ASA24 vs. Intake24 vs. MyFood24: A Comprehensive Validation and Comparative Review for Researchers

Abstract

This article provides a systematic comparison of three prominent online 24-hour dietary recall tools—ASA24, Intake24, and MyFood24—for researchers and clinical professionals. It explores the foundational principles and evolution of these automated systems, detailing their methodological approaches and application in diverse populations. The review synthesizes evidence from validation studies, including comparisons against interviewer-led recalls, weighed food records, and biomarker technologies, to assess the relative accuracy and measurement error of each tool. Furthermore, it addresses common challenges in implementation and offers optimization strategies for large-scale studies, clinical trials, and nutritional epidemiology, serving as a decision-making guide for selecting the most appropriate dietary assessment tool.

The Digital Shift in Dietary Assessment: Understanding ASA24, Intake24, and MyFood24

Accurate dietary assessment is fundamental to understanding the relationships between nutrition, health, and disease. For decades, research relied on traditional methods such as interviewer-administered 24-hour recalls, written food diaries, and food frequency questionnaires [1]. While these methods have contributed valuable data, they are constrained by significant limitations including high administrative costs, participant burden, memory dependency, and reactivity bias where participants may alter their intake because they are being observed [2] [1]. The digital transformation in research has catalyzed the development of automated, self-administered dietary assessment tools designed to overcome these challenges while maintaining, and in some cases improving, data accuracy.

This guide provides an objective comparison of three prominent web-based dietary assessment tools: ASA24, Intake24, and myfood24. Framed within a broader thesis on comparative validation, we examine their performance through experimental data, usability metrics, and implementation protocols to inform researchers, scientists, and drug development professionals in selecting the most appropriate tool for their specific study contexts.

The following table summarizes the origin, key characteristics, and primary applications of each automated dietary assessment tool.

Table 1: Overview of Automated Self-Administered Dietary Assessment Tools

Feature	ASA24	Intake24	myfood24
Developer	National Cancer Institute (NCI), USA [3]	Academic Consortium (Open Source) [4]	University of Leeds and Imperial College London [5]
Primary Model	24-hour recalls & food records [3]	24-hour dietary recall [4]	24-hour dietary recall & food diary [5] [6]
Database Foundation	USDA Automated Multiple-Pass Method (AMPM) [3]	Multiple-pass 24-hour recall [4]	Branded and generic food items [5]
Key Feature	Automated coding; Linked to US food/nutrient databases [3]	Open-source; Highly adaptable for different countries [4]	Unique branded foods database; Sustainability metrics (UK) [5]
Common Uses	Epidemiological, interventional, and clinical research [3]	Large-scale population dietary surveys [7] [4]	Academic research, clinical use, and teaching [5] [6]

Comparative Performance and Validation Data

Tool performance is measured through validity (how well the tool measures what it claims) and usability (user efficiency and satisfaction). The following data, drawn from controlled studies, provides a comparative lens.

Usability and User Preference

Usability is a critical driver of participant response rates and long-term adherence in cohort studies. The System Usability Scale (SUS) is a standard metric, where a score above 68 is considered satisfactory.

Table 2: Usability Comparison from Controlled Studies

Study Context	Tool A	SUS Score A	Tool B	SUS Score B	User Preference
French-Canadian Adults (n=68) [8]	R24W (Intake24 variant)	81 ± 2	ASA24-Canada-2018	58 ± 2	84% preferred R24W
Older Adults (60-74 years), Norway [6]	myfood24 (Norway)	56	Not Applicable	Not Applicable	(Only 27% found it satisfactory)
Healthy & Diabetic Adults (n=136) [1]	Keenoa (Image-assisted app)	77	ASA24-Canada	53	74.8% preferred Keenoa

Relative Validity and Nutrient Intake Correlation

Relative validity studies assess how well a new tool's intake estimates correlate with those from an established reference method.

Table 3: Relative Validity and Nutrient Agreement

Metric	ASA24 vs. Keenoa (Image-Assisted App) [1]	R24W (Intake24) vs. ASA24 [8]
Study Design	Randomized crossover, 4 days per tool	Randomized crossover, 2 non-consecutive days per tool
Energy Intake Agreement	No significant difference (P=0.38 men, P=0.61 women) [1]	No significant difference for energy and most nutrients
Correlation Coefficients	Macronutrients: r=0.48-0.73; Micronutrients: r=0.40-0.74 [1]	Not specified in excerpt
Statistical Agreement	Weighted Cohen κ: 0.30-0.52 (all P<.001) [1]	Not specified in excerpt
Underreporting Rate	8.8% for both tools [1]	Not specified in excerpt

Detailed Experimental Protocols

To ensure the reproducibility of validation studies and to aid researchers in designing their own trials, we detail the common methodologies employed in the cited research.

Randomized Crossover Design for Relative Validity

This robust design ensures each participant acts as their own control, reducing inter-individual variability.

Diagram 1: Crossover Study Workflow

Protocol for Localization and Cultural Adaptation

A key strength of open-source tools like Intake24 is their adaptability for diverse populations and food cultures, as demonstrated in the South Asia Biobank study [7].

Table 4: Key Reagents and Materials for Dietary Tool Localization

Research Reagent	Function in the Adaptation Process
Local Food Composition Table (FCT)	Serves as the definitive source for nutrient information of local foods, forming the backbone of the tool's database [7].
Comprehensive Food List	A curated list of commonly consumed foods, dishes, and beverages in the target region ensures the tool has high coverage and relevance [7].
Validated Portion Size Images	Standardized photographs aid users in estimating portion sizes more accurately than relying on memory or verbal descriptions alone [4].
Food Probes	Systematic prompts for forgotten foods (e.g., "Did you have any chutneys or pickles?") and details on preparation methods that improve recall completeness [7].

Diagram 2: Tool Localization Process

Discussion and Future Directions

The experimental data indicates that while automated tools like ASA24, Intake24, and myfood24 can generate comparable nutrient intake data [1] [8], usability is a significant differentiator. Tools developed or heavily adapted for specific populations (e.g., R24W in Québec, Intake24 in South Asia) tend to achieve higher usability scores and user preference, highlighting the importance of cultural and linguistic tailoring [7] [8].

Future directions in dietary assessment are likely to focus on:

• Integration of Artificial Intelligence (AI): AI-enhanced image recognition for food identification and portion size estimation, as seen in tools like Keenoa, can reduce user burden and potentially improve accuracy [1].
• Voice-Based Interfaces: Early pilot studies show that voice-based dietary recall tools are feasible and preferred by some populations, such as older adults, potentially overcoming challenges associated with screen-based interfaces [9].
• Multi-Method Approaches: Combining automated dietary data with other emerging technologies like VR for food behavior simulation or social media analytics can provide richer, real-time consumer insights [2].

For researchers, the choice of tool should be guided by the specific study population, the need for cultural adaptation, and the relative importance of high usability for long-term adherence versus the specific nutrient database and features required.

Accurate dietary assessment is fundamental to nutritional epidemiology, yet traditional methods like interviewer-led 24-hour recalls are resource-intensive and costly to implement at scale. The Automated Multiple-Pass Method (AMPM), developed by the United States Department of Agriculture (USDA), provides a structured, cognitive-based approach to enhance recall accuracy [10] [3]. This methodology has served as the foundational framework for several automated dietary assessment tools, including ASA24, INTAKE24, and myfood24, which have transformed data collection in large-scale studies [10] [11] [12].

These tools maintain the core AMPM structure while adapting it for self-administration, offering a viable combination of methodological rigor and operational feasibility. This guide objectively compares the performance, validation evidence, and implementation characteristics of these three platforms, providing researchers with evidence-based insights for tool selection.

The AMPM Core: Shared Methodological Foundation

The AMPM is engineered to mimic the cognitive processes of human memory through a structured series of passes or steps. This systematic approach minimizes omissions and enhances the detail of dietary recall.

