Characterizing high risk drop-out neighborhoodhelp households

Background: NeighborhoodHELP is a program that matches selected households in North Miami-Dade County with a student healthcare team. Along the first three years the program has experienced important attrition.

Objectives: To describe the frequency of attrition in the program and to explore households factors associated with it.

Design: Historical cohort using data obtained within a randomized trial assessing the impact of the program. The baseline assessment included 156 standardized questions pertaining household and individual health/wellness.

Setting: Low income areas in North Miami Dade County (Miami Gardens, Op Locka and Unincorporated Miami).

Participants: A total of 272 households were originally recruited and were randomized into Student (n=64) and Control groups (n=208).

Interventions: NA.

Main Outcome Measures: Household drop-out status and time from inclusion into the program to drop-out. Reasons for leaving the program were also explored.

Results: 120 (44%) households had left the program. Univariate analysis showed that intervention status, race/ethnicity, primary language of the households, time resided at the households and household ownership were all associated with leaving the program. Cox regression analysis combining these factors with employment status and household income showed that Spanish as primary language was the only predictor independently associated with leaving the program (HR 1.69; 95% CI 1.05 - 2.72; p = 0.03).

Conclusions: The results of the study suggest that Spanish speaking households was the only factor associated with an increased likelihood of leaving the program. As the program expands further analysis will be conducted to provide greater insight about its attrition.