Effect of a mindfulness program on stress, anxiety, depression, sleep quality, social support, and life satisfaction: a quasi-experimental study in college students

The study addresses rising mental health concerns among college students—particularly stress, anxiety, depression, poor sleep quality, limited social support, and reduced life satisfaction—which impact academic performance and well-being. These issues may be accentuated in lower-middle-income contexts and health sciences programs. Mindfulness, defined as nonjudgmental present-moment awareness, has theoretical bases in enhancing metacognitive and interoceptive awareness, reducing cognitive reactivity, and regulating HPA-axis stress responses. The research question was whether a structured mindfulness program improves stress, anxiety, depression, sleep quality, perceived social support, and life satisfaction in university students within a culturally diverse Peruvian context. The purpose was to evaluate program effects comprehensively and inform culturally adaptable mental health strategies in higher education.

Prior research demonstrates that mindfulness-based interventions can reduce stress, anxiety, depression, and insomnia in university students, though evidence for sleep quality has been mixed and often inconclusive. Mechanistic accounts suggest mindfulness reduces HPA-axis activation and cortisol, enhances prefrontal regulation, decreases amygdala reactivity, and improves emotional regulation and metacognitive awareness. Evidence also links mindfulness with increased empathy, interpersonal effectiveness, and perceived social support, as well as greater life satisfaction through positive cognitive reappraisal (mindfulness-to-meaning model). However, much literature derives from Western contexts, underscoring the need to examine cultural variability in acceptability and efficacy. This study responds to gaps regarding sleep quality effects, mediation/moderation mechanisms, and cultural generalizability, particularly in underrepresented settings such as Peru.

Design: Quasi-experimental pretest–posttest with a waiting-list control group. Power analysis (G*Power 3) assumed moderate effect size, α = 0.05, power = 0.80, yielding N = 128 (64 per group). Ethics approval: School of Medicine Research Ethics Committee, Universidad César Vallejo, Peru (021-CEI-EPM-UCV-2023; 26/04/2023). Participants: Health sciences students at a private Peruvian university. Recruitment followed an awareness session; sections B and D were selected (B = control; D = experimental) to ensure participation; differences in willingness were not statistically significant. Inclusion thresholds: PSS-10 > 9, SAS > 49, SDS > 49, PSQI > 5. Exclusion: prior mindfulness-related practices (meditation, tai chi, yoga), psychiatric treatment or psychotropic medication, active substance abuse, pregnancy, or concurrent psychological/wellness interventions. Procedures: Baseline assessments administered to both groups. The experimental group completed 12 weekly 60-minute mindfulness meditation sessions over 3 months, led by a qualified teacher. Compliance report forms and daily audio-guided meditations were provided. Post-tests used the same instruments for both groups. Instruments: PSS-10 (10 items; locally validated; split-half reliability 0.98), Zung SAS (20 items; split-half 0.89), Zung SDS (20 items; split-half 0.94), PSQI (18 items; 7 components; split-half 0.96), MOS-SS (20 items; split-half 0.96), SWLS (5 items; split-half 0.94). Data analysis: Descriptive statistics (means, SD; medians, IQR as non-normal). Group and phase differences tested with Mann–Whitney U and Wilcoxon tests. ANCOVA controlled for baselines (pre-scores as covariates). Robust HC3 standard errors addressed heteroscedasticity and non-normality. Mediation (PROCESS model 4) and moderation (model 1) used 5,000 bootstrap samples. Software: GraphPad Prism 8, SPSS v27, PROCESS v4.2.

