TL;DR

This comprehensive meta-analysis of 71 studies reveals that greater engagement with short-form video platforms (TikTok, Reels, Shorts) is significantly associated with poorer cognitive functioning—particularly attention and inhibitory control—and worse mental health outcomes including anxiety, stress, and depression. These patterns hold across both youth and adult populations, with addiction-based measures showing the strongest negative associations.

Executive Summary

The rapid proliferation of short-form videos (SFVs) across social media platforms has transformed digital engagement, yet their health implications remain poorly understood. This systematic review and meta-analysis synthesized data from 71 studies representing 98,299 participants to examine associations between SFV use and both cognitive and mental health outcomes. Results revealed moderate negative associations with cognitive functioning (r = −.34), particularly for attention and inhibitory control, and weak negative associations with mental health (r = −.21), especially anxiety and stress. Notably, SFV use showed no association with body image or self-esteem. Patterns were consistent across age groups (youth vs. adults) and platforms, though SFV addiction measures yielded stronger effects than duration or frequency measures. These correlational findings highlight the need for longitudinal research to clarify directionality and mechanisms.

Section-by-Section Analysis

Introduction

Summary: The authors establish that short-form videos (defined as content lasting seconds to minutes) have become dominant sources of digital engagement, popularized by TikTok/Douyin and now integrated across major platforms (Instagram Reels, YouTube Shorts). While initially entertainment-focused, SFVs have expanded into education, politics, and commerce. Their design features—infinite scrolling, algorithm-driven personalization, full-screen immersive playback—maximize engagement but raise concerns about addiction and health impacts. Despite growing research, findings remain mixed, necessitating comprehensive synthesis.

Key Points:

  • Platform evolution: SFVs first popularized by Vine (2012), revitalized by TikTok (2016), now ubiquitous across social media
  • Design characteristics: Infinite-scrolling interfaces, algorithmic personalization, immersive single-video playback
  • Engagement concerns: Features designed to maximize time-on-platform may contribute to problematic use patterns
  • Mixed evidence: Some studies report negative health associations, others find null or even positive effects
  • Research gap: No prior synthesis has comprehensively examined SFV use across multiple health domains, platforms, and age groups

SFVs and Cognition

Summary: The authors review evidence linking SFV use to attentional difficulties, interpreting findings through Groves and Thompson's dual theory of habituation and sensitization. Repeated exposure to fast-paced, highly rewarding content may desensitize users to slower cognitive tasks while sensitizing them to instant gratification, potentially undermining sustained attention and executive control. Neuroimaging studies show reduced P300 activity and structural differences in prefrontal and reward regions among heavy users. Evidence extends to inhibitory control, memory, and reasoning, though findings are mixed and some domains remain understudied.

Key Points:

  • Attention deficits: Higher SFV use linked to poorer sustained attention in youth and adults; neuroimaging shows reduced P300 responses
  • Theoretical mechanism: Habituation to high-stimulation content + sensitization to algorithmic rewards → reduced cognitive endurance
  • Executive functioning: Associations extend to inhibitory control, memory, and working memory
  • Mixed findings: Some longitudinal studies show cross-sectional but not experimental effects
  • Research gaps: Limited investigation of fluid intelligence, processing speed, visuospatial ability

SFVs and Mental Health

Summary: SFV use has been associated with elevated depression, anxiety, stress, and loneliness. The authors propose multiple mechanisms: dopaminergic reward system alterations that create reinforcement loops; social contagion through repeated exposure to mental health content; sleep disruption from blue light and overstimulation; and displacement of restorative offline activities. Interestingly, SFV use showed no consistent association with body image or self-esteem, possibly reflecting the diversity of content and creators on these platforms. The authors emphasize that content type—not just usage duration—likely shapes these associations.

Key Points:

  • Mood disorders: Associations with depression, anxiety, stress consistent across studies
  • Dopaminergic mechanisms: Algorithm-driven rewards may alter reward sensitivity, increasing vulnerability to distress
  • Social contagion: Exposure to mental health content may heighten symptom awareness or identification
  • Sleep disruption: Blue light exposure and cognitive arousal delay sleep onset, worsening mood
  • Body image paradox: No overall association with body/self-esteem, despite appearance-focused content
    • Likely reflects content diversity (body positivity vs. idealized portrayals)
    • Suggests individual differences and content exposure patterns matter
  • Loneliness: SFV use linked to social isolation through displacement of face-to-face interaction

Previous Syntheses of Literature

Summary: The authors critique prior meta-analyses for focusing exclusively on TikTok, limiting generalizability given multi-platform use patterns. Previous reviews also focused solely on adolescents or mental health, neglecting adults and cognitive outcomes. The current investigation addresses these gaps by including all SFV platforms, examining both cognitive and mental health domains, and testing age as a moderator.

