Types and characteristics of DATs

FFQs dominated the landscape, accounting for 77% of all identified tools. The high prevalence of FFQs is consistent with their logistical advantages in large-scale, low-resource settings, such as lower respondent burden, cost-effectiveness, and ease of administration.^{(Reference Bailey10)} While food records and 24 HR recalls may offer more accurate, real-time intake data, their high variability and participant burden make them less commonly used in large-scale studies involving adolescents.^{(Reference Naska, Lagiou and Lagiou41,Reference Boushey, Harray and Kerr42)} These findings emphasise the methodological impact on energy intake reporting in young populations. Also, it is important to note that FFQs have historically been the primary method in large cohort studies globally, due to these same strengths.^{(Reference Cade, Thompson, Burley and Warm12,Reference Tabacchi, Garbagnati, Wijnhoven and Cairella43)} However, the continued reliance on FFQs in Latin America contrasts with recent trends in high-income countries, where 24 HR recalls and digital self-report tools are becoming more widely adopted to improve precision and respondent engagement.^{(Reference Amoutzopoulos, Steer and Roberts16,Reference Illner, Freisling, Boeing, Huybrechts, Crispim and Slimani44)}

In contrast to FFQs, only 9% of the DATs used 24 HR and another 9% employed qualitative tools such as illustrated checklists. While qualitative tools may reduce cognitive demands for younger children, their inability to quantify intake restricts their utility for nutritional epidemiology and public health planning. Semi-quantitative methods to estimate the amount of food consumed by children or adolescents were common (50%), with 40% using standard portion size estimations (kitchen measurements), 30% using food images, and 30% failing to specify the estimation approach.

DAT validation practices

Validation practices varied widely across studies. Of the 43 DATs, 70% reported a validation status, most often as relative validity in comparison against 24 HR recalls. Only one tool used biomarkers as a reference method, reflecting a gap in the use of objective validation approaches. However, it is important to recognise that biomarkers have limitations, they can be influenced by factors beyond dietary intake, including genetics, metabolic variations, and health status. Moreover, biomarker-based validation is often costly, logistically complex, and less feasible in large-scale or low-resource settings.^{(Reference Landberg, Karra and Hoobler45)} Four DATs did not specify the validation method at all, indicating a lack of reporting rigour that could compromise interpretability.

Across age groups, parent-reported FFQs were more common in studies involving children under 11 years, while self-report or combined parent–child report tools were used more frequently for older adolescents. Interestingly, relatively higher agreement levels (e.g., kappa values) were observed in validation studies when children or adolescents reported the intake. This could indicate that children and adolescents may have better knowledge of what they eat and are able to provide more details about their food consumption.^{(Reference Burrows, Truby, Morgan, Callister, Davies and Collins46,Reference Foster and Bradley47)} Although some studies reported relatively higher kappa values for FFQs completed by children or adolescents than for those completed by parents, this pattern should be interpreted with caution.^{(Reference Foster and Bradley47,Reference Van De Gaar, Jansen, Van Der Kleij and Raat48)} The validation evidence identified in this review was based predominantly on relative validity designs, most commonly comparisons with other self-reported dietary methods rather than objective gold-standard measures. In addition, reported kappa values were generally low to moderate, and differences across studies may reflect variation in age group, administration procedures, reference methods, recall demands, and context of food consumption rather than true differences in reporting accuracy. Therefore, the findings suggest an observed pattern in the available literature, but they do not establish that child or adolescent self-report is inherently superior to parent report. For younger children under 10 years, observer-based or a combined parent–child reporting may offer more reliable intake data, supporting existing evidence that memory and cognitive maturity also impact reporting accuracy.^{(Reference Livingstone and Robson2)}

Digital tools and access barriers

Interviewer-administered tools were most prevalent (58%), followed by self-administered tools (27%), while 15% of studies did not report the administration method. Despite the growing global use of web-based 24 HRs with automated coding, image capture, and mobile apps, only 19% (n = 8) of the DATs in this review were administered online. Examples of global research tools in this domain include ASA24, myfood24, and Intake24, which combine automation with user-friendly interfaces to reduce burden and enhance accuracy.^{(Reference Albar, Carter, Alwan, Evans and Cade49,Reference Albar, Alwan, Evans, Greenwood and Cade50)}

In Latin America, however, limited online use likely reflects broader infrastructural and socioeconomic challenges. Although smartphone ownership and internet access have grown rapidly across the region, disparities persist by country, income level, and urban–rural location. For instance, while smartphone penetration is above 70% in Brazil and Mexico, it is lower in parts of Central America and among rural populations.^{(Reference Robb, Bay and Vennegaard51,Reference Canterbury52)} Moreover, digital dietary tools impose a considerable burden on researchers, requiring access to cloud infrastructure, software customisation, technical support, and trained staff to manage and interpret complex data. These barriers contribute to the slow adoption of digital methods in the region, potentially exacerbating health data inequalities and limiting the scalability of innovative surveillance systems.^{(Reference Boushey, Spoden, Zhu, Delp and Kerr53)}

Variability in energy intake reporting

The forest plot of mean energy intake from 41 studies revealed substantial heterogeneity, with estimates ranging from approximately 1,200 to over 3,000 kcal/day. This variation reflects multiple contributing factors: differences in DAT design, administration method, age group, and geographic setting, as well as inconsistencies in portion size estimation and energy conversion protocols. Studies using self-report FFQs with older adolescents tended to show the highest intakes (>2,800 kcal), while those employing interviewer-administered 24 HR recalls with younger children reported the lowest (<2,000 kcal). Narrower confidence intervals in some cases suggest greater precision, often associated with larger sample sizes and standardised tools. The values summarised in this review were derived from studies conducted in diverse populations with different age distributions, countries, settings, study objectives, respondent types, and analytic approaches. As such, cross-study differences in mean energy intake cannot be attributed to the DAT alone. Higher or lower reported energy intakes may reflect population characteristics, contextual factors, or study design features as much as, or more than, the method itself. Therefore, the energy intake patterns presented in this review should be understood as descriptive of the existing literature rather than as evidence of the comparative performance of one dietary assessment method over another.

Strengths and limitations of the review

A key strength of this scoping review lies in its comprehensive approach to mapping DATs used among children and adolescents in Latin America, offering a valuable regional snapshot of methodological practices and gaps. It includes diverse study designs, settings, and age groups, and provides a structured overview of tool features, validation approaches, and digital adoption. Furthermore, its search strategy was comprehensive, addressing the Latin American context by including studies in English, Spanish, and Portuguese. The use of a registered protocol enhanced transparency and methodological rigour, and grey literature was considered through cross-referencing to reduce potential publication bias. To our knowledge, this is the first scoping review to systematically map DATs for children and adolescents across Latin America, highlighting the novelty and relevance of this synthesis.

However, limitations should also be acknowledged. First, the concentration of studies from Brazil and Mexico limits the generalisability of findings to underrepresented countries in Central and South America. Second, not all studies provided sufficient methodological detail, particularly concerning portion size estimation and validation procedures, which may affect interpretability. In addition, no formal quality appraisal was conducted, in line with scoping review methodology. Nevertheless, validation findings were interpreted cautiously by considering the completeness of reporting, the type of reference method used, and the extent of methodological detail provided in the original studies. This was particularly important given the substantial variability in reporting of validation procedures and outcomes across the included literature. Third, information on the underlying food composition tables used for dietary assessment was rarely reported; in many cases these may have relied on USDA or other non-local databases, which could affect accuracy in reflecting regional diets. Finally, variability in how dietary intake and validation outcomes were reported limited the ability to conduct direct comparisons across tools.