Table 1: The Core AMPM Passes and Their Function

Pass Name	Primary Function	Implementation in Digital Tools
Quick List	Rapid, uninterrupted listing of all foods/beverages consumed.	Free-text entry; initial food search.
Forgotten Foods	Probing for commonly omitted items (e.g., sweets, beverages).	Automated prompts and checklists.
Time and Occasion	Chronological organization of intake events.	Sorting foods into eating occasions.
Detail Cycle	Collection of detailed descriptions and portion sizes.	Food matching, portion size selection via images.
Final Review	Opportunity to add or remove items.	Final summary screen for confirmation.

The transition from interviewer-led to self-administered recalls introduces unique challenges in user experience and data quality. ASA24 was directly modeled after the USDA's AMPM [3] [13]. INTAKE24 and myfood24 also adopt the principles of the AMPM, though myfood24 incorporates only some of these aspects [10]. This shared foundation allows for a consistent methodological approach while enabling adaptations for specific populations or research contexts.

Tool Comparison: Performance and Validation Data

Validation Against Traditional Methods

Each tool has been validated against established dietary assessment methods, demonstrating their capability to yield comparable data with the advantage of reduced administrative burden.

Table 2: Comparison of Validation Studies Against Interviewer-Led Recalls

Tool	Study Sample	Key Findings vs. Interviewer-Led Recall	Reference
INTAKE24	180 participants, aged 11-24	Mean energy intake underestimated by 1%; most macronutrients within 4%. Limits of agreement: -49% to +93%.	[10]
ASA24	1,077 adults, aged 50-74	Equivalence for most nutrients and food groups; higher completion rates than food records/FFQs.	[11] [13]
myfood24	Adults, aged 18-65	Nutrient estimates 10-20% lower than interviewer-based tool, with wide limits of agreement. ICCs: 0.4-0.5.	[12]

INTAKE24 demonstrated a mean energy intake underestimation of only 1% compared to an interviewer-led recall, with most macronutrient and micronutrient estimates falling within 4% of the standard method [10]. Similarly, the ASA24-2011 version showed equivalence to the interviewer-administered AMPM for most nutrients and food groups in a large sample of adults [13].

Validation Against Objective Biomarkers

Comparison with recovery biomarkers provides a more objective measure of validity, as biomarkers are not subject to the same self-reporting biases.

Table 3: Comparison of Validation Studies Against Biomarkers

Tool	Biomarker Comparison	Key Findings	Reference
myfood24	Protein (urinary nitrogen), Potassium, Sodium, Energy (accelerometry)	Attenuation factors: ~0.2-0.3. Partial correlation coefficients: ~0.3-0.4. Performance similar to interviewer-based tool.	[12]
ASA24	Energy (DLW), Protein, Sodium, Potassium (24-h urine)	Energy intake lower than expenditure. Reported intakes for protein, potassium, sodium closer to biomarkers for women than men.	[11]
INTAKE24	Energy (Fitbit Charge 2)	Energy intake underestimated by 33% on average vs. objective TEE.	[14]

A biomarker validation study for myfood24 found attenuation factors of approximately 0.2-0.3 and partial correlation coefficients of about 0.3-0.4 for protein, potassium, and sodium, which was comparable to an interviewer-based tool [12]. In the IDATA study, ASA24-derived energy intakes were consistently lower than energy expenditure measured by doubly labeled water, a finding consistent with the known limitation of self-reported dietary data [11].

Technical Specifications and Feasibility

Implementation feasibility, including completion time, usability, and cost, is a critical consideration for study design.

Table 4: Technical Specifications and Feasibility Metrics

Characteristic	ASA24	INTAKE24	myfood24
Primary Developer	National Cancer Institute (NCI), USA	Newcastle University, UK	University of Leeds, UK
Cost	Free for researchers	Free	Commercial/Licensed
Portion Size Method	Food photographs	>3000 food photographs	Photographic images, natural measures
Completion Time	41-58 minutes (declines with repeated use)	Not Specified	Not Specified
Usability in Older Adults	Appropriate for ≥5th grade reading level	Suitable for wide age range (11-88 years)	Low usability score (55.5 SUS) without guidance in 60-74 year-olds [6]
Food Database	USDA Food and Nutrient Database	UK NDNS Nutrient Databank	UK Composition of Foods, branded items

ASA24 demonstrates high feasibility, with completion times declining from approximately 55 to 41 minutes for men and 58 to 42 minutes for women over subsequent administrations [11]. Over 91% of men and 86% of women in the IDATA study completed at least three recalls, rates higher than for food records or FFQs [11]. However, usability can vary by demographic; the Norwegian version of myfood24 received a suboptimal System Usability Scale score of 55.5 in older adults (60-74 years) without guidance, suggesting this population may require additional support [6].

The Scientist's Toolkit: Key Research Reagents

• ASA24 Researcher Website: A web-based platform for researchers to manage study logistics, schedule recalls, and obtain cleaned nutrient and food group data files for analysis [3].
• INTAKE24 Food Photograph Atlas: A library of over 3000 food photographs developed based on portion sizes reported in UK National Diet and Nutrition Surveys, validated for portion size estimation [10].
• myfood24 Branded Food Database: An extensive database incorporating over 50,000 branded food items with nutrient content from packaging, complementing generic food composition data [12].
• Doubly Labeled Water (DLW): The gold standard method for measuring total energy expenditure in free-living individuals, used as a recovery biomarker to validate self-reported energy intake [11] [14].
• 24-Hour Urinary Collections: Used to measure urinary excretion of nitrogen (for protein), potassium, and sodium, serving as recovery biomarkers for validating nutrient intake reporting [11] [12].

ASA24, INTAKE24, and myfood24 demonstrate that the AMPM methodology can be successfully adapted for self-administered dietary assessment. The evidence indicates that these tools provide a balanced approach, offering data quality comparable to interviewer-led recalls with significantly reduced operational burden and cost [10] [11] [12].

Tool selection depends on specific research requirements. ASA24 offers the robustness of a federally developed tool and is freely available. INTAKE24 has been extensively validated in younger populations and is also free to use. myfood24 provides the advantage of a extensive branded food database and international adaptations. Researchers must consider population demographics, as older adults may require additional support for optimal use [6]. All self-report tools show systematic under-reporting compared to biomarkers, particularly for energy [14] [12], underscoring the importance of repeated measurements and methodological awareness in study design and interpretation.

This guide provides an objective comparison of three prominent, web-based 24-hour dietary recall tools: ASA24, Intake24, and myfood24. Aimed at researchers and professionals, it contrasts their origins, development, key features, and performance based on published validation studies to inform tool selection for research and clinical practice.

Feature	ASA24	Intake24	myfood24
Lead Institution	National Cancer Institute (NCI), USA [3]	Newcastle University, UK [15]	University of Leeds & Imperial College London, UK [5]
Primary Funding	Multiple NIH Institutes [3]	Information Not Available	Information Not Available
Core Methodology	Adapted USDA's Automated Multiple-Pass Method (AMPM) [3] [16]	Information Not Available	Web-based 24-hour recall/food record [5] [6]
Database Foundation	USDA Food and Nutrient Databases [3]	UK Composition of Foods [17]	UK Composition of Food Integrated Dataset (CoFID) & branded foods [5]
Key Feature	Highly standardized, interviewer-free AMPM [16]	Open-source platform	Includes extensive branded foods database [5]
Cost	Free for researchers [3] [18]	Information Not Available	Commercial (with demo available) [5]
International Adaptation	Canadian & Australian versions available [3]	Multiple international adaptations [17]	Multiple regional versions (e.g., Germany, France, Norway) [6]

Comparative Experimental Data on Tool Performance

The following table summarizes key quantitative findings from validation and usability studies for ASA24 and myfood24. Data for Intake24 was not available in the search results.

