Dataset Information

Predicting the Users' Level of Engagement with a Smartphone Application for Smoking Cessation: Randomized Trial and Machine Learning Analysis.

ABSTRACT:

Introduction

Studies of the users' engagement with smoking cessation application (apps) can help understand how these apps are used by smokers, in order to improve their reach and efficacy.

Objective

The present study aimed at identifying the best predictors of the users' level of engagement with a smartphone app for smoking cessation and at examining the relationships between predictors and outcomes related to the users' level of engagement with the app.

Methods

A secondary analysis of data from a randomized trial testing the efficacy of the Stop-Tabac smartphone app was used. The experimental group used the "full" app and the control group used a "dressed down" app. The study included a baseline and 1-month and 6-month follow-up questionnaires. A total of 5,293 participants answered at least the baseline questionnaires; however, in the current study, only the 1,861 participants who answered at least the baseline and the 1-month follow-up questionnaire were included. Predictors were measured at baseline and after 1 month and outcomes after 6 months. Data were analyzed using machine learning algorithms.

Results

The best predictors of the outcomes were, in decreasing order of importance, intention to stop smoking, dependence level, perceived helpfulness of the app, having quit smoking after 1 month, self-reported usage of the app after 1 month, belonging to the experimental group (vs. control group), age, and years of smoking. Most of these predictors were also significantly associated with the participants' level of engagement with the app.

Conclusions

This information can be used to further target the app to specific groups of users, to develop strategies to enroll more smokers, and to better adapt the app's content to the users' needs.

SUBMITTER: Vera Cruz G

PROVIDER: S-EPMC10389794 | biostudies-literature | 2023

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Predicting the Users' Level of Engagement with a Smartphone Application for Smoking Cessation: Randomized Trial and Machine Learning Analysis.

Vera Cruz Germano G Khazaal Yasser Y Etter Jean-François JF

European addiction research 20230425 3

<h4>Introduction</h4>Studies of the users' engagement with smoking cessation application (apps) can help understand how these apps are used by smokers, in order to improve their reach and efficacy.<h4>Objective</h4>The present study aimed at identifying the best predictors of the users' level of engagement with a smartphone app for smoking cessation and at examining the relationships between predictors and outcomes related to the users' level of engagement with the app.<h4>Methods</h4>A secondar ...[more]

PMID: 37166304

Similar Datasets

Project description:AimsMost cigarette smokers want to quit smoking and more than half make an attempt every year, but less than 10% remain abstinent for at least 6 months. Evidence-based tobacco use treatment improves the likelihood of quitting, but more than two-thirds of individuals relapse when provided even the most robust treatments. Identifying for whom treatment is effective will improve the success of our treatments and perhaps identify strategies for improving current approaches.MethodsTwo cohorts (training: N = 90, validation: N = 71) of cigarette smokers enrolled in group cognitive-behavioral therapy (CBT). Generalized estimating equations were used to identify baseline predictors of outcome, as defined by breath carbon monoxide and urine cotinine. Significant measures were entered as candidate variables to predict quit status. The resulting decision trees were used to predict cessation outcomes in a validation cohort.ResultsIn the training cohort, the decision trees significantly improved on chance classification of smoking status following treatment and at 6-month follow-up. The first split of all decision trees, which was delay discounting, significantly improved on chance classification rates in both the training and validation cohort. Delay discounting emerged as the single best predictor of group CBT treatment response with an average baseline discount rate of ln(k) = -7.1, correctly predicting smoking status of 80% of participants at posttreatment and 81% of participants at follow-up.ConclusionsThis study provides a first step toward personalized care for smoking cessation though future work is needed to identify individuals that are likely to be successful in treatments beyond group CBT.ImplicationsThis study provides a first step toward personalized care for smoking cessation. Using a novel machine-learning approach, baseline measures of clinical and executive functioning are used to predict smoking cessation outcomes following group CBT. A decision point is recommended for the single best predictor of treatment outcomes, delay discounting, to inform future research or clinical practice in an effort to better allocate patients to treatments that are likely to work.