Key Points:

  • Conte et al. (2025): Systematic review of TikTok and adolescent mental health (20 studies); no quantitative synthesis
  • Gabrielle et al. (2024): Meta-analysis showing weak negative association between TikTok use and adolescent mental health
  • Galanis et al. (2024): Meta-analysis (16 studies) linking TikTok to depression/anxiety in youth and adults; age not tested as moderator
  • Limitations of prior work: TikTok-only focus, adolescent-only samples, mental health-only outcomes, no moderator testing
  • Current contribution: Most comprehensive synthesis to date across platforms, age groups, and health domains

Analysis Framework

1. Research Foundation

Theoretical framework:

  • Groves & Thompson's dual process theory: Habituation to high-stimulation content + sensitization to immediate rewards → reduced cognitive endurance
  • Dopaminergic reward processing: Algorithmic curation creates unpredictable reinforcement, potentially altering reward sensitivity
  • Social comparison theory: Upward comparisons to idealized content may lower self-evaluation
  • Uses and gratifications theory: Motivations for use (entertainment, social connection, self-improvement) shape outcomes

Objectives: Provide comprehensive synthesis of SFV use and cognitive/mental health correlates; identify patterns, inconsistencies, and moderating factors; guide future research priorities

Hypotheses:

  • Greater SFV engagement will be associated with poorer cognitive functioning and mental health
  • Associations will vary by cognitive domain, mental health indicator, measurement type, platform, and age group

Gaps addressed:

  • Platform diversity (beyond TikTok)
  • Cognitive outcomes (understudied relative to mental health)
  • Age differences (youth vs. adults)
  • Measurement heterogeneity (addiction vs. duration vs. frequency)

2. Methodology

Study design: Systematic review and meta-analysis (PROSPERO: CRD42024587550)

Search strategy:

  • Databases: PsycInfo, PubMed, Scopus, Web of Science, ProQuest (gray literature)
  • Search terms: Short-form video platforms (TikTok, Reels, Douyin, Snapchat, Bilibili) + cognitive/mental health terms
  • No date/language restrictions; final search: October 28, 2024
  • Additional forward/backward citation searching and journal hand-searching

Inclusion criteria:

  • Empirical quantitative studies examining SFV use and cognitive/mental health correlates
  • SFV-specific platforms or features (e.g., Instagram Reels, not general Instagram use)
  • Any age group, no clinical status restrictions

Sample:

  • 71 studies (70 in meta-analysis, 1 qualitative only)
  • 98,299 total participants; mean sample size = 1,384 (SD = 2,968)
  • Geographic distribution: 74% Asia, 11% North America, 11% Europe, 3% Africa, 1% Central America
  • Age groups: 73% adults (M_age >18 years), 27% youths (M_age ≤18 years)
  • Gender: Average 60% female across studies
  • Study designs: 87% correlational, 13% group comparisons

Measures:

  • SFV engagement: Addiction (52%), duration (27%), usage (11%), intensity (10%), frequency (7%)
  • Platforms: General SFV use (52%), TikTok-specific (48%)
  • Cognitive domains: Attention, inhibitory control, memory, working memory, language, reasoning
  • Mental health domains: Depression, anxiety, stress, loneliness, sleep quality, well-being, affect, body image, self-esteem

Quality assessment: Mixed Methods Appraisal Tool (5 criteria); 59% rated high quality, 41% low quality

Statistical approach:

  • Random-effects models (accounts for heterogeneity)
  • Effect size: Pearson's r (converted from standardized mean differences when necessary)
  • Interpretation: r = .10 (weak), .30 (moderate), .50 (strong)
  • Heterogeneity: Q statistic, I² (<30% trivial, 30-50% moderate, 50-75% substantial, >75% considerable)
  • Publication bias: Funnel plots, Egger's test, trim-and-fill method
  • Moderators: Age group, SFV platform, measurement type, cognitive/mental health domain, covariate inclusion