- Sample: 128 allocated; analyzed n=64 per group (CG lost n=4; EG lost n=3). Groups were comparable on gender, age, marital status, and clinical treatment (p > 0.05). - Nonparametric pre/post within-group and between-group results: • Stress (PSS-10): No pretest difference (p = 0.770). EG decreased significantly post (p < 0.001), median 26.00 (IQR 22.75–29.00) to 19.00 (15.00–21.00); CG no change (p = 0.159). Post between-group difference significant (p < 0.001). • Anxiety (SAS): No pretest difference (p = 0.716). EG decreased (p < 0.001), median 59.50 (57.25–63.00) to 51.00 (48.00–55.75); CG no change (p = 0.061). Post between-group difference significant (p < 0.001). • Depression (SDS): No pretest difference (p = 0.694). EG decreased (p < 0.001), median 55.00 (50.00–61.00) to 50.00 (48.00–55.00); CG no change (p = 0.985). Post between-group difference significant (p < 0.001). • Sleep quality (PSQI): No pretest difference (p = 0.825). EG improved (p < 0.001), median 7.00 (4.00–8.75) to 4.00 (2.00–6.00); CG no change (p = 0.062). Post between-group difference significant (p < 0.001). • Social support (MOS-SS): No pretest difference (p = 0.201). EG increased (p < 0.001), median 30.00 (21.25–33.00) to 55.00 (46.00–61.00). Post between-group difference significant (p < 0.001). • Life satisfaction (SWLS): No pretest difference (p = 0.810). EG increased (p < 0.001), median 18.00 (14.25–19.00) to 25.00 (22.00–29.00); CG no change (p = 0.073). - ANCOVA and HC3 robust results (group effects, partial η²): • Stress: ANCOVA F = 76.940, p < 0.001, η² = 0.381; HC3 B = 8.090, p < 0.001, partial η² = 0.376. • Anxiety: ANCOVA F = 150.647, p < 0.001, η² = 0.547; HC3 B = 9.800, p < 0.001, partial η² = 0.538. • Depression: ANCOVA F = 12.679, p < 0.001, η² = 0.092; HC3 B = 3.616, p < 0.001, partial η² = 0.091. • Sleep quality: ANCOVA F = 52.601, p < 0.001, η² = 0.296; HC3 B = 2.760, p < 0.001, partial η² = 0.286. • Social support: ANCOVA F = 312.865, p < 0.001, η² = 0.715; HC3 B = −25.319, p < 0.001, partial η² = 0.704. • Life satisfaction: ANCOVA F = 132.425, p < 0.001, η² = 0.514; HC3 B = −8.047, p < 0.001, partial η² = 0.510. - Mediation/moderation: PROCESS analyses found no significant indirect effects for sleep quality mediating anxiety (p = 0.07), social support mediating stress was borderline (p = 0.05), and life satisfaction mediating depression was not significant (p = 0.15). No significant moderation by social support, life satisfaction, or sleep quality (all interaction p-values > 0.45). Overall, mindfulness produced direct improvements across outcomes, with largest effects for social support, anxiety, and life satisfaction.

The findings directly address the research question, showing that a structured 12-session mindfulness program substantially reduced stress and anxiety and improved sleep quality, social support, and life satisfaction, with a modest but significant reduction in depression. These results align with theoretical mechanisms whereby mindfulness enhances metacognitive awareness, reduces cognitive reactivity, and regulates HPA-axis activity, improving emotional regulation and reducing physiological stress responses. Improvements in social support and life satisfaction suggest mindfulness may enhance interpersonal sensitivity, empathy, and positive reappraisal, thereby strengthening social connectedness and overall well-being. The lack of robust mediation or moderation by sleep, social support, or life satisfaction indicates that, in this context, mindfulness primarily exerted direct effects on psychological outcomes rather than through these pathways. Cultural considerations in the Peruvian university setting may shape acceptability and efficacy, highlighting the adaptability of mindfulness beyond Western contexts while pointing to potential variability in effects on depression. Collectively, the results support implementing mindfulness in educational settings as part of holistic mental health strategies.

A 12-session mindfulness meditation program effectively improved multiple dimensions of university students’ mental health, notably reducing stress and anxiety and enhancing sleep quality, perceived social support, and life satisfaction, with smaller but significant effects on depression. Robust ANCOVA and HC3 analyses corroborated these effects, with the strongest improvements observed in social support. The absence of significant mediation or moderation by sleep quality, social support, or life satisfaction suggests predominantly direct effects. Future research should employ randomized designs, larger and more diverse samples, longer interventions and follow-ups, and include objective physiological measures. Culturally tailored adaptations and exploration of alternative mechanisms (e.g., rumination, self-compassion, attention regulation) are recommended to optimize impact and clarify causal pathways.

- Quasi-experimental design with non-random group assignment may introduce selection bias and limits causal inference. - Sample size met power requirements but larger samples could improve precision, especially for mediation/moderation analyses and generalizability. - Intervention duration (12 sessions over 3 months) may be insufficient for larger effects on depression; no long-term follow-up to assess durability. - Reliance on self-report measures raises potential for response and social desirability biases; no physiological markers (e.g., cortisol, HRV) were included. - Sections were selected based on willingness to participate (though differences were not statistically significant), potentially biasing engagement and limiting generalizability to less motivated populations. - Cultural specificity (coastal Peru, health sciences students) may limit generalizability to other regions or academic disciplines.