Project description:BackgroundFacebook can be a suitable platform for public health interventions. Facebook users can express their reaction to the given social media content in many ways using interaction buttons. The analysis of these interactions can be advantageous in increasing reach and engagement of public health interventions.ObjectiveThis research aimed at understanding how Facebook users' interactions correlate with organic reach and engagement regarding the same smoking cessation support contents.MethodsThe study population consisted of Facebook users who were reached by a public smoking cessation support page without advertising. We included 1025 nonpaid Facebook posts (N=1025) which used smoking cessation strategies based on a motivational interviewing counseling style. The following data were collected from the "Post Details": the number of people who saw the given nonpaid content (organic reach) which consisted of fan and nonfan reach according to previous "page like" activity; each rate of "engagement indicators" (such as the symbols of "like," "love," "haha," "wow," "sad," "angry"; or other interactions: "shares," "comments," "clicks"); and the rate of negative Facebook interactions (eg, "post hides" or "unlike of page"). Overall, these data were analyzed with the Spearman correlation method.ResultsSurprisingly, we found a significant negative correlation between organic reach and the "like" reaction (rs=-0.418; P<.001). The strongest significant positive correlations of organic reach were observed with the "haha" reaction (rs=0.396; P<.001), comments (rs=0.368; P<.001), and the "love" reaction (rs=0.264; P<.001). Furthermore, nonfan reach correlated positively with "shares" (rs=0.388; P<.001) and clicks (rs=0.135; P<.001), while fan reach correlated positively with the "haha" reaction (rs=0.457; P<.001), comments (rs=0.393; P<.001), and the "love" reaction (rs=0.310; P<.001). Contrary to expectations, the "like" reaction was sharply separated by significant negative correlations from "wow" (rs=-0.077; P=.013), "sad" (rs=-0.120; P<.001), "angry" reactions (rs=-0.136; P<.001), and comments (rs=-0.130; P<.001). Additionally, a high rate of negative Facebook interactions was significantly associated with "wow" (rs=0.076; P=.016) and "sad" reactions (rs=0.091; P=.003).ConclusionsThis study has shown that it is possible to hypothesize a disadvantage of the "like" reaction and advantages of other interactions (eg, the "haha" reaction or "comments") in content algorithmic ranking on Facebook. In addition, the correlational analysis revealed a need of a further categorization to fan-specific interactions (eg, "haha" or "love" reactions) and nonfan-specific interactions (eg, "shares" and "clicks"). Regarding the direction of the correlations, these findings suggest that some interactions (eg, negative Facebook interactions, "wow," "sad," and "angry" reactions) may decrease the engagement, while other interactions ("like," "love," "haha" reactions, "shares," and "clicks") may increase the engagement during Facebook-based smoking cessation interventions. This hypothesis-generating research offers an important insight into the relationship between organic reach, engagement, and Facebook users' interactions for public health professionals who design Facebook-based interventions.

Project description:BackgroundHeated tobacco products (HTPs) are widespread in Japan, and smoking cessation of such products has become an important issue owing to the spread of harmful effects from HTPs. The efficacy of online digital therapy has been reported in smoking cessation treatment; however, we have limited evidence of online smoking cessation programs for HTP users.ObjectiveIn this study, we evaluate the usefulness of the Ascure program for HTP users (defined as exclusive HTP use or dual use of HTP and cigarettes) compared with exclusive cigarette users.MethodsThis was a retrospective study. We recruited adult smokers participating in the Ascure online smoking cessation program in Japan from June 2019 to February 2021. The Ascure smartphone app provided four elements: (1) educational video tutorials to enhance the understanding of nicotine dependence, (2) a personalized to-do list for behavior change, (3) a digital diary for record keeping, and (4) interactive chat sessions for relief from cravings or withdrawal symptoms. The primary outcome was the continuous abstinence rate (CAR) at weeks 21 to 24, biochemically validated using salivary cotinine testing. We considered those who dropped out of the program as smoking cessation failures. We analyzed the primary outcome using inverse probability weighting against tobacco product type estimated by multinomial propensity scores. We also assessed CAR at weeks 9 to 12 and program adherence.ResultsWe analyzed data from 2952 participants, including 52% (1524/3478) in the cigarette group, 35% (1038/3478) in the HTP group, and 13% (390/3478) in the dual-use group, who had a mean age of 43.4 (SD 10.8) years and included 17% (513/2952) women. CAR at weeks 21 to 24 showed that exclusive HTP users were more likely to stop tobacco use than exclusive cigarette smokers (CAR 52.6% for cigarette users vs CAR 64.8% for HTP users; odds ratio [OR] 1.17, 95% CI 1.12-1.22; P<.001). There was no significant difference between the exclusive cigarette users and the dual users (CAR 52.6% for cigarette users vs CAR 48.7% for dual users; OR 0.99, 95% CI 0.93-1.05; P=.77). CAR at weeks 9 to 12 was 56.7% (95% CI 54.2%-59.2%) for the exclusive cigarette users, 68.3% (95% CI 65.5%-71.1%) for the exclusive HTP users, and 58.2% (95% CI 53.3%-63.1%) for the dual users. The program adherence rate at week 24 was 70.7% overall (68.4% for cigarette users, 75% for HTP users, and 67.9% for dual users).ConclusionsExclusive HTP users had higher CARs and adherence compared with exclusive cigarette users, indicating a higher affinity for the Ascure online smoking cessation program. This program might be a useful smoking cessation option for HTP users, as well as for cigarette smokers.

