Adapting GloboDiet: A Comprehensive Guide to Standardized 24-Hour Dietary Recall Methods for Global Research

Abstract

This article provides a detailed guide on the adaptation and implementation of the GloboDiet software, a standardized 24-hour dietary recall methodology developed by the International Agency for Research on Cancer (IARC). Tailored for researchers and biomedical professionals, it covers the foundational principles of GloboDiet, step-by-step methodological processes for customization in diverse populations, solutions for common implementation challenges, and rigorous validation protocols. By synthesizing recent global case studies from Europe, Asia, and Africa, and discussing integration with emerging technologies, this resource aims to support the collection of high-quality, comparable dietary data for nutritional surveillance, epidemiology, and clinical research worldwide.

Understanding GloboDiet: Core Principles and Global Relevance for Dietary Surveillance

The GloboDiet software (formerly known as EPIC-Soft) represents a cornerstone in the field of standardized dietary assessment. Developed by the International Agency for Research on Cancer (IARC), a specialized agency of the World Health Organization (WHO), this computer- and interview-based 24-hour dietary recall (24-HDR) methodology was designed to collect high-quality, comparable dietary data across diverse populations and geographical regions [1] [2]. Its creation addressed a critical scientific challenge: the inability to accurately compare dietary intake data between different countries and studies due to methodological inconsistencies, which hampered epidemiological research on diet-disease relationships, particularly in large multinational studies like the European Prospective Investigation into Cancer and Nutrition (EPIC) [2] [3]. The core innovation of GloboDiet lies in its standardization concept, which systematically minimizes interviewer bias and random measurement errors through a highly structured interview protocol and harmonized databases [1] [4] [2]. This article provides a comprehensive overview of GloboDiet's origins, its underlying principles, and its performance as an indispensable tool for research, surveillance, and nutritional policy planning.

Historical Development and Core Architecture

The development of GloboDiet was driven by the specific needs of the EPIC study, one of the largest prospective cohort studies investigating the relationships between diet, lifestyle, and cancer across multiple European countries. The initial software, EPIC-Soft, was developed in the 1990s to serve as a standardized reference method for calibrating country-specific dietary questionnaires used across the 23 EPIC centers [3] [5]. This calibration was essential to correct for measurement errors and enable valid cross-country comparisons of diet-disease associations. The software's robustness and flexibility led to its rebranding as GloboDiet, reflecting its expansion beyond the European context to become a global dietary assessment tool [2] [3].

The software's architecture is built around approximately seventy interconnected databases, which are categorized into common databases and country-specific databases [2] [6]. This structure ensures a balance between international standardization and local relevance, a key factor in its successful global adaptation.

Table: Core Database Components of the GloboDiet Software

Database Type	Description	Key Functions
Common Databases	Centralized libraries shared across all GloboDiet versions	Ensure methodological standardization and comparability
Food & Recipe Classification	General classification lists for foods and recipes	Translation and addition of new (sub)groups
Facets & Descriptors	Pre-defined questions (facets) and possible answers (descriptors)	Systematic description and classification of foods
Probing Questions	Lists of foods commonly consumed together	Aid respondent memory during the interview
Quantification Methods	Standardized approaches for estimating food amounts	Adaptation of tools like photo series, standard units
Country-Specific Databases	Customized components reflecting local context	Capture dietary diversity within and between countries
Food & Recipe Lists	Comprehensive lists of locally consumed items	Compiled considering local dietary habits
Synonym & Brand Name Lists	Local names and commercial product brands	Facilitate accurate reporting and identification
Picture Books	Photo series depicting local portion sizes	Enable visual quantification of consumed foods
Coefficient Files	Edible portion, density, and cooking conversion data	Ensure accurate nutrient intake calculations

The GloboDiet interview follows a strict sequence to maintain standardization: 1) collection of general non-dietary information, 2) a quick list of all consumed items, 3) detailed description and quantification of reported foods and recipes, and 4) information on dietary supplements [4] [2]. This stepwise approach ensures that every interview, regardless of the interviewer or country, is conducted in a consistent manner.

The Standardization Concept: Principles and Implementation

The GloboDiet standardization concept is the software's most critical feature, designed to minimize both systematic and random errors that typically plague dietary recall data. This is achieved through multiple integrated strategies.

Structured Interview Protocol

The interview is guided by the software through a fixed sequence of steps, ensuring that every respondent is asked the same questions in the same order. The use of facets and descriptors—pre-defined, closed-ended questions and answers—standardizes the description of foods and recipes, minimizing interviewer variability in probing and recording responses [1] [2]. For example, a reported apple would be systematically described using facets such as "type" (with descriptors like Golden Delicious, Granny Smith) and "processing" (raw, stewed, canned).

Standardized Quantification Methods

A major source of error in dietary recalls is the inaccurate estimation of portion sizes. GloboDiet employs a variety of validated quantification tools, including country-specific picture booklets with photo series of different portion sizes, standard household measures, standard units (e.g., slice of bread), and food models [1] [2] [5]. These tools are critically adapted to reflect the local food supply and common consumption practices.

Harmonized Food Composition Data

To calculate nutrient intakes, GloboDiet is linked to standardized nutrient databases. The compilation of these databases follows rigorous protocols to ensure nutrient values are comparable. As exemplified by the creation of a standardized folate database for EPIC, this involves identifying a reference analytical method (e.g., microbiological assay for folate), matching foods across countries, and applying standard conversion factors [3]. This process is crucial for valid epidemiological studies investigating nutrient-disease relationships.

Global Adaptations and Performance

The flexibility of the GloboDiet architecture has allowed for its successful adaptation beyond Europe, demonstrating its utility in diverse cultural and dietary contexts. The following table summarizes key international adaptation projects.

Table: Global Adaptations of the GloboDiet Software

Region/Country	Key Adaptation Features	Primary Application & Performance
Republic of Korea	Addition of new food (sub-)groups and descriptors for specific Korean foods; adaptation of quantification methods and development of a picture book with local portion sizes [1].	First adaptation in an Asian context, confirming the software's flexibility and robustness for research and dietary surveillance [1].
Latin America (Brazil & Mexico)	Customization of ~70 databases; new descriptors for local foods; adaptation of quantification methods using local food package sizes and identified photos for country-specific picture booklets [2] [6].	Aimed at enabling dietary comparisons within and between Latin American countries for surveillance and research [2] [7].
Africa (Multi-country)	Evaluated by a panel of 29 experts who highlighted needs like describing local foods/recipes, culinary patterns (e.g., mortar pounding), and quantifying shared-plate eating [4] [8].	Positively evaluated for potential application across diverse African settings, setting a platform for improved dietary monitoring [4] [8].
Germany	Recent update involved adding ~600 new foods (e.g., vegan products) and deleting 525 obsolete items; updating quantification methods (e.g., coffee-to-go cups) and standard recipes [5].	A 2023 validation study (ErNst) using urinary biomarkers showed valid estimates for protein intake, supporting its use in national nutrition monitoring [5].

The adaptation process follows a standardized workflow to maintain core principles while incorporating necessary local context, as illustrated below.

Experimental Validation and Application Protocols

The validity and reliability of GloboDiet are rigorously tested, often using biomarker-based protocols.

Biomarker Validation Protocol

A key protocol for validating GloboDiet, as implemented in the German ErNst study, involves comparing nutrient intake from the recall with biological markers in urine [5].

• Objective: To validate the updated German GloboDiet version by comparing its estimates of protein and potassium intake with urinary excretion of nitrogen and potassium, which are considered suitable validation markers [5].
• Study Population: 109 participants (57 women, 52 men) across adult age groups (18-79 years) [5].
• Methodology:
  • 24-hour Urine Collection: Participants collected urine over a full 24-hour period. Completeness was checked using creatinine quotients [5].
  • GloboDiet 24-HDR: On the same day as the urine collection, participants underwent a face-to-face 24-hour dietary recall conducted by a trained interviewer using the GloboDiet software [5].
  • Data Analysis: Nutrient intakes from GloboDiet were linked to the German Nutrient Database (BLS). Statistical comparisons between intake and excretion included Wilcoxon rank tests, Spearman correlations, and Bland-Altman plots [5].
• Outcome: The study concluded that the updated GloboDiet provided valid estimates for protein intake. Results for potassium were ambiguous, suggesting a potential for underestimation, but overall, the software was deemed valid for national nutrition monitoring [5].

Expert Evaluation Protocol for Feasibility

Before adaptation in new regions, the methodology is often evaluated for feasibility by expert panels, as was done for Africa [4] [8].

• Objective: To evaluate GloboDiet as a possible common methodology for research and surveillance across Africa [4] [8].
• Expert Panel: 29 African and international experts in dietary assessment participated in a series of e-workshops [4] [8].
• Methodology:
  • Preparation: Experts reviewed core documents, a software presentation, and an interview simulation video [4].
  • E-Questionnaire: Experts completed an in-depth online questionnaire evaluating all sections of the GloboDiet interview [4].
  • Analysis: Feedback was collected on the software's structure, data collection approach, and potential to address local needs, including specific challenges like interviewing populations with low literacy and quantifying shared-plate meals [4] [8].
• Outcome: The evaluation was overall positive, supporting the flexibility and potential applicability of GloboDiet in diverse African settings and setting the stage for future implementation [4] [8].

The Scientist's Toolkit: Key Research Reagents

Table: Essential Materials for GloboDiet Implementation and Validation

Tool / Reagent	Function in Dietary Assessment	Application Context
GloboDiet Software Suite	Core platform for conducting standardized 24-hour dietary recalls.	Mandatory for all data collection in research and surveillance [1] [2].
Country-Specific Picture Booklet	Visual aid for quantifying portion sizes of local foods and dishes.	Used during the interview to improve accuracy of portion size estimation [1] [2].
Facet & Descriptor Library	Pre-defined questions and answers for the systematic description of foods.	Ensures standardized food classification and minimizes interviewer bias [1] [2].
Local Food Composition Database	Provides nutrient composition data for converting consumed foods into nutrient intakes.	Essential for calculating energy and nutrient exposures; requires harmonization for cross-country studies [2] [3].
Urinary Biomarkers (Nitrogen, Potassium)	Objective biological markers used to validate reported intakes of protein and potassium.	Used in validation studies (e.g., ErNst study) to assess the accuracy of the dietary method [5].
Standard Operating Procedures (SOPs)	Detailed guidelines for customizing databases, conducting interviews, and managing data.	Ensures consistency and quality assurance throughout the adaptation and data collection processes [1] [2].

The GloboDiet software, with its origins in the IARC-WHO collaborative framework, represents a paradigm shift in dietary assessment methodology. Its core strength lies in a rigorous standardization concept that permeates every aspect of its design—from the structured interview protocol and facet-based description system to the harmonized quantification methods and nutrient databases. This foundation has proven to be both robust and adaptable, as evidenced by its successful customization and positive evaluation across continents, from Asia and Latin America to Africa. By enabling the collection of high-quality, comparable dietary data, GloboDiet serves as a powerful tool for researchers and public health professionals. It is instrumental in understanding the global nutrition transition, monitoring the double burden of malnutrition, and investigating the complex relationships between diet and non-communicable diseases, thereby informing effective prevention strategies and policies worldwide.

Accurate and comparable dietary data is a cornerstone of public health nutrition, enabling researchers to monitor population intakes, understand diet-disease relationships, and evaluate the impact of interventions. The global nutrition transition, characterized by a shift from traditional diets to those high in processed foods, saturated fats, and added sugars, has been intimately associated with the rising burden of non-communicable diseases (NCDs) such as cardiovascular diseases, diabetes, and cancer, particularly in low- and middle-income countries (LMICs) [4]. This tangled situation is further complicated by the persistent prevalence of micronutrient deficiencies, creating a double burden of malnutrition that challenges fragile health economies [4].

Within this context, the standardization of dietary assessment methods is not merely a methodological preference but a fundamental prerequisite for generating reliable, comparable data across different populations and time periods. Without standardization, findings from individual studies cannot be confidently synthesized to inform global health initiatives or dietary guidelines. This application note explores the critical role of standardization, focusing on the implementation and adaptation of the GloboDiet 24-hour dietary recall software as a paradigmatic tool for enhancing nutritional surveillance and chronic disease research.

The Imperative for Standardized Dietary Methodology

The lack of standardized dietary tools and supporting research infrastructure remains a major obstacle to implementing concerted and region-specific research and action plans worldwide [4]. Different dietary assessment methods, even when applied to the same population, can yield vastly different pictures of dietary intake.

• Inconsistencies in Dietary Surveillance: A comparative study in Switzerland highlighted the profound impact of methodological choice. When dietary intake was assessed via a short set of food group questions (Swiss Health Survey), 20% of respondents met four or more Swiss food-based dietary guidelines. However, when assessed using two non-consecutive 24-hour dietary recalls (menuCH survey), the proportion meeting the same guidelines was a mere 2% [9]. This discrepancy underscores how non-standardized, cruder assessment methods can significantly overestimate diet quality, potentially leading to misguided public health conclusions.
• The Challenge of Misreporting: All self-reported dietary data are susceptible to error. A systematic review of studies using the doubly labeled water (DLW) technique as a reference standard revealed that under-reporting of energy intake is a pervasive issue across most common dietary assessment methods [10]. The degree of under-reporting is highly variable, but it is more frequent among females and can substantially bias the observed relationships between diet and health outcomes [10]. Another review detailed specific contributors to this misestimation, noting that omissions of consumed items are particularly high for vegetables (2–85% of the time) and condiments (1–80%), while beverages are omitted less frequently (0–32%) [11]. Standardized methodologies like GloboDiet incorporate probing techniques and systematic controls to help minimize such errors.

Table 1: Key Contributors to Misestimation in Self-Reported Dietary Intake

Contributor to Error	Description	Examples from Literature
Omission	Failing to report a food or beverage that was consumed	Vegetables omitted 2-85% of the time; Condiments omitted 1-80% of the time [11]
Portion Size Misestimation	Inaccurately estimating the quantity of food consumed	Both under- and over-estimation common for most food groups [11]
Intrusion	Reporting a food or beverage that was not consumed	Less frequently reported than omissions [11]
Misclassification	Incorrectly describing a food item (e.g., full-fat vs. low-fat milk)	Can lead to errors in estimating nutrient intake [11]

GloboDiet: A Standardized Framework for Dietary Assessment

The GloboDiet software, developed by the International Agency for Research on Cancer (IARC/WHO), represents a comprehensive approach to standardizing the 24-hour dietary recall (24-HDR). Originally developed and validated in Europe (where it was known as EPIC-Soft), it has since been adapted for use in regions including Latin America, Korea, and Africa [4] [1] [12]. Its primary strength lies in its design to collect highly detailed and standardized food consumption data while minimizing interviewer-induced bias.

The GloboDiet Interview Structure

The GloboDiet methodology is built around a rigorous, stepwise interview process conducted by a trained interviewer [4]. The core structure is as follows:

• General Information: Collection of non-dietary data about the interviewee and the recalled day.
• Quick List: The respondent freely recalls all foods and beverages consumed in the preceding 24 hours.
• Description and Quantification: Each item from the quick list is systematically described using predefined "facets" (questions) and "descriptors" (answers), and its portion size is estimated using standardized tools.
• Probing Questions: Specific, neutral questions are asked to prompt the memory for commonly forgotten items (e.g., sweets, beverages, additions to bread).
• Final Controls and Dietary Supplements: Consistency checks are performed, and information on dietary supplements is collected.

This structured approach ensures that every food item is described with a consistent and high level of detail, which is crucial for accurate coding and nutrient analysis.

The Customization and Harmonization Process

A key feature of GloboDiet is its modular architecture, which allows for both standardization and necessary localization. The software comprises approximately seventy common and country-specific databases related to foods, recipes, dietary supplements, and quantification methods [12]. The process of adapting GloboDiet to a new context follows established Standard Operating Procedures (SOPs) to ensure harmonization is achieved without compromising the core standardizing principles [1].

The adaptation process typically involves:

• Food Classification Customization: Adding new (sub-)groups to the common food classification to capture local foods (e.g., specific varieties of root vegetables in Africa) [4].
• Descriptor Expansion: Incorporating new descriptors to accurately classify and describe unique culinary patterns, such as "mortar pounding" in West African food preparation [4].
• Quantification Method Adaptation: Critically evaluating and adapting portion size estimation methods, including developing country-specific picture books of foods and dishes with relevant portion sizes [1] [12].
• Addressing Local Challenges: Developing specific solutions for interviewing populations with low literacy and for quantifying intake from shared plates and bowls, which are common eating practices in many cultures [4].

Table 2: Core Components of the GloboDiet Software for Standardized Dietary Assessment

Component Category	Key Elements	Function in Standardization
Interview Framework	Stepwise interview protocol (Quick List, Description, Quantification, Probing)	Ensures a consistent data collection process, minimizing interviewer bias [4].
Food Description	Facets (questions) and Descriptors (answers)	Systematically captures detailed information about each food (e.g., type, processing, fat content) for accurate coding [4].
Quantification Tools	Picture books, household measures, standard units	Aids the respondent in estimating portion sizes in a standardized way, improving accuracy of consumption data [9] [1].
Databases	Common and country-specific food, recipe, and supplement databases	Ensures consistent food identification and nutrient calculation while accommodating local diets [12].

Experimental Validation and Quality Assurance

The validity and reliability of any dietary assessment method are paramount. The GloboDiet methodology and its derivatives undergo rigorous testing using a variety of experimental protocols.

Validation Against Objective Biomarkers

The gold standard for validating self-reported energy intake (EI) is comparison against total energy expenditure (TEE) measured by the doubly labeled water (DLW) technique.

• Protocol for DLW Validation: In a typical study, participants are administered a dose of water containing stable isotopes (deuterium and oxygen-18). Urine samples are collected over a period of 7-14 days, and the enrichment of the isotopes is analyzed using isotope ratio mass spectrometry. The difference in elimination rates of the two isotopes is used to calculate carbon dioxide production and thus TEE [13]. In weight-stable individuals, TEE is assumed to be equivalent to EI.
• Findings for 24-Hour Recalls: Studies comparing online self-reported 24-hour recalls (like Intake24, a system based on GloboDiet principles) to DLW have shown a level of under-reporting comparable to interviewer-led recalls. One study found participants under-reported EI by 25% on average in their first recall, with limits of agreement being very wide (-73% to +68%), though precision improved with multiple recalls [13]. A systematic review confirmed that most dietary assessment methods, including technology-based ones, demonstrate significant under-reporting when compared to DLW [10].

Reliability and Convergent Validity Testing

• Test-Retest Reliability: The reliability of a single 24-hour recall for estimating intake is generally low due to high day-to-day variability in an individual's diet. For energy intake, the intra-class correlation coefficient (ICC) for a single recall can be as low as 0.35, meaning multiple recalls (e.g., two or more) are needed to estimate habitual intake for individuals [13].
• Convergent Validity: This involves comparing the method against another established dietary assessment tool. For instance, Intake24 has been shown to yield estimates comparable to interviewer-led 24-hour recalls, supporting its use as a lower-cost, lower-burden alternative [13].

A Protocol for Adapting GloboDiet to a New Regional Context

The following workflow outlines the key stages in adapting the GloboDiet methodology for research and surveillance in a new country or region, based on documented experiences from Korea, Latin America, and Africa [4] [1] [12].

Table 3: Key Research Reagent Solutions for Standardized Dietary Assessment

Tool / Resource	Function	Application in GloboDiet/Protocol
GloboDiet Software	The core computerized interview platform for conducting standardized 24-hour dietary recalls.	Provides the structured framework for data collection, description, and quantification [4] [12].
Standard Operating Procedures (SOPs)	Detailed, step-by-step instructions for customizing and administering the dietary recall.	Ensures methodological consistency and harmonization across different adaptation sites and studies [1].
Doubly Labeled Water (DLW)	A stable isotope-based biomarker used to measure total energy expenditure (TEE).	Serves as a reference method for validating the accuracy of self-reported energy intake [13] [10].
Graduated Portion Size Picture Booklets	Visual aids featuring series of photographs of foods in different portion sizes.	Used during the interview to help respondents estimate and report the quantity of food consumed more accurately [9] [1].
Harmonized Food Composition Database	A database linking consumed foods to their nutrient content.	Essential for converting reported food consumption into estimates of nutrient intake; requires harmonization for cross-country comparisons [14].

Standardization in dietary assessment is a critical, non-negotiable element for advancing the fields of nutritional surveillance and chronic disease research. The adaptation and implementation of standardized tools like the GloboDiet 24-hour recall methodology provide a viable path toward generating high-quality, comparable dietary data across diverse populations. This, in turn, is fundamental for reliably monitoring the nutrition transition, understanding the complex etiology of diet-related NCDs, formulating evidence-based dietary guidelines, and evaluating the impact of public health interventions on a global scale. The ongoing efforts to adapt this methodology in Africa, Asia, and Latin America represent a significant step forward in building the robust, collaborative infrastructure needed to address the double burden of malnutrition worldwide.

GloboDiet, formerly known as EPIC-Soft, represents the most extensive international effort to standardize 24-hour dietary recall methodology for multinational research and surveillance. Initially developed for the European Prospective Investigation into Cancer and Nutrition (EPIC), this computerized 24-hour dietary recall interview program was specifically engineered to standardize interviews across multiple countries and cultures [15] [16]. The system was originally deployed across 22 European centers, translated into nine languages, and contained common rules to describe, quantify, and probe approximately 1500-2200 foods and 150-350 recipes [16]. This foundational work established GloboDiet as the only available computerized 24-hour diet recall system developed to provide comparable food consumption data between several European countries at that time [16].

Within the European Food Consumption Validation (EFCOVAL) project, EPIC-Soft was substantially adapted and further developed to optimize its utility for pan-European dietary monitoring and risk assessment [15]. The software underwent complete reprogramming and migration to a modern Windows environment, including significant changes to its internal architecture and user interface [15]. These enhancements facilitated easier adoption across different countries and study contexts while maintaining the core structure and standardization concept that made the original EPIC-Soft methodologically robust.

