Free scoping read. A written assessment of your manuscript in two business days, at no cost.

Start one
Contents

Academy · Research Data, Analytics & AI · Food Addiction Prevalence in Youth

A Meta-Analysis Evaluating the Prevalence of Food Addiction in Children and Adolescents

A 2021 systematic review and meta-analysis pooled 22 studies of more than 6,900 children and adolescents to estimate the prevalence of food addiction and the strength of its association with weight status.

Pubrica AcademyWritten by a research analyst Reviewed bya pediatric nutrition researcher Published10 October 2026 Last checked10 October 2026 Reading time6 minutes Sources10 peer-reviewed

Food addiction, as operationalized by the Yale Food Addiction Scale, describes a pattern of eating behavior that resembles substance-use disorder criteria: loss of control, continued use despite negative consequences, and withdrawal-like symptoms when certain foods are unavailable. Whether this pattern is common among children and adolescents — and whether it tracks with body weight — has been the subject of a growing number of cross-sectional studies using the child and adolescent version of the scale, the YFAS-C.

Yekaninejad and colleagues brought that literature together in a 2021 systematic review and meta-analysis published in Obesity Reviews, pooling data from 22 studies and 6,996 participants1. This article reports their findings.

Learning objectives

  • The pooled prevalence of food addiction found across 22 studies in children and adolescents
  • How prevalence differs between community samples and overweight/obese samples
  • Which measurement tool (YFAS-C) underlies these estimates, and its limitations
  • Why the review identified no longitudinal studies on this question
  • What the weight-status finding does and does not establish about causation

Common misconceptions

  • “Food addiction prevalence in children is a settled, precisely known number.” The pooled estimate carries a 95% confidence interval of 11-19%, and the review found only cross-sectional studies — no longitudinal data exist yet to show how prevalence changes over time or whether food addiction precedes or follows weight gain.
  • “Food addiction and obesity in youth are simply the same thing measured two ways.” The review found overweight/obese samples had a higher prevalence (19%) than community samples (12%), and weight status was associated with addiction severity — but weight status was only marginally associated with raw prevalence (p = 0.056), meaning the two measures are related but not interchangeable.

The study by Yekaninejad et al. (2021)

The meta-analysis in question — Yekaninejad MS, Badrooj N, Vosoughi F, Lin CY, Potenza MN, Pakpour AH, "Prevalence of food addiction in children and adolescents: A systematic review and meta-analysis," published in Obesity Reviews in 20211 — systematically searched the literature for cross-sectional studies that measured food addiction in children and adolescents using the Yale Food Addiction Scale for Children (YFAS-C) or its adapted versions. The final pool included 22 studies covering 6,996 participants. Notably, the authors reported that no longitudinal studies on the topic existed at the time of the review; every included study was a single cross-sectional assessment.

Pooled prevalence estimates

Across all 22 studies, the pooled prevalence of food addiction in children and adolescents was 15% (95% CI 11-19%)1. Splitting the pool by sample type produced two distinct estimates:

  • Community samples: 12% (95% CI 8-17%)
  • Overweight/obese samples: 19% (95% CI 14-26%)

In other words, roughly one in six to seven children in the general population met YFAS-C criteria for food addiction in these pooled studies, rising to roughly one in five among children already classified as overweight or obese.

Constructs measured by the YFAS-C

The Yale Food Addiction Scale for Children (YFAS-C) adapts diagnostic criteria modeled on substance-use disorder — loss of control over consumption, continued use despite negative consequences, tolerance, and withdrawal-like symptoms — to eating behavior in a pediatric population, typically reported by the child or a parent proxy depending on the study. Because every study identified by the review used a cross-sectional design, the YFAS-C-based prevalence figures describe symptom endorsement at a single point in time, not a clinical diagnosis confirmed through longitudinal follow-up.

Association with weight status

The clearest finding of the review concerned weight status. Weight status was significantly associated with food addiction severity (p = 0.002), and the association with food addiction prevalence specifically was close to, but did not cross, conventional statistical significance (p = 0.056)1. The authors' own conclusion, in their words, was that "healthcare providers should consider and address the high FA prevalence among pediatric population"1.

