Compliance with an EMA monitoring protocol and its relationship with participant and smoking characteristics

Abstract

Introduction: Arguably, the greatest advantage of ecological momentary assessment (EMA) studies is that data are collected repeatedly in real-time and real-world situations, which reduces recall and situational biases and thus improves the accuracy and validity of the data collected. However, the validity of EMA data is contingent upon compliance rates. If participant characteristics are related to missing data, analyses should control for these factors, or they should be targeted in EMA training sessions. This study evaluates the impact of demographic and smoking-related participant characteristics on compliance to an EMA smoking study protocol.

Methods: Prequit-day data were taken from the control arm of an ongoing randomized controlled trial of a smoking-cessation program. After training, 119 participants were asked to carry a mobile device with them at all times for ~6 days and to log every cigarette they smoked in addition to completing randomly scheduled assessments. Different types of compliance were assessed: the percentage of completed random prompts (signal-contingent compliance), the percentage of logged cigarettes per day compared to a timeline follow-back measure, and the correlation between logged cigarettes and a carbon monoxide assessment 2 hr later (both event-contingent compliance).

Results: Overall compliance rates were 78.48% for event-contingent and 72.17% for signal-contingent compliance. None of the demographic or smoking-related participant characteristics predicted signal-contingent compliance; however, female participants showed higher event-contingent compliance than male participants, and Caucasian participants showed higher event-contingent compliance than non-Caucasian participants.

Conclusions: Compliance did not depend on smoking-related characteristics. EMA is a valid method for assessing smoking behavior in real-time and real-world settings.