Table 1: Core Evolution of GloboDiet Software Platform

Development Phase	Primary Geographic Focus	Key Technological Advancements	Major Application Scope
Initial Development (1990s)	22 European centers	Standardized computerized 24-h recall; Multi-language support	EPIC study - diet-cancer relationships
EFCOVAL Enhancements (2000s)	Pan-European monitoring	Migration to modern Windows environment; Improved architecture	Dietary monitoring and risk assessment
Global Expansion (2010s+)	Africa, Asia, Latin America	Enhanced adaptability for diverse food cultures	Nutrition surveillance in LMICs

GloboDiet Methodology and Standardization Framework

Core Interview Structure

The GloboDiet methodology employs a highly structured interview process consisting of multiple sequential steps designed to maximize completeness and accuracy of dietary data collection. The main sections of the dietary interview include: (i) collection of general information on the interviewee, (ii) generation of a quick list of all foods and recipes consumed during the preceding day, (iii) detailed description of each food item using predefined facets (questions) and descriptors (answers), (iv) quantification of consumption amounts, (v) probing questions to capture potentially forgotten items, (vi) final quality controls, and (vii) documentation of dietary supplements consumed [4]. This stepwise approach ensures comprehensive coverage of all dietary intake while maintaining standardization across different interviewers, populations, and geographic regions.

The description phase utilizes a unique system of "facets" and "descriptors" to capture detailed characteristics of each food item consumed. Facets represent specific questions about food attributes (e.g., preparation method, fat content, brand), while descriptors constitute the predefined possible answers [15] [16]. This structured approach ensures that the same food item is described consistently across different interviews and countries, enabling meaningful data comparison and aggregation.

Quantification Methods

A critical component of the GloboDiet methodology is its multifaceted approach to quantifying food consumption. The system incorporates numerous quantification methods including standard units, household measures, food shapes, and picture books with photo series displaying different portion sizes [17] [5]. The German adaptation, for example, provides approximately 3550 standard units, 100 photo series, about 50 different household measures, and 24 food shapes to assist with accurate quantification [17]. This comprehensive approach allows interviewers to select the most appropriate quantification method based on the specific food type and the respondent's ability to estimate portions.

Diagram 1: GloboDiet Standardized Interview Workflow

Global Applications and Adaptations

European Implementation

The initial implementation of GloboDiet (then EPIC-Soft) across 22 European centers demonstrated the feasibility of standardized dietary assessment across diverse food cultures. This multicenter deployment established common methods to classify and export dietary data, facilitating data exchange, comparison, and analysis across European countries [16]. The successful European implementation provided the proof-of-concept necessary for subsequent global expansion and established core protocols for maintaining methodological consistency while allowing necessary regional adaptations.

The European implementation identified several critical success factors for multinational dietary assessment, including the importance of standardized interviewer training, harmonized food classification systems, and common quality control procedures. These elements would later form the foundation for GloboDiet's adaptation in other world regions, providing a template for balancing standardization with necessary localization.

African Adaptation and Evaluation

The evaluation of GloboDiet for potential application in African settings represents one of the most systematic assessments of the methodology's transferability to low and middle-income countries. Through a consultative panel of 29 African and international experts in dietary assessment, researchers conducted an in-depth evaluation of the GloboDiet methodology across six e-workshop sessions [4]. Experts completed detailed e-questionnaires to evaluate all aspects of the software before and after participating in the e-workshop, providing comprehensive feedback on its potential applicability in diverse African contexts.

The evaluation revealed generally positive assessments of GloboDiet's main structure, stepwise approach for data collection, and standardization concept [4]. However, experts identified several critical areas requiring adaptation for African settings, including the need for better description of local foods and recipes, accommodation of particular culinary patterns (e.g., mortar pounding), development of appropriate quantification methods for shared-plates and shared-bowls eating situations, and specialized interviewing approaches for populations with low literacy skills, especially in rural settings [4]. These findings highlighted both the flexibility and the adaptation requirements of GloboDiet for valid application in African contexts.

Table 2: Regional Adaptation Requirements for GloboDiet Implementation

Geographic Region	Key Adaptation Requirements	Implementation Challenges	Notable Successes
Africa	Additional local food descriptors; Modified quantification for shared plates; Adaptation for low-literacy populations	Rural accessibility; Diverse culinary traditions; Resource limitations	Positive expert evaluation; Framework for capacity building
Republic of Korea	Integration of Asian food patterns; Local food composition data	Fundamental dietary pattern differences	Successful adaptation and implementation
Latin America (Brazil)	Local food items; Traditional preparation methods	Regional food diversity	Demonstration of cross-continent transferability
Germany	Regular food list updates; New portion size images; Modern food trends (vegan, convenience)	Keeping pace with rapidly changing food market	Successful validation with biomarkers

Asian and Latin American Applications

Beyond Africa, GloboDiet has demonstrated remarkable adaptability across fundamentally different dietary cultures, including successful implementation in Brazil and the Republic of Korea [4]. The Korean adaptation of GloboDiet represented the methodology's first Asian implementation and required significant modifications to accommodate characteristically different food patterns, preparation methods, and eating behaviors compared to European contexts [4]. The successful adaptation in Korea provided evidence that GloboDiet could be effectively modified for Asian food cultures, expanding its potential global applicability.

Similarly, the Brazilian implementation demonstrated GloboDiet's transferability to Latin American food cultures, further establishing its capacity for global deployment. These successful adaptations across three continents (Europe, Asia, and Latin America) provided compelling evidence of GloboDiet's flexibility while maintaining core standardization principles that enable meaningful cross-country comparisons.

Validation Protocols and Performance Metrics

Biomarker Validation Studies

The validation of GloboDiet implementations represents a critical component of its scientific rigor, with biomarker studies serving as the gold standard for assessing validity. The German validation study (ErNst study) employed a cross-sectional design with 109 participants (57 women and 52 men) to compare nutrient intake estimates from GloboDiet 24-hour recalls with biomarker measurements from 24-hour urine samples [17] [5]. The protocol specifically compared protein and potassium intake—known as eligible validation markers—with measured urinary excretion of nitrogen and potassium [17].

The validation protocol incorporated multiple statistical methods to assess agreement between intake and excretion, including Wilcoxon rank tests, confidence intervals, Spearman correlations, and Bland-Altman plots [17]. Participants provided complete 24-hour urine samples, with urinary creatinine excretion measured to verify completeness of collection. The mean creatinine quotient was 87% for men and 78% for women, exceeding the 60% threshold for inclusion [17]. This methodological approach ensured the reliability of the biomarker comparison and demonstrated the comprehensive nature of GloboDiet validation protocols.

African Evaluation Methodology

The African evaluation employed a structured expert consultation methodology to assess GloboDiet's applicability across diverse African settings [4]. The evaluation involved 48 African and international experts from universities and research institutes selected based on accomplished knowledge in dietary assessment, with empirical decision to include at least one African expert from each UN sub-region [4]. Experts received comprehensive preparatory materials including scientific papers describing the GloboDiet software, methodology presentations, and a 30-minute video simulating interviewer-interviewee interaction.

The evaluation utilized a detailed e-questionnaire developed using Wepi, a simplified online questionnaire authoring and publishing application for health professionals [4]. The questionnaire included nine parts covering all aspects of the GloboDiet interview, employing dichotomous questions, Likert scales, and open-ended questions to gather both quantitative and qualitative feedback. This rigorous evaluation methodology ensured comprehensive assessment of GloboDiet's potential for African implementation while identifying specific adaptation requirements.

Research Reagent Solutions: The GloboDiet Toolkit

Table 3: Essential Research Components for GloboDiet Implementation

Component Category	Specific Elements	Function in Dietary Assessment	Implementation Examples
Software Infrastructure	GloboDiet program; Food description facets; Standardized probes	Standardizes interview process; Ensures data comparability	EPIC-Soft → GloboDiet migration; Windows environment upgrade [15]
Food Composition Databases	Country-specific food lists; Nutrient composition data; Recipe calculators	Converts food intake to nutrient intake	~2000 food items in German version; Regular updates for market changes [17]
Quantification Tools	Picture books; Household measures; Food models; Standard units	Enables accurate portion size estimation	100 photo series; 50 household measures; 24 food shapes in German version [17]
Quality Control Systems	Interview controls; Final checks; Data management rules	Maintains data integrity throughout process	Integrated quality checks in interview workflow [17]
Training Materials	Interviewer manuals; Training videos; Standardized protocols	Ensures consistent implementation across sites	30-minute training videos for African evaluation [4]
Biomarker Validation Tools	24-hour urine collection; Doubly labeled water; Accelerometry	Provides objective validation of self-reported intake	24-h urine samples for protein/potassium validation [17]

GloboDiet's evolution from its origins as EPIC-Soft in the European EPIC study to its current global applications demonstrates the feasibility of standardized dietary assessment across remarkably diverse food cultures. The methodology has proven adaptable across continents while maintaining core standardization principles that enable meaningful cross-country and cross-region comparisons. The systematic approach to adaptation—incorporating local food items, preparation methods, and eating behaviors while preserving standardized interview structures and quality control procedures—provides a validated template for global dietary monitoring.

The successful implementation of GloboDiet across European, African, Asian, and Latin American contexts highlights its unique position as a truly global dietary assessment platform. As dietary patterns continue to evolve and the need for comparable nutritional surveillance grows, GloboDiet's standardized yet adaptable framework offers researchers and public health professionals a robust tool for understanding dietary determinants of health across diverse populations worldwide.

The GloboDiet methodology represents the state-of-the-art in standardized 24-hour dietary recall, developed by the International Agency for Research on Cancer (IARC/WHO) [18]. Originally validated and implemented across 19 European countries, this computer-assisted interview methodology is designed to collect highly detailed and comparable individual food consumption data [18]. Its core innovation lies in a structured architecture of interview steps, facets, and descriptors that systematically deconstruct the complex process of dietary assessment while maintaining standardization across diverse populations and cultural contexts.

The flexibility of the GloboDiet structure has been demonstrated through successful adaptations beyond Europe, including Latin America (Brazil and Mexico), the Republic of Korea, and ongoing evaluations for African settings [6] [1] [19]. This document provides a detailed technical deconstruction of the GloboDiet interview methodology, with specific application notes for researchers engaged in adaptation and implementation for dietary monitoring, surveillance, and nutritional research, particularly within the context of drug development and public health initiatives.

The GloboDiet Interview Protocol: A Stepwise Deconstruction

The GloboDiet 24-hour dietary recall follows a strict sequential protocol to ensure comprehensive and standardized data collection. The interview is structured into seven distinct sections, each with a specific function in the data acquisition process [18].

Protocol Sequence and Functional Description

Table 1: Core GloboDiet Interview Steps and Their Functions

Interview Step	Primary Function	Data Output	Key Considerations
1. General Information	Collect participant identifiers and context for the recalled day [18].	Participant code, date of birth, sex, anthropometrics, interview date, wake-up time, special diets, and special days (e.g., illness, travel) [18].	Establishes the interview context and allows for stratification of intake data by relevant non-dietary variables.
2. Quick List	Open-ended, cognitive approach to list all foods/beverages consumed in the preceding 24 hours [18].	Chronological list of consumption occasions, times, and places (e.g., home, workplace) [18].	Serves as a memory prompt; the interviewer does not probe for details at this stage.
3. Description of Foods & Recipes	Detailed description of each consumed item using a structured system of facets and descriptors [18].	Fully characterized food items, including preparation method, fat content, physical state, etc. (See Table 2).	The core standardization step, critical for ensuring data comparability. Relies on customized food databases.
4. Quantification	Estimation of the amount consumed for each described food item [18].	Consumed quantity using standardized methods: photos, household measures (HHMs), standard units (SU), or weight/volume [18].	Adaptation requires developing region-specific picture booklets and portion size estimation tools [6] [1].
5. Probing Questions	Checklist to prompt recall of easily forgettable foods (e.g., additions to tea or bread) [18].	Supplementary food items not reported in the Quick List.	Enhances completeness of the recall. Probes are linked to specific food items already recorded.
6. Final Controls	Automated and manual checks for data completeness and plausibility [18].	System warnings for aberrant daily energy/macronutrient values or unusually high volumes of specific foods [18].	Internal quality control mechanism to identify potential omissions or errors before finalizing the recall.
7. Dietary Supplements	Recording of any dietary supplements consumed in the preceding 24 hours [18].	Type, brand, and dosage of supplements.	Maintained in a separate, complementary database to the main food list.

Visual Workflow of the GloboDiet Interview

The following diagram illustrates the sequential flow and logical relationships between the core interview steps.

The Facet-Descriptor System: The Core Standardization Engine

The description step is the cornerstone of GloboDiet's standardization, transforming a simple food name into a highly structured and codified data point. This is achieved through a hierarchical system of facets (standardized questions) and descriptors (predefined, mutually exclusive answers) [18].

Hierarchical Structure and Application

The system functions by interrogating each food item through a series of relevant facets. The selection of a food item from the database triggers a specific pathway of facets that must be answered by choosing from a closed list of descriptors. This process ensures that the same food, consumed in different contexts or forms, is described consistently across all interviews and participants.

Table 2: Core Facets and Descriptors in the GloboDiet Description Step

Facet (Question)	Descriptor Examples (Answers)	Functional Role	Adaptation Notes
Food Preparation & Purchase	Prepared at home, commercial, restaurant, vending machine, fast food [18].	Identifies the origin and processing level of the food, which influences nutrient composition and potential contaminants.	Critical to add local market and street food descriptors for African and Latin American contexts [18] [6].
Cooking Method	Raw, fried, battered and fried, baked, sautéed, stewed, boiled, barbecued, steamed [18].	Critical for accurate nutrient estimation, as cooking can alter fat, water, and vitamin content.	Must be expanded to include methods like mortar pounding, identified as specific to African culinary patterns [18].
Physical State	Liquid, powdered, reconstituted from powdered [18].	Further refines the food form for accurate linkage to composition data.	Standard set often sufficient, but new states may emerge for novel products.
Fat Content	Whole, fat-reduced, light [18].	Allows for differentiation within food groups (e.g., dairy, meats) for precise nutrient assignment.	Requires validation against local product formulations and labeling regulations.
Type of Sugar	White, brown, unrefined [18].	Provides granularity for sweeteners and sugar-containing foods.	May need expansion with local sweeteners like panela or jaggery [6].
Food Source	Animal, vegetable, mixed [18].	High-level classification supporting food grouping and analysis.	Generally stable across adaptations.

Visual Logic of the Facet-Descriptor System

The following diagram maps the logical decision process for describing a food item using the facet-descriptor system.

Adaptation Protocols for Global Implementation

The GloboDiet structure is designed for cultural and culinary adaptation without compromising its core standardizing principles. The process involves systematic customization of its underlying databases and tools.

Database Customization and Food Classification

A primary task in adaptation is the expansion of the common food classification and descriptor lists to capture local specificity [6] [1].

Table 3: Key Databases Requiring Adaptation

Database Component	Adaptation Requirement	Exemplar from Previous Adaptations
Food Classification	Adding new (sub-)groups for locally unique foods [6] [1].	In Korea and Latin America, new subgroups were added to the common classification to accommodate specific foods without disrupting the existing structure [1] [6].
Descriptor Lists	Introducing new descriptors for existing facets to capture local varieties and preparation methods [18].	Experts evaluating GloboDiet for Africa proposed adding descriptors for culinary patterns like "mortar pounding" [18].
Recipe Database	Compiling and decomposing common local composite dishes into their ingredients [20].	The South Asia Biobank adaptation of Intake24 (a similar tool) involved developing a database of 2,283 items, including common recipes [20].
Quantification Methods	Creating country-specific picture booklets with relevant portion sizes and household measures (HHMs) [6] [1].	The Korean adaptation involved preparing "a picture book of foods/dishes... including new pictures and food portion sizes relevant to Korean diet" [1].

Protocol for Methodological Adaptation

The adaptation of GloboDiet for a new region should follow a rigorous, multi-stage protocol to ensure success:

• Initial Evaluation & Scoping: Conduct a consultative workshop with local dietary and public health experts to evaluate the methodology and identify specific needs, barriers, and opportunities [18]. This stage should map the local food supply and eating culture.
• Database Expansion: Following standard operating procedures (SOPs), customize the approximately seventy common and country-specific databases. This includes:
  • Food List and Classification: Populate with local foods and create new subgroups if necessary [6] [1].
  • Facet-Descriptor System: Augment descriptor lists to fully characterize local foods and preparation methods [18].
  • Recipe Compilation: Collect standard recipes for composite dishes and enter them with ingredient-level decomposition [20].
• Tool Development: Create region-specific quantification aids, primarily a picture booklet with photographs of foods and dishes at different portion sizes, as well as standardized household measures [1].
• Translation and Localization: Translate the software interface and all dietary probes into the local language, ensuring cultural appropriateness of terms [6].
• Training and Capacity Building: Implement rigorous training for interviewers and project staff to ensure faithful application of the standardized method. Knowledge transfer is a critical success factor [18] [19].
• Pilot Testing and Validation: Conduct a pilot study to test the adapted version, evaluate data quality, and identify any remaining issues before full-scale deployment.

The Scientist's Toolkit: Essential Research Reagents

Implementation of the GloboDiet methodology for research and surveillance requires a set of essential "research reagents" and tools.

Table 4: Essential Materials and Tools for GloboDiet Implementation

Item / Tool	Category	Function in Research	Notes
GloboDiet Software	Software	The core computer-assisted platform for conducting standardized 24-hour dietary recalls [18].	Licensed from IARC. Requires customization of underlying databases for the target population [6] [1].
Food Composition Database (FCDB)	Database	Links consumed, described, and quantified foods to their nutrient content for intake analysis [20].	Must be compatible with the local food supply. Often requires merging with international tables (e.g., FAO/INFOODS) [18].
Quantification Picture Booklet	Research Tool	Aids the participant in estimating the volume or size of consumed foods using photographs of standard portions [1].	Must be developed specifically for the target population using locally relevant foods and portion sizes [6] [1].
Standard Operating Procedures (SOPs)	Protocol	Detailed guidelines for interviewer training, data collection, processing, and management to ensure protocol adherence and data quality [18].	IARC provides core SOPs, which may be supplemented with country-specific instructions.
Dietary Supplement Database	Database	Records type, brand, and dosage of dietary supplements consumed for comprehensive nutrient intake assessment [18].	Maintained separately from the main food list.
Quality Control (QC) Metrics	Analytical Tool	Predefined metrics (e.g., recall completion time, number of items, missing foods) to monitor data collection quality [20].	For example, in the South Asia Biobank, 99% of recalls included >8 items, and 8% had missing foods [20].

A Step-by-Step Framework for Adapting GloboDiet in Diverse Populations

The GloboDiet software, developed by the International Agency for Research on Cancer (IARC/WHO), represents the international gold standard for conducting standardized 24-hour dietary recalls (24-HDR) in epidemiological research and nutritional surveillance [4] [21]. Its core strength lies in a flexible architecture that permits methodological harmonization across countries while accommodating diverse local food cultures. The adaptation process is a critical prerequisite for its deployment in any new region, ensuring that collected dietary data is both culturally relevant and scientifically comparable across international studies [21] [12]. This phase focuses on customizing the foundational elements of the software—the food and recipe databases—which directly impacts the accuracy and precision of dietary exposure assessment for research and public health policy.

The necessity of this phase is underscored by global health challenges. As many countries undergo a rapid nutrition transition characterized by a shift towards Westernized diets, the accurate monitoring of food consumption becomes essential for understanding the associated rise in non-communicable diseases (NCDs) [4] [21] [12]. Implementing a standardized tool like GloboDiet enables reliable dietary monitoring and the development of evidence-based interventions. The customization process ensures that this tool can capture everything from traditional culinary practices, such as mortar pounding in some African cultures, to the growing market of vegan and vegetarian products in Germany [4] [5].

GloboDiet Database Architecture: Common and Country-Specific Elements

The GloboDiet methodology is built upon a structured system of approximately seventy interlinked databases [21] [12]. These are strategically divided into common (or core) databases and country-specific databases, a design that ensures inter-country comparability while capturing local dietary specifics.

Table: Key Databases in the GloboDiet Customization Process

Database Type	Description	Role in Standardization	Customization Tasks
Common Databases	Shared backbone for all GloboDiet versions [21] [12].	Ensures data comparability within and between countries and regions.	Adaptation of the common food classification to include new, locally relevant food (sub-)groups [21] [7].
Food & Recipe List	Comprehensive list of foods, dishes, and beverages consumed in a country.	Captures the local food supply and consumption habits.	Compilation of a representative list based on national food consumption data, market surveys, and existing food composition tables [5] [21].
Facets & Descriptors	Systematic questions (facets) and pre-defined answers (descriptors) to classify and describe foods [21].	Standardizes the description of food characteristics (e.g., cooking method, fat content).	Addition of new descriptors required for local foods (e.g., specific cooking methods like "mortar pounding") [4] [21].
Quantification Methods	Tools to convert consumed foods into gram weights.	Allows for accurate and standardized portion size estimation.	Critical evaluation and adaptation of standard units, household measures, food shapes, and photo series to reflect local packaging and serving sizes [5] [21] [12].
Recipe Databases	Compositions of commonly consumed mixed dishes.	Enables the disaggregation of recipes into ingredients for nutrient calculation.	Collection of standard recipes, definition of yield factors, and determination of nutrient retention factors during cooking [21].

The following workflow outlines the major stages and decision points in the database customization process:

Detailed Protocols for Key Customization Experiments

Protocol 1: Adapting the Common Food Classification and Descriptors

This protocol details the process of expanding GloboDiet's core classification system to incorporate foods and culinary practices unique to a local diet, as demonstrated in adaptations for Korea, Latin America, and Africa [4] [21] [12].