Project description:BackgroundLittle is known about how individuals engage over time with smartphone app interventions and whether this engagement predicts health outcomes.ObjectiveIn the context of a randomized trial comparing 2 smartphone apps for smoking cessation, this study aimed to determine distinct groups of smartphone app log-in trajectories over a 6-month period, their association with smoking cessation outcomes at 12 months, and baseline user characteristics that predict data-driven trajectory group membership.MethodsFunctional clustering of 182 consecutive days of smoothed log-in data from both arms of a large (N=2415) randomized trial of 2 smartphone apps for smoking cessation (iCanQuit and QuitGuide) was used to identify distinct trajectory groups. Logistic regression was used to determine the association of group membership with the primary outcome of 30-day point prevalence of smoking abstinence at 12 months. Finally, the baseline characteristics associated with group membership were examined using logistic and multinomial logistic regression. The analyses were conducted separately for each app.ResultsFor iCanQuit, participants were clustered into 3 groups: "1-week users" (610/1069, 57.06%), "4-week users" (303/1069, 28.34%), and "26-week users" (156/1069, 14.59%). For smoking cessation rates at the 12-month follow-up, compared with 1-week users, 4-week users had 50% higher odds of cessation (30% vs 23%; odds ratio [OR] 1.50, 95% CI 1.05-2.14; P=.03), whereas 26-week users had 397% higher odds (56% vs 23%; OR 4.97, 95% CI 3.31-7.52; P<.001). For QuitGuide, participants were clustered into 2 groups: "1-week users" (695/1064, 65.32%) and "3-week users" (369/1064, 34.68%). The difference in the odds of being abstinent at 12 months for 3-week users versus 1-week users was minimal (23% vs 21%; OR 1.16, 95% CI 0.84-1.62; P=.37). Different baseline characteristics predicted the trajectory group membership for each app.ConclusionsPatterns of 1-, 3-, and 4-week smartphone app use for smoking cessation may be common in how people engage in digital health interventions. There were significantly higher odds of quitting smoking among 4-week users and especially among 26-week users of the iCanQuit app. To improve study outcomes, strategies for detecting users who disengage early from these interventions (1-week users) and proactively offering them a more intensive intervention could be fruitful.

Project description:IntroductionSmoking lapses after the quit date often lead to full relapse. To inform the development of real time, tailored lapse prevention support, we used observational data from a popular smoking cessation app to develop supervised machine learning algorithms to distinguish lapse from non-lapse reports.Aims and methodsWe used data from app users with ≥20 unprompted data entries, which included information about craving severity, mood, activity, social context, and lapse incidence. A series of group-level supervised machine learning algorithms (eg, Random Forest, XGBoost) were trained and tested. Their ability to classify lapses for out-of-sample (1) observations and (2) individuals were evaluated. Next, a series of individual-level and hybrid algorithms were trained and tested.ResultsParticipants (N = 791) provided 37 002 data entries (7.6% lapses). The best-performing group-level algorithm had an area under the receiver operating characteristic curve (AUC) of 0.969 (95% confidence interval [CI] = 0.961 to 0.978). Its ability to classify lapses for out-of-sample individuals ranged from poor to excellent (AUC = 0.482-1.000). Individual-level algorithms could be constructed for 39/791 participants with sufficient data, with a median AUC of 0.938 (range: 0.518-1.000). Hybrid algorithms could be constructed for 184/791 participants and had a median AUC of 0.825 (range: 0.375-1.000).ConclusionsUsing unprompted app data appeared feasible for constructing a high-performing group-level lapse classification algorithm but its performance was variable when applied to unseen individuals. Algorithms trained on each individual's dataset, in addition to hybrid algorithms trained on the group plus a proportion of each individual's data, had improved performance but could only be constructed for a minority of participants.ImplicationsThis study used routinely collected data from a popular smartphone app to train and test a series of supervised machine learning algorithms to distinguish lapse from non-lapse events. Although a high-performing group-level algorithm was developed, it had variable performance when applied to new, unseen individuals. Individual-level and hybrid algorithms had somewhat greater performance but could not be constructed for all participants because of the lack of variability in the outcome measure. Triangulation of results with those from a prompted study design is recommended prior to intervention development, with real-world lapse prediction likely requiring a balance between unprompted and prompted app data.

Dataset Information

Predicting the Users' Level of Engagement with a Smartphone Application for Smoking Cessation: Randomized Trial and Machine Learning Analysis.

Introduction

Objective

Methods

Results

Conclusions

Publications

Predicting the Users' Level of Engagement with a Smartphone Application for Smoking Cessation: Randomized Trial and Machine Learning Analysis.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets