Modeling barriers to participation and retention in a longitudinal parenting intervention

A major threat to both the efficacy of interventions and the integrity of their evaluations is increasingly high non-response at recruitment and subsequent assessment. Parenting intervention research in high-risk maternal-child populations may be subject to unique threats to recruitment and retention, including custodial issues, persistent illness of the child or mother, and high mobility.

Legacy for ChildrenTM is a multi-site, randomized controlled trial of a parenting intervention for low-income mothers, with annual assessments and supplemental contact for retention prenatally to age five. This analysis sought to identify factors associated with initial assessment completion and subsequent retention.

Logistic regression modeled consent and assessment completion at baseline and retention through a year 2 assessment within each site as a function of demographic factors. Results indicated that multilingual women were significantly less likely to consent and ultimately complete the baseline assessment (p<0.05) and were less likely to complete a year 2 assessment, even among those completing the baseline assessment. Greater affective and instrumental social support (only available among initial participants) also predicted retention of initial participants through the two-year assessment. Employment, marital status, educational attainment, and health (overall and mental) were unrelated to retention.

These findings suggest additional emphasis on understanding and addressing the barriers to participation for multilingual and socially isolated, low-income mothers may further improve the effectiveness of the Legacy intervention and may benefit evaluations of maternal-child health interventions more generally.