Authors' Reply: Peer Mentorship's Paradox in Dialysis: Neutral Overall but Lifesaving for Black Patients: Toward Equity-Driven Trial Designs

In their Letter to the Editor, Zhang et al.1 provide valuable insight into the results of the PEER-HD trial published in JASN.2 Their suggested reasons for the differences in outcome found between two sites are pertinent. Their recommendation to prospectively evaluate context-specific drivers of preventable hospitalization among patients with dialysis-dependent kidney failure and tailor content of mentor intervention to mentee contexts in any future attempts at testing peer-led interventions is thoughtful. We also recognize the challenges in identifying context-specific drivers of preventable hospitalizations.3 We acknowledge that this trial was unique in that it provided mentors and mentees with practical tools, such as a better understanding of estimated dry weight metrics and tools to understand reasons for their symptoms paired with communication training to engage dialysis staff to best use available resources.4 These skills are not covered by traditional educational materials provided to patients and should be universally developed in all individuals with kidney failure. We recommend caution in assigning too much importance to the findings of the as-treated analyses as the 20-minute threshold, though prespecified by the investigators, likely represents the occurrence of contact versus no contact between mentor/mentee pairs over 3 months of intervention rather than any dose dependence of the mentor intervention along a continuously distributed duration of telephone contact. The distribution of mentor/mentee duration of telephone contact is represented in Supplemental Figure 1 and shows a bimodal pattern: with either minimal contact or long duration of contact.1 Our observations of mentor/mentee engagement during the trial suggest that initial rapport and emotional support provided by mentors played a role in the duration of mentor/mentee engagement, which could have influenced the as-treated outcomes. We measured rating of relatability and performance of the mentors by their mentees using an established survey tool to gain a better understanding of the nature of their engagement (Supplemental Table 2).2 While effectiveness of peer mentorship was demonstrated in Black participants, we agree with Zhang et al.1 to interpret these findings with caution because of the risk of type 1 error and because the interaction effect between Black race and randomization to peer mentor intervention with respect to outcome was nonsignificant. We also agree that the potential reduction in hospitalization with peer mentorship was not mediated by improvements in adherence behaviors and suggested that it likely represents increased comfort with self-managing some symptoms at home rather than visiting the emergency department.