• Objective: To integrate locally consumed foods and relevant descriptive characteristics into the standardized GloboDiet structure without compromising its harmonization principles.
• Materials and Reagents:
  • National food consumption survey data or dietary intake records from previous studies.
  • Existing national food composition tables (e.g., from the Korean Nutrition Society, Brazilian TBCA) [22] [21].
  • Inventory of traditional and commonly consumed mixed dishes with standard recipes.
  • GloboDiet common food classification and facet/descriptor library (from IARC).
• Experimental Procedure:
  • Food Item Compilation: Create a master list of all foods and beverages consumed locally, using national nutrition surveys, household budget surveys, and market surveys to ensure representativeness [21].
  • Classification Mapping: Systematically map each food item from the master list to the existing GloboDiet common food classification. For foods that do not fit into existing groups (e.g., unique fermented vegetables, specific cuts of meat), propose new (sub-)groups for approval and integration into the common system [21] [7].
  • Descriptor Gap Analysis: For each food, run through the standard GloboDiet interview facets (e.g., "cooking method," "preservation method," "fat content"). Identify any missing descriptors required to accurately describe local foods (e.g., "steamed" for Korean dishes, "mortar pounded" for African foods) [4] [21].
  • Descriptor Integration: Add the newly identified descriptors to the common facet system, ensuring they are mutually exclusive and comprehensively cover the local context.
  • Iterative Review: The adapted classification and descriptors are reviewed by a panel of local and international dietary assessment experts to validate comprehensiveness and maintain standardization [4].

Protocol 2: Customizing Food Quantification Methods

Accurate quantification of consumed portions is critical. This protocol outlines the development and adaptation of country-specific tools to estimate portion sizes, a task that has been central to adaptations in Germany, Korea, and Brazil [5] [21] [12].

• Objective: To develop and validate a set of quantification tools (photo atlas, standard units, household measures) that reflect the typical types and portion sizes of foods available in the local market.
• Materials and Reagents:
  • Digital camera and scale for high-resolution photography and accurate weight measurement.
  • Local market survey data on common food package sizes (e.g., mini, small, medium, extra-large dairy products) and produce sizes [5].
  • Standardized kitchenware and tableware (plates, bowls, cups, glasses, spoons) commonly used in households.
• Experimental Procedure:
  • Food Selection for Photography: Identify foods and mixed dishes that are frequently consumed and/or difficult to estimate visually. Prioritize foods with high variability in portion size.
  • Picture Booklet Development:
    • Prepare each food in a natural consumption state.
    • Photograph each food/dish in at least four different portion sizes (e.g., small, medium, large, extra-large) on typical local tableware [21] [12].
    • Weigh each portion to determine the exact gram weight it represents.
    • Compile the validated photographs into a country-specific picture booklet for use by interviewers during 24-HDR interviews [21] [23].
  • Standard Unit and Household Measure Cataloging:
    • Collect data on common, market-available standard units (e.g., a can, a slice of bread, a yogurt pot) and record their average weight [5] [21].
    • Catalogue common household measures (e.g., a cup, a tablespoon, a rice bowl) and determine their volume-to-weight conversion factors for key foods.
  • Database Linkage: Link each food item in the GloboDiet food list to the most appropriate quantification methods (photo series, standard units, household measures) within the software.

Protocol 3: Compiling and Managing Local Recipe Databases

Mixed dishes pose a significant challenge in dietary assessment. This protocol describes the creation of a standardized recipe database to ensure accurate nutrient intake calculations.

• Objective: To create a comprehensive database of standardized recipes for locally consumed mixed dishes, complete with yield factors, nutrient retention factors, and density coefficients.
• Materials and Reagents:
  • Recipe Manager application (a specialized module of GloboDiet) [21].
  • National recipe collections and cookbooks.
  • Food composition data for individual ingredients.
• Experimental Procedure:
  • Recipe Collection: Gather recipes for the most commonly consumed mixed dishes from national dietary surveys, cookbooks, and expert consultations.
  • Recipe Standardization: Define a standard recipe for each dish, including all ingredients, their quantities in grams (as purchased), and detailed preparation instructions.
  • Data Entry and Coefficient Calculation:
    • Enter the standardized recipes into the Recipe Manager.
    • Determine and assign appropriate coefficients for edible portion, weight change during cooking, and fat absorption where applicable [21].
    • Calculate yield factors to convert the weight of the raw ingredients to the weight of the final cooked dish.
  • Integration: The finalized recipes are integrated into the country-specific GloboDiet database, allowing interviewers to select a dish and have the software automatically disaggregate it into its constituent ingredients for nutrient analysis.

The Scientist's Toolkit: Essential Reagents and Materials

Table: Key Research Reagent Solutions for GloboDiet Database Customization

Item	Function in Customization	Exemplars from Search Results
National Food Composition Table (FCT)	Provides the nutrient profile for local foods; foundational for linking food consumption to nutrient intake.	Brazilian Food Composition Table (TBCA) [22] [23]; West-African Food Composition Table [4]; Korean Nutrition Society FCT [21].
Food Consumption Survey Data	Informs the selection of the most frequently consumed foods and portion sizes, ensuring the database is representative.	Data from previous national surveys used to prioritize foods in Brazil, Mexico, and Korea [21] [12].
Standardized Picture Booklet	Serves as a visual aid during the 24-HDR interview to improve the accuracy of portion size estimation.	Developed for Korea [21], Brazil [12] [7], and Germany [5] with locally relevant foods and portion sizes.
GloboDiet Software Suite	The core platform containing the common databases and structure to be customized. Includes the Recipe Manager for handling mixed dishes.	The IARC-WHO software, with versions adapted for Europe, Latin America, Asia, and Africa [4] [5] [21].
24-Hour Urine Collection	A biological validation method used to compare self-reported intake of specific nutrients (e.g., protein, potassium, sodium) with urinary excretion.	Used in the German ErNst study to validate protein and potassium intake estimated by GloboDiet [5].
Dietary Biomarkers	Objective measures of nutrient intake used to validate the dietary data collected by GloboDiet.	Protein (via urinary nitrogen), Potassium (via urinary excretion) used in validation studies [5].

Within the research on standardized 24-hour dietary recall (GloboDiet) adaptation methods, Phase 2 focuses on a critical and complex process: the customization of the core food classification system and food descriptors to achieve regional specificity. A GloboDiet adaptation is not merely a translation of the software interface but a profound methodological restructuring of its underlying databases to reflect the local food supply, culinary practices, and consumption habits. This phase ensures that the collected dietary data are accurate, relevant, and comparable both within the region and internationally. Failure to adequately adapt these elements introduces measurement error, misclassification of foods, and ultimately, compromises the validity of the dietary data for research, surveillance, and policy-making. This document outlines the detailed application notes and protocols for executing this phase, drawing from established adaptation projects in Latin America, Africa, and South Asia [6] [4] [7].

Core Adaptation Workflow

The process of adapting food classification and descriptors is systematic and iterative. The following workflow diagram illustrates the key stages and their relationships.

Quantitative Data from Prior Adaptations

The scope of adaptation varies significantly by region, necessitating careful resource planning. The following table summarizes the quantitative outputs from documented GloboDiet and similar 24-hour recall tool adaptation projects, providing a benchmark for researchers.

Table 1: Quantitative Benchmarks from Regional Adaptation Projects of 24-Hour Dietary Recall Tools

Project / Region	Adapted Food Items in Database	Key Adaptations and Additions	Primary Data Sources for Food List
GloboDiet Latin America (Brazil & Mexico) [6] [7]	~70 common and country-specific databases	New food subgroups and descriptors for local foods; adapted quantification methods; country-specific photo albums.	Local food composition tables, national dietary surveys, expert consultation.
Intake24 South Asia (Bangladesh, India, Pakistan, Sri Lanka) [20]	2,283 items	Comprehensive food database with portion sizes, food probes, and nutrient information reflective of diverse South Asian diets.	Local dietary studies, household data, expert consultation with nutritionists.
Intake24 New Zealand [24]	2,618 items	968 foods matched to the NZ Food Composition Database; 558 new recipes; inclusion of Māori, Pacific, and Asian foods; fortified food differentiation.	Australian Intake24 list, NZ Food Composition Database, national surveys, household purchasing data, ethnic community nutritionists.
GloboDiet Africa (Evaluation) [4]	Not specified (Pre-implementation)	Proposed: Description of local foods/recipes (e.g., mortar pounding); solutions for shared-plate eating and interviewing populations with low literacy.	Pan-African expert consultation via e-workshops and questionnaires.

Detailed Experimental Protocols

Protocol for Developing and Validating a Regional Food List

This protocol is foundational to the adaptation process, as demonstrated by the Intake24-New Zealand project [24].

Objective: To create a contemporary, representative food list that balances comprehensiveness with user burden, ensuring all commonly consumed foods are available for selection during the 24-hour recall.

Materials & Reagents:

• Baseline food list from a comparable region or previous national survey.
• Local Food Composition Database (e.g., New Zealand Food Composition Database).
• Data from previous national nutrition surveys or smaller dietary intake studies.
• Household food purchasing data (e.g., NielsenIQ Homescan).
• Access to supermarket websites and industry data for brand-specific information.

Procedure:

• Select Baseline List: Identify a suitable starting point. The New Zealand team used the Australian Intake24 food list due to similarities in food supply, which provided an efficient foundation [24].
• Categorize Foods: Review the food list at a category level (e.g., Dairy, Grains, Mixed Meals). This ensures systematic coverage across all food types.
• Identify Local Foods: Populate each category with region-specific foods. Utilize:
  • Local food composition databases.
  • Dietary intake studies from specific sub-populations.
  • Consultation with dietitians and nutritionists familiar with ethnic minority communities (e.g., Māori, Pacific, and Asian foods in New Zealand) [24].
  • Market sales data for fast-moving consumer goods like breakfast cereals.
• Refine and Consolidate: Critically evaluate the range of foods within each sub-category.
  • Add foods that are nutritionally distinct (e.g., plant-based milk alternatives with added calcium) or are major contributors to nutrient intake.
  • Merge similar foods into a single generic item (e.g., "red apples, not further defined") to reduce participant burden [24].
  • Remove foods not available in the local market or that are redundant.
• Link to Nutrient Data: Match each food item in the final list to a corresponding nutrient line in the local food composition database. This is a prerequisite for automated nutrient intake calculation.

Protocol for Reliability Testing of Food Classification (NOVA)

When integrating a food classification system like Nova, testing the reliability of the coding process is essential, as exemplified by Neri et al. [25].

Objective: To ensure that trained coders can consistently and accurately assign individual foods to the correct Nova processing category (unprocessed/minimally processed, processed culinary ingredients, processed foods, ultra-processed foods).

Materials & Reagents:

• A dataset of unique food items extracted from 24-hour dietary recalls.
• Detailed Nova classification guidelines (e.g., Monteiro et al., 2019 [25]).
• Trained and certified coder pairs.

Procedure:

• Coder Training: Train and certify coders on the Nova classification system using the standardized guidelines. This includes addressing nuanced definitions and complex multi-ingredient foods [25].
• Independent Categorization: Assign coder pairs to independently categorize the same set of unique food items into one of the four Nova categories.
• Assess Inter-rater Reliability:
  • Calculate the percent concordance between coder pairs (e.g., 88.3% concordance was achieved in a study on children's diets) [25].
  • Calculate Cohen’s κ coefficient to measure agreement beyond chance. A κ of 0.75, as reported in the same study, indicates substantial agreement [25].
• Resolve Discrepancies: Establish a protocol for resolving discordant codes, typically through expert consensus or a third coder.
• Establish Analytic Variables: Once reliable coding is achieved, merge the final Nova categories back into the full dietary dataset to create variables such as "number of ultra-processed calories per day" for subsequent analysis [25].

Protocol for Validating Adapted Tools with Biomarkers

After adaptation, the entire system's performance must be validated. The following protocol is based on the validation of the updated German GloboDiet version [5].

Objective: To validate the nutrient intake estimates generated by the adapted GloboDiet system by comparing them against objective biomarkers.

Materials & Reagents:

• Adapted GloboDiet software linked to the local nutrient database.
• 24-hour urine collection kits and containers.
• Accelerometers for physical activity measurement (for energy validation).
• Trained interviewers and laboratory facilities for urine analysis.

Procedure:

• Study Design: Conduct a cross-sectional study with approximately 100 participants, stratified by sex and age groups, to ensure a representative sample [5].
• Data Collection:
  • Conduct a face-to-face 24-hour dietary recall using the adapted GloboDiet tool on a pre-defined day.
  • Collect a 24-hour urine sample from the participants over the exact same 24-hour period.
  • Analyze urine for nitrogen (to validate protein intake) and potassium content [5].
• Data Analysis: Compare the estimated nutrient intake from GloboDiet with the urinary excretion values using:
  • Wilcoxon rank tests to check for significant differences.
  • Spearman correlations to assess the strength of the relationship.
  • Bland-Altman plots to evaluate the agreement and identify any bias [5].
• Interpretation: Valid protein intake is confirmed when no significant difference is found between intake and excretion, and statistical measures show good agreement. For potassium, results can be more ambiguous, and a combination of metrics should be used for evaluation [5].

The Scientist's Toolkit: Research Reagent Solutions

The following table details essential "reagents" – the core databases and classification systems – required for the successful adaptation of food classification and descriptors.

Table 2: Key Research Reagents for Food Classification and Descriptor Adaptation

Research Reagent	Function in Adaptation	Examples & Notes
Common Food Classification (e.g., GloboDiet's FCO)	Provides the international standardized structure for grouping and describing foods, enabling future harmonization [6] [7].	New sub-groups are added to this common framework to accommodate unique local foods without breaking the standardized core.
Local Food Composition Database (FCDB)	Serves as the authoritative source for nutrient profiles of local foods and is essential for linking the food list to nutrient data [26] [24].	E.g., New Zealand Food Composition Database [24], German Nutrient Database BLS [5]. Must be updated regularly to reflect market changes.
Nova Food Classification System	A framework for categorizing foods by level of industrial processing, used to investigate diet-health relationships [25] [27].	Requires reliable coding procedures. Tools like Nova24h are being developed to automatically classify foods within recalls [27].
Standardized Portion Size Measurement Tools	Critical for converting reported food consumption into quantifiable amounts. These must be culturally appropriate.	Includes country-specific picture booklets [6], photo albums [7], standard units, and household measures adapted to local tableware [5].
Reference Nutrient Database (e.g., USDA SR)	Often used as a primary source or for cross-referencing when local data is missing, but must be carefully matched to local foods [28].	An algorithm comparing energy and key nutrients can be used to select the closest match from the USDA database to a local food item [28].

Within the systematic adaptation of the GloboDiet 24-hour dietary recall methodology, the development of culturally relevant quantification methods is a critical phase that directly impacts the accuracy and reliability of collected dietary data [29]. This phase involves customizing tools like photo books and standard units to reflect the local food landscape, enabling respondents to report consumed portions accurately. Experience from adaptations in Europe, Latin America, and Asia demonstrates that meticulous customization of these components is essential for valid dietary intake assessment in diverse cultural contexts [1] [6] [5].

Core Components of Quantification Methods

The GloboDiet software utilizes a suite of complementary quantification aids. The table below outlines the key components and their functions.

Table 1: Core Components of the GloboDiet Quantification System

Component	Description	Primary Function	Cultural Adaptation Example
Photo Book	A collection of photo series showing various portion sizes for common foods and dishes [30] [5].	Helps respondents visualize and select the portion size they consumed.	Korea and Brazil developed new pictures of local dishes and portion sizes [1] [7].
Standard Units (SU)	Common, culturally recognizable units for foods (e.g., an apple, a slice of bread, a can) [29] [5].	Provides a quick, standardized way to quantify items often consumed in discrete units.	Germany added mini fruits and extra-large dairy products; Brazil customized units for local market foods [6] [5].
Household Measures (HHMs)	Measured utensils like graduated spoons, cups, bowls, and plates [29] [5].	Used to quantify foods and liquids consumed via household utensils.	Africa identified a need to adapt HHMs for shared-plate eating situations [29].
Food Shapes	Two-dimensional drawings representing geometric shapes of food portions (e.g., a wedge of pie) [29].	Aids in quantifying fractions of a whole food item.	A common method across adaptations, but the specific shapes may be updated [5].
Weight/Volume Method	Direct entry of the consumed amount in grams (g) or millilitres (ml), often from product labels [29].	Allows for precise quantification when the exact weight or volume is known.	Relies on local food packaging and common market weights [1].

Experimental Protocols for Development and Validation

Protocol for Developing a Country-Specific Photo Book

The development of a photo book for dietary assessment, as implemented in studies like menuCH and the German update, follows a rigorous multi-stage process [30] [5].

Objective: To create a validated visual aid for portion size estimation that reflects the typical foods, dishes, and serving sizes consumed by the target population.

Workflow: The following diagram illustrates the end-to-end workflow for developing a country-specific photo book.

Detailed Methodology:

• Food Item Selection: Analyze existing national dietary survey data (e.g., 24-hour recalls) to identify the foods and mixed dishes that contribute most significantly to energy and nutrient intake in the population [30]. For Switzerland's menuCH survey, this process identified 119 series of portion-size images [30].
• Portion Size Determination: For each selected food, calculate the distribution of consumed portion sizes. Typically, photographs represent a range from a small (e.g., 5th percentile) to a large (e.g., 95th percentile) portion, often based on the median and standard deviation of reported intakes [30].
• Photography Protocol: Conduct photoshoots under controlled, consistent lighting conditions. Use a color chart for calibration and include a reference object for scale. Each food is photographed at multiple portion sizes, which are precisely weighed and recorded [30].
• Validation: Conduct a study where participants serve themselves a typical meal, which is then photographed and weighed. Compare the estimated portion from the photo to the actual weight to assess accuracy and bias [30]. In Brazil, a specific validation study for its photo album was planned during the adaptation [7].
• Integration: The final, validated photo series are compiled into a book or digital database and linked directly to the corresponding food items within the GloboDiet software [30].

Protocol for Updating Standard Units and Household Measures

The German adaptation of GloboDiet provides a clear protocol for updating standard units and household measures [5].

Objective: To ensure the quantification methods reflect the current food market and common consumption practices.

Methodology:

• Market Survey: Systematically survey the food retail environment (supermarkets, local markets) to document common packaging sizes, unit sizes (e.g., fruits, baked goods), and the availability of new products (e.g., plant-based alternatives, coffee-to-go cups).
• Review of Existing Databases: Critically evaluate pre-existing GloboDiet SU and HHM databases against the market survey data. Flag items that are obsolete or no longer relevant.
• Add New Items: Introduce new SUs and HHMs to cover gaps. The German update, for instance, added mini and extra-large sizes for dairy products and fruits, and new international dishes like sushi [5].
• Link to Food Items: Associate the updated and new SUs and HHMs with the relevant foods in the GloboDiet food list. The final German version included approximately 3550 standard units and 50 different household measures [5].

Protocol for Validation Studies

Validation is crucial before deploying new quantification methods in large-scale studies. The ErNst study in Germany and performance studies in Brazil offer robust models [5] [31].

Objective: To assess the accuracy of the overall dietary intake data collected using the adapted GloboDiet version.

Methodology (Biomarker Validation):

• Design: A cross-sectional study where participants complete a 24-hour dietary recall using GloboDiet and concurrently collect a 24-hour urine sample [5].
• Participants: Recruit a convenience sample of healthy adults, stratified by age and sex. The ErNst study aimed for approximately 50 men and 50 women [5].
• Biomarkers:
  • Protein Intake: Validate against urinary nitrogen excretion, using a conversion factor (e.g., N × 6.25) [5].
  • Potassium Intake: Validate against urinary potassium excretion [5].
• Statistical Analysis:
  • Wilcoxon signed-rank test to check for significant differences between intake and excretion.
  • Spearman correlation coefficients to assess the strength of the relationship.
  • Bland-Altman plots to evaluate the limits of agreement and identify any systematic bias [5].

Table 2: Key Reagents and Materials for Validation Studies

Category	Item	Specification / Function
Biological Sample Collection	24-hour Urine Collection Kit	Includes containers, instruction sheet for participants, and cold-chain storage solutions.
Laboratory Analysis	Assay Kits for Urinary Nitrogen & Potassium	Standardized kits for colorimetric or ion-selective electrode analysis.
Dietary Assessment	GloboDiet Software & Adapted Tools	The fully adapted software, photo book, and SU/HHM databases.
Anthropometry	Digital Scales & Stadiometer	To measure body weight and height for BMI calculation and creatinine estimation.
Physical Activity Monitoring	Accelerometer (e.g., ActiGraph)	To estimate total energy expenditure for energy intake plausibility checks [5].

The quantitative outcomes from various adaptation projects highlight the scope of work involved in customizing quantification methods.

Table 3: Summary of Quantification Method Adaptations Across Countries

Country/Region	Photo Book Details	Standard Units & Household Measures	Key Adaptations & Outcomes
Germany (Updated) [5]	~100 photo series.	~3550 Standard Units, ~50 Household Measures, 24 Food Shapes.	Added vegan/vegetarian products, international dishes (sushi), various coffee cup sizes. Validation showed valid estimates for protein intake.
Switzerland (menuCH) [30]	119 series of 5-6 portion-size images.	~60 actual household measurements.	Used a data-driven approach from a national survey. A picture book was provided to participants for quantification.
Brazil [7]	Country-specific photo album developed and validated.	Customized based on POF 2008-2009 and ISA-Capital surveys.	Quantification methods adapted considering food packages available in the Brazilian market. Showed good agreement with NDSR software [31].
Korea [1]	New pictures developed with food portion sizes relevant to the Korean diet.	Quantification methods critically evaluated and adapted.	New descriptors and classification for Korean foods; confirmed flexibility of GloboDiet for Asian contexts.
Africa [29]	Identified as a necessary future development.	Not specified, but shared-plates and low-literacy settings noted as challenges.	Experts proposed adaptations for specific culinary patterns (e.g., mortar pounding).

Discussion

The consistent success of GloboDiet adaptations across continents underscores that while the core software structure and standardization principles are robust, the quantification methods must be highly localized [1] [29] [6]. A primary challenge, particularly evident in the African evaluation, is adapting methods for populations with low literacy and for eating cultures involving shared plates and bowls [29]. This may require developing additional interviewer training and novel visual aids beyond standard photo books.

Furthermore, quantification is not a one-time task. The German experience demonstrates the necessity of periodic updates to reflect a rapidly changing food supply, driven by factors like the rise of plant-based alternatives and international cuisine [5]. Ultimately, investing in this phase is foundational for generating high-quality, comparable dietary data that can inform public health policy, nutritional epidemiology, and the assessment of drug-diet interactions on a global scale.

