Inclusion criteria included couples that: 1) were at least 18 years old, 2) were HIV-discordant, where the female partner was HIV-negative and the male partner was HIV-positive, 3) were interested in conceiving, and 4) had visited a fertility clinic in Ontario to discuss sperm washing as a method to reduce horizontal HIV transmission. Recruitment was conducted through one of the only two clinics that offered sperm washing at the time in Ontario (both private clinics) as well as HIV care providers (from both academic and community practices) in Ontario. The second clinic offering sperm washing was also a study site, but no enrolment occurred through this site. Eligible HIV-discordant couples were initially contacted by their fertility clinic and/or HIV care provider. Participants expressing interest were subsequently contacted by a research team member who provided additional study information, determined eligibility, obtained informed consent and conducted the interview.

Retrospective chart reviews were conducted on health care records maintained by the participants’ HIV physicians and the participating fertility clinic, when available. Data collected via chart review were related to four areas for each couple: 1) demographics, 2) HIV history, 3) fertility history, and 4) other medical history. Data not found through chart reviews was collected through the interviews or a follow-up phone call.

Qualitative data was collected through in-depth semi-structured interviews with each couple, either in person or over the phone. In both cases, members of the couple were interviewed together. Questions were asked following a semi-structured interview guide, addressing the following areas: 1) desire to have children, 2) worries concerning conception, 3) feelings regarding fertility, 4) impact on lifestyle, 5) support networks, 6) physician counselling on pregnancy planning received, 7) sources of information on pregnancy planning in the context of HIV, 8) opinion on the current available support, and 9) satisfaction with fertility planning experience. Within each domain, participants were asked for recommendations of ways in which their experiences could have been improved. The interviews lasted between ninety minutes and two hours and were audio-recorded.

Discordant couples seeking fertility treatments make up a small and moreover, hidden, population due to stigma associated both with HIV status and the need for fertility treatments. In order to maximize recruitment, we adopted a two-pronged approach in which potential participants were recruited from: 1) one of the only two fertility centres in Ontario known to provide specific services to people living with HIV, and 2) from three large urban HIV clinics.

In addition to offering to conduct face-to-face interviews with potential study participants, we offered telephone interviews. This allowed for the inclusion of potential participants who could not participate in face-to-face interviews due to travel barriers.

In an attempt to address communication barriers between ourselves and clinic staff, in addition to regular phone and email contact, we introduced a team member located in the vicinity of one of the fertility clinics, whose responsibility it was to give a presentation, meet with staff and answer questions. This new member also served to conduct interviews with couples that lived closer to the fertility clinic.

In order to overcome the concerns of distance and the potential problems associated with scheduling interviews with couples, we ensured that all interview schedules were very flexible for couples. Interviewers offered to schedule interviews during evenings, weekends and holidays. In addition, interviews could be held in locations both convenient and comfortable for the couples, such as at home, or at clinics.

A key challenge in conducting research on HIV-discordant couples considering fertility treatment is that much of this population is hidden to researchers and health care providers. There are no formal databases of HIV-discordant couples wishing to conceive via fertility treatment, and the number of known couples is small. As the use of ART as an option for HIV-discordant couples has been limited, few health care providers ask about fertility desires [4,8]. Many couples do not disclose their HIV status due to stigma, making it difficult to identify those who have tried to access ART options. As a result, much of this population is not readily accessible through standard recruitment channels, and traditional random sampling methods are not appropriate.

Given the absence of a sampling frame for the HIV population in Canada or Ontario, and the sensitive nature of the study, non-probability or purposive sampling was appropriate and gatekeeper referral was used to make initial contact. For confidentiality reasons we—as research personnel—could not participate in the recruitment of potential participants at the fertility or HIV clinics. This created several challenges: 1) communication barriers between ourselves (the research personnel) and clinic staff delayed recruitment, 2) the fertility clinic did not have staff with salaried time protected for research, requiring them to assist with the study when they had additional time between clinical duties, and 3) clinical staff were put in a position where they had to balance the privacy of their clients with research interests. Indeed, the gatekeepers involved in our study often expressed concerns regarding the privacy of patients.

