Participants and Follow-ups
Over the period of 15 months (between June 2001 and August 2002), about 1,000 male inmates were approached and invited to join the study. A total of 689 male inmates was eligible and agreed to be part of the study. To be eligible for the study, the inmates must be between 20–50 years old at the enrolment, have at least 5 years of remaining sentence term from the date of study, not be seriously ill, and pass a test on basic knowledge about HIV/AIDS. The test on basic HIV knowledge was required as a part of participant's enrolment to ensure that the study did not enroll those who are totally ignorant about HIV and might not understand the significance of their participation in the project. The staff of the project explained the objectives and the nature of the study to the potential participants. If they fitted the eligibility criteria and agreed to join the study, a consent form was signed. They were then interviewed about their demographics and information related to their incarceration. There were no potential participants who did not pass the basic test on HIV knowledge. HIV risk behaviors before and during incarceration were ascertained at the date of enrolment. The risk behaviors were verified whether they were presented before the incarceration (Yes or No) and whether they were continued or initiated during the incarceration (Yes or No). The responses were then categorized into "No/No" (never), "Yes/No" (before incarceration only), "Yes/Yes" (continued into incarceration), and "No/Yes" (initiated during incarceration).
Blood was taken for determination of HIV-1 and other infectious profiles, complete blood count and blood chemistry. Urine specimens were collected for determination of opiates and metamphetamine. Pre-test and post-test counseling for anti-HIV testing was given to all inmates by experienced nurses.
The follow-up period was 5 months. As a part of experimentation of follow-up schedules, 500 inmates (72.6%) were followed up only once at the end of 5 months. The other 189 (27.4%) were followed up monthly for 5 times during the study period. The rationale behind more frequent visits in a part of the participants is to enable us to detect early HIV seroconverters (if any) and to study these early seroconverters in more details. However, financial limitation of the project did not allow us to follow up all participants on a monthly basis. For those who were followed up on a monthly basis, they were interviewed, counseled, physically examined, and blood tested at each follow-up visit as they were at baseline. Risk behaviors were assessed at the baseline and at the 5-month visit only. In this study, the results of anti-HIV testing were kept confidential and were not used to separate anti-HIV positives from those who were negative. At the time of the study, there were no programs that provide prophylaxis for opportunistic infections or offer anti-retroviral for HIV-infected prisoners.
There were no reported incidents of adverse consequences associated with notifying HIV seropositives or seroconverters of their status. We also specifically asked, at the end of the study, if the participants experienced such adverse consequences during their participation in the study and none reported such experiences.
Data management and statistical analysis
All personal data of the inmates including their identification, demographics, risk behaviors and laboratory findings were kept confidential. Codes were used to identify these subjects. All data were double entered by the investigators, using Microsoft Access version 97 (Microsoft Corporation, New York, USA). Potential HIV risk factors were examined using univariate and multivariate analyses. For association of categorical variables, Yate's corrected chi square test was used, except where the expected frequency is less than 5 and Fisher's exact test is recommended. Variables that were found statistically significant in the univariate analysis and biologically plausible, as determined by prior knowledge and suggested by literature, were included in the multivariate analysis. Multiple logistic regression analysis was used with anti-HIV status as the outcome variable (Intercooled version 6, Stata Corporation College Station, Texas, USA). Maximum likelihood ratio estimation was used to estimate the parameters and the goodness of fit was applied to assess various models during the model-fitting process. Likelihood ratio (LR) test was applied to assess statistical significance. Both crude odds ratio and adjusted odds ratios (ORs), with associated 95% confidence intervals (CIs) and p-values, were presented.
Drug use risks were assessed based on the inmates' injection history, history of attending drug abuse treatment clinic, and urine test for opiates at the time of enrolment. Injection history was further assessed based on duration of injection, presence or absence of infection scar(s), and history of sharing drug injection paraphernalia.
HIV seroconversion rate was also calculated among the inmates who were HIV negative at enrolment and turned HIV-positive during the follow-up period. Poisson estimation was used to determine the rate and it's associated 95% confidence interval.
At the enrolment, the subjects were blood tested for Anti-HIV antibody by HIV EIA (Uni-Form II plus O, Organon Techinka, Boxtel, The Netherlands) with HIV-1 confirmation by Western blot (HIV Blot 2.2, Gene Lab Diagnostic Pte Ltd., Science Park, Singapore). The hepatitis B profiles included HBsAg, HBc and HBs antibody EIA (ETI-MAK-4, ETI-AB COREK-2 and ETI-AB-AUK-3, Diasorin s.r.1, Vercelli, Italy). The hepatitis C antibody was determined by an EIA kit (ETI-AB-HCVK-3, Diasorin s.r.1, Vercelli, Italy). The syphilis serology included VDRL, and TPHA (VDRL, Syphscreen RPR, Porton Cambridge, Kennett, United Kingdom and Syphilis TPHA tests, Human, Wies baden, Germany). In addition, anti HSV-2 IgM EIA (Capita HSV-2 IgM, Trinity Biotech, New York, USA) and anti-chlamydia IgM EIA (Sero ELISA chlamydia true IgM, Savyon Diagnostics Ltd., Ashdod, Israel) were also carried out. Urine was tested for opiates and metamphetamine (Capita EIA test kit, Trinity Biotech, New York, USA). Complete blood count was analyzed by an automated analyzer (Counter-HmX, Florida, USA). Blood chemistry profiles included liver function test, lipid and renal profiles and were tested by an automated analyzer (Cobas Mira, Roche Boehringer Mannheim, Mannheim Germany). All laboratory tests were performed only at the enrolment, except for HIV and HCV serology that were tested at every visit.
The Ethical Review Committee for Research in Human Subjects of the Ministry of Public Health of Thailand approved the study protocol in March 2001.