This application note details the successful adaptation of the GloboDiet software, a standardized 24-hour dietary recall methodology, for research and surveillance in the Republic of Korea. The adaptation process demonstrated the flexibility and robustness of the GloboDiet system, marking its first successful expansion from a European to an Asian context. The project involved the comprehensive customization of approximately seventy common and country-specific databases to accurately capture the unique aspects of the Korean diet, which is characterized by specific staple foods, preparation methods, and eating habits. The key outcomes include a validated Korean version of GloboDiet and a set of protocols that can inform future adaptations of standardized dietary assessment tools in other non-European regions. This work underscores the critical importance of methodological standardization, coupled with strategic localization, for generating high-quality, comparable dietary data in nutritional epidemiology and public health surveillance.

The rapid nutritional and economic transition in the Republic of Korea, often termed 'the Miracle on the Han River,' has been accompanied by a significant shift in dietary patterns from traditional to more Westernized diets [21]. This transition is of considerable public health importance, as such dietary changes are frequently associated with an increased risk of cancer and other non-communicable diseases (NCDs) [1] [21]. Consequently, precise dietary monitoring is essential for understanding these associations and informing effective public health interventions.

A major challenge in tracking dietary changes across countries is the lack of standardized dietary assessment methodologies, which limits the ability to harmonize data and implement concerted international research and surveillance actions [21]. The GloboDiet software (formerly EPIC-Soft), developed at the International Agency for Research on Cancer (IARC), was designed to address this challenge. Initially deployed and validated across 23 centers in ten European countries, GloboDiet serves as a highly standardized, computer-driven 24-hour dietary recall program [1] [21].

This case study documents the process, key learnings, and outcomes of adapting the GloboDiet methodology for the Korean population. The project's primary objective was to evaluate the feasibility of customizing this international tool for an Asian context without compromising its core principle of standardization, thereby enabling the collection of dietary data that is both contextually relevant and internationally comparable [1].

Background: The Korean Dietary Context

The Korean diet has undergone a profound transition, moving beyond a simple binary of 'Traditional' and 'Western' patterns. Recent analyses of national data (2007–2022) have identified three major dietary patterns among Korean adults:

• Traditional Pattern: Characterized by high intake of white rice, vegetables, and kimchi [32].
• Red Meat & Alcohol Pattern.
• Flour-based Foods & Sweets Pattern [32].

Since 2013, the Flour-based Foods & Sweets pattern has become the most prevalent, particularly among younger adults aged 19-29 years and women [32]. This shift towards Westernized patterns has significant health implications; adherence to the Red Meat & Alcohol pattern in men is associated with significantly higher odds of cardiometabolic risk factors, including hyperglycemia, high blood pressure, and obesity [32]. This evolving dietary landscape creates an urgent need for precise assessment tools to monitor trends and associated health outcomes.

Table 1: Major Dietary Patterns in Korean Adults (KNHANES 2007-2022)

Dietary Pattern Name	Core Characteristics	Key Associated Foods	Population Groups Where More Prevalent
Traditional	High intake of staple Korean foods	White rice, vegetables, kimchi	Older age groups
Red Meat & Alcohol	High intake of animal proteins and alcohol	Red meat, alcoholic beverages	Men
Flour-based Foods & Sweets	High intake of refined carbohydrates and sugars	Flour-based foods, sweets, sugar-sweetened beverages	Adults 19-29 years, Women

Adaptation Methodology and Experimental Protocol

The adaptation of GloboDiet for Korea followed established international Standard Operating Procedures (SOPs) and guidelines to ensure standardization was maintained throughout the customization process [1] [21]. The work was conducted collaboratively between the International Agency for Research on Cancer and the National Cancer Center of Korea.

The adaptation process was systematic and multi-stage, as outlined below.

Protocol: Database Customization and Management

The backbone of the adaptation involved customizing and translating approximately seventy common and country-specific databases.

• Common Databases: These form the core of GloboDiet's standardization and include food and recipe classifications, facets, descriptors, and quantification methods. The existing common classification was adapted to accommodate Korean foods, and new descriptors were added to the pre-defined facets to capture specific local food characteristics [1] [21].
• Country-Specific Databases: These were developed de novo to capture the uniqueness of the Korean diet. The process involved:
  • Food List Generation: A comprehensive list of foods and recipes was prepared based on data from the Korean Nutrition Society, focusing on frequency of consumption [21].
  • Recipe Management: Traditional Korean recipes were broken down into their ingredients using the specialized GloboDiet "Recipe Manager" application. This was critical for accurately estimating nutrient intake from complex dishes [21].
  • Nutrient Databank Linkage: Each food and recipe in the list was linked to its corresponding nutrient composition data from the Korean Nutrient Database [21].

Protocol: Adaptation of Facets and Descriptors

Facets are systematic questions used to code food characteristics, while descriptors are the pre-defined answers. This structure is vital for standardizing the description of foods.

• Action: New descriptors were added to existing facets to classify and describe Korean-specific foods accurately. For example, the facet "cooking method" was expanded to include techniques prevalent in Korean cuisine [1] [21].
• Rationale: This step ensures that interviewers can consistently and accurately capture details about foods that were not previously part of the European-centric database.

Protocol: Adaptation of Quantification Methods

Accurate portion size estimation is a critical component of dietary recall. The quantification methods were rigorously evaluated and adapted to the Korean context.

• Household Measures: Local standard units and household measures (e.g., common bowl and cup sizes) were gathered from major Korean markets [21].
• Photographic Aids: A picture book of foods and dishes was developed, featuring new images and portion sizes relevant to the Korean diet. This visual aid helps respondents estimate the amount of food they consumed more accurately [1].
• Standard Units: Common commercial food packages available in the Korean market were added as standard units for quantification (e.g., a standard package of tofu or a can of a specific beverage) [21].

The Scientist's Toolkit: Key Research Reagents & Materials

Table 2: Essential Materials and Databases for GloboDiet Adaptation

Item Name / Category	Function / Role in Adaptation	Specific Korean Context Example
GloboDiet Software Platform	Core interactive 24-hour recall interview software; provides standardized structure and workflow.	Base software was customized without altering its core structure or standardization concepts [1].
Common Food & Recipe Classification	Standardized backbone for categorizing foods consistently across countries.	Existing classification was adapted with new (sub-)groups for Korean-specific foods [1].
Facet & Descriptor Library	Pre-defined questions and answers to systematically describe food characteristics (e.g., cooking method).	New descriptors were added to capture specific attributes of Korean foods [1] [21].
Recipe Manager Application	Specialized module for deconstructing complex recipes into individual ingredients and their proportions.	Used to input and manage traditional Korean multi-ingredient dishes (e.g., stews, mixed rice) [21].
Local Food Composition Table (FCT)	Provides nutrient composition data for foods and recipes consumed in the country.	Korean Nutrient Database from the Korean Nutrition Society was integrated [21].
Quantification Method Database	Library of methods for portion size estimation (photos, household measures, standard units).	Adapted with local household measures, market-surveyed standard units, and a new picture book with Korean portion sizes [1] [21].

Key Outcomes and Learnings

The following table summarizes the core components that were adapted during the process.

Table 3: Summary of Key Adaptations for the Korean GloboDiet

Adaptation Component	Key Actions and Outcomes
Software Structure	Core concept of standardization and software integrity was maintained despite extensive customization [1].
Food Classification	New (sub-)groups were added to the common classification to accommodate Korean foods [1].
Facets & Descriptors	New descriptors were added to existing facets to accurately describe and classify specific Korean foods [1] [21].
Quantification Methods	Critically evaluated and adapted using local household measures, standard units, and a newly developed picture book with Korean-specific food portion sizes [1] [21].
Recipe Management	Traditional Korean recipes were loaded into the Recipe Manager, specifying ingredients and their proportions to calculate accurate nutrient intake [21].

Critical Learnings and Implications for Future Research

• Feasibility of Global Standardization: The successful development of the Korean GloboDiet version confirmed that the IARC-WHO international dietary methodology is sufficiently flexible and robust for customization beyond Europe. It is possible to adapt the tool for an Asian context without compromising its core concept of standardization [1].
• Importance of Structured Localization: The adaptation is not a simple translation but a deep, structural customization of databases. The process highlighted that a strict adherence to established SOPs, combined with meticulous local data collection, is the key to achieving this balance [21].
• Foundation for Comparative Studies: By using a standardized methodology like GloboDiet, Korea is now positioned to participate in international comparative studies on diet and disease, which have previously been hampered by methodological heterogeneity [21].
• Model for Other Regions: The Korean adaptation serves as a model for subsequent projects. For instance, similar evaluations have been conducted for potential application in African settings, highlighting shared challenges such as describing local foods and addressing low literacy in rural areas [19]. More recently, the Intake24 tool was successfully adapted for South Asian populations, further validating this approach [20].

The adaptation of the GloboDiet software for the Korean diet stands as a seminal case study in the transfer of a standardized dietary assessment methodology to a new cultural and dietary context. The project successfully demonstrates that with a rigorous, protocol-driven approach, it is feasible to maintain international comparability while achieving local relevance. The key to success lies in the systematic customization of core and local databases, particularly food lists, facets/descriptors, and quantification methods. The resulting methodology provides researchers and public health professionals in Korea with a powerful tool for monitoring dietary trends, investigating diet-disease relationships, and evaluating the impact of public health policies. The learnings from this endeavor provide a valuable roadmap for future adaptations of standardized dietary recall tools in other low- and middle-income countries and diverse cultural settings worldwide.

The international standardized 24-hour dietary recall (GloboDiet) methodology represents a crucial advancement for addressing malnutrition and related comorbidities in Low- and Middle-Income Countries (LMICs). Developed by the International Agency for Research on Cancer (IARC/WHO), this software-based dietary assessment tool was originally validated in European contexts but has since been piloted for global application through the "Global Nutrition Surveillance initiative" [19] [33]. The pressing need for reliable and comparable individual food consumption data in African settings stems from the rapid nutritional transition characterized by increased consumption of energy-dense foods, saturated fatty acids, sugar, and salt, which has been frequently associated with cancer and other non-communicable diseases [19] [33]. Without standardized dietary tools and their related research support infrastructure, implementing concerted and region-specific research and action plans remains challenging across Africa [19].

The GloboDiet methodology provides a structured interview procedure for collecting detailed information about all foods and beverages consumed during the previous 24-hour period. Its core strength lies in the standardization of dietary data collection, which enables meaningful comparisons across different populations and geographical regions [1]. This standardization is particularly valuable for LMICs where heterogeneous dietary patterns, diverse food supplies, and varying food preparation methods have traditionally complicated nutritional surveillance and research [20]. The methodology follows a stepwise approach for data collection that includes establishing a quick list of consumed items, detailed description of foods and recipes, and comprehensive quantification methods [19] [17].

African Evaluation of GloboDiet: Key Findings and Methodologies

Expert Evaluation Methodology

The evaluation of GloboDiet for African applications employed a structured consultative approach involving African and international experts in dietary assessment. The methodological framework consisted of six e-workshop sessions complemented by an in-depth e-questionnaire administered both before and after the e-workshop participation [19]. This dual-method approach allowed researchers to capture initial impressions and refined evaluations following detailed exposure to the GloboDiet methodology.

A panel of 29 experts participated in this comprehensive evaluation, representing diverse specializations within nutritional sciences and public health. The assessment focused on multiple dimensions of the GloboDiet software, including its main structure, the stepwise approach for data collection, and the standardization concept [19]. Experts were asked to evaluate specific sections of the methodology, including general information collection, quick listing of consumed items, description of foods and recipes, quantification methods, probing questions, quality controls, and dietary supplements assessment [34].

The evaluation specifically investigated the feasibility of implementing GloboDiet in diverse African settings, with particular attention to methodological barriers and adaptation requirements. Experts assessed the software's capacity to address local specific needs while maintaining the integrity of standardized data collection [19]. This evaluation methodology established a foundation for identifying critical adaptation requirements while preserving the core standardized elements that enable international comparisons.

Quantitative Evaluation Results

Table 1: Expert Evaluation of GloboDiet Sections for African Application

GloboDiet Section	Overall Expert Evaluation	Specific Adaptation Suggestions
General Information	Adequate, useful, applicable, comprehensive, easily understandable [34]	Add dwelling place, marital status, number of children, education, physical activity, breastfeeding status, employment status [34]
Quick List	Easy, comprehensive, good, useful features, clear, appropriate [34]	Ask for daily activities to capture food consumption occasions; adapt food consumption occasions list [34]
Description of Foods/Recipes	Relevant aspects covered, detailed, comprehensive, clear [34]	Add sun drying, smoking, mortar pounding, stone grounding, sifting; contextualize food descriptors [34]
Quantification Methods	Satisfactory but requires significant adaptation [34]	Create local picture books with local household measures; use local standard units; develop new approaches for shared plates [19] [34]
Dietary Supplements	Relevant, adapted, necessary, good, helpful, appropriate [34]	Clearly define "tonic" and "energy-booster" plant by-products; establish classification as food or supplement [34]

The expert evaluation revealed an overall positive assessment of GloboDiet's potential applicability across African settings. Experts expressed satisfaction with the main structure of the software and the stepwise approach for data collection [19]. The standardization concept was particularly valued for its potential to generate comparable data across different regions and populations. However, the evaluation also identified significant gaps that require addressing before effective implementation can occur.

The data collected through this structured evaluation informed the development of a roadmap for implementation, emphasizing the necessity of rigorous capacity building and knowledge transfer to support a stepwise approach to methodological rollout across pilot African countries and regions [19]. The positive expert assessment sets a promising platform for improved dietary monitoring and surveillance, provided that the identified adaptations are systematically addressed.

Critical Adaptation Requirements for African Contexts

Food Description and Classification Adaptations

The African evaluation highlighted substantial needs for expanding food descriptors to adequately capture the diversity of local foods and preparation methods. Experts recommended adding new facets and descriptors to classify and describe specific African foods, noting that the existing European-centric classification required significant augmentation [19] [34]. Particular culinary patterns common across African regions, such as mortar pounding of staples and traditional fermentation methods, were identified as inadequately represented in the current GloboDiet structure [19].

The adaptation process requires developing comprehensive local food databases that reflect regional dietary patterns while maintaining standardized classification frameworks. This approach mirrors successful adaptations in other non-European contexts, such as the Republic of Korea, where new food subgroups were added to the existing common food classification, and new descriptors were created to characterize specific Korean foods [1]. For African implementations, additional descriptors were proposed for food processing techniques including sun drying, smoking, mortar and peddle pounding, stone grounding, and sifting [34]. The classification of foods obtained through gifts and home production also requires special consideration in many African contexts where these represent significant contributions to dietary intake [34].

Food Quantification Methodologies

Food quantification emerged as a particularly challenging aspect requiring substantial adaptation for African settings. The evaluation identified several critical areas where standard quantification methods proved inadequate, including:

• Shared-plate eating situations: Traditional quantification approaches assume individual portions, but many African eating cultures involve communal consumption from shared plates or bowls, necessitating novel quantification methodologies [19] [34].

• Local measurement units: Experts emphasized the need for local standard units that reflect commonly used household measures, such as specific varieties of cups, bowls, and spoons that vary across regions [34]. The use of handful measures for grains and leafy vegetables was specifically noted as requiring standardization [34].

• Visual estimation aids: The development of local picture books with photographs of foods without forks and knives, reflecting typical presentation methods, was recommended to improve portion size estimation [34]. These visual aids should represent local foods in typical serving arrangements and include reference objects appropriate to the context.

• Rural-urban distinctions: A clear distinction between urban and rural areas was recommended for quantification approaches, recognizing that packaging, purchasing patterns, and consumption practices often differ significantly [34].

The evaluation specifically highlighted the need to study and validate methods for "shared plate" quantification, acknowledging that this represents a particularly complex challenge that may require innovative data collection approaches [34].

Interviewing Populations with Low Literacy Skills

The African evaluation placed significant emphasis on methodological adaptations for populations with low literacy skills, especially in rural settings. This challenge was acknowledged as requiring specific considerations and appropriate solutions [19]. Research from Brazil using GloboDiet has demonstrated that establishing adequate communication during interviews is of utmost importance when working with low-educated populations [35]. Interviewers reported needing to pay particular attention to adjusting their questioning techniques to ensure comprehension among respondents with limited formal education [35].

The familiarity of individuals with food and nutrition concepts significantly influences reporting accuracy, with those having greater familiarity and interest in food-related topics demonstrating better recall abilities [35]. This finding suggests that pre-interview activities that enhance food awareness may improve data quality in low-literacy contexts. Additionally, the use of photograph manuals for food portion quantification was generally helpful, though further investigations to optimize their use for low-literacy populations were recommended [35] [34].

The socio-economic context of food access may also influence interview dynamics, as interviewees from economically vulnerable groups may experience discomfort reporting their food consumption due to concerns about exposing their economic situation [35]. Interviewer training must address these sensitivities to ensure accurate reporting across diverse socioeconomic groups.

Experimental Protocols for GloboDiet Validation and Adaptation

Protocol for GloboDiet Validation Using Biomarkers

Table 2: Validation Study Protocol Using Urinary Biomarkers

Study Component	Protocol Specifications	Quality Control Measures
Study Design	Cross-sectional design with 24-hour GloboDiet recall and 24-hour urine collection on same day [17]	Target sample size: 50 men and 50 women; convenience sampling with screening for exclusion criteria [17]
Participant Recruitment	Recruitment through institutional databases, internet, and local media; exclusion based on diseases/medications affecting nutrient metabolism [17]	Questionnaire sent prior to first visit gathering drug use and socio-economic information [17]
Data Collection	Face-to-face GloboDiet interviews; 24-hour urine collection; anthropometric measurements; accelerometer placement [17]	Urine completeness checked via creatinine quotient (>60% threshold); trained interviewers; standardized protocols [17]
Biomarker Analysis	Comparison of protein/potassium intake from GloboDiet with urinary nitrogen/potassium excretion [17]	Statistical analysis: Wilcoxon rank tests, confidence intervals, Spearman correlations, Bland-Altman plots [17]
Data Interpretation	Assessment of agreement between intake and excretion; identification of potential underestimation patterns [17]	Consideration of confounding factors (dietary composition, health status, microbiome) [17]

The validation protocol for dietary assessment tools requires robust biomarker comparison to establish methodological accuracy. The German validation study (ErNst study) provides a template for African applications, employing a cross-sectional design with 24-hour GloboDiet recalls and 24-hour urine collection on the same day [17]. This approach allows for direct comparison between estimated nutrient intake and urinary excretion of validation markers, particularly protein and potassium [17].

The validation process necessitates careful participant screening to exclude individuals with conditions or medications that might affect nutrient metabolism or excretion. The German study implemented strict exclusion criteria and verified the completeness of 24-hour urine collections using creatinine quotients, with mean values of 87% for men and 78% for women meeting the threshold for inclusion [17]. This quality control measure is particularly important in field conditions where compliance with complete urine collection may vary.

Statistical analysis for validation studies should employ multiple complementary approaches, including Wilcoxon rank tests for median differences, confidence intervals for estimation precision, Spearman correlations for relationship strength, and Bland-Altman plots for agreement assessment [17]. This multi-faceted statistical approach provides a comprehensive evaluation of the methodology's performance, as demonstrated in the German validation where protein intake showed valid estimates while potassium results were more ambiguous [17].

Protocol for Dietary Assessment in Low-Literacy Populations

Conducting valid dietary assessments in populations with low literacy requires specialized protocols that address the unique challenges of these contexts. Based on research from Brazil using GloboDiet, the following protocol elements are recommended:

• Interviewer training: Focus on communication adaptation skills, with emphasis on rephrasing questions without altering their substantive content. Interviewers should receive specific training in working with low-educated populations, including techniques for building rapport and reducing social desirability biases [35].

• Interview structure: Incorporate memory aids and contextual cues, such as asking about daily activities to help recall food consumption occasions [34]. The multiple-pass method inherent in GloboDiet provides a structured framework that can be particularly beneficial for systematic recall in low-literacy populations.

• Visual aids development: Create photograph manuals with portion sizes relevant to local foods and eating practices. These visual aids should depict foods without utensils when appropriate and reflect typical serving methods [35] [34].

• Pilot testing: Conduct comprehensive feasibility testing with the target population to identify comprehension challenges and refine probing questions. This testing should specifically evaluate the understanding of portion size photographs and household measures [35].

• Contextual data collection: Gather information on participants' familiarity with food and nutrition concepts, as this factor has been shown to influence reporting accuracy [35]. This information can help interpret data quality and identify potential systematic biases.

Implementation of these protocol adaptations requires careful documentation to maintain standardization while allowing for necessary contextualization. The goal is to achieve equivalent data quality across diverse population groups despite literacy differences.

Figure 1: GloboDiet Adaptation Workflow for LMICs - This diagram illustrates the stepwise approach for adapting and implementing the standardized 24-hour dietary recall methodology in African settings, highlighting critical adaptation areas throughout the process [19] [33].

Research Reagents and Essential Materials

Table 3: Essential Research Reagents and Tools for GloboDiet Implementation

Tool Category	Specific Items	Function in Dietary Assessment
Software Platforms	GloboDiet Software [19] [17]	Standardized 24-hour dietary recall administration with structured interview procedure and quality checks
Food Composition Databases	Local Food Composition Tables [20] [36]	Provide nutrient composition data for local foods and recipes for intake calculations
Visual Quantification Aids	Local Picture Books [34]Food Models [34]Salted Replicas [34]	Assist participants in estimating portion sizes through visual references relevant to local context
Biomarker Validation Tools	24-hour Urine Collection Kits [17]Accelerometers [17]	Objective validation of nutrient intake (protein, potassium) and energy expenditure measurement
Interviewer Training Materials	Standard Operating Procedures [33]Training Manuals [19]	Ensure standardized administration across interviewers and research sites
Data Management Systems	GloboDiet-Research Infrastructure [33]	Web-based infrastructure support for data storage, processing, and standardization

Successful implementation of GloboDiet in LMICs requires access to specialized research reagents and tools that support standardized data collection while accommodating local contexts. The core software platform provides the foundation for dietary assessment, enabling structured interviews through a computer-based interface that includes built-in quality controls [19] [17]. This software must be complemented with comprehensive food databases that reflect local food supplies while maintaining standardized classification frameworks.

