Smartphone-derived person-level lifestyle profiles differed significantly only in positive functioning, not in overall mental well-being, positive affect, or satisfying interpersonal relationships, with the physically active and unplugged profile showing higher positive functioning than the mobile and always-on social profile.
Key Findings
Results
Multilevel latent profile analysis identified eight day-level profiles and seven person-level profiles from smartphone-sensing data in German adults.
The study used a 2-week prospective longitudinal cohort design with 553 German adults (mean age 42.12, SD 12.89 years; 44.65% female).
Ten smartphone-sensing indicators captured 5 domains: communication and social media app use, mobility, physical activity, environmental context, and phone-use intensity.
Day-level profiles reflected distinct combinations of smartphone-sensing indicators, while person-level profiles represented different distributions of these daily patterns.
The sample was drawn from an initial sample recruited according to quotas designed to reflect the German population.
Results
Person-level lifestyle profiles differed significantly only in positive functioning, not in overall mental well-being, positive affect, or satisfying interpersonal relationships.
The omnibus test for positive functioning was significant (Wald χ²6=13.39; P=.04).
No significant omnibus differences were found for overall mental well-being, positive affect, or satisfying interpersonal relationships.
Mental well-being was assessed using the Warwick-Edinburgh Mental Well-Being Scale.
Associations were tested using classification-error-adjusted mean comparisons and omnibus Wald tests.
Results
The physically active and unplugged profile had significantly higher positive functioning than the mobile and always-on social profile.
Mean positive functioning scores were 3.94 (SD 0.63) for the physically active and unplugged profile versus 3.61 (SD 0.74) for the mobile and always-on social profile.
The effect size was Cohen d=0.47 (95% CI 0.21-0.73), indicating a moderate difference.
No other pairwise profile differences in positive functioning were statistically significant.
This was the only significant pairwise comparison among the seven person-level profiles.
Results
Big Five personality traits did not significantly moderate the associations between lifestyle profiles and any mental well-being outcome.
Personality was assessed using the 15-item Big Five Inventory-2 Extra-Short Form, covering extraversion, agreeableness, conscientiousness, openness, and negative emotionality.
Moderation was examined using hierarchical regressions comparing models with and without profile-by-personality interactions.
Personality-by-profile interactions did not significantly improve prediction for any well-being outcome.
This null finding applied across all five Big Five traits and all four well-being outcomes.
Results
Sensitivity analyses excluding the smallest person-level profile produced comparable results, supporting the robustness of the primary findings.
Sensitivity analyses were conducted by excluding the smallest profile from the analyses.
Results were described as 'comparable,' supporting robustness of the primary findings.
This approach addressed potential concerns about the stability of findings driven by small subgroups.
Discussion
The authors propose that interpretable person-centered digital phenotypes could support monitoring tools and inform personalized multibehavior interventions targeting combinations of behaviors.
The authors note that causal conclusions cannot be drawn from the current study design.
They suggest the profiles could 'support understandable monitoring tools' in real-world settings.
The authors recommend prospective replication and validation before clinical application.
They argue that targeting combinations of behaviors rather than single behaviors in isolation may be a more effective intervention approach.
What This Means
This research suggests that patterns of everyday smartphone use — including how much people move around, use social media, make calls, and interact with their environment — can be used to group people into distinct lifestyle profiles. Using two weeks of passively collected smartphone data from 553 German adults, the researchers identified seven overall lifestyle profiles (and eight types of daily behavioral patterns within them). These profiles captured meaningful combinations of behaviors rather than looking at single behaviors in isolation, which is a more realistic picture of how people actually live.
When the researchers compared these lifestyle profiles to measures of mental well-being, they found that the profiles only differed meaningfully on one specific aspect: 'positive functioning,' which reflects things like feeling good and living well. People in the 'physically active and unplugged' profile scored moderately higher on positive functioning than people in the 'mobile and always-on social' profile. However, the profiles did not differ on overall mental well-being, positive emotions, or satisfaction in relationships. Additionally, personality traits like extraversion or conscientiousness did not change these associations.
This research suggests that grouping people by their combined behavioral patterns — rather than single behaviors — can reveal meaningful differences in at least some aspects of mental well-being. The findings are exploratory and cannot establish cause and effect, but they point toward the potential for smartphone-based lifestyle profiles to serve as practical monitoring tools or to guide interventions that address multiple behaviors together rather than one at a time. The researchers emphasize that further replication and validation studies are needed before such approaches could be used in clinical or public health settings.
Zhu N, Schoedel R, Sust L, Bühner M, Terhorst Y. (2026). Digital Phenotyping of Lifestyle Profiles and Mental Well-Being in German Adults: Prospective Longitudinal Cohort Study.. Journal of medical Internet research. https://doi.org/10.2196/101850