistic regression analysis for human immunodeficiency virus (HIV) infection among 781 heroin users in Central Taiwan, 2006e2010.
Variables Odds ratio p Variance inflation factor Estimate (95% CI)
Onset age of heroin use (y)
0.95 0.91e0.98 0.001 1.04 Vessel or muscle injection of heroin 1.95 1.01e3.49 0.025 1.04 Criminal convictions (nonsubstance causes) 1.95 1.09e3.49 0.005 1.01
Needle sharing 2.11 1.17e3.80 0.013 1.45 Paraphernalia
sharing
5.17 3.18e8.41 0.000 1.45
general population in Taiwan (0.082% in 2009)[4]. The HIV rate among those who used heroin by injection was 23%. Among the general population in Taiwan, the prevalence of chronic HBV carriers is estimated to be 15e20% and the prevalence of HCV infection 3e4%[34]. Our findings show a high prevalence of HBV (66.2%) and HCV (90.1%) infection among heroin users. The high prevalence of HIV infection among this study population should be viewed in the context of the findings from other Asian countries. The HIV prevalence among injecting heroin users is as low as 2.6% in Australia, and as high as 28% in Indonesia [2] or 52% in Thailand[2]. The rate in other countries ranges from 1% in New Mexico, USA[35]; 20.4% in Vancouver, Canada[13,36]; to 31.7% in Madrid and Barcelona, Spain [13]. In Taiwan, Chen et al. [7]reported a higher HIV-infection frequency (38%) than that of our study, possibly because their patients belonged to a higher-risk population (male incarcerated IDUs).
Our results also indicated that HIV-positive heroin users were more likely to be those who injected heroin, and shared needles and other injection paraphernalia.
A previous study in Taiwan also found that a shift in heroin administration from smoking to injection, and increased needle-sharing behaviors further exacerbated the HIV epidemic[7]. HIV prevalence among drug addicts has reduced by half within the past 4 years. However, among the HIV-infected population, the proportion of drug users remains approximately one-third [4]. Our study re-sults indicated that drug injection is still a crucial factor associated with HIV infection in Taiwan.
Our results are consistent with the results of several other studies regarding the strong association between sharing needles and/or injection paraphernalia and HIV infection [11]. Needle-sharing behaviors are typically associated with outbreaks of HIV/AIDS in Thailand, India, Malaysia, Burma, China, and Vietnam[31]. The prevalence of injecting and sharing injection equipment becomes lower in patients admitted to addiction treatment service programs[37]. The proportion of drug users remains high
among the HIV-infected population[4]. Our findings suggest that sharing needles/injections and/or paraphernalia remain important factors relating to HIV infection. Coun-seling for IDUs should emphasize the risks of indirect sharing and use of injection paraphernalia[13].
Heroin users in our study group first used heroin at a mean age of 23.9 years (SD 5.7). Patients whose initial use of heroin occurred at a younger age were more likely to test positive for HIV infection. This finding is consistent with the findings of other studies conducted in India[11], indicating that earlier heroin use was related to injecting heroin at an earlier age[38], and younger IDUs were re-ported to borrow current injection equipment more often than older IDUs[39]. Greater attention should be focused on the needs of younger heroin users to help delay or prevent the onset of drug injecting[39]. HIV-positive her-oin users were more likely to have nondrug criminal con-victions in this study group. Although other studies have examined drug-related criminal convictions and HIV infection [18,23], none has investigated nondrug-related criminal convictions among heroin users and HIV infec-tion. Our study is the first in Asia to examine the percep-tions of HIV status among heroin users, and our results show that more than 90% of HIV-positive patients know their status. Because similar studies on heroin users do not exist, a comparison was not feasible. Compared with the findings from other populations, our findings are similar to previous reports on the general population in Malawi (71%)
[22] and on men who have sex with men in Brazil (kappaZ 0.88)[40], but contrary to those on the general population of Nigerians[41]. The high-risk perception may be because of good public health education in Taiwan. We did not have information on the influence of their risk perception on their own heroin-related behaviors, such as engaging in fewer unprotected sexual acts, and sharing needles or paraphernalia. Further research is necessary to explore this issue.