Tool & Study Focus	Methodology Summary	Key Comparative Results
ASA24: Supplement Reporting [16]	1076 participants randomly assigned to complete ASA24 or interviewer-administered AMPM recalls. Reported supplement use was compared.	Proportions reporting supplement use were equivalent between ASA24 (46%) and AMPM (43%). Concluded little difference due to mode of administration [16].
ASA24: Usability in Low-Income Populations [19]	Study assessed if assistance impacted recall accuracy among low-income women.	The provision of assistance did not substantially impact the accuracy of 24-hour dietary recalls completed using ASA24 [19].
myfood24: Usability in Older Adults [6]	60 older adults (60-74 yrs) completed a 24-hour recall using the Norwegian myfood24 with no guidance. Usability was measured via System Usability Scale (SUS).	Mean SUS score was 55.5, below the satisfactory threshold (68). 27% of participants had satisfactory scores. Higher usability was linked to younger age within the cohort [6].
myfood24 vs. ASA24: Voice-Based Recall [9]	20 older adults (>65 yrs) tested a voice-based tool (DataBoard) and ASA24. Users rated feasibility, acceptability, and ease on a 1-10 scale.	Voice-based recall (DataBoard) was rated easier to use (6.7/10) than ASA24. Participants also felt they could use the voice-based tool more frequently for food reporting (7.2/10) [9].

Experimental Protocols in Dietary Assessment Validation

Researchers employ specific methodological protocols and reagents to validate dietary assessment tools.

• Validation Against Biomarkers: The most robust validation involves comparing nutrient intakes from the tool against objective biomarkers in blood or urine. For example, myfood24 has been validated against a suite of independent nutrient biomarkers [5], and ASA24 total water intake has been validated using doubly labeled water [19].
• Comparison to Traditional Methods: A common protocol involves having the same participants complete both the web-based tool and an established, often interviewer-administered, 24-hour recall (like the USDA's AMPM). Intakes for energy, nutrients, and food groups are then statistically compared for equivalence [16] [17].
• Usability Testing: Researchers assess the user-friendliness of a tool using standardized questionnaires like the System Usability Scale (SUS). Participants rate their agreement with statements about the tool's ease of use, complexity, and need for support. Scores are calculated to determine overall usability [6].

Research Reagent Solutions

Item	Function in Dietary Assessment Research
Doubly Labeled Water	A gold-standard biomarker used to validate total energy expenditure and, by extension, energy intake reported by dietary tools [19].
Nutritional Biomarkers	Objective measures of nutrient intake or status in biological samples (e.g., blood, urine) used to validate self-reported dietary data [5].
System Usability Scale (SUS)	A standardized 10-item questionnaire that provides a quick and reliable measure of a tool's perceived usability and user-friendliness [6].
Automated Multiple-Pass Method (AMPM)	A structured, five-step interview protocol developed by the USDA to enhance recall and reduce reporting error, serving as a benchmark for new tools [3] [16].

Dietary Recall Assessment Workflow

The following diagram illustrates the general workflow for a self-administered 24-hour dietary recall, which is shared by tools like ASA24, Intake24, and myfood24.

Key Takeaways for Tool Selection

• For maximum standardization and public health alignment: ASA24, developed and maintained by the NCI with NIH funding, offers a rigorously evaluated tool that aligns closely with major US national surveys [3] [16].
• For research requiring branded food data: myfood24 has a unique, expert-curated database of branded food items, which may enhance accuracy in populations that rely heavily on pre-packaged foods [5].
• For specific demographic considerations: While all tools can be used with adults, usability varies by age group. Older adult populations may require support or alternative interfaces, as evidenced by lower usability scores for web-based tools in this demographic [6] [9].

Accurate dietary assessment is fundamental for elucidating diet-disease relationships, yet traditional methods face significant challenges regarding cost, participant burden, and measurement error [20]. While 24-hour dietary recalls provide detailed intake data, traditional interviewer-administered recalls are resource-intensive, requiring trained personnel and substantial time commitments [21]. This limitation has driven the development of automated self-administered systems—ASA24, Intake24, and myfood24—which promise enhanced feasibility for large-scale studies and repeated measures collection [3] [22] [23].

These tools address critical research gaps by potentially reducing administrative costs while facilitating the collection of multiple dietary assessments essential for estimating usual intake. However, their comparative performance across different populations and contexts requires thorough examination. This guide objectively compares these three prominent systems using validation evidence, focusing on their practicality for research applications requiring repeated measures.

The table below summarizes the origin, core features, and validation status of the three dietary assessment tools.

Table 1: Key Characteristics of Automated 24-Hour Dietary Recall Tools

Feature	ASA24	Intake24	myfood24
Developer	National Cancer Institute (USA)	Newcastle University (UK)	University of Leeds (UK)
Cost Model	Free	Open-source	Commercial/Licensed
Primary Method	Automated Multiple-Pass Method (AMPM)	Multiple-pass 24-hour recall	24-hour recall or food record
Food Database	USDA-based, updated biennially	Customizable for different countries	UK Composition of Foods plus branded items
Portion Size Estimation	Photographs, household measures	Validated portion size images	Photographic aids, portion images
Adaptation Framework	Available for specific countries (CA, AU)	Designed for localization	Requires validation for new populations

ASA24, developed by the National Cancer Institute, is a freely available web-based tool that has been used to collect over 1.14 million recall days as of June 2025 [3]. Intake24, an open-source system originally developed in the UK, has been adapted for multiple countries including New Zealand, Denmark, and Portugal [23] [24]. Myfood24 is a commercial dietary assessment tool supporting both 24-hour recalls and food records, with adaptations beyond its original UK version [22] [20].

Comparative Validation Evidence

Criterion Validity Against Objective Measures

The most rigorous validation studies compare self-reported intake against objective biological markers or known intake in controlled settings.

Table 2: Criterion Validity Against Objective Reference Measures

Tool	Reference Standard	Energy Reporting Accuracy	Nutrient Correlation Findings
ASA24	Feeding study (weighed true intake)	Reported 80% of items consumed (vs. 83% for interviewer-AMPM) [25]	Comparable accuracy to interviewer-administered recalls for nutrients and food groups [25]
Intake24	Doubly Labeled Water (DLW)	Under-reported EI by 25% (single recall) to 25% (three recalls) [23]	Moderate correlation with TEE (ρ=0.31-0.47) [23]
myfood24	Urinary biomarkers, accelerometry	Attenuation factors of 0.2-0.3 for protein, potassium [20]	Partial correlation ~0.3-0.4 with biomarkers; comparable to interviewer recalls [20]

ASA24 demonstrated strong performance in a feeding study where true intake was known, with minimal difference in reporting accuracy compared to interviewer-administered recalls [25]. Intake24 shows comparable under-reporting to traditional methods when validated against doubly labeled water, with moderate correlation to total energy expenditure [23]. Myfood24 demonstrates similar attenuation to interviewer-led methods when compared to urinary biomarkers, with partial correlation coefficients of approximately 0.3-0.4 for protein and potassium [20].

Reliability and Repeatability

For research investigating diet-disease relationships, the ability to reliably rank individuals by intake is often more critical than absolute accuracy.

Table 3: Reliability and Repeatability Metrics

Tool	Study Design	Reliability Findings	Implications for Research
Intake24	Test-retest (4+ recalls)	ICC=0.35 (single recall) to 0.52 (pairs of recalls) for energy [23]	Multiple recalls improve ranking ability; suitable for epidemiological studies
myfood24	Repeated 7-day WFR, 4 weeks apart	Strong reproducibility (ρ≥0.50) for most nutrients; highest for folate (ρ=0.84) [22]	Effective for tracking dietary changes in intervention studies
ASA24	Data quality assessment	Multiple imputation poorly estimated missing nutrients (mean ρ≈0.24) [21]	Highlights importance of data completeness over statistical correction

Myfood24 demonstrates strong reproducibility for most nutrients when assessments are repeated after four weeks, particularly for folate and vegetable intake [22]. Intake24 shows moderate reliability with intraclass correlation coefficients improving when using multiple recalls, supporting its use for ranking participants by nutrient intake [23].

Implementation Practicality: Cost, Burden and Adaptability

Administrative and Participant Burden

The reduction in administrative requirements represents a significant advantage of automated systems. ASA24 eliminates interviewer costs and enables simultaneous data collection from multiple participants [21]. Intake24 offers rapid completion with median recall time of 13 minutes in large-scale implementation [7], significantly reducing participant time commitment compared to traditional recalls.