Access to research participants can be prevented or significantly delayed if gatekeepers are not adequately involved in the research process. Therefore, a crucial aspect of successful recruitment is the development of a strong relationship between researchers and gatekeepers [9]. Researchers often have a vested interest in the study well before recruitment staff is involved: they conduct literature reviews, formulate hypotheses, develop the study design, and often apply for funding opportunities. Thomas [9] reflected on his team’s experiences between research staff and gatekeepers (study site therapists), identifying the importance of respecting the gatekeepers’ unique cultures, recognizing their concerns, allowing and encouraging them to collaborate in study design to increase their sense of ownership, and ensuring they understand how they and their clients will benefit from the research. Witte et al. [10] also noted the benefits of taking the initiative to encourage closer relationships between researchers and gatekeepers.

In order to access as many potential participants as possible, we utilized a two-pronged approach to recruitment, recruiting through both the primary HIV physician in addition to fertility clinic staff. While fertility clinics do not usually remain in regular contact with clients after treatment, they have a greater number of eligible participant couples than the HIV clinics since every HIV-discordant couple seeking ART must contact them at some point to receive services. Furthermore, our research found that not all HIV physicians ask their patients about fertility desires. Hence, we anticipated that the majority of our recruits would come to us from fertility clinics, and the HIV clinics were involved only to cast as wide a net as possible. A surprising finding of our study was that it was actually the HIV clinics that made initial contact more often. This may be due to factors including more frequent medical visits with eligible couples, an active interest in academic research, a greater capacity to allocate resources to the project, or, more likely, a combination of these factors.

Before the introduction of an additional team member to act as a liaison and improve communication between the study and the research staffs, there was no recruitment via the fertility clinic. The addition of this liaison gave us someone who was very familiar with the study protocol to meet with staff members at the clinic, explain the purpose of the study, answer questions and be a contact person for them. This made it easier for staff to voice their concerns and increased the rate at which potential participants were identified and contacted. On the basis of discussions with fertility clinic staff members upon completion of the study, we conclude that more could have been done to improve the crucial gatekeeper-researcher relationship. Strategies for doing this are addressed in the Discussion.

Once potential participants (couples who expressed interest in participating in the study) were identified by the HIV care providers and/or fertility clinic, they were contacted by a member of the research team to receive more information. This aspect of recruitment represented a barrier for many potential participants due to the stigma associated both with living with HIV and pursuing fertility treatments (rendering them “doubly-vulnerable”). Early on in the study, we learned that some couples did not disclose their HIV status to other family members or friends. Furthermore, they did not share with others that they were undergoing fertility treatments for fear that this would reveal their HIV status. When family members or friends had asked them why they were having fertility treatments, they felt forced to either answer the question vaguely or create false reasons to prevent inadvertent HIV disclosure. The desire for discretion concerning fertility treatments was especially prominent in couples where the perceived suspicion of having HIV was already present, for example among men who were known by friends and family to be haemophiliac. Thus, we tried to be as sensitive as we could in our approach to making contact with potential participants. This may have been an undetected but significant barrier to recruitment, since communication with potential participants in the course of recruitment was vague and recipients may have had difficulty discerning the purpose of the communication.

The stigma of living with HIV has long been considered a great obstacle in treatment and prevention success, impacting testing and treatment uptake, and access to services [11,12]. Fertility treatments are clearly affected by this stigma, as evidenced by the limited services available to people living with HIV [13-16]. Furthermore, the limited available data describing the experience of those undergoing such treatments suggest that stigma is a major factor for these couples, both in terms of accessing options for having children, as well as the process of undergoing treatments without identifying to others why they are seeking fertility services [14].

Researchers contacting potential participants took every reasonable precaution to ensure privacy and disguised the reasons for calling from everyone except the potential participants. Our careful approach was successful in maintaining confidentiality and at no point during recruitment was any information passed inadvertently to the wrong party. Despite this, as noted above, the necessary vagueness required in calling potential participants at home may have acted as an additional barrier to recruitment. Some strategies for circumventing this are outlined in the Discussion.

Another challenge in this study was related to the disparate geographic locations of study site staff, researchers, participants and subsequent long travel distances required for the interviews. This issue is particularly acute for a small patient population such as ours: During study enrolment, only two fertility centres in Canada offered fertility services to HIV-discordant couples. As both were located in southern Ontario, those interested in sperm washing would have to travel to one of these centres to access this service. In addition, since access to HIV care tended to be concentrated in larger urban settings, many potential participants from more rural areas travelled to urban centres for HIV care. Our solution to this problem—enhanced flexibility on the part of interviewers and the use of telephone interviews—introduced its own set of challenges. Advantages to telephone interviewing include lower costs, greater access to people living far away, less space requirements, ability to take notes unobtrusively, greater anonymity, ability for participants to be comfortable in their environment, potential increased safety for interviewers and participants, and potential for better facilitation of disclosure of more sensitive information [17,18]. Disadvantages to telephone interviews include lack of telephone access for some participants, lack of visual cues, potential distractions, and potentially shorter interview times. Novick [18] suggests that there is bias against telephone interviews among researchers, as even researchers who found phone interviews to be effective in their studies suggested that the telephone was a good second choice to face-to-face interviews.