3. Findings

Cognitive Correlates

Overall effect: r = −.34 (95% CI [−.42, −.26], p < .001) — moderate negative association

  • Greater SFV engagement linked to poorer cognitive performance
  • Publication bias minimal (Egger's test: β = 1.40, p = .636)
  • Considerable heterogeneity (I² = 95.73%)

By cognitive domain (significant moderator, Q(5) = 59.63, p < .001):

  • Inhibitory control: r = −.41 (moderate)
  • Attention: r = −.38 (moderate)
  • Working memory: r = −.21 (weak)
  • Language: r = −.16 (weak)
  • Memory: r = −.14 (weak)
  • Reasoning: r = −.13 (non-significant)

By SFV measurement (significant moderator, Q(2) = 60.63, p < .001):

  • Intensity: r = −.55 (strong; 1 study)
  • Addiction: r = −.37 (moderate)
  • Duration: r = −.20 (weak)

By platform (non-significant moderator):

  • General SFV: r = −.35
  • TikTok: r = −.26

By age group (non-significant moderator):

  • Adults: r = −.32
  • Youths: r = −.36

Mental Health Correlates

Overall effect: r = −.21 (95% CI [−.25, −.17], p < .001) — weak negative association

  • Greater SFV engagement linked to poorer mental health
  • No publication bias detected
  • Considerable heterogeneity (I² = 96.43%)

By mental health domain (significant moderator, Q(8) = 44.61, p < .001):

  • Stress: r = −.34 (moderate)
  • Anxiety: r = −.33 (moderate)
  • Depression: r = −.23 (weak)
  • Loneliness: r = −.23 (weak)
  • Sleep quality: r = −.22 (weak)
  • Well-being: r = −.14 (weak)
  • Affect: r = −.13 (weak)
  • Body image: r = −.10 (non-significant)
  • Self-esteem: r = −.08 (non-significant)

By SFV measurement (significant moderator, Q(4) = 44.64, p < .001):

  • Addiction: r = −.32 (moderate)
  • Intensity: r = −.14 (non-significant)
  • Usage: r = −.13 (weak)
  • Duration: r = −.10 (weak)
  • Frequency: r = −.05 (non-significant)

By platform (significant moderator, Q(1) = 9.48, p = .002):

  • General SFV: r = −.27
  • TikTok: r = −.15

By age group (non-significant moderator):

  • Adults: r = −.21
  • Youths: r = −.21

Covariate inclusion: Non-significant moderator for both cognitive and mental health outcomes

4. Critical Evaluation

Limitations

Study-level:

  • Cross-sectional dominance (87%): Cannot infer causality or directionality
    • Possible reverse causality: individuals with attention difficulties or distress may gravitate toward SFVs
    • Unmeasured confounds (e.g., personality, emotion regulation) may drive both SFV use and outcomes
  • Geographic concentration: 74% from Asia limits cultural generalizability
  • Platform focus: Only TikTok examined as specific platform; no studies on Instagram Reels, YouTube Shorts, etc.
  • Measurement variability:
    • SFV addiction scales most common but assess problematic use, not typical engagement
    • Frequency/intensity measures often adapted from other platforms or unvalidated
  • Content blindness: Studies assess use quantity but not content type (educational vs. entertainment vs. appearance-focused)
  • Covariate control: Only ~48% controlled for confounds; few (<10%) controlled for general social media use
  • Publication language: Primarily English-language sources may miss regional research

Cognitive outcomes:

  • Attention and inhibitory control over-represented; memory, reasoning, processing speed understudied
  • Few neuroimaging studies to elucidate mechanisms

Mental health outcomes:

  • Body image/self-esteem findings may reflect insufficient examination of content exposure patterns
  • Limited investigation of clinical populations

Methodological quality:

  • 41% rated low quality, primarily due to sampling issues (representativeness, nonresponse bias)

Strengths

  • Comprehensive scope: First meta-analysis to examine both cognitive and mental health correlates
  • Large sample: 98,299 participants across 71 studies
  • Platform inclusion: Beyond TikTok-only focus of prior reviews
  • Moderator testing: Age, platform, measurement type, health domain
  • Rigorous methods: Pre-registered protocol, systematic search, quality assessment, publication bias testing
  • Consistent patterns: Effects robust across sensitivity analyses (fixed vs. random effects, quality ratings, study design)