Absence of longitudinal studies in the evidence base

It is instructive to consider why every study identified by the review was cross-sectional, since this is itself informative about the maturity of the research area. A prevalence question (“how common is this pattern right now”) can be answered from a single-timepoint survey, which is comparatively quick and inexpensive to conduct in a school or clinic sample; a question about trajectory or causation requires a cohort followed over years, which is far more expensive and slower to produce publishable results. Research areas typically generate cross-sectional prevalence studies first, and longitudinal designs only once the prevalence question has established that the phenomenon is sufficiently common and stable to warrant follow-up. On the Joanna Briggs Institute’s evidence hierarchy, cross-sectional studies sit toward the lower, descriptive end, below cohort and case-control designs and well below any experimental design2 — which is a reason for appropriate caution, not for dismissing the 15% estimate: a well-conducted, adequately powered cross-sectional survey is still the correct design for a pure prevalence question, even though it cannot address what comes next. Individually, each of the 22 cross-sectional surveys is itself best reported against the STROBE checklist for observational studies, which specifies how a cross-sectional sample's recruitment, response rate and definition of exposure (here, YFAS-C-defined food addiction) should be described so a later meta-analyst can judge its quality at all 10.

Reporting and appraisal of prevalence reviews

A systematic review and meta-analysis of prevalence studies is expected to follow the same PRISMA reporting framework used for intervention reviews — a documented search strategy, a flow diagram tracking records from identification through final inclusion, and a pre-specified method for pooling the prevalence estimates across studies3. A reader deciding how much weight to give the pooled 15% figure can apply the Critical Appraisal Skills Programme's checklist for systematic reviews to check specifically whether the search was likely to have found all the relevant surveys, whether heterogeneity between the 22 studies' prevalence estimates was assessed, and whether the resulting confidence interval (11-19%) is precise enough to be clinically useful4. The GRADE approach supplies the complementary vocabulary for stating, explicitly, how much the cross-sectional-only evidence base and any inconsistency across the 22 studies should lower confidence in the pooled number, separate from whether the number itself is statistically precise5.

Position of pooled prevalence estimates in the evidence hierarchy

A systematic review and meta-analysis of cross-sectional prevalence studies still ranks above any single cross-sectional study in that same hierarchy, because pooling across 22 independent samples and roughly 7,000 children reduces sampling noise and tests whether the 15% figure replicates across different countries and recruitment settings — which several of the included studies in fact spanned. What pooling cannot do is upgrade the underlying design: no amount of meta-analytic pooling converts cross-sectional prevalence data into longitudinal or causal evidence. The International Committee of Medical Journal Editors' general guidance on abstracts applies squarely here: a report of this kind should state its main findings "with specific effect sizes and their significance" while explicitly "not overinterpret[ing] findings"6 — precisely the discipline Yekaninejad and colleagues followed in confining their own conclusion to a prevalence statement rather than a causal one.

Limitations of the current evidence

Two gaps merit attention from authors citing this evidence base. First, the absence of longitudinal studies means that the review cannot address whether food-addiction symptoms in childhood predict later obesity, persist into adulthood, or resolve spontaneously; it addresses only prevalence at a single point in time. Second, the marginal (not significant) association between weight status and raw prevalence (p = 0.056), alongside the significant association with severity, suggests that weight status and food-addiction symptoms are related but should not be treated as proxies for one another in study design or clinical screening.

Implications of a future intervention trial

Should this prevalence finding eventually motivate an intervention trial (for example, one testing a behavioral program aimed at reducing food-addiction symptoms in overweight adolescents), that trial would need to be designed and reported to an entirely different standard, following the CONSORT checklist for randomized trials rather than the observational conventions that govern a prevalence survey7. That shift, from describing a pattern to testing whether modifying it is effective, is precisely the gap that the current cross-sectional evidence base cannot close on its own, however precisely it measures the pattern itself. The CONSORT checklist is one of the design-specific reporting guidelines catalogued by the EQUATOR Network8, and whichever checklist a future trial follows, the Committee on Publication Ethics would expect its methods and any conflicts of interest around industry-funded nutrition research to be reported with the same transparency as the underlying science9.

Key points

Key points

Across 22 studies and nearly 7,000 children and adolescents, approximately 15% met YFAS-C criteria for food addiction, with higher prevalence in overweight or obese samples; however, every included study was cross-sectional, so the direction of any causal relationship remains unestablished.

Practical application

Practical application

  1. Prevalence should be reported by sample type rather than as a single overall figure. The distinction between 12% (community) and 19% (overweight/obese) carries more clinical and research meaning than a single pooled figure of 15% quoted alone.
  2. The cross-sectional limitation should be stated explicitly in any citing manuscript. Because the underlying evidence base contains no longitudinal studies, phrasing that implies a trajectory or causal sequence between food addiction and weight status should be avoided.
  3. Severity findings should be distinguished from prevalence findings in summaries. The weight-status association was significant for severity (p = 0.002) but only marginal for prevalence (p = 0.056); conflating the two overstates the strength of the prevalence association.