User experience studies reveal important considerations for implementation. In usability testing of Intake24-NZ, 84% of participants reported the tool was easy to use, though challenges emerged with search functions and portion size estimation [24]. These findings highlight that even well-designed systems require optimization for specific populations.

Adaptation for Diverse Populations

The flexibility of these tools to accommodate different cultural contexts and food systems varies substantially.

Diagram 1: Tool Adaptation Requirements

Intake24 has demonstrated substantial adaptability, with the South Asian version incorporating 2,283 food items reflecting local dietary patterns [7]. The New Zealand adaptation required modifications to food lists, terminology, portion images, and prompts to suit the local context [24]. ASA24 maintains separate versions for different countries but follows a standardized framework [3]. Myfood24 requires revalidation when adapted for new populations, as functionality may be affected by changes to underlying food composition databases [22].

Research Applications and Recommendations

The Researcher's Toolkit: Essential Methodological Components

Table 4: Key Research Reagents and Methodological Components

Component	Function in Dietary Assessment	Examples from Validation Studies
Recovery Biomarkers	Objective validation of nutrient intake	Doubly labeled water (energy), urinary nitrogen (protein), urinary potassium [23] [20]
Food Composition Databases	Convert reported foods to nutrient intakes	UK Composition of Foods, USDA FNDDS, country-specific databases [3] [20]
Portion Size Estimation Aids	Improve quantification of amounts consumed	Validated portion images, household measures, food models [25] [24]
Usability Testing Protocols	Identify practical implementation barriers	Think-aloud methods, screen recording, participant surveys [24]

Selection Guidelines for Research Scenarios

Diagram 2: Tool Selection by Research Scenario

For budget-constrained studies, ASA24 offers a no-cost solution with validated methodology comparable to interviewer-led recalls [3] [25]. For cross-cultural research or studies requiring localization, Intake24 provides demonstrated adaptability and has been successfully implemented in diverse populations [7] [24]. For studies prioritizing biomarker-validated precision, myfood24 has extensive validation against urinary biomarkers and blood measures [22] [20].

All three tools successfully address the fundamental research gap of enabling feasible repeated measures collection, which is essential for estimating usual intake and reducing measurement error in diet-disease association studies [23] [20].

Automated dietary assessment tools represent a significant advancement in nutritional epidemiology, directly addressing critical research gaps related to cost, participant burden, and the need for repeated measures. ASA24, Intake24, and myfood24 all demonstrate reasonable validity compared to both objective biomarkers and traditional interviewer-administered recalls, while offering substantial practical advantages.

The choice between these systems should be guided by specific research constraints and objectives: ASA24 for budget-sensitive projects, Intake24 for culturally diverse populations, and myfood24 for studies prioritizing extensive biomarker validation. All three tools enable the collection of repeated measures essential for characterizing usual intake, thereby strengthening nutritional epidemiology and public health research.

Tool-Specific Methodologies and Deployment in Research Populations

Automated dietary assessment tools have revolutionized nutritional epidemiology by enabling scalable, cost-effective data collection. Among the leading platforms, the Automated Self-Administered 24-Hour (ASA24) Dietary Assessment Tool has emerged as a widely adopted solution developed by the National Cancer Institute (NCI) [3]. This guide provides a structured comparison of ASA24 against two other prominent tools—Intake24 and myfood24—focusing on their technical architectures, validation evidence, and implementation capabilities to inform researcher selection.

ASA24, Intake24, and myfood24 share the common purpose of automating 24-hour dietary recalls but differ significantly in their underlying structure, availability, and target populations.

ASA24 is a free, web-based tool that enables automatically coded, self-administered 24-hour diet recalls and food records [3]. Its core structure adapts the United States Department of Agriculture's (USDA) Automated Multiple-Pass Method (AMPM), a validated five-step approach designed to enhance memory and reduce systematic recall error [3]. The system maintains a robust technical infrastructure capable of supporting 800 concurrent users without compromising performance [26].

Table 1: Fundamental Structural Characteristics

Feature	ASA24	Intake24	myfood24
Availability	Free web-based platform [3]	Open-source system [4]	Commercial product with research applications [5]
Core Methodology	Adapted USDA AMPM [3]	Multiple-pass 24-hour recall [4]	Flexible 24-hour recall or food diary [27]
Primary Developer	National Cancer Institute (NCI) [3]	Academic consortium [7]	University of Leeds/Imperial College London [5]
Food Database	USDA Food and Nutrient Database for Dietary Studies [3]	Customizable local databases [7]	Extensive branded and generic UK foods (~45,000 items) [27]
Concurrent User Capacity	800 simultaneous respondents [26]	Information not specified in sources	Information not specified in sources

The multi-lingual capabilities of ASA24 include support for English and Spanish in the US version, with the Canadian version supporting English and French [3]. This international adaptability extends to country-specific versions for Australia and Canada, though these are updated less frequently than the US version [3].

Multi-Pass Methodology: Experimental Protocols and Validation

The multi-pass methodology forms the scientific foundation of automated 24-hour recalls, employing sequential cognitive steps to enhance recall accuracy.

Figure 1: The Multi-Pass Recall Methodology. This sequential approach systematically guides respondents through recall completion to enhance accuracy and completeness [3].

ASA24's implementation of the AMPM has been extensively validated. In the Interactive Diet and Activity Tracking in AARP (IDATA) Study, ASA24 demonstrated comparable performance to interviewer-administered recalls when evaluated against recovery biomarkers [19]. The system's portion size estimation incorporates digital images, which research shows "may lead to less misestimation of portion size compared to traditional interviewer-administered recalls" [19].

Completion time metrics further reflect the efficiency of these automated systems. ASA24 recalls average 24 minutes to complete, with most respondents finishing within 17-34 minutes [26]. Intake24 demonstrates even faster median completion times of 13 minutes in large-scale implementations like the South Asia Biobank study [7].

Table 2: Validation Metrics Against Biomarkers

Validation Measure	ASA24 Performance	myfood24 Performance	Intake24 Status
Energy Intake Validation	Attenuation factors ~0.2-0.3 against DLW [19]	Attenuation factors ~0.2-0.3 against biomarkers [12]	Protocol published for biomarker comparison [19]
Protein Intake Correlation	Moderate correlation with urinary nitrogen [19]	Partial correlation ~0.3-0.4 with urinary biomarkers [12]	Information not specified in sources
Usability Scores (SUS)	53/100 in comparative study [1]	55.5-77/100 across studies [1] [6]	Information not specified in sources
Underreporting Rate	8.8% in controlled trial [1]	Information not specified in sources	Information not specified in sources

Implementation Considerations for Research

Technical Requirements and Accessibility

ASA24 operates as a web-based application requiring HTML5-compliant browsers on computers or mobile devices with internet connectivity [26]. The system explicitly cannot be used offline, presenting limitations for populations with unreliable internet access [26]. Intake24 shares similar web-based architecture but emphasizes localization capabilities for diverse cultural contexts [4] [7].

Population appropriateness varies across tools. ASA24 is most suitable for those with at least a fifth-grade reading level and comfort with digital devices [3]. The system has been successfully implemented in low-income populations, though literacy challenges may necessitate interviewer assistance in some cases [3] [26]. Myfood24 has demonstrated variable usability across age groups, with older adults (60-74 years) reporting lower System Usability Scale scores (55.5/100) without guidance [6].

Data Output and Analytical Capabilities

ASA24 provides researchers with comprehensive nutrient and food group data files through its researcher portal [3]. The system generates Respondent Nutrition Reports that compare intake against dietary guidelines, which can be shared with participants [3] [26]. Myfood24 distinguishes itself with sustainability metrics (greenhouse gas emissions, water and land use) for UK foods and extensive branded food coverage [5].