Implications

Clinical/Practical:

  • Cognitive health: Sustained attention and impulse control may be vulnerable to heavy SFV use
    • Implications for academic performance, occupational functioning, goal-directed behavior
    • May be relevant for individuals with ADHD or executive function difficulties
  • Mental health: Anxiety and stress show strongest associations
    • Clinicians should assess SFV use patterns when treating mood/anxiety disorders
    • Content exposure and user motivations likely shape outcomes
  • Public health: Findings support emerging concerns about digital media design and well-being
    • Platform-level interventions: usage reminders, content diversification, reduced algorithmic intensity
    • Digital literacy education: mindful engagement, awareness of design features
  • Body image: Lack of overall association suggests content matters more than platform
    • Body positivity content may buffer negative effects
    • Vulnerable individuals may need targeted support

Theoretical:

  • Habituation-sensitization framework supported for cognitive outcomes
  • Dopaminergic reward mechanisms plausible but require neuroimaging confirmation
  • Social comparison and contagion may operate differently on SFV platforms vs. photo-based social media
  • Uses and gratifications theory highlights importance of motivations and content choices

Policy:

  • Age restrictions (e.g., Australia's proposed minimum age) reflect growing regulatory concern
  • Evidence base insufficient to support blanket bans but warrants continued monitoring
  • Platform transparency and user control over algorithms may support healthier engagement

5. Visual Elements

Figure 1 (PRISMA flowchart): Shows systematic review process

  • 4,803 database records + 9 other sources → 2,495 after deduplication
  • 117 full-text articles assessed → 71 included (70 in meta-analysis)
  • Exclusions: no SFV focus (n=16), no relevant outcomes (n=21), no low/high comparison (n=5), no analyzable statistics (n=4)

Figure 2 (Forest plot: Cognitive domains):

  • Inhibitory control shows tightest confidence intervals and largest effects
  • Attention shows moderate heterogeneity across studies
  • Memory, working memory, language show weaker, more variable effects
  • Reasoning non-significant with wide confidence interval (1 study)

Figure 3 (Forest plot: Stress, Anxiety, Depression):

  • Stress and anxiety show consistent moderate negative effects across studies
  • Depression shows more heterogeneity but consistent direction
  • Confidence intervals largely non-overlapping with zero

Figure 4 (Forest plot: Sleep, Loneliness, Well-being, Affect):

  • Sleep quality shows moderate negative association
  • Loneliness shows weak to moderate effects with heterogeneity
  • Well-being and affect show weak, more variable associations

Figure 5 (Forest plot: Self-esteem, Body image):

  • Confidence intervals for both domains overlap zero
  • Body image shows slight positive skew (some studies show positive associations)
  • Self-esteem highly variable across studies

Table 1 (Summary of effect sizes and moderators):

  • Comprehensive overview of all meta-analytic findings
  • Presents mean r, 95% CI, p-values, and moderator Q statistics
  • Shows effect sizes stratified by age, platform, measurement, and domain

