Factors associated with probability of personal digital assistant-based dietary self-monitoring in those with type 2 diabetes

Abstract

Knowledge of factors associated with the use of technology could inform the design of technology-based behavioral interventions. This study examined modifiable and nonmodifiable factors associated with technology-based self-monitoring. 123 participants with type 2 diabetes self-monitored diet using a personal digital assistant in a 6-month behavioral intervention. Multinomial logistic regression was used to examine probability of nonadherent and suboptimally adherent behavior relative to adherent behavior. Sociodemographic characteristics were not associated with probability of self-monitoring. Probability of adherence generally was greater in the weeks preceding no group session, and lower in the weeks following no group session or following skipped sessions. Non-modifiable factors suggested by the literature to be associated with poorer access to technology (lower income, older age, minority race, and lower education) were not associated with probability of self-monitoring in this population.