Visual quantification aids represent critical tools for accurate portion size estimation, particularly in populations with varying literacy levels. The African evaluation specifically recommended creating local picture books with foods photographed without forks and knives, reflecting typical consumption patterns [34]. Additional quantification aids may include food models, salted replicas, and standardized household measures that correspond to locally used utensils and containers [34].

For validation studies, biomarker collection kits are essential for establishing the methodological validity of dietary recalls. The German validation protocol utilized 24-hour urine collection materials and accelerometers to objectively measure physical activity and energy expenditure [17]. These biological samples enable comparison between reported nutrient intake and urinary excretion of specific markers, particularly protein and potassium [17].

The evaluation of GloboDiet for African settings demonstrates both the feasibility and necessity of adapting standardized international dietary assessment methodologies for LMICs. The expert assessments confirm the methodology's potential flexibility while highlighting critical adaptation requirements specific to African contexts [19]. Successful implementation requires addressing key challenges in food description, quantification methods, and interviewing protocols for populations with low literacy skills.

The proposed implementation follows a stepwise approach that begins with comprehensive preparation and tool development, progresses through rigorous validation, and culminates in broader expansion across regions [33]. This phased implementation strategy allows for necessary contextual adaptations while maintaining the standardization essential for cross-national comparisons. The establishment of regional hubs and systematic capacity building will be crucial for long-term sustainability, particularly in resource-constrained settings [19] [33].

The adaptation of GloboDiet for African use holds significant promise for strengthening nutritional surveillance, informing food and nutrition policies, and developing targeted interventions to address the dual burden of malnutrition in LMICs. By generating reliable and comparable dietary data, this initiative can contribute substantially to understanding nutritional transitions and their health implications across diverse African populations [19] [33]. The insights from African evaluations provide a valuable roadmap for similar adaptations in other LMIC regions facing comparable dietary assessment challenges.

Navigating Real-World Challenges in GloboDiet Implementation

The standardized 24-hour dietary recall (24-HDR) methodology, particularly the GloboDiet software, represents a cornerstone for collecting high-quality comparable dietary data in epidemiological research and national surveillance [2]. Developed by the International Agency for Research on Cancer (IARC), this computer- and interview-based tool is designed to minimize systematic errors and has been successfully adapted across Europe, Asia, and Latin America [2] [1]. However, its application in diverse cultural and socio-economic contexts presents specific methodological challenges related to common eating practices, such as consumption from shared plates, interviewing populations with low literacy, and capturing intake of locally specific unfamiliar foods [4]. This document outlines standardized protocols and application notes to address these complexities, ensuring the integrity and comparability of dietary data collected in varied settings as part of a broader thesis on GloboDiet adaptation methods.

Application Note 1: Assessing Dietary Intake from Shared Plates and Bowls

Background and Complexity

The standard GloboDiet interview is structured around reporting individual food items and their quantities. This model faces significant challenges in cultures where eating from communal platters is a common practice. In such settings, individuals often serve themselves multiple times from a central bowl or plate, making it difficult to recall and quantify the total amount of a specific food item they have consumed [4]. This eating practice can lead to significant under-reporting or misestimation of actual intake if not properly addressed by the methodology.

Proposed Methodological Adaptations and Protocol

To maintain data standardization while accommodating this practice, the following additions to the standard GloboDiet probing and quantification steps are recommended.

• Enhanced Probing and Description: Interviewers must receive specific training to probe for shared eating situations. When a composite dish or a food item is reported, a mandatory follow-up question should be: "Was this dish consumed from a shared plate or bowl?" If confirmed, the interviewer must then document the total number of people sharing the meal and the approximate total volume or quantity of the food item in the shared container.
• Quantification Strategy: The preferred method is to estimate the total cooked volume of the dish (e.g., using common household containers like a "medium-sized bowl" as a reference) and then calculate the individual's share by dividing this total volume by the number of people sharing. The GloboDiet software's portion size estimation aids, such as photograph series of full dishes in common serving bowls, should be expanded to include images of these shared containers to aid in total volume estimation [2] [4].
• Data Handling: A specific facet or descriptor should be activated for foods identified as being consumed from a shared plate. This allows for later data quality checks and analytical adjustments if needed.

Table 1: Protocol for Assessing Intake from Shared Plates

Step	Action	Tools Required	Data Recorded in GloboDiet
1. Identification	Probe: "Was this dish consumed from a shared plate/bowl?"	Pre-defined probing question in software	Yes/No flag
2. Context	If Yes: "How many people shared this dish?"	Interviewer training	Number of individuals
3. Total Quantification	"What was the total amount of [food] in the shared container?"	Picture book with images of common serving bowls/platters filled with food; Household measures	Total quantity (e.g., one medium bowl)
4. Individual Calculation	Software or interviewer calculates individual share: Total Quantity / Number of People	GloboDiet's internal calculation function	Final consumed amount

Workflow Diagram

The following diagram illustrates the logical workflow for integrating the assessment of shared plate consumption into the standard GloboDiet interview structure.

Application Note 2: Interviewing Populations with Low Literacy

Background and Complexity

A key principle of GloboDiet is its design to be independent of a respondent's literacy level [2]. However, populations with low formal education (often defined as less than 9 years of schooling) may face additional challenges. These include difficulty in understanding the abstract nature of some questions, discomfort in reporting their socio-economic situation through their diet, and greater challenges in estimating food portion sizes [35] [4]. Studies have shown larger quantification errors among low-educated individuals, highlighting the need for tailored interviewer approaches [35].

Proposed Methodological Adaptations and Protocol

The adaptation focuses on interviewer communication skills and the enhanced use of visual aids.

• Interviewer Training and Communication: Interviewers must be trained to rephrase complex questions into simpler, more concrete language without breaking the standardization protocol. The establishment of adequate communication and a comfortable rapport is paramount to reduce the respondent's discomfort in reporting their actual food intake [35]. Training should include role-playing scenarios with individuals from varied educational backgrounds.
• Leveraging Familiarity: The interview should begin by leveraging the respondent's inherent familiarity with food and its preparation. Asking about who cooked the meal or where the food was purchased can serve as a less intimidating entry point to the detailed recall [35].
• Enhanced Visual Aids: The use of the photograph manual for food portion quantification is especially crucial for this population [35]. The picture book should be made central to the interview process. Interviewers should be trained to ensure the respondent is physically interacting with the picture book, pointing to portion sizes, rather than just verbally describing them.

Table 2: Protocol for Interviewing Low-Literacy Populations

Challenge	Adapted Protocol	Key Tools & Reagents
Understanding Questions	Use simplified language and concrete examples. Avoid technical terms. Confirm understanding.	Pre-defined, validated alternative phrasings for standard questions.
Portion Size Estimation	Heavy reliance on visual aids. Encourage physical interaction with the picture book.	GloboDiet Picture Booklet with portion size photos [35] [2].
Reporting Bias	Build rapport and reassure about confidentiality. Frame questions neutrally to reduce socio-economic discomfort.	Interviewer training modules on building rapport and ethical conduct.
Recall of Consumed Items	Use a conversational, multi-pass approach. Link foods to daily routines and events.	GloboDiet's quick list and detailed description steps, flexibly applied.

Research Reagent Solutions

The following key materials are essential for the effective implementation of the GloboDiet methodology in challenging contexts.

Table 3: Essential Research Reagents for GloboDiet Adaptation

Reagent / Tool	Primary Function	Application in Addressing Complexities
GloboDiet Software Suite	Standardized data capture and coding of 24-HDR interviews.	Core platform; ensures harmonization across adapted protocols.
Country-Specific Picture Booklet	Visual aid for food identification and portion size quantification.	Critical for low-literacy populations and for quantifying shared dishes [35] [2].
Food & Recipe Database	Library of local foods, descriptors, and facets for classification.	Must be expanded to include unfamiliar, local, and traditional foods [2] [4].
Facet & Descriptor Library	Structured questions and answers to describe food details (e.g., processing, fat content).	Allows precise description of unique local food preparation methods.
Household Measure Database	Common local utensils (cups, bowls, spoons) for quantification.	Essential for quantifying food from shared containers using familiar objects.
Trained Interviewers	Conduct the standardized interview with flexibility and empathy.	The most crucial "reagent" for engaging low-literacy populations and navigating complex eating scenarios [35].

Application Note 3: Capturing Unfamiliar and Local Foods

Background and Complexity

The global expansion of GloboDiet requires that its common food classification system and descriptive facets be adaptable to local food supplies. Standardized classifications must accommodate specific Brazilian and Mexican foods, Korean side dishes, or African staples without compromising the ability for between-country comparisons [2] [4] [1]. This involves adding new food subgroups and creating new descriptors for specific culinary practices (e.g., "mortar pounding" as noted in African contexts) [4].

Proposed Methodological Adaptations and Protocol

This process is foundational and occurs during the initial customization of GloboDiet for a new country or region.

• Food List and Classification Expansion: The common food classification is used as a template. New country-specific (sub)groups are added where necessary. For example, the Korean adaptation required new subgroups for specific types of Kimchi and other fermented vegetables [1]. Similarly, the Latin American versions added new subgroups and descriptors for local fruits, tubers, and maize-based preparations [2].
• Development of New Descriptors: Existing facets (questions) are reviewed, and new descriptors (answers) are created to accurately describe local foods. This ensures that a locally unique food item can be described with the same level of detail and precision as a common European food, using the same structured framework.
• Quantification Method Adaptation: The available quantification methods (standard units, household measures, photo series) are critically evaluated and expanded. This includes identifying and photographing common portion sizes of the new local foods and adding common local household measures (e.g., a specific type of clay pot or a traditionally used cup) to the database [2] [1].

Workflow Diagram

The diagram below outlines the logical process for integrating new and unfamiliar foods into the GloboDiet structure, from identification to full classification.

The GloboDiet methodology demonstrates significant flexibility and robustness, proving adaptable to diverse eating cultures and population subgroups without sacrificing its core principle of standardization [1]. The successful addressing of complexities related to shared plates, low literacy, and unfamiliar foods hinges on a systematic approach that includes: 1) strategic enhancements to probing and quantification protocols, 2) intensive and empathetic interviewer training, 3) and the rigorous expansion of the underlying food and descriptor databases using local expertise. By implementing these detailed application notes and protocols, researchers can ensure the collection of high-quality, comparable dietary intake data crucial for advancing research and informing public health policy across all populations.

Standardized 24-hour dietary recall (24-HDR) methods are cornerstone tools for nutritional epidemiology, public health monitoring, and clinical research. Their scientific validity hinges on the ability of underlying food lists and nutrient databases to accurately reflect the current food supply and consumption patterns. The rapid expansion of plant-based and vegan food products represents a significant market trend that challenges the relevance of these static instruments. This application note details a structured protocol for the timely adaptation of the GloboDiet 24-HDR methodology, using the surge in vegan products as a paradigm for maintaining methodological rigor in the face of a dynamic food environment.

Current Market Dynamics of Vegan and Plant-Based Foods

The plant-based food market has undergone substantial growth and transformation. Understanding these trends is critical for informing the strategic update of dietary assessment tools.

The market data underscores the necessity of incorporating these products into dietary monitoring frameworks. The following table summarizes key quantitative insights from recent market research:

Table 1: Quantitative Overview of the Plant-Based Food Market (2024-2025)

Metric	Value / Finding	Source / Year	Implication for Dietary Assessment
U.S. Retail Market Size	$8.1 billion [37]	2024	Significant market presence cannot be ignored.
Global Market Projection	$61.35 billion by 2028 [38]	2021-2028 Forecast	Highlights long-term, sustained growth trend.
Household Penetration (U.S.)	59% of households purchased plant-based foods [37]	2024	Moves beyond niche category into mainstream consumption.
Cross-Purchasing Behavior	96% of plant-based meat buyers also purchased animal-based meat [37]	2024	Critical for recall probing questions; indicates flexitarian patterns.
Top Consumer Demand Factors	Taste, Price, Health, Naturalness [39] [37]	2025	Drives product innovation and formulation changes.
Key Product Development Trend	Shift from "replacement" to "standalone" plant-based products [39]	2025	Requires new descriptive facets beyond analogies to animal products.
Emerging Protein Sources	Fava bean, lentil, sunflower protein [39]	2025	Expands the diversity of ingredients requiring nutrient composition data.

Evolving Consumer Expectations and Product Formulation

Beyond market size, the nature of plant-based products is evolving, directly impacting how they must be classified and described in GloboDiet. A significant trend is "Rethinking Plants," which emphasizes more natural and authentic formulations [39]. Consumers increasingly seek products with recognizable plant ingredients, shorter ingredient lists, and fewer additives, moving away from highly processed analogues [39]. This shift necessitates new descriptors and facets within GloboDiet to capture product characteristics related to naturalness and processing degree.

Protocol for Adapting GloboDiet Food Lists and Databases

This protocol provides a step-by-step methodology for updating the GloboDiet system to incorporate new food trends, ensuring standardized and comparable data collection across time and regions.

Phase 1: Market Surveillance and Food Identification

Objective: To systematically identify and characterize new vegan products entering the food supply.

Workflow:

• Source Identification: Establish a routine scan of:
  • Retailer Data: Leverage point-of-sale scanning data (e.g., SPINS, IRI) to identify high-growth categories and new product launches [37].
  • Industry Reports: Monitor publications from market research firms (e.g., Innova Market Insights, GFI) and food industry associations [39] [38].
  • Scientific Literature: Track publications on dietary patterns and novel foods.
• Product Procurement and Documentation: For a representative sample of identified products, physically procure items. Document key attributes:
  • Full product name and brand.
  • Complete ingredient list.
  • Nutritional facts panel data.
  • Packaging images (for quantification aids).
  • Sample products for nutrient analysis if required.

Phase 2: Food Description and Classification within GloboDiet

Objective: To integrate new foods into the GloboDiet structure using standardized facets and descriptors.

Workflow:

• Food Classification: Place the new food item into the hierarchical GloboDiet food classification system. This may require creating new sub-groups for novel product categories (e.g., "Plant-based cheese alternatives," "Fermented plant-based products") [6].
• Facet and Descriptor Assignment: Describe the product using GloboDiet's predefined facets. For vegan products, key facets include [40] [6]:
  • Food source & type: e.g., "Plant-based," "Meat analogue," "Dairy alternative".
  • Main protein source: e.g., "Soy," "Pea," "Wheat," "Fava bean," "Lentil" [39].
  • Fat content: e.g., "Full-fat," "Reduced-fat," "Light".
  • Processing degree: e.g., "Fresh," "Frozen," "Shelf-stable".
  • Flavoring: e.g., "Natural flavor," "Herbs," "Spiced".
• Recipe Creation: For complex prepared foods or dishes, create a standard recipe in the GloboDiet recipe module, breaking it down into its constituent ingredients with respective quantities [40].

Diagram: Logical Workflow for Integrating New Foods into GloboDiet

Phase 3: Quantification Method Development

Objective: To define and implement accurate methods for quantifying portion sizes of new food items.

Workflow:

• Method Selection: Determine the most appropriate quantification method(s) for the food item, which may include [17] [40]:
  • Standard Units (SU): e.g., "1 patty," "1 sausage."
  • Household Measures (HHMs): e.g., "cups," "tablespoons" (using graduated spoons/cups).
  • Photo Series: Develop new or adapt existing photo series showing different portion sizes. This is crucial for items like plant-based meats or cheese alternatives that come in variable forms [17].
  • Food Shapes: For geometrically regular items.
  • Weight: Using a food scale (as a reference method).
• Picture Booklet Update: Integrate new photo series into the standardized picture booklet used by interviewers to aid participant portion size estimation [17] [6].

Phase 4: Nutrient Data Bank Linkage

Objective: To ensure each food item is linked to accurate nutrient composition data.

Workflow:

• Data Source Prioritization:
  • Direct Chemical Analysis: The gold standard, used for commonly consumed new products or when composition is uncertain.
  • Brand-Specific Data: Use data from manufacturer's labels, validated where possible.
  • Generic Food Composition Data: Use data from a similar, well-established food item as a proxy, with clear documentation of the substitution.
  • Recipe Calculation: For complex products, calculate nutrient composition based on standard recipe ingredients and cooking methods.
• Database Update: Systematically add new foods and their nutrient profiles to the national nutrient database (e.g., the German Nutrient Database BLS) linked to GloboDiet [17]. Key nutrients of interest for plant-based products include protein quality, iron, calcium, zinc, and vitamin B12 [41].

Phase 5: Interviewer Training and Probing Question Updates

Objective: To ensure interviewers can effectively elicit and code consumption of new food categories.

Workflow:

• Probing Question Updates: Revise the checklist of probing questions used during the "Quick List" and "Description" stages of the GloboDiet interview to include new product categories [40]. For example, after a participant mentions "cheese," the probe should include "Was this dairy cheese or a plant-based alternative?".
• Specialized Training: Train interviewers on the new food categories, descriptors, and common brand names to ensure accurate and consistent coding during interviews [20].

Essential Research Reagent Solutions and Materials

The following table outlines key resources required for the successful adaptation and implementation of the updated GloboDiet methodology.

Table 2: Essential Research Reagents and Materials for Dietary Recall Adaptation

Item / Tool	Function / Application	Specification / Example
GloboDiet Software	Standardized 24-HDR data collection platform.	Country-specific version with administrator access for updating food lists and facets [17] [6].
National Nutrient Database	Provides nutrient composition data for consumed foods.	e.g., German BLS; requires protocol for adding new foods and recipes [17].
Standardized Picture Booklet	Aids participants in estimating portion sizes during the interview.	Must be updated with photo series for new product forms (e.g., plant-based meat cuts, cheese shreds) [17] [40].
Market Data Subscription	Provides quantitative data on product launches and sales trends.	e.g., SPINS, IRI, or Innova Market Insights data to prioritize update efforts [39] [37].
Dietary Biomarkers	Provides objective measures for validating self-reported intake.	e.g., 24-hour urinary nitrogen for protein intake validation, used to check plausibility of new data [17].
Food Sampling Kit	For direct analysis of novel foods.	Includes tools for sample collection, homogenization, and shipment to analytical labs for nutrient analysis.

The dynamic nature of the global food supply, exemplified by the rapid rise of vegan products, demands a proactive and systematic approach to maintaining the validity of standardized dietary assessment tools like GloboDiet. The protocol outlined herein—encompassing structured market surveillance, detailed food description, rigorous quantification, and robust nutrient data linkage—provides a replicable framework for researchers. By implementing these application notes, national nutrition monitoring systems can ensure they continue to generate high-quality, comparable data essential for understanding dietary patterns, assessing nutrient adequacy, and informing public health policy in a changing world.

High-quality data collection is the cornerstone of valid nutritional research and surveillance. For studies utilizing 24-hour dietary recalls, ensuring interviewer competence and protocol adherence through systematic training and quality control (QC) is paramount, as interviewer-related errors can significantly impact data accuracy and comparability [42]. This is particularly critical within the context of standardizing dietary assessment tools like GloboDiet for international adaptation, where consistent application across diverse populations and cultures is essential for data validity [1] [4]. This application note provides detailed protocols for interviewer training and QC procedures to ensure data quality in studies employing standardized 24-hour dietary recalls.

Pre-Study Preparation and Interviewer Training

A rigorous and structured training program is fundamental to preparing interviewers for high-quality data collection.

2.1 Core Training Components Interviewer training should encompass multiple facets to build proficiency [42]:

• Review of Written Protocols: Thorough instruction on the structured interview protocol, including the multiple-pass method.
• Role-Playing: Simulated interviews to practice technique and problem-solving.
• Practice Interviews with Target Population: Conducting interviews with individuals similar to the study subjects (e.g., children for pediatric studies) to gain real-world experience.
• Audio Recording and Transcription: Practice interviews should be audio-recorded and transcribed. This provides interviewers with immediate feedback and reinforces the habit of detailed documentation, a key component for subsequent QC [42].

2.2 GloboDiet-Specific Customization When adapting the GloboDiet methodology, training must extend to its unique features [1] [4]:

• Software Proficiency: Comprehensive training on the GloboDiet software interface, navigation, and data entry procedures.
• Food Description Facets: Instruction on using the predefined "facets" (questions) and "descriptors" (answers) to classify and describe foods and recipes consistently.
• Local Food Database and Quantification Methods: Familiarization with the customized food list, recipe database, and local quantification tools (e.g., picture books, household measures, food shapes) developed during the adaptation process [1] [5].

Quality Control Procedures During Data Collection

Ongoing quality control during the data collection phase is critical to maintain interviewer adherence to the protocol and identify issues promptly.

3.1 Real-Time QC Monitoring The recommended strategy involves continuous, unobtrusive monitoring of all interviews [42]:

• Audio Recording of All Interviews: Every dietary recall interview should be audio-recorded. This practice allows for random selection of interviews for review without the interviewer's prior knowledge, discouraging deviation from the protocol and enabling retrospective verification [42].
• Random Selection for Review: One interview per research dietitian should be randomly selected for QC review on a regular basis (e.g., weekly or daily) [42] [43].
• Structured QC Checklist: A standardized QC checklist, based on the interview protocol, should be used by a second research dietitian to review the selected audio recordings and their corresponding transcripts. This checklist assesses adherence to key interviewing steps, probing questions, and data recording procedures [42].

3.2 Multi-Level Data Review (Tiered QC) For large-scale studies, especially multisite trials, a tiered review process can further enhance data quality [44]. The table below summarizes a four-phase model and its impact.

Table 1: Multi-Level Quality Control Review Process and Impact

Review Phase	Responsible Personnel	Primary Actions	Impact on Data (from empirical evaluation) [44]
Phase 1: Initial Review	Dietary Interviewer	Review and edit the dietary recall after the participant has left.	Serves as the first line of defense against obvious errors.
Phase 2: Local Review	Lead Nutritionist at Field Center	Review and edit the recall as deemed appropriate.	Further reduces random errors and inconsistencies within a site.
Phase 3: External Review	Nutrition Coordinating Center (NCC)	Independent, external quality assurance review using expert knowledge and the database.	Identifies subtle errors missed in local review; primarily reduces variance of nutrients rather than shifting group means.
Phase 4: Reconciliation	Lead Nutritionist & NCC	Resolve any differences identified between Phase 2 and Phase 3.	Finalizes the dataset, ensuring consensus and highest data quality.