Key Concepts Reference

Name Definition Example
Short-form video (SFV) Video content lasting seconds to several minutes, typically delivered via infinite-scrolling interfaces with algorithmic personalization TikTok videos, Instagram Reels, YouTube Shorts
Habituation Reduced responsiveness to repeated exposure to a stimulus; in this context, desensitization to slower, effortful cognitive tasks after frequent high-stimulation content Heavy SFV users may find reading or studying less engaging due to habituation to rapid content switching
Sensitization Increased responsiveness to a stimulus through repeated exposure; here, heightened reward response to algorithmic content delivery Users become more drawn to checking SFVs for novelty and dopamine hits
P300 event-related potential Electrophysiological brain response (300ms post-stimulus) indexing attention and cognitive processing; reduced P300 suggests impaired attentional allocation Heavy SFV users show lower P300 amplitude during attention tasks, indicating reduced neural engagement
Inhibitory control Executive function ability to suppress prepotent responses and resist distractions Ability to stop scrolling when intending to study or resist checking SFVs during focused work
Dopaminergic reward system Brain circuitry (ventral tegmental area, nucleus accumbens) mediating motivation, pleasure, and reinforcement learning Unpredictable, personalized content triggers dopamine release, reinforcing repeated checking behavior
Social contagion Spread of emotions, behaviors, or beliefs through social networks or repeated exposure Viewing mental health content may normalize symptom expression or increase self-identification with disorders
Upward social comparison Comparing oneself to others perceived as superior, often leading to negative self-evaluation Comparing oneself to influencers with idealized appearances or lifestyles
Algorithmic curation Content recommendation driven by machine learning based on user engagement patterns TikTok's "For You" page delivers personalized videos optimized for maximal engagement
Goldilocks hypothesis Theory that moderate (not high or low) digital media use optimizes well-being Balanced SFV use may offer benefits (connection, entertainment) without displacing restorative activities
Effect size (r) Standardized measure of association strength; r = .10 (weak), .30 (moderate), .50 (strong) Cognitive: r = −.34 (moderate); Mental health: r = −.21 (weak)
Heterogeneity (I²) Proportion of variance in effect sizes due to true differences vs. sampling error I² = 95% indicates substantial true variability across studies
Publication bias Tendency for significant findings to be published more readily, skewing meta-analytic estimates Tested via funnel plots and Egger's test; minimal bias detected in this study
SFV addiction Problematic, compulsive SFV use characterized by loss of control, withdrawal, preoccupation, and functional impairment Measured via scales like Bergen Social Media Addiction Scale adapted for TikTok

Your Insights

How convincing is the evidence?

The evidence is moderately convincing for associations but weak for causality:

Strengths:

  • Large, diverse sample (98k+ participants across 71 studies)
  • Consistent pattern across multiple cognitive and mental health domains
  • Effects robust across sensitivity analyses
  • Dose-response gradient (addiction > duration > frequency)
  • Neuroimaging support for attention/reward mechanisms

Limitations:

  • Directionality ambiguity: Cross-sectional dominance cannot rule out reverse causality (people with poor attention/mood may seek SFVs)
  • Confounding: Few studies controlled for general social media use or personality traits that may drive both SFV use and outcomes
  • Effect sizes: Weak to moderate effects suggest SFV use is one factor among many
  • Content blindness: Not accounting for what users watch limits interpretability
  • Platform specificity: Results may not generalize to emerging SFV platforms

Overall: The consistency of findings across studies, age groups, and platforms provides reasonable confidence that associations are real, but the correlational nature means we cannot conclude SFV use causes cognitive or mental health problems. Longitudinal and experimental work is critical.

How does this fit with or challenge existing literature?

Fits with:

  • Social media research: Echoes findings linking general social media use to attention difficulties, sleep problems, and mood symptoms
  • Screen time literature: Aligns with evidence that excessive screen time displaces restorative activities and disrupts circadian rhythms
  • Dopamine/reward literature: Consistent with research on intermittent reinforcement and behavioral addiction

Challenges/Extends:

  • Platform differences: General SFV use showed stronger mental health associations than TikTok-specific use, suggesting multi-platform engagement may be riskier
  • Body image null finding: Contrasts with Instagram research linking photo-based platforms to body dissatisfaction. SFV content diversity (body positivity + idealization) may buffer effects
  • Age equivalence: No moderation by age challenges assumptions that youth are uniquely vulnerable. Adults and youths showed similar cognitive and mental health associations
  • Attention specificity: Stronger effects for attention/inhibitory control than memory/reasoning suggests SFVs may not uniformly impair all cognitive domains

Theoretical contributions:

  • Applies habituation-sensitization theory to digital media context
  • Distinguishes SFVs from traditional social media (photo/text-based)
  • Highlights importance of content diversity and user motivations

What assumptions underlie the work?

  1. SFV use is measurable and comparable: Assumes self-reported addiction scales, duration, and frequency capture meaningful engagement patterns across platforms
  2. Health indicators are unidirectional: Treats cognitive/mental health as outcomes, assuming SFV use precedes impairment (directionality assumption)
  3. Platform homogeneity: Grouping "general SFV use" assumes TikTok, Reels, Shorts function similarly
  4. Linear relationships: Assumes more SFV use = worse outcomes (doesn't test curvilinear/threshold effects)
  5. Content neutrality: Most studies don't measure content type, implicitly assuming usage quantity matters more than content quality
  6. Western constructs: Cognitive and mental health measures developed in Western contexts applied globally (74% Asian samples)
  7. User agency: Implicitly frames users as passive recipients of algorithmic influence, downplaying active content choices

What alternative explanations exist?