Our study findings should be interpreted according to certain study design limitations. The cross-sectional study design is limited to determining the causality relationship among variables; the measurement of psychosocial factors is based on self-reported behaviors. The potential for social desirability bias exists, such as underestimating impulsivity characteristics. The sample was not selected randomly and was recruited from a psychiatric center in Central Taiwan; therefore, our results might not be applicable to other settings. These heroin users may be a highly select group with greater dependence severity and higher risk behav-iors. The prevalence of HIV infection may be overestimated among this population. The information regarding sexual behavior is insufficient; in certain countries such as China, heterosexual transmission has become the dominant mode of HIV transmission[42].
Our study also has certain strengths. It is the first study to examine risk perceptions of HIV status among heroin users in Asia. Of the studies related to heroin use and HIV infection in Taiwan[7,43], our study contains the largest group sample and is the only study investigating the popu-lation of both injecting and noninjecting heroin users. We explored psychosocial and heroin-related behaviors comprehensively in this study to provide more complete information on the current trends of HIV infection among
Table 3 Multivariate logistic regression analysis for nee-dle sharing among 781 heroin users in Central Taiwan, 2006e2010.
Variables Odds ratio p Variance inflation factor Estimate (95% CI) Severity of heroin dependence (SDS, total score) 1.17 1.05e1.32 0.006 1.03 Family support (APGAR) 0.91 0.85e0.98 0.01 1.01 Paraphernalia sharing 20.59 10.78e39.32 <0.001 1.19 HIV infection (serostatus) 2.75 1.47e5.15 0.002 1.17
APGARZ adaptation, partnership, growth, affection, resolve score; CIZ Confidence interval; SDS Z Severity of Dependence Scale.
heroin users in Taiwan. Our study is beneficial for advancing further the effectiveness of current treatment programs, including maintenance-related programs.
In conclusion, we identified the current factors associ-ated with HIV infection among heroin users in Taiwan. Our results are helpful for formulating and modifying preven-tion programs in the country. Prevenpreven-tion programs should emphasize on disseminating knowledge about drug addic-tion and HIV infecaddic-tion in adults and younger groups prior to people using illegal drugs. Although a harm-reduction pro-gram has been pursued for approximately 4 years in Taiwan, needle- and injection paraphernalia-sharing be-haviors are strongly associated with HIV infection among heroin users in the country. Educational programs should apply greater emphasis on perceiving the risks of drug-use behavioral patterns to increase the motivation for behav-ioral change. Providing specific advice on avoiding alcohol or heavy drinking is critical because of the high rate of viral hepatitis among HIV-infected patients. Establishing a close liaison between MMT services and medical services is also vital to provide antiretroviral or liver disease treatments.
Acknowledgments
This research was supported by a grant from the National Science Council (NSC 97-2314-B-040-007), Taiwan.
References
[1] UNAIDS/WHO. Core slides: global summary of the HIV and AIDS epidemic, 2009 AIDS epidemic update. http://www.unaids. org/en/dataanalysis/knowyourepidemic/epidemiologypublic-ations/2009aidsepidemicupdate/; 2010.
[2]Lawrinson P, Ali R, Buavirat A, Chiamwongpaet S, Dvoryak S, Habrat B, et al. Key findings from the WHO collaborative study on substitution therapy for opioid dependence and HIV/AIDS. Addiction 2008;103:1484e92.
[3] HIV/AIDS epidemic in Asia. Available at,http://www.kff.org; 2008.
[4]CDC-Taiwan. HIV/AIDS monthly report. Taiwan: CDC; 2009. [5]Yang CH, Yang SY, Shen MH, Kuo HS. The changing
epidemi-ology of prevalent diagnosed HIV infections in Taiwan, 1984e2005. Int J Drug Policy 2008;19:317e23.
[6]CDC-Taiwan. HIV/AIDS 1984e2006/12/31. Taiwan: CDC; 2006. [7]Chen CH, Ko WC, Lee HC, Hsu KL, Ko NY. Risky behaviors for HIV infection among male incarcerated injection drug users in Taiwan: a caseecontrol study. AIDS Care 2008;20:1251e7. [8]CDC-Taiwan. Substance abuse prevention teaching package.