Evaluation of this model shows that while correlations between phases are high (≥0.96), the external review (Phase 3) is particularly effective at reducing the variance of certain nutrients like energy, folate, and fiber, thereby increasing the precision of the intake data [44].

Post-Data Collection Validation

After data collection, validation studies are essential to assess the overall quality of the dietary intake estimates.

4.1 Biochemical Validation Comparing self-reported intake with nutritional biomarkers provides an objective measure of validity [5]:

• Protocol: Collect 24-hour urine samples from participants concurrent with the 24-hour dietary recall.
• Analysis: Measure urinary nitrogen and potassium excretion.
• Comparison: Compare the estimated intake of protein (from urinary nitrogen) and potassium from the GloboDiet recall with the measured urinary excretion using statistical methods such as Wilcoxon rank tests, correlation analyses (e.g., Spearman), and Bland-Altman plots [5].
• Outcome: This method has been used to validate GloboDiet in international settings, confirming that it provides valid estimates for protein intake, with more ambiguous but generally acceptable results for potassium [5].

The Researcher's Toolkit: Essential Reagents and Materials

Table 2: Key Research Reagent Solutions for Dietary Recall Quality Control

Item	Function in QC Process	Specific Examples / Notes
GloboDiet Software	Standardized 24-hour dietary recall data collection and coding.	Includes country-specific food databases, recipe modules, and quantification tools. Central to ensuring data standardization [1] [4].
Digital Audio Recorder	Records interviews for subsequent QC review.	Enables random, retrospective QC checks without interviewer foreknowledge, minimizing behavioral bias [42].
Standardized QC Checklist	Provides a structured tool for evaluating interviewer performance.	Based on the study protocol; used to systematically review audio recordings and transcripts for adherence [42] [43].
Quantification Aids	Assist respondents in estimating portion sizes.	Picture books, food models, household measures, and food shapes. Critically updated for the local context during GloboDiet adaptation [1] [5].
Structured Training Materials	Used for initial and refresher interviewer training.	Includes written protocols, training videos, and practice datasets [42] [4].
Biochemical Validation Kits	For objective validation of self-reported intake.	24-hour urine collection kits and laboratory analysis for biomarkers like nitrogen and potassium [5].

Workflow and Protocol Diagrams

The following diagram illustrates the comprehensive, integrated workflow for ensuring data quality from interviewer preparation through final validation.

Figure 1: Integrated Workflow for Interviewer Training and Quality Control

The specific post-interview QC review process, a critical component of the data collection phase, is detailed in the following protocol.

Figure 2: Post-Interview Quality Control Review Protocol

Implementing a multi-faceted system of interviewer training and quality control is non-negotiable for generating high-quality, reliable data in dietary research. The strategies outlined—comprehensive training, real-time monitoring via audio recording and random review, tiered data checks, and biochemical validation—create a robust framework that minimizes interviewer-related error and bias. Adhering to these standardized protocols is especially critical for the successful adaptation and application of tools like GloboDiet across different countries and cultures, ultimately ensuring that the collected data is valid, comparable, and fit for purpose in research and public health surveillance.

The adoption of standardized dietary assessment tools is critical for generating reliable, comparable data to inform public health policy and nutritional research. This document outlines application notes and protocols for the adaptation of GloboDiet, a standardized 24-hour dietary recall tool, drawing on the evaluative framework of the African Consultative Panel. The recommendations are contextualized within broader thesis research on standardized dietary recall adaptation methods, providing researchers, scientists, and drug development professionals with a structured approach for implementing this methodology in diverse settings, with particular emphasis on the African context.

GloboDiet Adaptation Framework

Core Principles of Adaptation

The adaptation of GloboDiet for a new regional context is a systematic process that requires the customization of approximately seventy common and country-specific databases to reflect local dietary habits while maintaining methodological standardization essential for intra- and inter-country comparisons [2]. This process ensures that the tool captures dietary particularities without compromising data comparability.

The African Consultative Panel emphasizes building on existing structures and knowledge. As highlighted during the Africa launch of the Global Evaluation Initiative (GEI), effective monitoring and evaluation systems must "build on what is already there," leveraging local capacities, experiences, and knowledge while matching these with coordinated support and global resources [45]. This approach ensures sustainability and cultural relevance.

Key Adaptation Components

Table 1: Essential Databases for GloboDiet Adaptation

Database Type	Description	Customization Tasks
Common Databases	Standardized across all versions	Translation and harmonization
Food & Recipe Classifications	Hierarchical categorization of foods	Add new (sub)groups relevant to local diet [2]
Facets & Descriptors	Attributes describing foods and recipes	Select, add, and translate descriptors [2]
Probing Questions	Prompts to recall forgotten foods	Translate and adapt to local eating patterns [2]
Quantification Methods	Tools for estimating portion sizes	Adapt methods considering local consumption [2]
Country-Specific Databases	Unique to each dietary context	Development from local data sources
Food & Recipe Lists	Comprehensive inventory of consumed items	Compile from local surveys, FCDBs, and expert input [2] [24]
Synonym Lists	Variant names for the same food	Collect colloquial and regional terms [2]
Brand Name Lists	Commercial product identifiers	Document most common brands by food group [2]
Picture Book	Visual aids for portion size estimation	Develop photos of local foods and serving sizes [2]
Coefficient Files	Conversion factors (e.g., raw-to-cooked)	Calculate using local food preparation data [2]

The adaptation process requires meticulous compilation of local foods and recipes. For instance, the Intake24 adaptation for South Asian populations resulted in a database of 2,283 food items [20], while the New Zealand version contained 2,618 foods [24]. This comprehensive coverage is essential for accurate dietary assessment.

Experimental Protocols for Tool Validation

Biomarker Validation Protocol

Validation against biological markers provides the most robust assessment of dietary assessment tool accuracy. The following protocol, adapted from the ErNst study validating the German GloboDiet version, outlines procedures for validating protein and potassium intake [5].

Objective: To validate nutrient intake estimates from GloboDiet by comparison with urinary biomarkers.

Sample Population:

• Recruitment of approximately 100 participants (50 men, 50 women)
• Equal distribution across age groups (e.g., 18-39, 40-59, 60-79 years)
• Exclusion criteria: diseases or medications affecting nutrient metabolism or excretion

Study Timeline:

• Day 1: Participant enrollment, anthropometric measurements (height, weight, BMI), bioelectrical impedance analysis, accelerometer fitting
• Day 2: 24-hour urine collection, GloboDiet 24-hour recall administration, accelerometer removal

Urine Collection and Analysis:

• Participants receive detailed instructions for 24-hour urine collection
• Urine volume is measured, and aliquots are stored at -20°C until analysis
• Urinary nitrogen (for protein) and potassium are analyzed
• Creatinine excretion is measured to assess completeness of urine collection using established formulae [5]

Dietary Assessment:

• Trained interviewers conduct face-to-face 24-hour recalls using GloboDiet
• Recalls cover the same 24-hour period as urine collection
• Interviews follow standardized GloboDiet protocol with detailed description and quantification of all consumed foods and beverages

Statistical Analysis:

• Wilcoxon rank tests to assess differences between intake and excretion
• Confidence intervals for mean differences
• Spearman correlations between intake and excretion
• Bland-Altman plots to assess agreement

Quality Control:

• Interviewer training and standardization
• Completeness checks for urine collections (creatinine quotient >60%)
• Accelerometer data validation for energy expenditure comparison

Food List Development Protocol

Development of a comprehensive, culturally appropriate food list is fundamental to dietary assessment tool adaptation, as demonstrated in the Intake24 adaptation for New Zealand [24].

Objective: To create a representative food list that reflects the local diet and ethnic diversity.

Procedure:

• Select Baseline Food List: Identify an appropriate starting point (e.g., similar country's validated list)
• Category Review: Systematically review food categories (e.g., dairy, grains, fruits, mixed dishes)
• Identify Local Foods: Compile foods from multiple sources:
  • National food composition databases
  • Previous dietary surveys
  • Supermarket websites and sales data
  • Consultation with nutritionists from diverse ethnic communities
• Optimize Item Selection: Balance comprehensiveness with user burden by:
  • Combining similar foods (e.g., multiple apple varieties)
  • Separating nutritionally distinct items (e.g., fortified vs. non-fortified)
  • Including generic "not further defined" options
• Link to Nutrient Data: Match food items to appropriate nutrient composition databases

Table 2: Essential Research Reagent Solutions for Dietary Assessment Adaptation

Reagent/Category	Function/Application	Specifications
GloboDiet Software	Standardized 24-hour dietary recall platform	Includes ~70 common and country-specific databases [2]
Food Composition Database	Nutrient calculation for reported foods	National database (e.g., German BLS, New Zealand FCDB) [24] [5]
Picture Booklet	Portion size estimation aid	Contains standardized photo series of local foods and serving sizes [2]
Urinary Nitrogen Assay	Biomarker analysis for protein validation	Quantifies urinary nitrogen for protein intake validation [5]
Urinary Potassium Assay	Biomarker analysis for potassium validation	Quantifies urinary potassium for intake validation [5]
Accelerometer	Physical activity measurement	Estimates total energy expenditure for energy intake validation [5]
Quality Control Databases	Data quality assurance	Includes maximum portion sizes and nutrient limits for plausibility checks [2]

Workflow Visualization

Implementation Considerations for African Contexts

The African Consultative Panel emphasizes context-specific implementation strategies:

Building on Existing Infrastructure:

• Leverage and strengthen existing monitoring and evaluation systems rather than creating parallel structures [45]
• Identify and collaborate with national evaluation associations and networks
• Utilize available data sources while building capacity for enhanced data processing

Cultivating Evaluation Culture:

• Support the institutionalization of evaluation in public policy [45]
• Engage young and emerging evaluators as "fertile soil to embed the seeds" of evaluation capacity [45]
• Foster peer learning and shared progress among African nations

Strategic Partnerships:

• Collaborate with regional organizations like the African Evaluation Association (AfrEA) [45]
• Develop databases of local evaluators to build sustainable capacity
• Ensure coordination between national, regional, and global knowledge resources

The adaptation of standardized 24-hour dietary recall tools like GloboDiet requires a meticulous, systematic approach that balances international standardization with local dietary relevance. The protocols outlined herein, informed by the African Consultative Panel's emphasis on building existing capacity and cultivating evaluation culture, provide researchers with a validated roadmap for implementation. By following these application notes and maintaining fidelity to the core principles of adaptation and validation, researchers can generate high-quality, comparable dietary data essential for informing public health nutrition policies and understanding nutritional transitions in diverse populations.

Ensuring Data Integrity: Validation Protocols and Comparative Analysis with Modern Tools

Within research on standardizing 24-hour dietary recall (24-HDR) methods, such as the adaptation and implementation of the GloboDiet software, validating the collected data against objective biomarkers is a critical step. This protocol details a established methodology for comparing self-reported nutrient intake from 24-HDRs with urinary biomarkers, specifically for protein and potassium. This process is essential for quantifying measurement error, assessing the validity of dietary data, and ensuring the reliability of nutrition surveillance and research [46] [5].

Experimental Protocols

Core Protocol: 24-Hour Urine Collection and 24-Hour Dietary Recall

This foundational protocol involves the simultaneous collection of urinary biomarkers and dietary intake data for the same 24-hour period.

• Objective: To validate the assessment of protein and potassium intake from a 24-HDR by comparing it with urinary excretion levels.
• Principle: Urinary nitrogen and potassium are considered recovery biomarkers. Under metabolic equilibrium, a known and relatively constant proportion of ingested protein (as nitrogen) and potassium is excreted in urine over 24 hours, providing an objective measure of intake [46] [47] [5].

Materials and Reagents

• Study Participants: Recruited from the general population, typically adults who are metabolically stable, not pregnant, and not on medically prescribed diets that would alter nutrient excretion [48] [5].
• 24-Hour Dietary Recall Tool: A standardized tool, such as the GloboDiet software, which is administered by a trained interviewer [5].
• Urine Collection Kit: Includes large, pre-weighed collection containers, secondary tubes for aliquoting, a cold pack or cooler for storage, and a detailed instruction sheet and log sheet for participants.
• Para-aminobenzoic acid (PABA) tablets: (Optional) Used to verify the completeness of the 24-hour urine collection [46].
• Laboratory Equipment:
  • For Nitrogen Analysis: Equipment for the Dumas method (e.g., rapid N exceed) or Kjeldahl method [49].
  • For Potassium Analysis: Atomic absorption spectroscopy or flame photometry [49].
  • For Creatinine Analysis: Analyzer using the Jaffé reaction or other standardized methods [49].

Step-by-Step Procedure

• Participant Preparation and Briefing:
  • Schedule the participant's 24-HDR interview for the day immediately following the 24-hour urine collection period.
  • Provide comprehensive verbal and written instructions for the urine collection. Emphasize the importance of collecting every urine sample over the full 24 hours.
  • Instruct the participant to discard the first urine of the day (e.g., the first morning void) and note this as the start time. For the next 24 hours, all urine must be collected into the provided container, which should be kept cool (e.g., in a refrigerator or with a cold pack) [49] [5].
  • If using PABA for compliance, provide the tablets with precise instructions on when to take them [46].

• 24-Hour Urine Collection:

  • The participant collects all urine for a full 24-hour period, following the provided instructions.
  • The participant records the start and end times, and notes any missed collections or spills on the log sheet.
  • Upon completion, the participant returns the collection container and log sheet to the research team.

• Urine Sample Processing:

  • Upon receipt, the total volume of the 24-hour urine is measured and recorded.
  • The urine is mixed thoroughly, and aliquots are taken for analysis of nitrogen, potassium, and creatinine. Aliquots are typically stored at ≤ -80°C until analysis [49].
  • Completeness Check: The completeness of the collection is verified. This can be done by:
    • PABA Check: Measuring PABA recovery; collections with <50% recovery are often excluded [46].
    • Creatinine Index: Calculating the ratio of observed to expected creatinine excretion based on body weight and sex. A ratio of <60% may indicate an incomplete collection [5].

• 24-Hour Dietary Recall Administration:

  • On the day the urine sample is returned, a trained interviewer conducts a face-to-face or telephone 24-HDR using the standardized GloboDiet software.
  • The recall should cover the exact same 24-hour period as the urine collection.
  • The interview follows a structured protocol, including a quick list, detailed description (using facets and descriptors), and quantification of all foods and beverages consumed [4] [5].

• Laboratory Analysis:

  • Analyze the urine aliquots for nitrogen, potassium, and creatinine concentration using the standardized laboratory methods listed in the materials section.

• Data Processing and Calculation:

  • Biomarker-Estimated Intake:
    • Protein (from Nitrogen): Total urinary nitrogen (g) is converted to protein intake using a conversion factor. A common calculation is: Protein intake (g) = (Urinary Nitrogen (g) × 6.25) / 0.8. The divisor 0.8 accounts for the assumption that approximately 80% of ingested nitrogen is excreted in urine over 24 hours [49] [5].
    • Potassium: Total urinary potassium (mg or mmol) is used to estimate intake, assuming about 80% of dietary potassium is excreted in urine: Potassium intake (mg) = Urinary Potassium (mg) / 0.8 [49].
  • Self-Reported Intake: Nutrient intake from the 24-HDR is calculated by the GloboDiet software linked to a national food composition database (e.g., the German Nutrient Database BLS) [5].

Extended Validation Protocol with Multiple Methods and Biomarkers

For a more comprehensive validation, the core protocol can be expanded to include multiple dietary assessment tools and additional biomarkers, as demonstrated in large-scale studies [48] [47].

• Objective: To compare the performance of different self-reported dietary assessment tools (e.g., 24-HDR, food records, FFQs) against a suite of recovery biomarkers.
• Study Design: A prospective study where participants complete multiple dietary assessments over several weeks alongside the collection of biomarker data.

Procedure

• Baseline Phase: Collect demographic, anthropometric (weight, height), and socioeconomic data.
• Dietary Assessment Phase: Participants are asked to complete:
  • Multiple 24-HDRs (e.g., 3 non-consecutive days) using the tool under investigation (e.g., GloboDiet, myfood24) [48] [47] [50].
  • Other dietary assessment methods, such as food frequency questionnaires (FFQs) or food records, for comparative purposes [47].
• Biomarker Collection Phase: Running concurrently with the dietary assessment.
  • Urine Collection: Multiple 24-hour urine samples are collected, ideally on days matching the 24-HDRs [47].
  • Doubly Labeled Water (DLW): Administered to measure total energy expenditure (TEE) as a biomarker for energy intake [48] [47].
  • Blood Collection: For the analysis of other biomarkers, such as serum carotenoids (for fruit/vegetable intake) or erythrocyte membrane fatty acids (for fatty acid intake) [48].

Data Analysis and Statistical Methods

The following statistical approaches are recommended to assess the agreement between self-reported intake and biomarker-based intake [48] [5].

• Descriptive Statistics: Report means, medians, and standard deviations for both self-reported and biomarker-estimated intakes.
• Wilcoxon Signed-Rank Test: A non-parametric test to determine if there is a systematic difference (bias) between the two methods at the group level [5].
• Correlation Analysis: Spearman's rank correlation coefficients are calculated to assess the ability of the dietary method to rank individuals according to their intake [48] [5].
• Bland-Altman Plots: Used to visualize the agreement between the two methods by plotting the difference between them against their mean. This helps identify systematic bias and whether the bias changes with the level of intake [48] [5].
• Method of Triads: This technique uses the three measurements (self-report, biomarker, and another reference method) to estimate the correlation between each method and the true, unobserved intake [48].

The workflow for the experimental and analytical process is summarized below.

Key Research Reagent Solutions

The following table lists essential materials and their functions for implementing this protocol.

Item	Function/Application in Protocol
GloboDiet Software	Standardized, interviewer-administered 24-HDR software. Ensures structured data collection and nutrient calculation via linked food composition databases [1] [4] [5].
Para-aminobenzoic acid (PABA)	Urinary recovery biomarker used to verify the completeness of a 24-hour urine collection. Incomplete collections (<50% recovery) are typically excluded from analysis [46].
Doubly Labeled Water (DLW)	The gold-standard recovery biomarker for total energy expenditure (TEE) in free-living individuals. Serves as an objective reference for validating self-reported energy intake [48] [47].
Food Composition Database	A country-specific nutrient database (e.g., German BLS, UK Composition of Foods) linked to the dietary assessment software to convert reported food consumption into nutrient intakes [49] [5] [50].
Creatinine Assay Kit	Used to measure urinary creatinine concentration, which serves as an auxiliary check for the completeness of a 24-hour urine collection via the creatinine index [49] [5].

Representative Data and Validation Outcomes

Validation studies consistently show that 24-HDRs tend to underestimate absolute intake compared to biomarkers, but provide reasonable ranking for protein.

Table 1: Comparison of Selected Validation Studies for 24-Hour Recalls Against Urinary Biomarkers

Study & Tool	Nutrient	Mean Bias (Self-report - Biomarker)	Correlation with Biomarker (Spearman's ρ)	Key Findings
ErNst (GloboDiet, Germany) [5]	Protein	Not significantly different	0.37	The updated GloboDiet provided valid estimates for protein intake at the group level.
	Potassium	Suggested underestimation	0.24	Ambiguous results for potassium; correlation was weak.
IDATA (ASA24, USA) [47]	Protein	-15% to -17% (underestimate)	~0.3-0.4	Multiple ASA24s provided better estimates of absolute intake than FFQs.
	Potassium	-15% to -17% (underestimate)	~0.3-0.4	Misreporting was present but less severe than in FFQs.
myfood24 (UK) [50]	Protein	Attenuation factor*: ~0.2-0.3	~0.3-0.4	Performance was broadly similar to a traditional interviewer-based recall.
	Potassium	Attenuation factor*: ~0.2-0.3	~0.3-0.4	The online tool was a feasible alternative for large-scale studies.
EFCOVAL/EPIC (EPIC-Soft, Europe) [46]	Protein	Underestimation: 2% to 13% across centers	Not Reported	BMI and study design aspects (e.g., day of week) influenced bias across centers.

*The attenuation factor describes the degree to which the true diet-disease relationship is diluted by measurement error; a value closer to 1.0 indicates less error.

Troubleshooting and Technical Notes

• Incomplete Urine Collections: This is a major source of error. Strictly enforce and verify collection protocols using PABA or creatinine checks. Exclude incomplete samples from analysis [46].
• Conversion Factors: Be aware that the proportion of ingested nutrient excreted in urine (e.g., the 80% factor for protein and potassium) is an estimate. Variations can occur due to individual physiology and diet composition [49].
• Day-to-Day Variation: A single 24-HDR and urine collection does not represent habitual intake. Multiple, non-consecutive measurements are required to estimate usual intake for individuals or populations [48] [47].
• Influence of Participant Characteristics: Factors such as higher Body Mass Index (BMI) are consistently associated with greater underreporting of energy and nutrients in self-reported dietary data [46] [47].

Within the broader research on standardized 24-hour dietary recall adaptation methodologies, the validation of country-specific versions of the GloboDiet software is a critical process. As national food markets evolve, with products being introduced or discontinued, dietary assessment tools must be updated and revalidated to ensure they continue to provide valid data for nutrition monitoring and research [17] [5]. This case study details the protocol and outcomes of the ErNst study (Erfassung der Energie- und Nährstoffzufuhr), which was conducted to validate the intensively updated German version of GloboDiet prior to its deployment in the German National Nutrition Monitoring [17] [51] [5].

Experimental Protocol and Study Methodology

Study Design and Participant Recruitment

The ErNst study was conducted as a cross-sectional observational study at the Max Rubner-Institut (MRI) in Karlsruhe, Germany, between October and December 2018 [17] [51]. The study employed a convenience sampling strategy, recruiting participants through an MRI-internal database, internet announcements, and local media [17] [5].

Inclusion Criteria: The study enrolled 117 participants, with 109 providing complete data for the GloboDiet validation [51]. Participants were healthy adults aged 18-79 years, equally distributed by sex (57 women, 52 men) and across three age groups (18-39, 40-59, and 60-79 years) [17] [51]. Individuals with diseases or medications known to affect nutrient or energy intake were excluded [17].

Sample Size Calculation: A necessary sample size of 50 men and 50 women was calculated a priori to achieve meaningful results, with test strength calculations performed by the Leibnitz Institute for Social Science in Mannheim, Germany [17] [5].