We found that flexibility had a positive effect on recruitment and retention, as we had great success in recruiting participants who would otherwise have had difficulties meeting with us. Our experience with telephone interviews (three couples) was very positive as the couples who chose this mode felt it was much more convenient for them. They thought it was easier than meeting in person and were often interviewed after they had put their children to bed. One drawback we encountered was technical difficulties with recording the phone interviews, which may have affected the quality of data collection; however, this problem improved as our experience with telephone interviews increased.

Fertility treatments such as sperm washing and IUI are an inherently couple-oriented process. The key question that arises in interviewing couples for a research study is whether it is better to interview each member of the couple alone, interview them together, or interview them individually and then together as a couple. The benefits of interviewing couples (or other key stakeholders) together include retrieving more comprehensive data, allowing the couple to fill in one another’s gaps, and observation of the dynamics of the relationship [19]. The couples interviewed together can use each other to recall forgotten stories and spark spontaneous discussion, allowing more information to be recalled and explored [20]. It can allow for the identification of discrepancies between each individual’s memory of events, and determine their different perceptions of their stories [19,20]. A joint interview may give a more reliable and comprehensive picture than either member’s individual account since a bias in one direction may be balanced by another version from the partner [21].

While disagreements between a couple’s accounts of a situation or philosophy in a joint interview may allow for a better understanding of their ideas by allowing the interviewer to explore these differences in depth with them, there can be significant disadvantages to interviewing couples together. It may bring to the surface tensions in the couple that may persist well after the interview concludes [19,20]. Participants provide consent to participate in an interview, but neither they, nor the interviewer, can anticipate what sort of information, “forgotten and well hidden resentments or past half-truths”, may emerge in the process, nor their impact on the relationship [22]. This should be considered for sensitive topics and if couples are to be interviewed together, the risk of creating tensions should be addressed in the informed consent process [22]. There is the chance that an individual may withhold details or feel uncomfortable in the presence of the other member. There is also the phenomenon of trying to present a “united front” [20], wherein a story is adjusted to avoid giving evidence of relationship discord, or to prevent giving what they feel is a negative perception of their relationship. There is also a concern that if one person dominates the interview, the experience may not have the richness of two accounts of the same events [19]. Some authors have found that individual interviews may facilitate more discussion on sensitive matters [21], and there is the consideration that in sensitive situations couples may refuse to participate in the study when they are asked to be interviewed together.

There is no consensus in the literature of what is the best way to interview couples (or household members): together, alone or both. When addressing the advantages and disadvantages of the various approaches to interviewing, authors should think carefully about their research area and identify which method is appropriate for their study while addressing logistical concerns [19-23]. In the context of our study, we felt it was important to be consistent in interviewing all couples in the same manner and, due to the highly sensitive nature of the study, wanted to reduce the risk of exposing “secrets” to the other partner if interviews were conducted separately or separately and then together. Furthermore, since the goal of the study was to document the joint experience of couples, the decision was taken to interview couples together. Even in this situation, however, the possibility of secrets being exposed during the interview process still existed. Thus, we made it clear from the outset (both on the information sheet and during the informed consent process) that if there were topics which they did not want raised, or were very sensitive between them as a couple, they should consider very carefully whether the study was a good idea for them.

Couples were interviewed together with the understanding that they would be able to end the interview at any time, refuse to answer particular questions, or inform us of feelings of discomfort with a line of questioning or a topic. Each couple—including one that had separated—was comfortable being interviewed together and the interviews proceeded very smoothly. While this worked very well for our study, it is possible that by restricting ourselves to only interviewing couples together, we may have lost access to some information that individuals did not want revealed to their partner, or prevented potential participants from joining because of this criterion. Most of the couples in our study reported a very positive experience with their sperm washing fertility services, but it is possible that couples with negative experiences or significant intra-couple conflict chose to not participate, resulting in sampling bias. For a better understanding of these couples’ experiences, future studies require a broader sample population.