Frequently asked questions

22 cross-sectional studies covering 6,996 children and adolescents, all using the Yale Food Addiction Scale for Children (YFAS-C) or an adapted version of it.

15% (95% CI 11-19%) across all samples, with 12% in community samples and 19% in overweight/obese samples.

The review cannot answer this. Every included study was cross-sectional, so no longitudinal data exist yet to establish which comes first or whether the relationship is causal in either direction.

References

  1. Yekaninejad MS, Badrooj N, Vosoughi F, Lin CY, Potenza MN, Pakpour AH. "Prevalence of food addiction in children and adolescents: A systematic review and meta-analysis." Obesity Reviews. 2021;22(6):e13183. read 10 October 2026. View on PubMed Central
  2. Joanna Briggs Institute Levels of Evidence and Grades of Recommendation Working Party. JBI Levels of Evidence. 2014. read 10 October 2026. View source
  3. Page MJ, McKenzie JE, Bossuyt PM, et al. The PRISMA 2020 statement, hosted by the EQUATOR Network. read 10 October 2026. View on EQUATOR Network
  4. Critical Appraisal Skills Programme (CASP). CASP Systematic Review Checklist for appraising systematic reviews and meta-analyses of RCTs. read 10 October 2026. View source
  5. GRADE Working Group. Grading quality of evidence and strength of recommendations in health care. read 10 October 2026. View source
  6. International Committee of Medical Journal Editors (ICMJE). Recommendations for the Conduct, Reporting, Editing, and Publication of Scholarly Work in Medical Journals. Updated January 2024. read 10 October 2026. View source
  7. Hopewell S, Chan AW, Collins GS, et al. CONSORT statement: reporting guideline for randomised trials, hosted by the EQUATOR Network. read 10 October 2026. View on EQUATOR Network
  8. EQUATOR Network. Enhancing the QUAlity and Transparency Of health Research — library of reporting guidelines. read 10 October 2026. View source
  9. Committee on Publication Ethics (COPE). Core practices and guidance on research and publication integrity. read 10 October 2026. View source
  10. von Elm E, Altman DG, Egger M, Pocock SJ, Gøtzsche PC, Vandenbroucke JP; STROBE Initiative. The STROBE Statement, hosted by the EQUATOR Network. read 10 October 2026. View on EQUATOR Network

Every source above is a peer-reviewed article, a publisher’s record of one, or a standard issued by the body that owns it. Each carries the date it was last read. Where a standard moves, this article is revised rather than left up with a note.

How to cite this article

Pick your referencing style and copy it. Your university or target journal may vary the punctuation — check its guide if it has one.

Pubrica Academy. (2026, October 10). A Meta-Analysis Evaluating the Prevalence of Food Addiction in Children and Adolescents. Pubrica. https://pubrica.com/academy/concepts-definitions/a-meta-analysis-evaluating-the-prevalence-of-food-addiction-in-children-and-adolescents/

This article is written to be learned from. Cite it if you are referring to what it says; if you are using the evidence behind it, cite the sources listed above instead.

Citing This Meta-Analysis Correctly

A brief checklist for the accurate use of these figures.

[ ] Cite the pooled 15% figure with its 95% CI (11-19%), not as a bare point estimate
[ ] Specify community (12%) vs. overweight/obese (19%) subgroup when relevant
[ ] Note all 22 included studies were cross-sectional — no causal or longitudinal claims
[ ] Distinguish the severity association (p = 0.002) from the prevalence association (p = 0.056)
[ ] Attribute the measurement tool as YFAS-C, not a generic "addiction scale"

Pubrica AcademyWritten and maintained by Pubrica’s subject matter experts. Articles are revised when the underlying standards change, and each article states the date on which it was last reviewed.

How Pubrica can help

Pubrica’s research analysts verify every prevalence figure, confidence interval and p-value against the original source before inclusion in a manuscript, so that secondary citations of meta-analyses such as this one withstand reviewer scrutiny.

Find something else

Search the Pubrica Academy

Access is free and requires no account. Enter a topic, or open the full library to filter by category, sub-topic or resource type.

Verification of cited prevalence data against the original source

A scoping assessment is free and takes two business days. It provides a written assessment of what the work requires and which specialist would undertake it.

Get a free quote

Regulations

Check your project against this rule

  1. 1Your work
  2. 2Your details
  3. 3Send
What do you need?

Material you send is held under signed confidentiality, with access limited to the assigned team and encrypted transfer. It is not used to train anything and is not passed to third-party tools. Ask for the agreement before you send anything you would rather cover first.