Table 3: Key Research Reagent Solutions

Resource	Function	Implementation Example
ASA24 Demonstration Site	Free practice environment without data saving [3]	Pilot testing with study populations prior to actual data collection
Researcher Test Accounts	Controlled evaluation of study setup and output files [26]	Verification of analysis pipelines before participant enrollment
Sleep Module	Assessment of sleep-diet relationships within 24-hour cycle [26]	Integrated data collection on timing, quantity, and quality of sleep
Biomarker Validation Protocols	Objective assessment of tool accuracy against physiological measures [19] [12]	Doubly labeled water for energy intake, urinary nitrogen for protein
Localization Frameworks	Cultural and linguistic adaptation of food databases and interfaces [7]	South Asia Biobank implementation with 2,283 food items [7]

ASA24's structure as a freely available, multi-pass, multi-lingual dietary assessment tool provides researchers with a rigorously validated platform suitable for large-scale epidemiological studies. Its strongest advantages include zero cost implementation, extensive validation history, and robust technical infrastructure supporting hundreds of concurrent users. However, researchers working with specific populations may need to consider alternatives—Intake24 for extensive localization requirements in diverse cultural contexts, or myfood24 for studies prioritizing branded food composition or sustainability metrics. The selection decision ultimately hinges on study-specific requirements regarding population characteristics, cultural context, analytical needs, and resource constraints, with all three tools representing significant advancements over traditional dietary assessment methods.

Accurate dietary assessment is fundamental for public health monitoring, nutritional epidemiology, and understanding diet-disease relationships, particularly in younger populations whose dietary habits are still forming. The development of digital dietary recall tools has revolutionized data collection by offering scalable, cost-effective alternatives to traditional interviewer-led methods. Among these tools, Intake24 stands out for its specific design focus on young populations, employing rigorous usability testing and validation protocols to ensure data quality and user engagement. This guide objectively compares Intake24's performance with other prominent automated dietary assessment systems—ASA24 (Automated Self-Administered Dietary Assessment Tool) and myfood24—within the context of a broader thesis on comparative validation. For researchers, scientists, and drug development professionals, understanding the technical specifications, validation evidence, and usability metrics of these tools is crucial for selecting appropriate methodologies for nutritional surveillance and intervention studies.

Intake24 is an open-source, web-based 24-hour dietary recall system originally developed at Newcastle University for national nutrition surveys. Its core design philosophy centers on creating a low-burden, engaging experience specifically for younger users (aged 11-24 years), while maintaining data quality comparable to traditional methods [10] [28]. The system implements an automated multiple-pass 24-hour recall method, featuring a predefined food database, portion size estimation aids using validated food photographs, and automated linkage to food composition databases for nutrient analysis [24].

The tool was developed through an iterative design process involving four cycles of user testing, evaluation, and system refinement [10]. This user-centered approach focused heavily on usability and user experience, making the tool particularly suitable for younger populations who may be more technologically adept but have different cognitive abilities and engagement needs compared to adults.

ASA24, developed by the US National Cancer Institute, is another widely used automated self-administered 24-hour dietary recall system based on the Automated Multiple-Pass Method used in the National Health and Nutrition Examination Survey [29]. In contrast, myfood24 is a UK-developed online dietary assessment system that incorporates both 24-hour recall and food diary functionality, with a database encompassing generic foods, branded items, and food outlet products [12].

Experimental Protocols and Validation Methodologies

Relative Validation Against Interviewer-Led Recalls

A key validation study for Intake24 involved a comparison with interviewer-led multiple-pass 24-hour recalls in 180 participants aged 11-24 years [10] [28]. The study employed a cross-over design where each participant completed both an Intake24 recall and an interviewer-led recall on the same day across four occasions over one month. A weighted randomization was implemented, with 75% of participants completing Intake24 first and 25% completing the interviewer-led recall first, serving as a methodological check for order effects [10]. Nutritional intake data for energy, macronutrients, and micronutrients derived from both methods were statistically compared using mean differences, correlation analyses, and limits of agreement.

Criterion Validation Against Doubly Labeled Water

The gold standard for validating energy intake assessment involves comparison with total energy expenditure measured using the doubly labeled water (DLW) technique. A study with 98 UK adults aged 40-65 years assessed the validity of energy intake reported using Intake24 against objectively measured energy expenditure [23] [30]. Participants completed Intake24 recalls during a 9-10 day DLW measurement period, enabling direct comparison of reported energy intake with measured energy expenditure using Bland-Altman analysis to assess accuracy (mean bias) and precision (limits of agreement), along with correlation analyses to quantify ranking ability [23].

Usability Testing Protocols

Comprehensive usability studies for Intake24 have employed mixed-methods approaches. The New Zealand adaptation (Intake24-NZ) was evaluated through a study where 37 participants aged ≥11 years completed a 24-hour dietary recall while providing real-time feedback through screen observation recordings and verbal commentary [24]. This was supplemented with a structured usability survey capturing user experiences, challenges, and satisfaction metrics. Similar field testing for the UK version collected feedback through online surveys following multiple recall completions, with sentiment analysis performed on free-text comments [31].

Table 1: Key Validation Studies for Intake24

Study Focus	Population	Sample Size	Design	Key Metrics
Relative Validation [10] [28]	Ages 11-24 years	180	4 non-consecutive days, crossover with interviewer-led recall	Mean difference in energy/nutrient intake, correlation coefficients, limits of agreement
Criterion Validation [23] [30]	Adults 40-65 years	98	Comparison with doubly labeled water measurement	Energy intake under-reporting rate, correlation with TEE, Bland-Altman limits of agreement
Usability Testing [24]	Ages ≥11 years (including Māori)	37	Mixed methods: observation + survey	Completion rates, ease-of-use ratings, challenge identification, user satisfaction
Field Testing [31]	Ages 11-65+ years	230 (completed ≥1 recall)	Real-world implementation with feedback survey	Completion rates, user-friendliness ratings, reasons for non-completion

Comparative Performance Data

Accuracy of Nutrient Intake Estimation

When compared against interviewer-led recalls in young populations, Intake24 demonstrated comparable accuracy for most nutrients. The system underestimated energy intake by just 1% on average, with limits of agreement ranging from -49% to +93% [10] [28]. Mean intakes of all macronutrients and most micronutrients were within 4% of the interviewer-led recall, with the exception of non-milk extrinsic sugars which showed slightly greater variation [28].

In criterion validation against doubly labeled water, Intake24 demonstrated a higher under-reporting rate of 25% for a single recall, which improved slightly to 22% when the average of two recalls was considered [23]. The correlation between reported energy intake and measured energy expenditure was 0.31 for a single recall, improving to 0.47 when using the average of two recalls [23]. This under-reporting magnitude is consistent with that observed in interviewer-led 24-hour recalls and estimated weight food diaries, suggesting Intake24 performs similarly to traditional methods despite its self-administered format [23].

Table 2: Comparison of Automated 24-Hour Dietary Recall Tools

Feature	Intake24	ASA24	myfood24
Origin	Newcastle University, UK	National Cancer Institute, USA	University of Leeds, UK
Target Population	Originally 11-24 years, later expanded	Adults and children (separate versions)	Adults and adolescents
Food Database	~2,500 foods (UK version) [23]	Extensive USDA-based database	>50,000 branded + 3,000 generic foods [12]
Portion Size Estimation	Validated food photographs (2,400+ images) [31]	Food model booklet + images	Photographic images (6,000+ foods) [12]
Validation Evidence	Interviewer-led recalls, DLW, weighed records [10] [23] [28]	Feeding studies, interviewer-led recalls [29]	Biomarkers, interviewer-led recalls [12]
Usability Testing	4 cycles of iterative testing [10], mixed methods evaluation [24]	Cognitive and usability testing [29]	User evaluation studies
Reliability (ICC for Energy)	0.35 (single recall), 0.52 (two recalls) [23]	Data not specified in sources	Data not specified in sources
Adaptations	New Zealand, Portugal, Denmark, UAE, South Asia [24] [7]	Canada, Australia	United Kingdom

Usability and User Engagement Metrics

Usability testing of Intake24 reveals generally positive user reception, particularly among younger populations. In field testing, 44% of users agreed they would like to use Intake24 often, compared to 15% who disagreed, and over 75% felt the system accurately captured their dietary intakes [31]. The tool was described as "user-friendly, enjoyable to use, and easy to follow and understand" [31].

