Retention strategies and predictors of attrition in an urban pediatric asthma study

The Urban Environment and Childhood Asthma (URECA) study is a multicenter prospective birth cohort study designed to examine factors related to the development of childhood asthma and allergies in an inner-city population. The retention of these participants has been challenging due to high mobility, inconsistent phone service, custody issues, and stressful life situations.

In this article, we describe the specific retention challenges we encountered during the first 2 years of follow-up in URECA and the strategies we utilized to address them. We also examine how selected maternal characteristics and other factors are related to retention and missed study visits.

Strategies implemented to engage participants included: collecting updated and alternative contact information, after-hours phone calls to participants, culturally competent staff, flexible study event scheduling, clinic visit transportation, quarterly newsletters, retention events, drop-in home visits, and cell phone reimbursements. An internally developed web-based data management system enabled close monitoring by site teams and the coordinating center. The rate of deactivations was calculated using survival analysis. Characteristics of active and deactivated participants were compared using the chi-squared test with a Cochran-Mantel - Haenszel adjustment for study site. The proportion of missed visits of the total expected in the first 2 years was calculated and compared by family characteristics using an ANOVA model or a trend test controlling for study site. All analyses were performed using SAS version 9.1 (Cary, NC).

The 2-year retention rate was 89%. Participation in the first study event predicted subsequent engagement in study activities. Mothers who did not complete the first visit were more likely to miss future events (46.1% vs. 8.9%, p<0.0001) and to be deactivated (38.5% vs. 4.5%, p<0.0001). Mothers under 18 years of age were more likely to leave the study compared to older mothers (22.7% vs. 10.1%, p = 0.02). Also, mothers who were married missed fewer events than those not married (8.8% vs. 15.6%, p = 0.01). In addition, deactivations were more common when the child had entered daycare by 3 months of age (10.9% vs. 3.6%, p = 0.05).

The URECA population is predominantly minority, thus our findings might not be generalizable to other populations. Furthermore, we may not be able to observe the effects that might exist in a more diverse population. For example, 86% of the mothers are unmarried, making it difficult to reliably examine the effect of marital status.

In research, successfully engaging and retaining participants is essential for achieving the study objectives. Identifying factors related to missed visits and deactivations are the initial step in recognizing the potential at-risk participants and can enable the design of targeted strategies to retain participants.