Taiwan: CDC; 2009.
[9]Wu M-J. Association between medical compliance and other factors and the retention time of methadone maintenance treatment among intravenous drug abusers in Taiwan. Taipei: Public Health, National Yang-Ming University; 2009.
[10] Barrio G, De La Fuente L, Toro C, Brugal TM, Soriano V, Gonzalez F, et al. Prevalence of HIV infection among young adult injecting and non-injecting heroin users in Spain in the era of harm reduction programmes: gender differences and other related factors. Epidemiol Infect 2007;135:592e603. [11] Mahanta J, Medhi GK, Paranjape RS, Roy N, Kohli A,
Akoijam BS, et al. Injecting and sexual risk behaviours, sexually transmitted infections and HIV prevalence in inject-ing drug users in three states in India. AIDS Dec 2008;22(Suppl. 5):S59e68.
[12] Vallejo F, Toro C, de la Fuente L, Brugal MT, Soriano V, Silva TC, et al. Prevalence of and risk factors for hepatitis B virus infec-tion among street-recruited young injecinfec-tion and non-injecinfec-tion heroin users in Barcelona, Madrid and Seville. Eur Addict Res 2008;14:116e24.
[13] Bravo MJ, Royuela L, De la Fuente L, Brugal MT, Barrio G, Domingo-Salvany A. Use of supervised injection facilities and injection risk behaviours among young drug injectors. Addic-tion 2009;104:614e9.
[14] Li H, Goggins W, Lee SS. Multilevel analysis of HIV related risk behaviors among heroin users in a low prevalence community. BMC Public Health 2009;9:137.
[15] Rondinelli AJ, Ouellet LJ, Strathdee SA, Latka MH, Hudson SM, Hagan H, et al. Young adult injection drug users in the United States continue to practice HIV risk behaviors. Drug Alcohol Depend 2009;104:167e74.
[16] Strathdee SA, Galai N, Safaiean M, Celentano DD, Vlahov D, Johnson L, et al. Sex differences in risk factors for HIV sero-conversion among injection drug users: a 10-year perspective. Arch Intern Med 2001;161:1281e8.
[17] Schonnesson LN, Atkinson J, Williams ML, Bowen A, Ross MW, Timpson SC. A cluster analysis of drug use and sexual HIV risks and their correlates in a sample of African-American crack cocaine smokers with HIV infection. Drug Alcohol Depend 2008;97:44e53.
[18] Racz J, Gyarmathy VA, Neaigus A, Ujhelyi E. Injecting equip-ment sharing and perception of HIV and hepatitis risk among injecting drug users in Budapest. AIDS Care 2007;19:59e66. [19] Cheng Y, Sherman SG, Srirat N, Vongchak T, Kawichai S,
Jittiwutikarn J, et al. Risk factors associated with injection initiation among drug users in Northern Thailand. Harm Reduct J 2006;3:10.
[20] Risser J, Cates A, Rehman H, Risser W. Gender differences in social support and depression among injection drug users in Houston, Texas. Am J Drug Alcohol Abuse 2010;36:18e24. [21] Meade CS, Graff FS, Griffin ML, Weiss RD. HIV risk behavior
among patients with co-occurring bipolar and substance use disorders: associations with mania and drug abuse. Drug Alcohol Depend 2008;92:296e300.
[22] Bignami-Van Assche S, Chao LW, Anglewicz P, Chilongozi D, Bula A. The validity of self-reported likelihood of HIV infection among the general population in rural Malawi. Sex Transm Infect 2007;83:35e40.
[23] Chiang SC, Chen SJ, Sun HJ, Chan HY, Chen WJ. Heroin use among youths incarcerated for illicit drug use: psychosocial environment, substance use history, psychiatric comorbidity, and route of administration. Am J Addict 2006;15:233e41. [24] Chien CP, Cheng TA. Depression in Taiwan: epidemiological
survey utilizing CES-D. Seishin Shinkeigaku Zasshi 1985;87: 335e8.
[25] Radloff L. The CES-D Scale: a self-report depression scale for research in the general population. Appl Psychol Meas 1977;1: 385e401.