Data Collection Workflow

The study employed a comprehensive multi-method assessment protocol conducted over two visits to the study center, with data collection procedures summarized in Table 1.

Table 1: Data Collection Methods and Metrics in the ErNst Study

Assessment Method	Key Metrics Collected	Participants (n)
GloboDiet 24-h Recall	Food & beverage consumption, nutrient intake	109 [17]
24-h Urine Collection	Nitrogen, potassium, sodium, creatinine excretion	107 [17]
Accelerometry	Physical activity, total energy expenditure (TEE)	82 [17]
Anthropometry	Body height, weight, BMI, body composition (BIA)	117 [51]
Stool Sample	Microbiome composition (16S rRNA sequencing)	106 [51]
FFQ (30-day)	Habitual diet (for microbiome association analysis)	106 [51]

GloboDiet 24-Hour Dietary Recall Administration

The updated German GloboDiet version used in the study featured substantial revisions from previous iterations, including:

• Food List Updates: 600 new foods added and 525 obsolete foods deleted, resulting in a final list of approximately 2000 food items [17] [5]. Additions included vegan and vegetarian products (e.g., cereal drinks), while deletions included foods no longer available on the German market (e.g., specific fish types due to fishing bans) [5].
• Quantification Methods: Actualization of standard units, household measures, food shapes, and a picture book with approximately 100 photo series of different portion sizes [17] [5]. The system was extended to account for diverse product sizes (e.g., dairy products in mini to extra-large sizes, coffee-to-go cups) [5].
• Recipe Database: Updates to standard recipes, including addition of international dishes (e.g., sushi) and removal of less popular traditional dishes (e.g., those containing offal) [17] [5].

Trained interviewers conducted face-to-face 24-hour recalls using GloboDiet on Mondays, Wednesdays, or Thursdays, covering food consumption on Sundays, Tuesdays, or Wednesdays, respectively [17] [5]. The interviews followed the standardized GloboDiet structure: (1) collection of general participant information, (2) creation of a chronological "quick list" of consumed foods and beverages, (3) detailed description and quantification of food items, and (4) quality control checks [17] [4]. Nutrient intake was calculated by linking the food consumption data to the German Nutrient Database BLS (Bundeslebensmittelschlüssel) [17].

Biomarker Collection and Analysis

24-Hour Urine Collection: Participants collected urine over a 24-hour period, with completeness verified through creatinine excretion analysis [17] [5]. The following formulas were applied, with mean creatinine quotients of 87% for men and 78% for women, confirming completeness [17]:

• Women: Creatinine-ratio (%) = (100 × creatinine [mg/d]) / (21 × weight [kg])
• Men: Creatinine-ratio (%) = (100 × creatinine [mg/d]) / (24 × weight [kg])

Urinary nitrogen excretion was used to validate protein intake, while potassium excretion validated potassium intake [17] [5]. Sodium excretion was also measured, though a gap with estimated intake was expected due to difficulties in quantifying added salt [17].

Physical Activity Monitoring: Participants wore accelerometers for at least 24 hours to measure physical activity and calculate total energy expenditure (TEE) for comparison with energy intake from GloboDiet [17] [51].

Statistical Analysis

The agreement between dietary intake (GloboDiet) and biomarker measurements was assessed using multiple statistical approaches [17] [5]:

• Wilcoxon rank tests
• Confidence intervals
• Spearman correlations
• Bland-Altman plots

This multi-method approach provided comprehensive insights into the validity of the GloboDiet estimates [17].

Key Validation Results

Protein and Potassium Validation

The validation of GloboDiet against urinary biomarkers yielded distinct results for protein and potassium, as summarized in Table 2.

Table 2: Validation Results for Protein and Potassium Intake

Nutrient	Validation Biomarker	Correlation with Urinary Excretion	Key Findings
Protein	Nitrogen in 24-h urine	Statistically significant correlation [17]	Overall valid estimates of protein intake [17] [5]
Potassium	Potassium in 24-h urine	Weak correlation [17]	Ambiguous results: good agreement in Bland-Altman plots but potential underestimation in 24-h recalls [17] [5]

For potassium, the different statistical methods yielded conflicting interpretations: while Bland-Altman plots showed good agreement between 24-hour recalls and urine samples, the weak correlation suggested that 24-hour recalls might underestimate true potassium intake [17] [5].

Additional Analytical Findings

Energy Intake Validation: Energy intake estimated through GloboDiet was compared with total energy expenditure (TEE) measured by accelerometry in 82 participants [17]. As energy intake and expenditure undergo daily fluctuations, this comparison served only as a rough estimate of agreement at the group level in this short-term analysis [17].

Sodium Intake: A discrepancy was anticipated between estimated sodium intake from GloboDiet and sodium excretion measured in urine, primarily due to the difficulty in quantifying salt used during food preparation and added at the table [17] [5].

Visual Experimental Workflow

Diagram 1: ErNst Study Experimental Workflow

The Researcher's Toolkit: Essential Research Reagents and Materials

Table 3: Key Research Materials and Methods in the ErNst Study

Item/Resource	Specification/Version	Primary Function in Study
GloboDiet Software	Updated German version	Standardized 24-h dietary recall administration [17]
German Nutrient Database (BLS)	Current version	Nutrient intake calculation from food consumption data [17]
24-h Urine Collection Kit	Standardized containers	Collection of urinary biomarkers for validation [17]
Accelerometer	Not specified in sources	Objective measurement of physical activity and TEE [17]
Bioelectrical Impedance Analyzer	Not specified in sources	Assessment of body composition [51]
Food Picture Book	~100 photo series	Portion size estimation during dietary recall [17]

Discussion and Implications

Despite partially ambiguous results for potassium, the ErNst study concluded that the updated German GloboDiet version, when linked to the current German Nutrient Database, provides valid estimates of nutrient intake overall [17] [5]. This successful validation allowed for its planned implementation in the German National Nutrition Monitoring prepared to launch in 2024 [17].

This case study exemplifies the rigorous validation process required when adapting standardized dietary assessment tools like GloboDiet to evolving food environments. The methodology demonstrates the importance of using recovery biomarkers like urinary nitrogen and potassium to objectively validate self-reported dietary intake [17] [5]. The multi-method statistical approach provides a comprehensive framework for interpreting validation results, particularly when different statistical methods yield conflicting conclusions, as observed with potassium assessment.

The ErNst study protocol contributes significantly to the broader thesis on GloboDiet adaptation methodologies by providing a template for validation studies in other contexts, demonstrating the necessary sample sizes, biomarker selection, and statistical approaches required to ensure the validity of updated dietary assessment tools in national nutrition monitoring systems [17] [51] [5].

The accurate assessment of dietary intake is a cornerstone of nutritional epidemiology, public health monitoring, and the study of diet-disease relationships. Among the various methods available, 24-hour dietary recalls are widely valued for their ability to capture detailed quantitative intake data without altering habitual eating patterns. GloboDiet, developed under the auspices of the International Agency for Research on Cancer (IARC), represents one of the most standardized international methodologies for this purpose [1]. Its primary strength lies in a highly standardized interview structure and detailed food classification system, enabling valid cross-country comparisons.

Recently, the research landscape has witnessed the emergence of new priorities, particularly the need to assess food intake through the lens of processing. The NOVA classification system, developed by researchers at the University of São Paulo, has gained significant traction for categorizing foods into four groups based on the nature, extent, and purpose of industrial processing: (1) unprocessed or minimally processed foods, (2) processed culinary ingredients, (3) processed foods, and (4) ultra-processed foods (UPFs) [27] [52] [25]. Concurrently, the field has seen a shift towards web-based, self-administered tools like Intake24 and ASA24, which offer a cost-effective and scalable alternative to interviewer-led recalls.

This Application Note provides a comparative analysis of how GloboDiet, Intake24, and other major tools integrate the NOVA classification system. We present structured data, detailed protocols, and visual workflows to guide researchers in selecting and implementing the most appropriate methodology for investigating the role of food processing in health and disease, within the broader context of standardized dietary recall adaptation.

Core Characteristics of Major Dietary Assessment Tools

Table 1: Core characteristics and NOVA integration of major 24-hour dietary recall tools.

Feature	GloboDiet	Intake24	ASA24 (Automated Self-Administered 24-Hour Recall)	Nova24h
Primary Format	Interviewer-led, computer-assisted software [1]	Self-completed, web-based [53] [54]	Self-completed, web-based [27]	Self-completed, web-based [27]
Standardization	High; international standardized methodology (SOPs) [1]	Adaptable open-source platform [53] [54]	Standardized for the US context [52]	Developed specifically for the NutriNet Brasil cohort [27]
Core Food Classification	Extensive, hierarchical food and recipe database [1]	Country-specific food databases [20] [53]	USDA Food and Nutrient Database for Dietary Studies (FNDDS) [52]	Food list based on Brazilian national survey [27]
NOVA Integration Method	Post-hoc manual or semi-automated linkage of food codes [1]	Post-hoc application to food codes; requires customization [20]	Pre-defined linkage files between FNDDS codes and NOVA groups [52]	Fully integrated; automatic classification during data collection [27]
Inherent NOVA Output	No	No	Yes (via supplementary files) [52]	Yes (primary output) [27]

Quantitative Performance Metrics

Table 2: Documented performance metrics related to tool application and NOVA classification.

Metric	Intake24 (South Asia Biobank)	Nova24h (NutriNet Brasil)	Manual Nova Coding (US Study)
Recall Completion Time	Median: 13 minutes [20]	~15 minutes [27]	Not Applicable (Coder task)
Food Database Size	2,283 items [20]	526 food items capturing all variations [27]	3,099 unique foods coded [25]
Agreement with Reference	N/A	Intraclass Correlation Coefficients (ICCs): 0.54 - 0.78 for Nova groups [27]	Inter-coder concordance: 88.3%; Cohen’s κ: 0.75 [25]
Energy Contribution from UPFs	N/A	19.0% (Nova24h) vs. 20.9% (reference tool) [27]	62% of daily calories for US children [25]

Integration Methodologies and Experimental Protocols

Protocol 1: Adapting a Standardized Tool (GloboDiet) for a New Population

The adaptation of GloboDiet for the Korean national survey exemplifies a rigorous protocol for implementing a standardized recall tool in a new cultural and dietary context [1].

• Objective: To create a Korean version of the GloboDiet software and its complementary tools, ensuring it captures the specificities of the Korean diet while maintaining international standardizability.
• Materials:
  • GloboDiet Software Core Structure: The international software framework from IARC.
  • Local Food Composition Tables: Comprehensive databases of Korean foods and beverages.
  • Dietary Survey Data: Existing national survey data to identify commonly consumed foods.
• Procedure:
  • Database Customization: Translate and customize approximately seventy common and country-specific databases. This includes:
    • Food List: Populate with foods and culinary preparations specific to Korea (e.g., Kimchi, specific types of rice cakes).
    • Descriptor Facets: Adapt existing facets or create new ones to accurately describe specific Korean foods (e.g., fermentation status, specific cooking methods like "Bokkeum" stir-frying).
    • Recipe Compilation: Document standard recipes for common Korean mixed dishes.
  • Quantification Method Adaptation: Critically evaluate and adapt quantification methods.
    • Develop a picture book of foods/dishes with portion sizes relevant to the Korean diet.
    • Incorporate household measures and standard units commonly used in Korea.
  • Software and Protocol Finalization:
    • Integrate all customized databases and quantification aids into the software.
    • Train interviewers on the use of the Korean GloboDiet version and the standardized interviewing protocol.
• Outcome: The successful development of the Korean GloboDiet version confirmed the methodology's flexibility for adaptation in an Asian context, providing a model for other populations [1].

Protocol 2: Implementing NOVA Classification with Dietary Recalls

This protocol details a validated method for training coders to apply the NOVA system to foods collected via 24-hour recalls, a common requirement for tools like GloboDiet and Intake24 [25].

• Objective: To reliably and validly assign NOVA categories to individual foods from 24-hour dietary recalls.
• Materials:
  • Dietary Dataset: A dataset of unique foods identified from 24-hour recalls (e.g., from NDSR, GloboDiet, or other systems).
  • Reference Documents: The most recent NOVA classification publications detailing definitions and examples for all four groups [27] [52] [25].
  • Coding Guide: A project-specific guide with decision trees for ambiguous or multi-ingredient foods.
• Procedure:
  • Coder Training:
    • Conduct interactive training sessions using the NOVA reference documents.
    • Practice categorizing a wide range of foods, with a focus on challenging items (e.g., multi-ingredient foods, brands).
    • Discuss discrepancies as a group to establish consensus.
  • Certification:
    • Require coders to independently categorize a standardized set of foods.
    • Achieve a pre-specified agreement threshold (e.g., >85% concordance with a master coder) before proceeding to study data.
  • Independent Coding and Reliability Assessment:
    • Assign each unique food in the dataset to two independent, certified coders.
    • Calculate inter-rater reliability using percent concordance and Cohen’s κ coefficient.
    • Resolve discrepancies through consensus or by a third, senior coder.
  • Validity Check:
    • Merge the finalized NOVA categories back into the full dietary dataset.
    • Assess construct validity by comparing the average daily macronutrient content (e.g., % energy from added sugar, protein) across NOVA groups. Expect UPFs to contribute disproportionately to added sugars and processed culinary ingredients to fats [25].
• Outcome: A reliably coded dataset where each food item is assigned a NOVA category, enabling the analysis of dietary intake by level of food processing.

The Scientist's Toolkit: Research Reagent Solutions

Table 3: Essential resources for implementing NOVA classification in dietary surveys.

Item / Resource	Function / Description	Example / Source
GloboDiet Software	Standardized, interviewer-led 24-hour recall software platform for international studies.	International Agency for Research on Cancer (IARC) [1]
Intake24	Open-source, web-based, self-completed 24-hour recall system for scalable dietary assessment.	Newcastle University / University of Cambridge [53] [54]
ASA24	Automated, self-administered 24-hour recall tool developed by the National Cancer Institute (NCI).	National Cancer Institute (NCI), USA [27] [52]
NOVA Classification Reference	Authoritative documents defining the four food processing groups and providing categorization criteria.	Monteiro et al. (2019) [52] [25]
USDA FNDDS Linkage Files	Pre-defined files linking USDA food codes to NOVA groups for use with ASA24 and NHANES data.	Available via proposal to NCI's Risk Factor Assessment Branch [52]
Food Propensity Questionnaire (FPQ)	A questionnaire to assess a participant's usual frequency of consuming specific foods over time.	Used in NHANES and other national surveys to complement 24-hour recalls [55]
Standardized Portion Size Aids	Visual aids (e.g., picture books, photographs) to improve the accuracy of portion size estimation.	Intake24 includes >2400 photographs; GloboDiet uses custom picture books [1] [56]

Comparative Workflow: From Data Collection to NOVA Analysis

The pathway from collecting dietary data to generating NOVA-based findings differs significantly depending on the tool chosen. The following diagram contrasts the workflows for a tool with integrated NOVA classification (Nova24h) versus tools requiring post-hoc classification (GloboDiet, Intake24).

The integration of the NOVA classification system into 24-hour dietary recalls marks a significant advancement in nutritional epidemiology, enabling critical research on the health impacts of food processing. The choice of dietary assessment tool involves a strategic trade-off.

GloboDiet offers unparalleled standardization and detail, making it ideal for multinational studies, but requires significant resource investment for adaptation and post-hoc NOVA coding. Intake24 provides a cost-effective, scalable, and flexible open-source solution, though it similarly necessitates post-hoc integration of NOVA. In contrast, ASA24 benefits from pre-existing, standardized linkage files for the US context, streamlining the process for American researchers. The specialized Nova24h tool demonstrates the feasibility and efficiency of a fully integrated system, though it is currently tailored to a specific population.

For researchers framing their work within the context of standardized adaptation methodology, GloboDiet provides a proven, rigorous model. However, for large-scale studies where cost and scalability are primary concerns, self-administered tools like Intake24 and ASA24, coupled with a robust protocol for NOVA implementation, offer a powerful and valid alternative. The decision ultimately hinges on the specific research question, population, and available resources.

The accurate assessment of dietary exposure is a fundamental prerequisite for robust nutritional epidemiology, enabling the linkage of food consumption to health outcomes and disease burden. Standardized 24-hour dietary recall (24-HDR) methodologies provide the precise quantitative data necessary to investigate these relationships. Within this context, GloboDiet (formerly EPIC-Soft), developed by the International Agency for Research on Cancer (IARC/WHO), represents a state-of-the-art, computer-assisted 24-HDR tool designed for international standardization [4] [1]. Its structured approach minimizes random and systematic errors, thereby enhancing data quality and comparability across diverse populations [17] [36].

This protocol details the application of GloboDiet data to attribute disease burden, using the association between ultra-processed foods (UPFs) and type 2 diabetes as a primary example. We frame this within the broader methodological thesis of adapting GloboDiet for specific populations, a process that ensures the food database, quantification methods, and dietary interview protocols are contextually relevant [20] [1]. The precise and standardized dietary data generated by GloboDiet is crucial for establishing reliable exposure-disease relationships, which can inform public health policies and nutritional guidelines.

GloboDiet Methodology and Adaptation Framework

GloboDiet operates as a highly structured interview, guided by standardized software, which systematically collects detailed information on all foods and beverages consumed in the preceding 24 hours. The core of its methodology lies in the multi-step interview process and the extensive, adaptable databases that support it.

The Structured Interview Process

The GloboDiet interview is meticulously designed to enhance memory and standardize data collection across participants and interviewers [4] [40]. The process consists of five key stages, as illustrated in the workflow below:

Stage 1: General Information and Quick List. The interview begins by collecting non-dietary information about the participant and the recalled day. The respondent then freely recalls all consumed foods and beverages chronologically, creating a "quick list" without any quantification [40].

Stage 2: Description and Facet-Based Probing. Each food item from the quick list is described in detail using a system of predefined "facets" (questions) and "descriptors" (answers). These facets cover critical attributes such as food source, cooking method (e.g., fried, baked, steamed), fat content, and processing level—a key facet for identifying UPFs [4] [40]. This step is crucial for standardizing the description of foods like a "plant-based burger" by specifying its degree of processing.

Stage 3: Quantification. The consumed amount of each described food is estimated using multiple complementary methods. These include photographic atlases of portion sizes, standard household measures (e.g., spoons, cups), standard units (e.g., one apple), and food shapes [17] [1]. The adaptation of these quantification tools to local tableware and portion sizes is essential for accuracy [1].

Stage 4: Quality Control and Final Checks. Integrated probing questions prompt the interviewer to ask about commonly forgotten items (e.g., sugars in tea, fats used in cooking). Automated checks flag potential errors, such as an implausibly high or low daily energy intake [17] [40].

Stage 5: Data Export and Nutrient Calculation. The fully described and quantified food consumption data are linked to a compatible food composition database (FCDB) to calculate the intake of nutrients, food groups, and other bioactive compounds [17].

Adaptation for Research and Surveillance

A core strength of GloboDiet is its flexibility for adaptation to different countries and food cultures without compromising standardization. The process involves several key components, whose quantitative scope is summarized in the table below.

Table 1: Scope of GloboDiet Database Adaptation in Select Countries/Regions

Adaptation Component	South Asia Biobank [20]	Republic of Korea [1]	Germany (Updated Version) [17]
Total Food Items in Database	2,283 items	Not Specified	~2,000 items
New Foods Added	Context-specific foods	Country-specific foods	~600 items (e.g., vegan products)
Foods Removed	Not Specified	Not Specified	525 items (obsolete foods)
Quantification Methods	Local tools adapted	Local picture book, packages	~3,550 standard units, 100 photo series
Key Adaptation Features	Local food probes, portion sizes	New food subgroups, descriptors	Updated recipes, new dish types (e.g., sushi)

The adaptation process is comprehensive. It involves expanding the food classification to accommodate local food subgroups and adding new descriptors to existing facets to capture the unique characteristics of regional foods [1]. Furthermore, quantification methods are critically evaluated and adapted using pictures of local tableware, market food packages, and portion sizes relevant to the dietary habits of the target population [17] [1]. This ensures that the tool has good coverage of the local food supply, as demonstrated by the South Asia Biobank adaptation, which showed 99% of recalls included more than 8 items, with only 8% having missing foods [20].

Experimental Protocol: Linking GloboDiet Data to Diabetes Outcomes

This section outlines a detailed protocol for a hypothetical cohort study investigating the association between UPF consumption, assessed by GloboDiet, and the incidence of type 2 diabetes.

Study Design and Population

• Design: Prospective cohort study.
• Participants: Adults aged 40-79 years, free of diabetes at baseline.
• Sample Size: To be determined by power calculation, aiming to detect a predefined hazard ratio for diabetes per 10% increase in UPF consumption.
• Recruitment: Recruit a population-based sample, stratified by age and sex to ensure representativeness. Exclusion criteria typically include pre-existing diabetes, cancer, cardiovascular disease, or conditions leading to severe dietary restrictions [57].

Data Collection Workflow

The data collection procedure for linking dietary exposure to disease outcome involves multiple timed assessments, as depicted below.

Baseline Exposure Assessment:

• GloboDiet Interviews: Conduct at least two, non-consecutive 24-HDR interviews using the adapted GloboDiet software, administered by trained interviewers. One recall should ideally cover a weekend day [36]. Interviews can be performed face-to-face or by telephone.
• Ultra-Processed Food Classification: The detailed food description from GloboDiet (facets and descriptors) allows for the post-hoc classification of foods according to the NOVA system, which categorizes foods based on the extent and purpose of industrial processing. A food's processing level can be determined based on facets describing physical state, cooking method, and ingredient composition [4].
• Calculation of UPF Exposure: The proportion of total daily energy intake derived from UPF items is calculated for each participant. This serves as the primary exposure variable.

Covariate Assessment:

• Anthropometrics: Measure height, weight, and waist circumference using standardized protocols to calculate Body Mass Index (BMI).
• Lifestyle and Socio-demographics: Administer questionnaires to collect data on physical activity, smoking status, educational attainment, and income.
• Biomarkers: Collect fasting blood and 24-hour urine samples. Urinary nitrogen and potassium can be used as recovery biomarkers to validate the assessment of protein and potassium intake, thereby evaluating the accuracy of self-reported dietary data [17] [57].

