EMASENS
Developing a just-in-time adaptive intervention to prevent smoking lapses
The aim of the EMASENS project is to develop a smartphone-based stop smoking tool that provides personalised support to smokers in real time, when they most need it (a ‘just-in-time adaptive intervention’, or JITAI). To provide this kind of support, we first need to understand when to send support and what type of support to send.
Stage 1: Predicting smoking lapses
In the first stage of EMASENS, we aimed to gain a better understanding of when and why people are at risk of lapsing (i.e., smoking cigarettes after the quit date). To build a lapse prediction algorithm, we triangulated findings from two studies that took slightly different approaches.
Study 1: Unprompted data from a smoking cessation app
In the first study, we used observational data from a popular smoking cessation app to develop supervised machine learning algorithms that distinguish lapse from non-lapse reports. We used data from app users with 20 or more unprompted entries, covering craving severity, mood, activity, social context and lapse incidence. A series of group-level algorithms (e.g., Random Forest, XGBoost) were trained and tested, and their ability to classify lapses for out-of-sample observations and out-of-sample individuals was evaluated. Individual-level and hybrid algorithms were then trained and tested.
Using unprompted app data appeared feasible for constructing a high-performing group-level lapse classification algorithm, but performance was variable when applied to unseen individuals. Algorithms trained on each individual’s own data, and hybrid algorithms trained on the group data plus a proportion of each individual’s data, performed better but could only be built for a minority of participants. It was therefore important to repeat the analyses with data from a prompted study design (using both signal- and event-contingent Ecological Momentary Assessments) and to triangulate the results.
Study 2: Prompted data and smartwatch sensors
In the second study, we used self-reported data from brief, hourly surveys sent directly to people’s smartphones, together with passively collected smartwatch data, to train and test group-level, individual-level and hybrid lapse prediction algorithms. Individual-level and hybrid algorithms performed better than group-level algorithms, particularly when the passively collected sensor data were included. However, multiple success criteria (e.g., acceptability, scalability, technical feasibility) need to be carefully balanced against algorithm performance in the development and implementation of a JITAI.
The key output of the first stage was the selection of a best-performing algorithm to take forward as the basis of a smoking cessation JITAI.
Stage 2: Developing and evaluating the intervention
The second stage of EMASENS, which forms part of Corinna Leppin’s PhD project, is nearing completion. It focuses on developing the intervention messages using theory and behaviour change tools (e.g., the Behaviour Change Wheel) combined with extensive user testing, building a bank of messages tailored to each identified lapse risk factor, and evaluating the effectiveness of the JITAI.
Publications from Corinna Leppin’s PhD
Health Psychology Review (2024)
PLOS Digital Health (2025)