The final sample included 32 cases and 65 controls. Participants were about 31 years old on average (full sample: M = 30.9, SD = 4.2; cases: M = 31.3, SD = 4.2; controls: M = 30.1, SD = 4.2) and had approximately two children (full sample: M = 1.8, SD = 0.9; cases: M = 1.6, SD = 0.8; controls M = 1.9, SD = 1.0). Most participants were White (61%), with smaller numbers of Indigenous (18%) and Asian (11%) participants represented. At the time of the study, most participants indicated they were on “some type of leave (e.g., maternity, disability, sick leave, etc.)” (63%) from their job. Smaller numbers of participants reported being employed full-time (14%), part-time (7%), or unemployed (12%). Participating children had an average age of about 8 months (full sample: M = 7.6, SD = 4.3; cases: M = 8.2, SD = 4.6; controls: M = 7.2, SD = 4.1) and included roughly equal numbers of males and females in the case and control groups (cases: 40% females; controls: 51% females; Please see Table 1 for a full summary of the demographic findings). A series of independent samples t-tests and χ2 tests revealed cases and controls did not significantly differ on any of the demographic variables apart from employment status. As such, we controlled for the participant’s work status in each of the substantive analyses.

Cases largely rated their child’s AD as mild to moderate in severity as only one participant fell in the severe range (PO-SCORAD > 57; M = 25.4, SD = 13.8). On the PBQ, mothers with a child with AD were found to have mean scores of 5.8 (SD = 3.95), 2.3 (SD = 1.8), 2.2 (SD = 2.4), and 0 (SD = 0) on the impaired bonding, anxiety with caregiving, pathological rejection/anger, and risk of abuse scales, respectively. By comparison, mothers in the control group reported higher scores on each of these scales (impaired bonding: M = 7.1, SD = 4.4; anxiety with caregiving: M = 3.5, SD = 2.3; rejection/anger: M = 3.6, SD = 3.3; risk of abuse: M = 0.03, SD = 0.2). Across both groups, the majority of mothers had no scores on any of the four PBQ subscales that exceeded the cut-offs indicative of disordered bonding (cases = 78%; controls = 86%).

Results from a series of multiple linear regression analyses revealed significant differences between cases and controls on two of the four PBQ subscales. Relative to controls, mothers with a child with AD reported significantly lower scores on the scales related to anxiety about caregiving (b = − 1.47, 95% CI = − 2.49, − 0.45, p < 0.01) and pathological rejection/anger (b = − 1.74, 95% CI = − 3.17, − 0.31, p = 0.02). Cases did not, however, significantly differ from controls on the subscales related to impaired bonding or risk of abuse (Please see Table 2 for a full summary). While mothers with a child with AD reported higher quality bonds on some subscales, the severity of a child’s AD was not found to significantly predict bonding quality on any of the PBQ subscales (all p > 0.05).

The final qualitative sample included ten mothers with infants with AD. Most participants were in their late 20 s and early 30 s (M = 29.2, SD = 5.0) and had one child in the home (M = 1.4, SD = 0.5). Similar to the quantitative findings, 60% of participants were White, 20% were Indigenous, and 10% were Asian, with the remainder reporting their race as “other”. The vast majority of participants were married (90%) at the time of the study and reported living with just one other adult (70%). At the time of the interview, most participants were on “leave” from their job (70%), with smaller numbers working full-time or part-time (20%). Participants tended to have a post-secondary education (70%) and half reported an annual household income of $100,000 Canadian dollars or more. Their infants with AD were mostly male (80%) and ranged in age from 1 to 18 months (M = 8.5 months, SD = 4.9). The majority of these children had PO-SCORAD scores consistent with mild (30%) to moderate AD (60%).

Through the qualitative interviews, one overarching theme related to maternal-infant bonding was identified: AD can have a positive influence on bonding.

Contrary to expectations, mothers did not describe AD as having a negative impact on their bond. Surprisingly, participants either indicated AD had no impact or they emphasized the perceived positive impacts on bonding. Rather than describing the additional caregiving needs of their infant as solely burdensome, many noted how they provided opportunities for nurturance that would not have been available otherwise. In particular, some mothers described how the need to regularly bath and moisturize their baby promoted a greater degree of physical and emotional closeness: “I think eczema makes us closer in some ways, making sure that I am putting his cream on, and we always have our daily massages with his lotion and stuff like that, just to make sure that he is being taken care of.”—Participant 1, male child, 1 month old.

Emotional closeness was not only predicated on physical touch, however, as participants also described how the discomfort caused by their infant’s AD evoked feelings of tenderness and an empathic desire to alleviate their child’s suffering. For some mothers, the alleviation of their infant’s suffering brought with it a satisfying sense of purpose as their infant was dependent on them for relief: “I am taking care of him more, it is a better relationship because I always have to look after him and I always have to think about him, in a good way. ‘Cause I know that he needs me, he needs me to take care of him more.”—Participant 3, male child, 7 months.

While a number of mothers emphasized the positive contributions of AD on bonding, other participants explained that their child’s condition had little impact on their emotional or physical connection: “I would say that we have a really good relationship, regardless of the eczema.”—Participant 4, female child, 8 months. Interestingly, none of the mothers described AD as having a negative impact on their relationship with their baby. That being said, one mother did indicate that she was hesitant to caress her infant’s face out of fear of irritating his inflamed skin. However, this limitation was not perceived as having a negative effect on their bond (See Table 3 for a table of qualitative codes and representative quotations).