Gender and academic year as moderators of the efficacy of mobile app interventions to promote physical activity in adolescents: a randomized controlled trial

Adolescent physical inactivity has increased, exacerbated by the COVID-19 pandemic, with associated adverse effects on kinanthropometric variables (higher fat mass) and fitness (lower cardiorespiratory and muscular fitness). Given links between adolescent body composition/fitness and later disease risk, effective interventions are needed. Schools are promising settings, and mobile apps are widely used by adolescents. However, adherence to app-based interventions tends to diminish over time and little is known about whether efficacy differs by gender or academic year. This study aimed to determine: (a) differences in school-promoted step-tracker app use by gender and academic year; (b) efficacy of the intervention on physical activity, kinanthropometric and derived variables, and fitness by gender; and (c) efficacy by academic year.

- Post-pandemic decreases in adolescent physical activity persist and relate to unhealthy behaviors and poorer fitness. Higher fat percentage in adolescence predicts future cardiometabolic disease risk. Fitness relates to cognitive performance and later disease risk. - School-based interventions (2–20 months) can increase activity and fitness; embedding academic rewards may enhance adherence. - mHealth interventions (web, SMS, apps) are promising; mobile apps are most common. Initial gains may wane with adherence decline over weeks. - Recent school-promoted step-tracker interventions with academic rewards maintained adherence and improved out-of-school activity, body composition, and fitness. - Gender/age differences: males typically more active; activity declines with age; females and older adolescents use mobile phones more, though evidence on step-tracker app use by gender is mixed. No prior education-promoted app studies analyzed moderation by gender/academic year, indicating a research gap addressed here.

Design: Cluster randomized controlled trial with pre- and post-test (10-week interval). Classes within two large secondary schools in Murcia (Spain) were randomized to intervention or control. Intervention classes were further randomized to one app: Strava, Pacer, MapMyWalk, or Pokémon Go. CONSORT guidelines followed; ethics approval CE022102; trial registration NCT04860128. Participants: Convenience sample; 400 adolescents (210 males, 190 females), ages 12–16 (mean 13.96 ± 1.21). Inclusion: enrolled in compulsory secondary education, age 12–16, completed pre/post measures, medically able. Exclusion: no/irregular phone access; school change; <80% PE attendance; starting or ceasing organized PA during intervention (walking/running allowed). Randomization/blinding: Cluster randomization by class to intervention/control; intervention classes randomized to app. Outcome assessors were blinded to group and prior scores; eligibility assessed blinded to allocation. Intervention: 10 weeks; after-school app use ≥3 times/week. Progressive targets: ~7000 steps/day (~4.57 km) in week 1 up to 12,500 steps/day (~8 km) by week 10. Apps counted steps/distance; Pokémon Go recorded distance automatically. Academic reward in PE provided for adherence. Controls continued usual PE and activities without apps. Measurements: Pre/post in both groups. - Physical activity: PAQ-A (Spanish validated), 8 Likert items averaged; 9th item dichotomous barrier. - Kinanthropometrics: mass, height, sitting height; skinfolds (triceps, thigh, calf); girths (relaxed arm, waist, hips, thigh, calf) by ISAK protocols. Derived variables: BMI; muscle mass (Poortmans), fat mass (Slaughter), sum of 3 skinfolds; waist-to-hip ratio; corrected girths (arm, thigh, calf). Duplicate measures (third if >5% folds or >1% others); TEM reported (low intra- and inter-evaluator error). Calibrated instruments (Harpenden caliper, TANITA BC 418-MA scale, SECA 213 stadiometer, Lufkin tape). - Fitness: 20-m shuttle run (VO2max via Léger formula), handgrip strength (Takei Tkk5401, arm extended), countermovement jump (force platform, 200 Hz), curl-up test. Two trials for tests (best score), except shuttle run once at end; standardized warm-up and rest intervals per NSCA. Procedure: PAQ-A, then anthropometrics, familiarization, warm-up, strength tests, curl-ups, shuttle run. Statistical analysis: Normality via Kolmogorov–Smirnov; parametrics applied. Descriptives mean ± SD. Chi-square tested app use by gender and academic year; corrected standardized residuals (±1.96 threshold); effect sizes via Cramer's V or contingency coefficient. One-way ANCOVA for pre-post differences by app use with moderators (gender, academic year). Two MANOVAs evaluated pre-post differences by app use × gender and app use × academic year; Bonferroni post-hoc tests. Partial eta squared for effect sizes (p<0.05). SPSS v25 used.

