Table 1 Overview of ethical challenges and design responses

Definition of study population

Pregnant women and women with a history of domestic conflict and violence are vulnerable groups

Antenatal care offers a potentially safe opportunity to intervene on violence; Interventions to mitigate family conflict and violence may have beneficial impact on pregnancy outcomes

Study advertisement

Disruption of family relationships because of family members’ perception of the intervention as being focused on DV or because of resistance to the types of social and behavioral change encouraged by the intervention, and risk of backlash from family and community

Study framed as testing a family health intervention to promote the health of younger and older women and children; Community advisory group comprising community elders and leaders (male and female) formed and community meetings held to enlist support for the study and introduce research team

Participant recruitment

Limited decision-making autonomy among young women; Loss of confidentiality due to screening for domestic conflict and violence; Loss of confidentiality during participant follow-up

Multistep recruitment process used beginning with DILs; Completion of eligibility screening in private room at health center; Ascertainment of acceptable modes of contacting participants for follow-up; Use of standardized scripts to respond to questions from family/community members and avoidance of discussions about study specifics with non-participants; Avoidance of conversations that may jeopardize confidentiality in public settings

Participant retention and follow-up

Loss of confidentiality

Use of participant tracking system with modes of contact approved by participant; Staff trained not to discuss sensitive study information with participant or others in public settings; Use of standardized scripts for participant follow-up

Safety Monitoring

Defining safety in a context of high prevalence of DV and in a population with prior history of family conflict and violence; Monitoring safety

Classification of safety-related events as “safety alerts” or “adverse events,” with alerts comprising incidents that are highly unlikely to be related to study participation and/or not considered to be severe and adverse events comprising incidents that are unexpected, temporally associated with and likely to be related to study participation and severe; Establishment of rapport with participants and creating a safe environment at the health center by offering information on DV-related resources at every visit, establishing an information desk at study health centers, and providing staff contact information; Presence of on-site counselor

Randomization

Confusion regarding or lack of understanding of randomization

Explanation of randomization in terms of a lottery; Opening of envelope with treatment assignment in front of dyad

Intervention implementation

Stress, discomfort or distress as a result of intervention participation; Loss of confidentiality in group education sessions

Establishing ground rules for group session participation including respect for each other’s views; Inclusion of activities to promote peer support and dialogue, such as co-counseling; Staff trained to handle difficult situations arising during group sessions; Counseling available on site and referral information offered; Child care and food offered; Participants and staff take a pledge of confidentiality regarding personal information shared in the group

Data collection

Stress, discomfort or distress caused by interview questions; Loss of confidentiality

Interviewers trained to conduct interviews with sensitivity and empathy; Use of statements such as, “some women can feel sad or upset by these questions—do remember that you can decide not to answer any question” repeated at strategic points of interview; Counseling available on site and referral information offered; All interviews conducted in closed room at health center; Data identified only through unique numeric identifiers