However, usability studies have identified consistent challenges with the food search functionality. Data from recorded observations and usability surveys revealed difficulties with correct search term usage, search results obtained (type and order of foods displayed), portion size estimation, and responding to associated food prompts [24]. Despite these challenges, most participants (84%) still reported that Intake24 was easy to use and navigate [24].

Completion rates in field studies provide insight into practical usability. In one study, 60% of participants who agreed to take part completed at least one recall, with 34% completing all four requested recalls [31]. The median recall completion time was approximately 13 minutes in large-scale implementations [7], indicating reasonable respondent burden.

Research Reagent Solutions: Essential Methodological Components

Table 3: Essential Research Components for Dietary Assessment Validation

Component	Function in Validation	Implementation Examples
Doubly Labeled Water (DLW)	Criterion method for validating energy intake reporting by measuring total energy expenditure [23]	Participants ingest isotopic water (²H₂¹⁸O); urine samples collected over 9-14 days analyze isotope elimination rates [23]
Interviewer-Led Multiple-Pass Recalls	Reference method for relative validation of food and nutrient reporting [10]	Structured interviews using standardized protocols with portion size aids; considered enhanced comparison method [10]
Food Photography Atlas	Standardizes portion size estimation across different assessment methods [10]	Series of images showing progressively increasing portion sizes; validated against weighed food portions [24]
Biomarker Analysis	Provides objective validation of specific nutrient intakes [12]	Urinary nitrogen (protein), potassium, sodium; plasma vitamin levels; not reliant on self-report [12]
Usability Survey Instruments	Quantifies user experience, acceptability, and practical challenges [24] [31]	Structured questionnaires covering ease of use, navigation, satisfaction; often with open-ended feedback options [24]
Controlled Feeding Studies	Provides known intake reference for validation in controlled conditions [29]	Participants consume provided meals with documented nutrient composition; reported intake compared to known values

Technical Workflow and System Architecture

The following diagram illustrates the comprehensive validation workflow for Intake24, incorporating multiple methodological approaches:

Figure 1: Comprehensive validation workflow for Intake24, integrating multiple methodological approaches to system evaluation and refinement.

Intake24 represents a validated, cost-effective alternative to traditional dietary assessment methods, with particular strengths in usability for young populations. The evidence from multiple validation studies demonstrates that while the tool shows comparable under-reporting patterns to interviewer-led methods, it offers significant advantages in scalability, reduced administrative burden, and user engagement—particularly valuable for large-scale surveillance and studies with younger participants.

For researchers selecting dietary assessment tools, Intake24 provides robust validation evidence through multiple approaches, including comparison with interviewer-led recalls, doubly labeled water, and comprehensive usability testing. The tool's iterative development process and focus on user experience have resulted in a system specifically optimized for self-administration by younger populations. Meanwhile, ongoing comparative studies like the one protocolized by Kerr et al. [29] will provide further insights into the relative accuracy and cost-effectiveness of Intake24 compared to other technology-assisted methods like ASA24 and image-assisted records.

When implementing Intake24 in research settings, particularly with young populations, researchers should anticipate challenges related to food search functionality and portion size estimation, and consider implementing supplementary instructions or training to mitigate these issues. The collection of multiple recalls (at least two) significantly improves reliability estimates, and the system's adaptability across different countries and food cultures enhances its utility for multinational studies. As digital dietary assessment continues to evolve, Intake24's development pathway offers a model for balancing scientific rigor with practical usability in research tools targeting diverse populations.

Accurate dietary assessment is a cornerstone of nutritional epidemiology, public health research, and clinical studies. The validity of any dietary assessment tool is fundamentally dependent on the comprehensiveness, accuracy, and relevance of its underlying food composition database (FCDB). These databases provide the essential nutrient values for foods and beverages consumed, transforming qualitative food intake data into quantitative nutrient estimates. In an era of rapidly evolving global food systems, characterized by an explosion of branded products and diverse dietary patterns, traditional FCDBs limited to generic foods face significant limitations [32]. Researchers require tools that can accurately capture the nutritional intake from both generic whole foods and the branded products that dominate modern food environments.

This analysis examines the database framework of myfood24, objectively comparing its integrated approach to branded and generic food data against two other widely used online 24-hour dietary recall tools: ASA24 and Intake24. By evaluating their respective database structures, mapping methodologies, and validation evidence, this guide aims to equip researchers with the critical information needed to select the most appropriate dietary assessment tool for their specific study contexts and populations.

Comparative Analysis of Database Architectures

Database Scope and Composition

The three dietary assessment tools demonstrate markedly different approaches to database construction and scope, reflecting their distinct developmental origins and target applications.

Table 1: Core Database Characteristics Comparison

Feature	myfood24	ASA24	Intake24
Total Food Items	~488,987 total items [33]	Not explicitly stated	Not explicitly stated
USDA Branded Items	255,399 [33]	Utilizes USDA databases [3]	Not specified
USDA Generic Items	14,770 [33]	Utilizes USDA FNDDS [34]	Not specified
UK Generic Items	3,334 (McCance & Widdowson) [33]	Not applicable	Linked to UK NDNS Nutrient Databank [35]
Primary Coverage	International, with strong UK/US focus [33]	United States [3]	United Kingdom [35]
International Versions	Multiple (e.g., Germany, France, Australia, Middle East) [33]	Canadian and Australian versions available [3]	Adapted for use in New Zealand surveys [36]
Specialist Databases	Children's databases, medical foods, supplements [37] [33]	ASA24-Kids for children [3]	Not specified

myfood24 employs a dual-component architecture, systematically integrating both comprehensive branded product data and extensive generic food tables. Its most distinctive feature is the massive scale of its branded food database, which contains over 255,000 USDA-branded items and an additional 123,000 UK-branded items [33]. This is complemented by a wide array of international datasets from Germany, France, Australia, the Middle East, and other regions, making it particularly suitable for multi-national studies or research involving diverse ethnic populations [33].

ASA24 leverages the authoritative USDA Food and Nutrient Database for Dietary Studies (FNDDS) and other USDA resources as its core data foundation [34]. While the exact number of food items is not explicitly detailed in the search results, the system is known for its rigorous connection to these officially recognized U.S. databases. ASA24 also features specialized versions for different populations, including a Canadian adaptation and ASA24-Kids for use with children [3].

Intake24 links its food intake data directly to the UK National Diet and Nutrition Survey (NDNS) Nutrient Databank [35]. This connection ensures alignment with national dietary survey methodologies in the UK. The tool has demonstrated adaptability, having been successfully modified for use in New Zealand's national nutrition surveys [36].

Technical Methodologies: Database Mapping and Quality Assurance

The processes by which these tools build and maintain their databases significantly impact data quality and nutrient coverage.

MyFood24's Unique Mapping and Overstamping Process MyFood24 employs a sophisticated semi-automated mapping methodology to address a critical challenge in branded food composition data: incomplete nutrient information on product packaging. The process involves matching branded food products to generic food codes from established composition tables (e.g., McCance and Widdowson's in the UK). This allows for the imputation of missing micronutrient values that are not mandatory on nutrition labels, such as vitamin B6, iron, and zinc [38] [33]. Furthermore, myfood24 utilizes an "overstamping" process to actively infill nutrient gaps or "blanks" found in source data, thereby providing a more complete nutrient profile. For example, their May 2025 update overstamped 26 nutrient variables in USDA databases, including minerals like calcium and iron, vitamins like vitamin C and D, and fatty acids [37]. All items undergo rigorous quality checks, including identifying foods with nutrient values outside expected ranges for their category [38].

ASA24's Foundation on National Dietary Surveillance ASA24's methodology is built around its connection to the gold-standard Automated Multiple-Pass Method (AMPM) developed by the USDA [3]. Its databases are centrally managed and updated, with new U.S. versions typically released every two years to incorporate biennial food and nutrient database updates [3]. This ensures alignment with the evolving U.S. food supply and the latest national nutrition monitoring efforts.

