JMIR Publications

ABSTRACT

Background:

Using contextual cues for performing daily behaviors is one proven method for establishing healthy habits, but objective measures of this habit formation process do not currently exist. Mobile health (mHealth) tools provide the detailed, longitudinal data necessary for constructing objective measures of habit strength, which can improve our understanding of habit formation and help design more effective mHealth interventions for promoting healthier habits.

Objective:

Use behavioral data from a commercial mindfulness mediation mobile phone app to construct a measure of meditation habit strength and estimate the association between meditation habits and app users’ perceived mental health benefits.

Methods:

App usage data were analyzed for 2,771 paying subscribers of a meditation mobile phone app who also volunteered to complete a survey assessing their perceived changes in physical and mental health from using the app. An objective measure of meditation habit strength was calculated based on the similarity in the timing of app usage between consecutive days. Receiver operating characteristic (ROC) curve analysis was used to validate the habit measure as a strong predictor of users’ future behavior, and variable importance statistics from random forest models were used to corroborate these findings. Logistic regressions were used to estimate the association between the meditation habit measure and self-reported physical and mental health benefits.

Results:

The temporal similarity in users’ daily app use before completing the survey, as measured by the dynamic time warping (DTW) distance between app usage on consecutive days, significantly predicted app usage 28 days and six months after the survey, even after controlling for users’ demographic and socioeconomic characteristics, total app sessions, duration of app use, and number of days with any app use. Additionally, the temporal similarity measure significantly increased in the area under the ROC curve (AUC) for models predicting any future app use in 28 days (AUC = 0.868 with DTW and AUC = 0.850 without DTW [P < 0.001]) and for models predicting any app use in six months (AUC = 0.821 with DTW and AUC = 0.802 without DTW [P < 0.001]). Finally, a one-percent increase in the temporal similarity of users’ daily meditation practice with the app over the six weeks prior to the survey was associated with an increased odds of reporting mental health improvements by 2.94 OR (95% CI: 1.832 – 6.369).

Conclusions:

The temporal similarity of meditation app use was a significant predictor of future behavior, which suggests that this measure captures important components of the cued habit formation process. Additionally, temporal similarity was associated with greater perceived mental health benefits, which demonstrates additional mental health benefits may be derived from forming strong meditation habits. Future research should examine how temporal similarity measures characterize habits using mHealth data in other behavioral settings.