Informed Consent Process

During the informed consent process, research staff took great care to be transparent regarding all aspects of the research study. All participants were told we were specifically studying HIV, including it's potential to be sexually transmitted, and provided counseling around HIV testing. We explained to IDUs the reason for recruitment and all relevant study procedures. But it was less clear what to disclose to the recruited sex partners. To fully inform sexual partners of the study's purpose and specific reason for their study eligibility might require a disclosure of the index-participants' drug use and heightened risk of HIV.

Thus, the crux of the ethical dilemma was the informed consent process for the sexual partners, and specifically the explanations about partner selection and the study purpose. Our team generated multiple conceptual options for informed consent statements, with the varying levels of detail in italics for comparison:

1. "You were selected to participate in this study on relationships and risk behaviors because your partner volunteered for the study and was asked to invite his/her sex partner."

2. "You were selected to participate in this study of injection drug users and their sex partners because your partner volunteered for the study and was asked to invite his/her sex partner."

3. "You were selected to participate in this study of injection drug users and their sex partners because your partner volunteered for this study and was asked to invite his/her sex partner. About half of injection drug users are infected with HIV and we are particularly interested in understanding risk behaviors of injection drug users who are infected with HIV and their partners. In particular, we want to learn what behaviors result in the highest risk of HIV transmission and better ways to reduce risk of transmission."

The first statement provides no information to the partners recruited for the study about their risk of infection now or in the near future. The second statement differs from the first by specifying that the partner is an IDU. This information is especially revealing because it is common knowledge among residents of Estonia that many IDUs are HIV positive. The third statement clearly states partners' potential risk of HIV if they are not actively preventing transmission (e.g., using condoms).

The research team was unable to use the third statement, due to concerns of the local ethics committee about such specific, yet ambiguous communication that could further stigmatize IDUs. While it may appear arbitrary that one stigma (i.e., IDU status) can be shared while another (i.e., HIV status) can not, it is not uncommon that inconsistencies in sensitivity about exposures exist in many countries. The research team debated the first two options at length. The decision would have health, scientific, and ethical consequences. In the following we describe the process we underwent and factors that informed our decision.

The Process of Ethical Deliberation

Investigators originally planned to use the first statement, that is, to not disclose the study's primary focus on IDUs and HIV transmission risk to their sexual partners. Driven by our prior experience with vulnerable populations, we designed the study to fully protect IDUs while maximizing their comfort with participating in the research project. However, further reflection on the public health and ethical implications of this approach to disclosure led the research team to reconsider. We sought input from multiple sources for over a year prior to study implementation, including informal conversations with IDUs, NGOs (e.g., needle exchange programs), researchers, ethicists, friends and family members of IDUs, and others

As the discussion unfolded, substantial concern focused on the asymmetrical knowledge between investigators and study participants. The researchers would know more about the risk of continued unprotected intercourse with the index-participant (e.g., discordant HIV status, rare or no use of condoms) than the partner. Proponents of partner notification advocate a defined legal limit on confidentiality of the physician-patient relationship, the value of providing life-saving treatment for infected partners, and overall benefit to the public's health. Challengers to partner notification question whether mandatory partner notification laws for HIV will have an overall negative impact on both medical care and public health due to the loss of patients' ability to trust their physicians with sensitive information. [61-64] Investigators must consider many factors when developing a research study that figure into designing the process of informed consent. The key factors that arose in our research deliberations included protection of vulnerable populations, confidentiality waivers, validity of scientific results, differential risk for partners based on informed consent, and research logistics, all of which are discussed below.

Protection of Vulnerable Populations

In Estonia, research ethics committees consider IDUs among the vulnerable potential study participants. Known IDUs are marginalized for several reasons, including: a) their (illegal) behavior, b) most are members of an ethnic minority (about 80% are Russian-speaking), c) many have served time in jail or prison, d) many are poor, and e) about half are known to be HIV positive [34]

Sex partners of IDUs, who are not themselves IDUs, typically share only two of these vulnerabilities: Russian-speaking minority and low socioeconomic status. Another important factor to consider related to vulnerability is that divorce in Estonia is not uncommon. [65] As partners learn of the index-participant's HIV/IDU status, they may consider the option of divorce. Divorce is a socially acceptable option, and viable since sufficient housing exists in Estonia. By contrast, leaving a sexual partner is not a viable cultural or economic option in many other countries.

Confidentiality Waiver

For the IDU participants, requiring disclosure about drug use may result in feelings of vulnerability or frustration if they want to participate but do not want disclosure. Ensuring the confidentiality of study participants is a paramount human subjects projection. However, other equally pressing human subjects protections – informing partners of why they are being recruited into the study, must be balanced against protecting the confidentiality of the IDU. Accordingly, research participants commonly provide limited waivers of confidentiality for many studies. These waivers may be for protected information to be provided to researchers (e.g., medical records) or disclosed among research participants (e.g., research on couples or family counseling). In these situations, participants typically provide waivers of confidentiality that are limited to requirements for good scientific research. Indeed, in our study the index-participant willingly disclosed their sexual relationship with the referred partner(s).

Is there an ethical difference between requiring information disclosure about a person's drug use or HIV infection compared to disclosing information about a person's sexual relationship in order to participate in the study? Disclosure of IDU in Estonia to a partner could present significant harms to the IDU. In Estonia, not only is IDU illegal, but it is also illegal for a person who knows he/she is HIV-positive to transmit HIV to a person who was not aware of this risk, and criminal convictions have occurred. [66]

Conversely, waiving limited confidentiality is commonly required in research. If the requirements for research participation are unacceptable, then individuals can choose not to participate. In Estonia, the same HIV testing is free and available as is the counseling provided to study participants, which is comparable to that provided in the wider community. Thus, there is no overly compelling benefit for individuals to participate. The study's modest compensation was attractive to community members but not excessive such that a typical low-income IDU would unduly agree to unacceptable conditions. (This assessment was validated by the fact that some IDUs did not participate and others did not refer partners.)

Validity of Research Findings

For a research study to be ethical it must be scientifically valid. No risk is warranted if the study cannot be informative. From an epidemiologic perspective, the greatest risk to generalizability regarding informed recruitment is selection bias. If index-participants know that their sex partner(s) will be told about their drug use, then only partners who already know about this risk behavior will likely be referred for participation. Limitations on study participation driven by the consent process would, in turn, limit the effectiveness of resultant interventions to those with characteristics similar to the group studied. While our study was cross-sectional, the same issue applies to longitudinal studies designed to estimate HIV transmission rates more broadly.

Based on interviews with IDUs, we hypothesize that drug users who do not want their partners to know about their drug use are unlikely to recruit them for a study on HIV risk and transmission. It is not only the researchers who the IDU must trust, but also the other individuals the partner may encounter at the research location. Thus, the primary disadvantage of informed recruitment – selection bias – may likely be introduced by study participants, and that additional selection bias introduced by an in-depth informed consent is minimal. While empirical research is needed to formally test this hypothesis, it is supported by other researchers' experience working with IDUs (D. Des Jarlais, personal communication, 2008).

Risk

Arguments against and in favor of disclosure of the index participant's IDU status can be made depending on different interpretations of risk of HIV transmission to the public's health. A key argument against disclosure is that research participation does not increase the risk of HIV transmission/infection by participants. Even if the sex partners were not informed of the IDU status of the index participants during the recruitment process, all participants received state-of-the art education about behavioral risk factors and harm reduction techniques, which should minimize their risk of transmission. Another argument is that providing additional risk information does not necessarily enhance the study participant's protection. Empirical research shows that these educational and behavioral risk reduction interventions minimally affect sexual behavior. [51-54] That is, many partners continue to undertake risky sexual relations with IDUs after being informed. Accordingly, if the study provides limited risk reduction benefit, why introduce the potential for selection bias that would jeopardize the results? Such empirical evidence may lead some to conclude that disclosure is not necessary. For these and other reasons, nondisclosure of the index participant's IDU status is still commonly practiced in the US and abroad.

Nevertheless, there are substantial arguments in favor of disclosure. First, informing partners, who are study participants, enables them to become aware of the urgency with which to undertake precautions. It is imperative to ensure that potential study participants make informed choices, even if how they choose to act thereafter may not be healthful. Thus, doing so minimizes harms to study participants. Another justification for disclosure may be drawn from interviews with the IDUs in Estonia. Many partners leave the relationship upon learning about drug use. Because HIV is substantially more prevalent in the IDU community, leaving a relationship with an IDU does reduce a partner's risk of HIV infection. Thus, from the partners' perspective, this is an effective method of risk reduction that they can take and many are willing to take.

Moreover, nondisclosure may pose the potential for emotional harm to partners and the broader community if it became known that researchers knew some HIV-negative participants were having unprotected sex with HIV-positive individuals. Partners may feel betrayed and the community may feel manipulated. Such feelings could easily transform into distrust and reluctance to participate in research in the future, as has occurred among African Americans in the US. [67,68]

Logistics

Paper forms were used for most data collection, causing a delay in data entry. While the index-participant's IDU status was known, most partners were enrolled before information about HIV status was available. Nonetheless, studies of HIV transmission that include diverse risk groups and follow the transmission risk to multiple sex partners may benefit from rapid tests and real time data processing to provide appropriate information about disease risk to recruited individuals. Without these technologies, general statements related to risk of HIV can be made to all participants.

Resolution

For our study, we could not add IDU information to the partner's informed consent form because of the legal concerns about revealing HIV status. Thus, our pragmatic compromise was to tell all index participants to only recruit sex partners who knew about their drug use.