Outcome Ascertainment:

• Follow-up: Conduct biennial follow-ups for a minimum of 10 years.
• Diabetes Diagnosis: Identify incident type 2 diabetes cases through a combination of methods: follow-up questionnaires, validated by medical records; fasting blood glucose measurements; glycated hemoglobin (HbA1c) levels; and use of anti-diabetic medication.

Data Analysis Plan

• Primary Analysis: Use Cox proportional hazards regression models to calculate hazard ratios (HR) and 95% confidence intervals (CI) for the association between the energy proportion of UPFs (as a continuous variable) and incident diabetes, adjusting for potential confounders such as age, sex, BMI, total energy intake, physical activity, and smoking.
• Secondary Analyses: Categorize UPF intake into quartiles to explore non-linear relationships. Conduct stratified analyses by sex, BMI, and genetic predisposition to diabetes.
• Measurement Error Correction: Where possible, use the data from recovery biomarkers (e.g., urinary nitrogen) in measurement error models to correct relative risk estimates for the measurement error inherent in self-reported dietary data [17] [36].

The Scientist's Toolkit: Key Research Reagents and Materials

The following table details essential materials and tools required for implementing the GloboDiet methodology and validating its data in epidemiological research.

Table 2: Essential Research Reagents and Solutions for GloboDiet-Based Studies

Tool/Reagent	Function in Protocol	Specifications & Examples
GloboDiet Software	Core platform for conducting standardized 24-hour dietary recalls.	Includes the interview software, country-specific food list, facet descriptors, and quantification methods [17] [4].
Food Composition Database (FCDB)	Converts food consumption data into nutrient intake values.	Must be compatible with GloboDiet food codes. Examples: German Nutrient Database (BLS), West-African Food Composition Table [17] [36].
Portion Size Estimation Aids	Visual aids to improve the accuracy of food amount quantification.	Picture books, photographs of portion sizes, images of household measures, food models, and shape dimension cards [17] [1].
Urinary Biomarkers (Nitrogen, Potassium)	Objective validation markers for assessing the accuracy of self-reported intake of protein and potassium.	24-hour urine collections; analysis of nitrogen (for protein) and potassium concentrations [17] [57].
Accelerometer	Device to measure physical activity and estimate total energy expenditure.	Used as an objective measure to compare against self-reported energy intake for identifying under- or over-reporters [57].

The GloboDiet methodology provides a robust and standardized framework for collecting high-quality dietary data that is capable of being linked to disease endpoints such as type 2 diabetes. Its rigorous, multi-faceted approach to food description and quantification minimizes measurement error and allows for the precise classification of exposures like ultra-processed foods. The successful adaptation of GloboDiet across diverse cultural and food environments [20] [1] underscores its utility as a core tool for global nutritional surveillance and etiological research. By following the detailed application notes and protocols outlined in this document, researchers can generate reliable and comparable evidence on the dietary determinants of disease burden, ultimately informing effective public health strategies.

Standardized 24-hour dietary recalls represent a cornerstone in nutritional epidemiology, providing the foundational data for understanding population dietary habits and their relationship to health outcomes. Within this landscape, GloboDiet (formerly EPIC-Soft), developed by the International Agency for Research on Cancer (IARC), has emerged as a premier interviewer-administered software that enables standardized collection of dietary data across different countries and populations [5] [58]. This application note explores the integration of GloboDiet's robust dietary data with advanced machine learning (ML) methodologies to uncover complex dietary patterns, moving beyond traditional analysis approaches to enable more precise public health interventions and personalized nutrition strategies.

The adaptation of GloboDiet for national nutrition monitoring, as demonstrated by the recent German validation study, ensures that the software remains current with evolving food supplies, including the addition of plant-based alternatives and international dishes while removing obsolete items [5]. This continuous refinement process guarantees that ML algorithms are applied to nutritionally relevant and contemporary data, enhancing the real-world applicability of generated insights.

GloboDiet as a Standardized Data Collection Platform

Core Architecture and Validation

GloboDiet employs a structured interview procedure with integrated quality assurance mechanisms, ensuring standardized data collection across diverse populations [5]. The software's architecture incorporates:

• Comprehensive food classification with approximately 2,000 food items
• Detailed quantification methods including standard units, household measures, and picture books
• Recipe modification capabilities allowing for both preset standard recipes and custom ingredient entry
• Regular database updates reflecting changing food markets and consumption patterns

The recent validation of the updated German GloboDiet version demonstrated strong agreement between protein intake estimates from GloboDiet recalls and urinary nitrogen excretion, establishing its reliability for nutritional assessment [5]. While potassium validation showed more ambiguous results, the overall findings support GloboDiet's validity for estimating nutrient intake when linked to appropriate food composition databases.

Adaptation Frameworks

The adaptation of GloboDiet for different populations follows methodological frameworks similar to those used for other dietary assessment tools like Intake24. As demonstrated in the New Zealand adaptation, this process involves [24]:

• Baseline food list selection based on countries with similar food supplies
• Comprehensive review at food-category level with consideration of public health priorities
• Addition of country-specific foods identified through multiple data sources
• Nutrient database linkage to local food composition tables
• Portion size customization with local household measures and photo series

This systematic adaptation approach ensures that GloboDiet data captures culturally relevant dietary patterns while maintaining standardization necessary for cross-population comparisons and machine learning applications.

Machine Learning Approaches for Dietary Pattern Discovery

Comparative Analysis of Clustering Algorithms

Table 1: Performance comparison of clustering algorithms for dietary pattern identification

Algorithm	Key Parameters	Validation Metrics	Performance Notes	Implementation Considerations
K-means	Number of clusters (K), distance metric	Silhouette Index: -1 to 1 (closer to 1 better)	Identified Traditional vs. Health-conscious patterns in Dutch population [59]	Sensitive to initial centroids; requires standardized data
K-medoids	Number of clusters, distance metric	Davies-Bouldin Index: 0 to ∞ (lower better)	More robust to noise than K-means	Computationally intensive for large datasets
Hierarchical	Linkage method, distance threshold	Dunn Index: 0 to ∞ (higher better)	Provides dendrogram for cluster number selection	Memory intensive for large samples
Density-based	Epsilon neighborhood, minimum points	Calinski-Harabasz Index: higher values preferable	Identifies arbitrary shaped clusters	Struggles with varying densities

A recent study utilizing Dutch National Food Consumption Survey data demonstrated the systematic comparison of these clustering methods, with K-means emerging as the optimal approach for identifying two distinct dietary patterns in both sexes: "Traditional" (characterized by higher energy intake, bread, potatoes, red/processed meat, coffee, fats/oils, and sugary drinks) and "Health-conscious" (characterized by higher consumption of fruits, vegetables, tea, nuts, seeds, and breakfast cereals) [59].

Classification Algorithms for Dietary Pattern Prediction

Table 2: Classification algorithms for predicting dietary patterns based on sociodemographic factors

Classifier	Key Features	Accuracy Range	Important Predictors	Implementation Considerations
Naïve Bayes	Probabilistic, based on Bayes theorem	60-68% [59]	Education level, age, BMI	Works well with high-dimensional data
K-Nearest Neighbors	Instance-based learning	60-68% [59]	Education level, age, BMI	Sensitive to feature scaling and distance metrics
Decision Tree	White-box model, interpretable rules	60-68% [59]	Education level, age, BMI	Prone to overfitting without regularization
Random Forest	Ensemble of decision trees	60-68% [59]	Education level, age, BMI	Reduces overfitting, provides feature importance
Support Vector Machine	Finds optimal hyperplane	60-68% [59]	Education level, age, BMI	Effective in high-dimensional spaces
XGBoost	Gradient boosting framework	60-68% [59]	Education level, age, BMI	Handles missing values, often top performer

The comparative analysis of these classifiers demonstrated moderate predictive accuracies (60-68%) for identifying dietary patterns based on sociodemographic and lifestyle factors, with education level, age, and BMI consistently emerging as the most important predictors across algorithms [59].

Integrated Experimental Protocol for ML-Driven Dietary Pattern Analysis

Data Preprocessing and Feature Engineering

Protocol 4.1.1: GloboDiet Data Standardization

• Food Group Aggregation: Convert individual food items from GloboDiet recalls into standardized food groups (e.g., 29 food groups as used in the Dutch study) [59]
• Nutrient Calculation: Link consumed foods to appropriate food composition databases (e.g., German Nutrient Database BLS) for nutrient intake estimation [5]
• Energy Adjustment: Apply density methods or residual methods to account for variations in total energy intake
• Covariate Processing: Standardize sociodemographic (age, education, urbanization) and lifestyle (physical activity, smoking status) variables for classification tasks

Protocol 4.1.2: Cluster Validation and Selection

• Internal Validation: Calculate multiple internal validation metrics including:
  • Silhouette Index (target: >0.5)
  • Davies-Bouldin Index (target: lower values)
  • Dunn Index (target: higher values)
  • Calinski-Harabasz Index (target: higher values) [59]
• Cluster Stability: Assess stability through resampling techniques
• Biological Plausibility: Evaluate interpretability and alignment with existing nutritional knowledge

Machine Learning Implementation Workflow

Diagram 1: ML-driven dietary pattern discovery workflow

Table 3: Research reagents and computational tools for ML-based dietary analysis

Resource Category	Specific Tools/Platforms	Application in Dietary Pattern Analysis	Access Considerations
Dietary Assessment Platforms	GloboDiet, ASA24, Intake24	Standardized 24-hour recall data collection	GloboDiet licensed through IARC; ASA24 free for research [60]
Food Composition Databases	German BLS, USDA FNDDS, New Zealand FCDB	Nutrient calculation from food intake data	Country-specific access; requires regular updates [5] [24]
Clustering Algorithms	K-means, K-medoids, Hierarchical, DBSCAN	Unsupervised dietary pattern identification	Available in R (cluster, factoextra) and Python (scikit-learn) [59]
Classification Algorithms	Random Forest, XGBoost, SVM, Naïve Bayes	Predicting dietary patterns from sociodemographics	Open-source implementations widely available [59]
Validation Metrics	Silhouette Index, Davies-Bouldin, Dunn Index	Cluster quality assessment	Specialized R packages (clValid, NbClust) [59]
Image Recognition Tools	MyFoodRepo, Mask R-CNN, DeepLab V3	Automated food identification from images	MyFoodRepo dataset publicly available for research [61]

Validation and Interpretation Framework

Biomarker Validation Protocols

Protocol 6.1.1: Urinary Biomarker Validation

• Sample Collection: Collect 24-hour urine samples concurrently with GloboDiet 24-hour recalls [5]
• Biomarker Analysis: Measure urinary nitrogen (for protein validation) and potassium excretion
• Statistical Comparison:
  • Apply Wilcoxon rank tests for intake-excretion differences
  • Calculate confidence intervals for mean differences
  • Generate Bland-Altman plots for agreement assessment
  • Compute Spearman correlations for relationship strength [5]

Protocol 6.1.2: Model Performance Validation

• Train-Test Split: Reserve 20-30% of data for model validation
• Cross-Validation: Implement k-fold cross-validation (typically k=5 or k=10)
• Metric Selection: Use accuracy, precision, recall, F1-score for classification; internal validation metrics for clustering
• Feature Importance: Calculate and interpret feature importance scores from tree-based models

Interpretation and Translation Guidelines

The interpretation of ML-derived dietary patterns requires both statistical rigor and nutritional expertise:

• Pattern Labeling: Assign descriptive labels (e.g., "Traditional," "Health-conscious") based on predominant food groups with strong factor loadings or cluster characteristics [59]
• Nutrient Profiling: Analyze associated nutrient intakes to assess nutritional adequacy or excess
• Population Segmentation: Identify sociodemographic correlates to target public health interventions
• Temporal Tracking: Monitor pattern stability or shifts over time, including in response to global events like COVID-19 [59]

The integration of GloboDiet data with machine learning methodologies represents a paradigm shift in nutritional epidemiology, enabling more sophisticated dietary pattern discovery that accounts for the complex, synergistic nature of dietary intake. The structured protocols outlined in this application note provide researchers with a comprehensive framework for implementing these advanced analytical approaches while maintaining scientific rigor. As dietary assessment technologies continue to evolve, including the incorporation of image recognition and mobile health platforms, the potential for machine learning to unravel the complexities of diet-disease relationships will only expand, ultimately supporting more effective, personalized public health nutrition strategies.

Conclusion

The adaptation of the GloboDiet methodology provides a robust, standardized framework for collecting high-quality and comparable dietary data across diverse global populations, which is fundamental for advancing nutritional epidemiology and public health. The key takeaways underscore the critical importance of meticulous customization of food databases and quantification tools, the necessity of rigorous validation against biomarkers, and the effective troubleshooting of field challenges. For future biomedical and clinical research, the integration of GloboDiet with novel technologies—such as wearable sensors for passive monitoring and machine learning for advanced dietary pattern analysis—promises to deepen our understanding of diet-disease relationships. Furthermore, expanding its implementation in underrepresented regions will be crucial for building a truly global evidence base to inform targeted interventions and policies against the burden of non-communicable diseases.

References

• 1. Adapting a standardised international 24 h dietary recall ... [https://pubmed.ncbi.nlm.nih.gov/25899045/]
• 2. and interview-based 24 h dietary recall method (GloboDiet) for ... [https://pmc.ncbi.nlm.nih.gov/articles/PMC10261577/]
• 3. Compilation of a standardised international folate database ... [https://www.sciencedirect.com/science/article/abs/pii/S0308814614017737]
• 4. Evaluation of the international standardized 24-h dietary recall ... [https://globalizationandhealth.biomedcentral.com/articles/10.1186/s12992-017-0260-6]
• 5. Validation of the Updated GloboDiet Version by Protein ... [https://www.mdpi.com/2072-6643/15/20/4418]
• 6. and interview-based 24 h dietary recall method (GloboDiet ... [https://pubmed.ncbi.nlm.nih.gov/28803598/]
• 7. Comunicaciones Orales [https://www.alanrevista.org/ediciones/2015/suplemento-2/art-229/]
• 8. Evaluation of the international standardized 24-h dietary recall ... [https://experts.umn.edu/en/publications/evaluation-of-the-international-standardized-24-h-dietary-recall-/]
• 9. Different dietary assessment methods, similar conclusions? ... [https://pmc.ncbi.nlm.nih.gov/articles/PMC9991603/]
• 10. Validity of Dietary Assessment Methods When Compared ... [https://www.frontiersin.org/journals/endocrinology/articles/10.3389/fendo.2019.00850/full]
• 11. A Systematic Review Examining Contributors to ... [https://www.sciencedirect.com/science/article/pii/S2161831323000947]
• 12. cambridge.org/core/journals/public-health-nutrition/article/adapting-the... [https://www.cambridge.org/core/journals/public-health-nutrition/article/adapting-the-standardised-computer-and-interviewbased-24-h-dietary-recall-method-globodiet-for-dietary-monitoring-in-latin-america/443B9BF21952A518705F42542EA4631D]
• 13. Validity and reliability of an online self-report 24-h dietary ... [https://pmc.ncbi.nlm.nih.gov/articles/PMC6722486/]
• 14. Data Analysis for the 2025 Guidelines Committee Report [https://www.dietaryguidelines.gov/2025-advisory-committee-report/data-analysis]
• 15. The standardized computerized 24-h dietary recall method ... [https://pubmed.ncbi.nlm.nih.gov/21731006/]
• 16. Structure of the standardized computerized 24-h diet recall ... [https://pubmed.ncbi.nlm.nih.gov/10094230/]
• 17. Validation of the Updated GloboDiet Version by Protein ... [https://pmc.ncbi.nlm.nih.gov/articles/PMC10609925/]
• 18. Evaluation of the international standardized 24-h dietary recall methodology (GloboDiet) for potential application in research and surveillance within African settings - PMC [https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5477249/]
• 19. Evaluation of the international standardized 24-h dietary recall methodology (GloboDiet) for potential application in research and surveillance within African settings - PubMed [https://pubmed.ncbi.nlm.nih.gov/28629424/]
• 20. The Adaptation, Implementation, and Performance ... [https://www.sciencedirect.com/science/article/pii/S2475299125000022]
• 21. Adapting a standardised international 24 h dietary recall ... [https://www.cambridge.org/core/journals/british-journal-of-nutrition/article/adapting-a-standardised-international-24-h-dietary-recall-methodology-globodiet-software-for-research-and-dietary-surveillance-in-korea/359E2CCEF9FA86F90D4D2AFA2BA5EB16]
• 22. Improving Pregnant Women's Iodine Intake Estimates and ... [https://pmc.ncbi.nlm.nih.gov/articles/PMC9966772/]
• 23. Association of Maternal Consumption of Ultra-Processed ... [https://www.mdpi.com/1660-4601/22/4/608]
• 24. Developing a food list for a new 24‐h dietary recall tool for ... [https://pmc.ncbi.nlm.nih.gov/articles/PMC12168050/]
• 25. Reliability and validity of assigning ultraprocessed food ... [https://pmc.ncbi.nlm.nih.gov/articles/PMC10196599/]
• 26. Food Composition Databases - an overview [https://www.sciencedirect.com/topics/food-science/food-composition-databases]
• 27. A novel web-based 24-h dietary recall tool in line with the ... [https://www.cambridge.org/core/journals/public-health-nutrition/article/novel-webbased-24h-dietary-recall-tool-in-line-with-the-nova-food-processing-classification-description-and-evaluation/23D1A51DDB40190F0FBCFE93D9B41DE9]
• 28. Food composition database development for between country ... [https://pmc.ncbi.nlm.nih.gov/articles/PMC1388230/]
• 29. Evaluation of the international standardized 24-h dietary ... [https://pmc.ncbi.nlm.nih.gov/articles/PMC5477249/]
• 30. Development of a Multilingual Web-Based Food Frequency ... [https://pmc.ncbi.nlm.nih.gov/articles/PMC10610353/]
• 31. Performance of the Brazilian version of GloboDiet software ... [https://link.springer.com/article/10.1186/s41110-018-0072-3]
• 32. Secular trends in dietary patterns among Korean adults [https://nutritionj.biomedcentral.com/articles/10.1186/s12937-025-01235-0]
• 33. International agency for research on cancer-World Health ... [https://www.alanrevista.org/ediciones/2015/suplemento-1/art-236/]
• 34. Table 3 Experts evaluation of GloboDiet sections and ... [https://globalizationandhealth.biomedcentral.com/articles/10.1186/s12992-017-0260-6/tables/3]
• 35. Perspectives from individuals with low education and ... [https://pmc.ncbi.nlm.nih.gov/articles/PMC7132555/]
• 36. Measurement Errors in Dietary Assessment Using Self- ... [https://www.sciencedirect.com/science/article/pii/S2161831322008274]
• 37. Plant-based retail market overview | GFI [https://gfi.org/marketresearch/]
• 38. Market & Trends: Latest News 2025 [https://vegconomist.com/category/market-and-trends/]
• 39. Global plant-based trends 2025. Plant-based foods and ... [https://www.innovamarketinsights.com/trends/plant-based-trends-2025/]
• 40. Table 1 Structure of an interview by GloboDiet [https://globalizationandhealth.biomedcentral.com/articles/10.1186/s12992-017-0260-6/tables/1]
• 41. A vegan dietary pattern is associated with high prevalence ... [https://www.sciencedirect.com/science/article/pii/S1279770725000600]
• 42. Quality Control for Interviews to Obtain Dietary Recalls from ... [https://pmc.ncbi.nlm.nih.gov/articles/PMC1435375/]
• 43. Quality control for interviews to obtain dietary recalls from ... [https://www.sciencedirect.com/science/article/abs/pii/S0002822304012507]
• 44. Evaluation of quality control procedures for 24-h dietary ... [https://www.sciencedirect.com/science/article/abs/pii/S0091743504000453]
• 45. Africa launch of the Global Evaluation Initiative | undp-nec [https://nec.undp.org/event/building-what-already-there-africa-launch-global-evaluation-initiative]
• 46. Bias in protein and potassium intake collected with 24-h ... [https://pmc.ncbi.nlm.nih.gov/articles/PMC3496541/]
• 47. Comparison of self-reported dietary intakes from the ... [https://www.sciencedirect.com/science/article/pii/S0002916522027459]
• 48. Protocol of the validation of the experience sampling-based ... [https://pmc.ncbi.nlm.nih.gov/articles/PMC12587619/]
• 49. Validation of the web-based self-administered 24-h dietary ... [https://pmc.ncbi.nlm.nih.gov/articles/PMC8437850/]
• 50. Validity of an online 24-h recall tool (myfood24) for dietary ... [https://bmcmedicine.biomedcentral.com/articles/10.1186/s12916-018-1113-8]
• 51. Protocol Assessment of Energy and Nutrient Intake and the ... [https://www.sciencedirect.com/org/science/article/pii/S1929074823000094]
• 52. Using Nova with ASA24 & NHANES | EGRP/DCCPS/NCI/NIH [https://epi.grants.cancer.gov/nova/]
• 53. Intake24 [https://intake24.org/]
• 54. Intake24 - MRC Epidemiology Unit - University of Cambridge [https://www.mrc-epid.cam.ac.uk/research/measurement-platform/dietary-assessment/intake24/]
• 55. National Nutrition Surveys Applying Dietary Records or 24- ... [https://www.mdpi.com/2072-6643/15/22/4739]
• 56. Dietary data collection using Intake24 [https://bio-protocol.org/exchange/minidetail?id=6176388&type=30]
• 57. Assessment of Energy and Nutrient Intake and the ... [https://www.researchprotocols.org/2023/1/e42529]
• 58. Scaling up Dietary Data for Decision-Making in Low-Income ... [https://pmc.ncbi.nlm.nih.gov/articles/PMC5683006/]
• 59. Identifying and predicting dietary patterns in the Dutch ... [https://link.springer.com/article/10.1007/s00394-025-03817-4]
• 60. ASA24® Dietary Assessment Tool | EGRP/DCCPS/NCI/NIH [https://epi.grants.cancer.gov/asa24/]
• 61. The Food Recognition Benchmark: Using Deep Learning ... [https://www.frontiersin.org/journals/nutrition/articles/10.3389/fnut.2022.875143/full]