Participant attrition is a major influence on the effectiveness of evidence-based interventions. Assessing predictors of participant attrition and nurse and site characteristics associated with it could lay a foundation for increasing retention and engagement. We examined this issue in the national expansion of the Nurse-Family Partnership, an evidence-based program of prenatal and infancy home visiting for low-income, first-time mothers, their children, and families. Using a mixed methods approach, we examined participant, nurse, and site predictors of participant attrition and completed home visits. We used mixed multivariate regression models to identify participant, nurse, program, and site predictors of addressable attrition and completed home visits during pregnancy and the first year of the child's life for 10,367 participants at 66 implementation sites. We then conducted semi-structured interviews with nurse home visitors and supervisors at selected sites with the highest (N = 5 sites) and lowest (N = 6 sites) rates of participant addressable attrition and employed qualitative methods to synthesize themes that emerged in nurses' descriptions of the strategies they used to retain participants. Mothers who were younger, unmarried, African American, and visited by nurses who ceased employment had higher rates of attrition and fewer home visits. Hispanic mothers, those living with partners, and those employed at registration had lower rates of attrition. Those who were living with partners and employed had more home visits. Nurses in high retention sites adapted the program to their clients' needs, were less directive, and more collaborative with them. Increasing nurses' flexibility in adapting this structured, evidence-based program to families' needs may increase participant retention and completed home visits.