- App use uptake (n=240 users; n=160 non-users): • Gender: Females 71.1% users vs 28.9% non-users; Males 50.0% users vs 50.0% non-users; χ²=16.558, p<0.001; Cramer’s V=0.203; residuals significant (±4.1). • Academic year: 1st 53.8% users; 2nd 44.3%; 3rd 70.0%; 4th 74.4%; χ²=19.179, p<0.001; contingency coefficient=0.214; residuals significant from ±1.7 to ±3.1. - Overall pre-post within-group changes (Table 2): Experimental group showed significant improvements in most variables (e.g., body mass −0.918 kg; BMI −0.645 kg/m²; muscle mass +0.698 kg; VO2max +1.05 ml/kg/min; curl-ups +3.95; CMJ +1.26 cm), except sitting height and waist girth (ns). Control group had fewer/no significant changes in PA level, sitting height, sum of 3 skinfolds, waist girth, fat mass, CMJ, and curl-ups. - Moderation by gender (ANCOVA/MANOVA): Significant for BMI (p<0.001), corrected calf girth (p=0.008), fat mass (p=0.025), handgrip right (p=0.002) and left (p=0.002), and CMJ (p=0.005). Reported mean differences included: BMI (−0.352), corrected calf girth (−0.498), fat mass (0.748), handgrip right (−1.359), handgrip left (−1.103), CMJ (−2.456). • Post-hoc: BMI increased in males regardless of app use (p=0.009–0.014), but not in females who used apps (p=0.238). Corrected calf girth changes significant only in males (intervention p=0.011; control p=0.008). Fat mass decreased significantly in males who used apps (p=0.025). Handgrip increased in males in both groups (intervention p<0.001–0.011; control p=0.002) and in intervention females only (p=0.005–0.011). CMJ improved only in intervention males (p=0.002). - Moderation by academic year: Significant for height (p<0.001–0.044; mean diff −1.099 to −0.509), sum of 3 skinfolds (p=0.046–0.047; mean diff −3.255), waist girth (p=0.048; mean diff 0.584), hip girth (p<0.001–0.008; mean diff −1.461 to −0.777), corrected calf girth (p=0.019–0.029; mean diff −0.539 to −0.482), and fat mass (p=0.025–0.046; mean diff 1.011 to 1.392). • Post-hoc: Height increased in all years except 4th (ns). In 4th-year, sum of 3 skinfolds and fat mass increased in controls (p=0.046; p=0.046) and decreased in intervention (p=0.047; p=0.025). 4th-year intervention group showed decreased waist girth (p=0.048) and was the only group without significant hip girth increase (p=0.475). Corrected calf girth increased in all intervention years (p=0.019–0.029) except 4th (ns), with no significant changes in controls.

Findings demonstrate that when a mobile step-tracker program is embedded in PE with academic incentives, females and older adolescents are more likely to adopt app use. Potential drivers include females’ stronger emphasis on academic performance and older adolescents’ greater integration of smartphones into daily life. Physiologically and behaviorally, app use mitigated increases in fat-related measures, with especially clear benefits in females (maintenance/reduction of BMI and fat mass) and in 4th-year students (reductions in fat mass, sum of skinfolds, and waist girth versus increases in controls). Strength and power gains were more evident in males (greater handgrip and CMJ improvements), likely reflecting maturational and hormonal differences (e.g., testosterone) and neuromuscular adaptations typical of adolescence. The academic-year moderation suggests older adolescents—despite generally lower habitual activity and higher sedentary time—can benefit when app-based activity is mandated and supported within school, potentially offsetting age-related fat accrual. Collectively, results address the research question by identifying gender and academic year as moderators of efficacy, informing tailoring of app-based school interventions.

This RCT shows that school-promoted mobile apps are effective tools to promote physical activity and beneficial changes in body composition and selected fitness outcomes in adolescents, with efficacy moderated by gender and academic year. Females and older adolescents used apps more; app use was associated with preventing increases in BMI and fat mass, particularly in females, and with reductions in fat-related measures among 4th-year students. Fitness gains were more marked in males (e.g., CMJ, handgrip). Future research should consider training intensity/volume within app programs, incorporate maturation assessments, explore motivations/barriers by gender and academic year, and examine differential engagement with gamified versus non-gamified apps to optimize tailoring and adherence.

- Biological maturation not assessed; maturational differences could influence body composition and fitness changes. - Physical activity level measured via self-report (PAQ-A), not by objective devices (e.g., accelerometers). - Training variables within the app program (volume, intensity, adherence granularity) were not fully characterized. - Motivations/barriers for app use by gender and academic year were not measured. - Potential contamination/awareness bias in control classes aware of the intervention. - Dietary habits were not monitored, limiting interpretation of body composition changes. - Generalizability limited to similar school settings and populations.