Can an Emoji a Day Keep the Doctor Away? An Explorative Mixed-Methods Feasibility Study to Develop a Self-Help App for Youth With Mental Health Problems

Levi van Dam, Sianne Rietstra, Eva Van der Drift, Geert Jan J. M. Stams, Rob van der Mei, Maria Mahfoud, Arne Popma, Eric Schlossberg, Alex Pentland, Todd G. ReidFrontiers in Psychiatry 2019 · DOI

Abstract

Today's smartphones allow for a wide range of "big data" measurement, for example, ecological momentary assessment (EMA), whereby behaviours are repeatedly assessed within a person's natural environment. With this type of data, we can better understand - and predict - risk for behavioral and health issues and opportunities for (self-monitoring) interventions. In this mixed-methods feasibility study, through convenience sampling we collected data from 32 participants (aged 16-24) over a period of three months. To gain more insight into the app experiences of youth with mental health problems, we interviewed a subsample of 10 adolescents who received psychological treatment. The results from this feasibility study indicate that emojis can be used to identify positive and negative feelings, and individual pattern analyses of emojis may be useful for clinical purposes. While adolescents receiving mental health care are positive about future applications, these findings also highlight some caveats, such as possible drawback of inaccurate representation and incorrect predictions of emotional states. Therefore, at this stage, the app should always be combined with professional counseling. Results from this small pilot study warrant replication with studies of substantially larger sample size.

Introduction

Traditional mental health assessments often rely on retrospective questionnaires, which are susceptible to recall bias and provide only a static "snapshot" of a patient's condition. For youth—a demographic of "digital natives"—communicating complex emotions through formal surveys can be a significant barrier to care. This study explores the "datafication of health" by investigating whether emojis can serve as a low-friction, high-frequency proxy for emotional states via Ecological Momentary Assessment (EMA). By assessing moods repeatedly in a natural environment, clinicians may eventually provide "just-in-time" interventions before crises occur.

Methodology

The researchers employed a mixed-methods approach over three months with 32 participants (aged 16–24). The cohort included a "clinical" group (n=13) receiving mental health care and a "non-clinical" group (n=19).

Participants used the "G-Moji" Android app to answer a daily prompt: "How are you feeling today?" by selecting one of 14 emojis. The app also provided feedback through a monthly summary and activity overview.

The G-Moji app interface: (A) Daily mood check-in and (B) Monthly summary and activity overview.

To analyze the data, the team used Principal Component Analysis (PCA) with oblimin rotation to determine if emoji selections naturally clustered into psychological dimensions. Standardized metrics, including the SCL-90 (Symptom Checklist) and CYRM-12 (resilience), were used to validate the clinical status of the sample. Additionally, semi-structured interviews were conducted with 10 clinical participants to explore user experience and privacy concerns.

Key Results

Emoji Dimensionality and Engagement

PCA successfully reduced 12 emojis into two interpretable dimensions: Positive (e.g., ecstatic, confident, happy) and Negative (e.g., down, confused, anxious). These dimensions accounted for 50% of the total variance, suggesting emojis are a reliable scale for emotional valence. Engagement was moderate, with a 67% average response rate, though "app fatigue" was noted as usage dropped from 42% in the first month to 25% by the third.

Individual vs. Group Analysis

While group-level emoji frequencies did not significantly distinguish clinical from healthy groups, individual case analysis revealed critical insights. For instance, specific shifts in emoji patterns preceded a suicide attempt in one participant—a signal that would be obscured in group averages. This underscores the potential for personalized digital psychiatry.

Longitudinal tracking of emotional valence for two individual participants over three months.

Clinical Validation

The clinical group showed significantly higher psycho-neuroticism ($d = 1.70$) and lower resilience ($d = 0.63$), confirming the app's relevance for populations with complex needs. Negative emoji use was also significantly negatively correlated with positive emoji use ($r = -0.66$).

Discussion and Limitations

Youth participants reported that the app acted as a "mirror," enhancing self-awareness of emotional patterns. However, they expressed concerns regarding "algorithmic gaslighting"—the risk that inaccurate automated predictions (e.g., misinterpreting tiredness as sadness) could trigger negative emotional spirals.

The study is limited by its small sample size (n=32) and platform constraints (Android-only). Furthermore, while the "sick" emoji (with a thermometer) was included, users perceived it as a physical rather than emotional state, suggesting it should be excluded from future psychological valence scales. Future research should combine active emoji tracking with passive digital phenotyping (e.g., GPS, sleep patterns) to improve predictive accuracy.

Conclusion

Emojis offer a promising, low-friction method for tracking youth mental health. While not a replacement for professional care, emoji-based EMA can provide valuable longitudinal data for individual pattern analysis and clinical intervention.

Summary updated

Abstract

Today's smartphones allow for a wide range of "big data" measurement, for example, ecological momentary assessment (EMA), whereby behaviours are repeatedly assessed within a person's natural environment. With this type of data, we can better understand - and predict - risk for behavioral and health issues and opportunities for (self-monitoring) interventions. In this mixed-methods feasibility study, through convenience sampling we collected data from 32 participants (aged 16-24) over a period of three months. To gain more insight into the app experiences of youth with mental health problems, we interviewed a subsample of 10 adolescents who received psycthological treatment. The results from this feasibility study indicate that emojis) can be used to identify positive and negative feelings, and individual pattern analyses of emojis may be useful for clinical purposes. While adolescents receiving mental health care are positive about future applications, these findings also highlight some caveats, such as possible drawback of inaccurate representation and incorrect predictions of emotional states. Therefore, at this stage, the app should always be combined with professional counseling. Results from this small pilot study warrant replication with studies of substantially larger sample size.
BibTeX
@inproceedings{dam2019can,
  author = {Levi van Dam and Sianne Rietstra and Eva Van der Drift and Geert Jan J. M. Stams and Rob van der Mei and Maria Mahfoud and Arne Popma and Eric Schlossberg and Alex Pentland and Todd G. Reid},
  title = {{Can an Emoji a Day Keep the Doctor Away? An Explorative Mixed-Methods Feasibility Study to Develop a Self-Help App for Youth With Mental Health Problems}},
  booktitle = {Frontiers in Psychiatry},
  year = {2019},
  doi = {10.3389/fpsyt.2019.00593},
}
Figure