[26] Gossop M, Darke S, Griffiths P, Hando J, Powis B, Hall W, et al. The Severity of Dependence Scale (SDS): psychometric properties of the SDS in English and Australian samples of heroin, cocaine and amphetamine users. Addiction 1995;90: 607e14.
[27] Chen VC, Chen H, Lin TY, Chou HH, Lai TJ, Ferri CP, et al. Severity of heroin dependence in Taiwan: reliability and val-idity of the Chinese version of the Severity of Dependence Scale (SDS[Ch]). Addict Behav 2008;33:1590e3.
[28] Chau TT, Hsiao TM, Huang CT, Liu HW. A preliminary study of family Apgar index in the Chinese. Gaoxiong Yi Xue Ke Xue Za Zhi 1991;7:27e31.
[29] Smilkstein G, Ashworth C, Montano D. Validity and reliability of the family APGAR as a test of family function. J Fam Pract 1982;15:303e11.
[30] Evans JL, Hahn JA, Lum PJ, Stein ES, Page K. Predictors of injection drug use cessation and relapse in a prospective cohort of young injection drug users in San Francisco, CA (UFO Study). Drug Alcohol Depend 2009;101:152e7.
[31] Beyrer C, Razak MH, Lisam K, Chen J, Lui W, Yu XF. Overland heroin trafficking routes and HIV-1 spread in south and south-east Asia. AIDS 2000;14:75e83.
[32] Hsieh FY, Lavori PW, Cohen HJ, Feussner JR. An overview of variance inflation factors for sample-size calculation. Eval Health Prof 2003;26:239e57.
[33] Fleiss JL. Statistical methods for rates and proportions. New York: John Wiley; 1981.
[34] CDC-Taiwan. Statistics of communicable diseases and sur-veillance reports in Taiwan area. Taipei: Centers for Disease Control, Department of Health, Executive Yuan, Taiwan; 2010.
[35] Baumbach JP, Foster LN, Mueller M, Cruz MF, Arbona S, Melville S, et al. Seroprevalence of select bloodborne patho-gens and associated risk behaviors among injection drug users in the Paso del Norte region of the United StateseMexico border. Harm Reduct J 2008;5:33.
[36] Werb D, Kerr T, Small W, Li K, Montaner J, Wood E. HIV risks associated with incarceration among injection drug users: implications for prison-based public health strategies. J Public Health (Oxf) 2008;30:126e32.
[37] Gossop M, Marsden J, Stewart D, Treacy S. Reduced injection risk and sexual risk behaviours after drug misuse treatment: results from the National Treatment Outcome Research Study. AIDS Care 2002;14:77e93.
[38] Fuller CM, Vlahov D, Arria AM, Ompad DC, Garfein R, Strathdee SA. Factors associated with adolescent initiation of injection drug use. Public Health Rep 2001;116(Suppl. 1):136e45.
[39] Fennema JS, Van Ameijden EJ, Van Den Hoek A, Coutinho RA. Young and recent-onset injecting drug users are at higher risk for HIV. Addiction 1997;92:1457e65.
[40] Salani Mota RM, Sansigolo Kerr LR, Kendall C, Pinho A, de Mello MB, Dourado I, et al. Reliability of self-report of HIV status among men who have sex with men in Brazil. J Acquir Immune Defic Syndr 2011;57(Suppl. 3):S153e6.
[41] Fagbamigbe AF, Akinyemi JO, Adedokun BO, Bamgboye EA. Gender variation in self-reported likelihood of HIV infection in comparison with HIV test results in rural and urban Nigeria. AIDS Res 2011;8:44.
[42] Wang LWN, Wang L, Li D, Jia M, Gao X, Qu S, et al. The 2007 es-timates for people at risk for and living with HIV in China: progress and challenges. J Acquir Immune Defic Syndr 2009;50:414e8. [43] Cheng AT, Hawton K, Chen TH, Yen AM, Chang JC, Chong MY,
et al. The influence of media reporting of a celebrity suicide on suicidal behavior in patients with a history of depressive disorder. J Affect Disord 2007;103:69e75.
