Intervention completion rates among African Americans in a randomized effectiveness trial for diet and physical activity changes.

Halbert CH1, Bellamy S2, Briggs V3, Bowman M4, Delmoor E5, Johnson JC6, Kumanyika S2, Melvin C7, Purnell J8, Rogers R9, Weathers B10.

Author information

Abstract

BACKGROUND:

The intervention completion rate is an important metric in behavioral and intervention research; trials with limited intervention completion rates may have reduced internal validity. We examined intervention completion rates among 530 African Americans who had been randomized to an integrated (INT) or disease-specific (DSE) risk education protocol as part of a comparative effectiveness trial from September 2009 to August 2012.

METHODS:

The interventions were developed by an academic-community partnership using community-based participatory research. Intervention completion rates were determined based on attendance at all four intervention sessions. Intervention completers were participants who completed all four sessions and noncompleters were those who did not complete any session or only completed one to three sessions following randomization.

RESULTS:

Seventy-three percent of participants were intervention completers and 27% were noncompleters. There were no differences in intervention completion based on randomization to INT (72%) or DSE (75%), sociodemographic factors, or body mass index (BMI) in the total sample. Different factors were associated significantly with intervention completion within study groups. Among participants randomized to INT, the odds of intervention completion were greater with higher levels of intrinsic motivation, less exposure to information about diet and cardiovascular disease, and greater BMI. Among participants randomized to DSE, the odds of completing the intervention were associated significantly with older age and greater dietary self-efficacy.

CONCLUSIONS:

Many African Americans are likely to complete risk education interventions.

IMPACT:

Psychologic characteristics should be considered when determining intervention completion rates following randomization in behavioral and intervention trials.