Intake24's Portion Size Estimation System Intake24's distinctive methodological strength lies in its visual portion size estimation using a library of over 3,000 food photographs derived from portion sizes reported in UK National Diet and Nutrition Surveys. These images have been validated in both feeding studies and against weighed food intakes [35]. This approach is particularly valuable for enhancing the accuracy of self-reported portion sizes, a known source of measurement error in dietary assessment.

Figure 1: MyFood24's unique branded food mapping and enhancement workflow, demonstrating how limited back-of-pack data is transformed into a complete nutrient profile.

Experimental Validation and Comparative Performance

Validation Study Designs and Key Findings

Robust validation against established methods is crucial for evaluating the relative performance of dietary assessment tools.

Table 2: Validation Evidence from Comparative Studies

Tool	Validation Study Design	Key Findings	Reference
myfood24	Compared with interviewer-led 24-h recall in adolescents (11-18 years)	"Validated against face-to-face interviewer-led recalls" [35]	[35]
ASA24	(1) Feeding study with 81 adults; (2) Validation against school lunch observations (ASA24-Kids)	"Validated in a feeding study with 81 adults" [35]	[35] [3]
Intake24	Comparison with interviewer-led 24-h recall in 180 participants aged 11-24 (4 occasions each)	Energy intake underestimated by 1% vs interviewer-led recall; most macronutrients within 4%	[35]

The 2016 study by Foster et al. provides a direct comparison between Intake24 and interviewer-led recalls. In this study, 180 participants aged 11-24 completed both Intake24 and interviewer-led recalls on four separate days. The results demonstrated that Intake24 produced mean intake estimates for energy and most nutrients that were within 4% of the interviewer-led method, though with wide limits of agreement for energy (-49% to +93%), indicating substantial individual variation in reporting accuracy [35].

A 2024 review evaluating tools for national nutrition surveys shortlisted ASA24, Intake24, and myfood24 as the three tools with the highest suitability scores based on criteria including validation evidence, previous use in national surveys, and adaptability [36]. This suggests that all three tools are considered methodologically sound for large-scale research applications.

The Researcher's Toolkit: Essential Components for Dietary Assessment Validation

Table 3: Key Research Reagents and Methodologies for Dietary Tool Validation

Component	Function in Validation	Example Application
Interviewer-Led 24-h Recall	Serves as the comparison method in relative validation studies.	Used as benchmark in Intake24 validation [35].
Doubly Labeled Water (DLW)	Provides objective measure of total energy expenditure to assess energy intake underreporting.	Cited as gold standard for energy validation [35].
Portion Size Image Atlas	Standardizes self-reported portion size estimates to reduce measurement error.	Intake24 uses >3,000 validated food photographs [35].
Food and Nutrient Database	Converts reported food consumption into estimated nutrient intakes.	myfood24's mapped database provides missing micronutrient values [38].
Biomarker Measurements	Offers objective measures of nutrient status for validating nutrient intake estimates.	myfood24 has been "validated against a suite of independent nutrient biomarkers" [5].

Figure 2: Dietary assessment validation framework showing the relationship between assessment tools, food composition databases, and objective validation methodologies.

Discussion: Strategic Implications for Research Applications

Comparative Advantages and Application-Specific Recommendations

Each tool offers distinct advantages that make it particularly suitable for specific research contexts:

• myfood24 demonstrates clear superiority for studies where branded food consumption is prevalent or where research requires detailed micronutrient analysis beyond basic macronutrients. Its unique mapping methodology addresses the critical gap in micronutrient data for processed and packaged foods, which constitute a substantial portion of modern diets [38] [33]. This makes it particularly valuable for research on nutritionally vulnerable populations with high consumption of branded products. The tool's extensive international datasets further support cross-cultural research and studies involving immigrant populations with diverse dietary patterns.

• ASA24 represents the optimal choice for large-scale epidemiological studies in the United States, particularly those requiring alignment with national dietary surveillance data and patterns. Its direct connection to the USDA's FNDDS ensures compatibility with data from the National Health and Nutrition Examination Survey (NHANES) [3] [34]. As a free tool, it offers significant advantages for projects with limited budgets, though its updating schedule means it may not capture the very latest food products entering the market.

• Intake24 offers a robust solution for national nutrition surveys and population-level monitoring, particularly in the UK context. Its validated portion size image system enhances the accuracy of self-reported amounts, addressing a key source of measurement error in dietary assessment [35]. The tool's demonstrated adaptability for use in other countries (e.g., New Zealand) suggests utility for public health nutrition monitoring in various English-speaking populations [36].

Limitations and Future Directions in Dietary Assessment Technology

Despite significant advancements, important limitations persist across all three platforms. A 2025 review of food composition databases highlights substantial variability in scope, content, and updating frequency across platforms, with significant gaps in the representation of traditional, indigenous, and culturally-specific foods [32]. This lack of comprehensive coverage disproportionately affects the dietary assessment accuracy for populations consuming these foods.

Future development priorities should focus on expanding database diversity to better capture global edible biodiversity, enhancing the cultural appropriateness of portion size estimation methods, and improving the interoperability between different dietary assessment platforms through standardized data attributes and FAIR (Findable, Accessible, Interoperable, and Reusable) data principles [32].

The integration of extensive branded and generic food databases within myfood24's framework represents a significant methodological advancement in dietary assessment technology. By systematically addressing the critical challenge of incomplete micronutrient data in branded food products through its sophisticated mapping and overstamping processes, myfood24 offers researchers a uniquely detailed lens for examining nutritional intake in contemporary food environments.

The selection between myfood24, ASA24, and Intake24 should be guided by specific research requirements: myfood24 for studies demanding comprehensive branded food coverage and micronutrient detail; ASA24 for US-focused epidemiological research requiring alignment with national surveillance data; and Intake24 for population-level surveys prioritizing validated portion size estimation. As global food systems continue to evolve, ongoing refinement of these platforms—particularly through expanded cultural food representation and enhanced interoperability—will further strengthen their utility for generating robust evidence to inform public health nutrition policy and practice.

This guide provides a comparative analysis of three prominent digital dietary assessment tools—ASA24, INTAKE24, and MyFood24—evaluating their performance and suitability across different age groups and ethnic populations. Evidence from validation studies indicates that all three tools can be viable alternatives to traditional interviewer-led recalls for collecting dietary data in large-scale studies. However, their performance varies, with key differences emerging in their adaptation for diverse ethnic groups, usability across age ranges, and specific validation protocols. The choice of tool depends heavily on the target population and the specific nutrient or dietary component of interest.

Comparative Performance Metrics at a Glance

Table 1: Key Validation Findings for ASA24, INTAKE24, and MyFood24

Tool & Validated Population	Comparison Method	Key Findings on Energy & Nutrient Intake	Usability & Completion
ASA24 (Adults) [16] [39]	Interviewer-administered AMPM	Equivalence for most nutrients and dietary supplement use prevalence [16]. Nutrient amounts from supplements can vary (e.g., vitamin D) [39].	61-88% completion rates in clinic settings; older adults and those with lower tech literacy may require assistance [40].
INTAKE24 (Ages 11-24) [35]	Interviewer-led 24-h recall	Mean intakes for energy and most nutrients within 4% of interviewer-led recall. Energy underestimated by 1% on average [35].	Developed specifically for 11-24 year-olds; considered more engaging than paper methods [35].
MyFood24 (Adults) [22] [41]	Biomarkers & Interviewer-led recall	Strong correlation (ρ=0.62) between estimated folate intake and serum folate [22]. Good agreement with interviewer-led recalls for most nutrients [41].	Demonstrated "good" usability and feasibility in groups from adolescents to the elderly [22] [41].
INTAKE24 (South Asian Adults) [7]	Not directly compared; performance evaluated internally.	Enabled analysis of energy intake patterns across age, sex, and BMI (e.g., higher median energy in younger males) [7].	Median recall time: 13 minutes; 99% of recalls included >8 food items in the South Asia Biobank [7].

