Ecological Momentary Assessment of Smoking Triggers, Needs, and Cigarette Forgoing (spark Forgo)

This study uses a within-person, phased ecological momentary assessment (EMA) and ecological momentary intervention (EMI) design to characterize momentary factors associated with cigarette smoking and smoking forgoing and to evaluate non-adaptive and adaptive motivational messaging delivered via the *spark* smartphone platform.

During all study phases, participants will complete four randomly timed EMA assessments per day during predefined waking hours and a daily evening survey. EMA assessments will assess momentary contextual factors, including current activity or routine, physical environment, social context, affective and cognitive states, smoking urges or craving, cigarette use since the previous assessment, and instances in which a smoking urge occurred without smoking (i.e., smoking forgoing). Evening surveys will assess daily smoking-related attitudes, cognitions, and behaviors.

Participants may also initiate an event-contingent entry whenever they smoke a cigarette. Each self-initiated entry will record the smoking event. Following each cigarette log, the app will randomly assign, with equal probability, whether the participant completes additional questions about the smoking context and feeling states or only logs the cigarette. To minimize assessment burden, randomly timed EMA prompts will not be delivered until at least one hour after completion of a self-initiated cigarette survey.

The study includes three phases. During the EMA Observation Phase, participants will complete EMA assessments, evening surveys, and event-contingent cigarette logs; no intervention messages will be delivered. During the Non-Adaptive EMI Phase, participants will continue these assessments and receive motivational messages delivered at randomly selected times, without tailoring to participant behavior or context. During the Adaptive EMI Phase, participants will continue all assessments and receive motivational messages at times determined by predictive models developed using data from the first two phases. Message content will be adapted based on the estimated likelihood of successful smoking forgoing in response to momentary contexts and internal states.

Where technically feasible and permitted by participant consent, passive smartphone-based data may be collected, including geolocation-derived patterns, time-stamped device-activity indicators, and environmental-context variables such as local weather conditions. These data will be used to enhance characterization of contextual factors associated with smoking urges, smoking behavior, smoking forgoing, and responsiveness to motivational intervention delivery.