  1. Reverse causality:
    • People with ADHD or attention difficulties may prefer fast-paced SFVs
    • Anxious/depressed individuals may use SFVs for escape or emotional regulation
    • Those with lower executive function may struggle to disengage
  2. Third-variable confounds:
    • Personality: Low conscientiousness or high neuroticism may predict both SFV use and poor outcomes
    • Socioeconomic status: Stress and limited resources may drive both SFV escapism and mental health problems
    • Life circumstances: Academic pressure, social isolation, or trauma may precede and explain both
  3. Measurement artifacts:
    • Self-report bias: People experiencing distress may over-report problematic behaviors
    • Social desirability: Underreporting of SFV use due to perceived stigma
    • Recall bias: Heavy users may misestimate duration
  4. Selection effects:
    • Certain personality types gravitate toward SFV platforms
    • Cultural factors (e.g., collectivism, digital norms) shape both SFV adoption and well-being
  5. Content-specific mechanisms:
    • Educational/inspirational content may benefit cognition and mood
    • Appearance-focused or distressing content may harm specific individuals
    • Lack of content analysis means observed effects may be driven by subset of users/content
  6. Displacement vs. direct effects:
    • SFV time may displace sleep, exercise, socializing (indirect pathway)
    • Direct neurobiological effects (blue light, dopamine) may operate independently
    • Studies rarely disentangle these pathways

Future Research Directions

Gaps and Opportunities

Methodology

  1. Longitudinal designs: Track cognitive and mental health changes over time to establish temporal precedence
  2. Experimental studies: Manipulate SFV exposure (e.g., abstinence, restricted use) to test causality
  3. Neuroimaging: fMRI, EEG studies to clarify mechanisms (reward processing, attention networks, prefrontal regulation)
  4. Ecological momentary assessment: Real-time tracking of SFV use, mood, and cognition to capture dynamic relationships
  5. Content analysis: Code videos for themes (educational, entertainment, appearance-focused) and link to outcomes
  6. Multi-platform studies: Assess cumulative exposure across TikTok, Reels, Shorts simultaneously

Understudied Outcomes

  1. Cognitive domains: Fluid intelligence, processing speed, visuospatial ability, problem-solving
  2. Physical health: Physical activity, sedentary behavior, diet, BMI (preliminary evidence in supplemental materials)
  3. Sleep architecture: Polysomnography to measure REM, deep sleep disruption
  4. Academic/occupational functioning: GPA, work productivity, goal attainment
  5. Clinical populations: ADHD, anxiety disorders, depression (clinical samples vs. community samples)

Moderators and Mechanisms

  1. Content type: Educational vs. entertainment vs. appearance-focused; body positivity vs. idealized portrayals
  2. User motivations: Escapism, social connection, information-seeking, boredom relief
  3. Individual differences: Personality (conscientiousness, neuroticism), baseline cognitive function, emotion regulation
  4. Cultural context: Individualism vs. collectivism, digital norms, platform restrictions (e.g., China's Douyin vs. global TikTok)
  5. Active vs. passive use: Content creation vs. consumption; social interaction vs. scrolling

Interventions

  1. Platform-level: Test efficacy of usage reminders, time limits, content diversification, algorithm transparency
  2. User-level: Digital literacy programs, mindfulness-based interventions, cognitive-behavioral approaches
  3. Policy: Evaluate age restrictions, platform regulations, advertising limits

Theoretical Development

  1. Refine habituation-sensitization model: Test predictions about cognitive endurance, reward sensitivity
  2. Integrate uses and gratifications: Examine how motivations shape outcomes
  3. Develop SFV-specific theories: Distinguish from photo/text-based social media

Keywords and Research Areas

Primary terms: short-form video, TikTok, Instagram Reels, YouTube Shorts, Douyin, attention, executive function, depression, anxiety, social media addiction

Related areas:

  • Digital media and cognition
  • Dopaminergic reward processing
  • Algorithmic curation and behavior
  • Social comparison and body image
  • Sleep and circadian rhythms
  • Digital detox and media literacy
  • Adolescent development and digital media
  • Clinical applications of social media research
  • Neuroimaging of media use
  • Platform design and public health