Table 2: Adaptation for Diverse and Global Populations

Tool	Evidence of Cross-Cultural Adaptation	Key Adaptation Features
ASA24	Used in US studies with diverse ethnic samples (e.g., Non-Hispanic Black, Hispanic) [16] [39].	Primarily relies on its extensive US-based food database. Less documentation on fundamental database customization for other cultures.
INTAKE24	Successfully adapted for the South Asia Biobank (Bangladesh, India, Pakistan, Sri Lanka) [7].	Development of a comprehensive, bespoke food database with 2,283 local food items, linked to local portion sizes and nutrient information [7].
MyFood24	Adapted for multiple countries including Germany, France, Nigeria, and for Indigenous communities in the Peruvian Amazon [41].	Customization of the underlying food composition database (FCDB) is a core part of the adaptation process for each new population [41].
Foodbook24	Adapted for Polish and Brazilian populations in Ireland [17].	Food list expansion with 546 foods, translation into Polish and Portuguese, and use of home country nutrient databases for culturally specific foods [17].

Detailed Experimental Protocols and Methodologies

INTAKE24 Validation in Adolescents and Young Adults

A key validation study for INTAKE24 recruited 180 participants aged 11-24 years [35].

• Protocol: Each participant completed both an INTAKE24 recall and an interviewer-led 24-hour recall on the same day. This was repeated on four separate occasions over one month [35].
• Methodology: The interviewer-led recalls followed the multiple-pass method used in the UK Low Income Diet and Nutrition Survey (LIDNS). INTAKE24 uses a free-text entry for foods, which are matched to a database, and portion size is estimated using over 3,000 food photographs. A weighted randomization was used, with 75% of participants completing INTAKE24 first to test it under the most critical conditions [35].

MyFood24 Validation Against Biomarkers

A 2025 study assessed the validity of MyFood24 against objective biomarkers in 71 healthy Danish adults [22].

• Protocol: The study used a repeated cross-sectional design. Participants completed a 7-day weighed food record (WFR) using MyFood24 at baseline and again four weeks later [22].
• Biomarker Comparison: At the end of each 7-day WFR, the following biomarkers were collected:
  • Energy: Resting energy expenditure was measured via indirect calorimetry. The Goldberg cut-off was applied to identify acceptable energy reporters [22].
  • Protein: Estimated intake was compared against urinary urea excretion [22].
  • Folate: Total folate intake from the tool was correlated with serum folate levels [22].
  • Potassium: Estimated intake was compared against urinary potassium excretion [22].

Demographic Suitability Analysis

Suitability by Age Group

• Adolescents and Young Adults (11-24 years): INTAKE24 was developed specifically for this demographic. Its design through iterative user studies makes it particularly suitable and engaging for this tech-savvy age group [35]. MyFood24 has also been validated in British adolescents aged 11–18 years, showing agreement with interviewer-led recalls [41].
• Adults (General): All three tools have been validated in adult populations. ASA24 and MyFood24 have extensive validation data for this group [22] [16]. MyFood24's validity in adults is supported by strong correlations with biomarkers like serum folate [22].
• Older Adults (65+ years): A 2025 pilot study suggests that voice-based dietary recall tools may be preferable for older adults, who can find screen-based interfaces challenging [9]. The study found older adults rated a voice-based tool as easier and preferred it over ASA24 [9]. Another study confirmed that fewer older adults could complete ASA24 independently, indicating a potential need for technological support for this demographic [40].

Adaptation for Ethnic and Cultural Diversity

A tool's suitability depends heavily on how well its food list and nutrient database reflect the target population's diet.

• INTAKE24 and MyFood24 demonstrate a proven, structured approach for cross-cultural adaptation, centered on building or modifying the core food composition database [7] [41]. INTAKE24's adaptation for the South Asia Biobank involved creating a new database with 2,283 local items, ensuring good coverage and accurate portion sizes for that population [7].
• ASA24, while used in diverse U.S. cohorts, is primarily built around a U.S. food database [16] [39]. The search results do not provide evidence of fundamental database restructuring for other cultural diets, which may limit its application for non-U.S. populations without significant supplemental data collection.
• A critical finding from the Foodbook24 expansion was that despite adding hundreds of foods, Brazilian participants still omitted a higher percentage of foods (24%) in self-administered recalls compared to an Irish cohort (13%), highlighting the persistent challenge of achieving complete dietary coverage for all ethnic groups [17].

The Researcher's Toolkit: Dietary Assessment Workflow

The following diagram illustrates the key stages and decision points in adapting and deploying a digital dietary recall tool for a diverse population, based on methodologies from the cited research.

Essential Research Reagents and Solutions

Table 3: Key Materials and Methods for Dietary Assessment Validation

Item / Method	Function in Validation Research
24-Hour Urine Collection	Provides objective biomarkers for validating nutrient intake, specifically for potassium (using potassium excretion) and protein (using urinary urea/ nitrogen) [22].
Fasting Blood Samples	Allows measurement of concentration biomarkers to validate intake of specific nutrients, such as serum folate for folate intake [22].
Indirect Calorimetry	Measures resting energy expenditure (REE), which is used with the Goldberg cut-off to identify under- or over-reporters of energy intake in a sample [22].
Doubly Labelled Water (DLW)	Considered the gold standard for measuring total energy expenditure in free-living individuals; used as a reference method for validating reported energy intake (not used in the cited studies but referenced as a standard) [35].
Standardized Portion Size Images	Visual aids (e.g., food atlases, image series) critical for helping participants estimate the quantity of food consumed, thereby improving the accuracy of portion size reporting [35] [17].
Local Food Composition Databases (FCDB)	The foundation for accurate nutrient calculation. Adapting a tool for a new population requires developing or integrating a FCDB that reflects local foods and their nutrient profiles [7] [41] [17].

Navigating Limitations and Enhancing Data Quality in Practice

Accurate dietary assessment is a cornerstone of nutritional epidemiology, public health monitoring, and clinical research. However, two persistent challenges significantly compromise data quality: the underreporting of energy intake and inadequate representation of ethnic and cultural foods. As research populations become increasingly diverse, these pitfalls can introduce substantial measurement error, bias study findings, and marginalize specific demographic groups from nutritional science. The transition to web-based, automated 24-hour dietary recall tools like ASA24, Intake24, and myfood24 offers potential solutions through standardized methodology and enhanced adaptability. This guide provides a comparative analysis of these three major platforms, evaluating their performance in mitigating these critical issues through experimental data and validation studies. Understanding their relative strengths and limitations empowers researchers to select the most appropriate tool for studies involving diverse populations, thereby improving the accuracy and equity of dietary assessment.

Platform Origins and Key Characteristics

ASA24 (Automated Self-Administered 24-Hour Dietary Assessment Tool) is a freely available, web-based tool developed by the National Cancer Institute (NCI) in the United States. It adapts the USDA's Automated Multiple-Pass Method and is designed for large-scale epidemiologic research [3] [11].

Intake24 is an open-source, web-based 24-hour dietary recall system originally developed in the United Kingdom. Its design emphasizes cost-effectiveness for national nutrition surveys and it has been adapted for use in several countries, including New Zealand and across South Asia [7] [24].

myfood24 is a commercially supported online dietary assessment tool developed by academics at the University of Leeds. It markets itself as the only tool validated against nutrient biomarkers and supports both 24-hour recalls and food diaries, underpinned by a extensive branded food database [22] [42].

Table 1: Key Characteristics of Web-Based Dietary Assessment Tools

Feature	ASA24	Intake24	myfood24
Primary Developer	National Cancer Institute (USA)	Newcastle University (UK)	University of Leeds (UK)
Cost	Free	Free (open-source)	Commercial
Core Methodology	24-hour recall, Food record	24-hour recall	24-hour recall, Food diary
Adaptability for International Use	Canadian & Australian versions available [3]	Adapted for NZ, Bangladesh, India, Pakistan, Sri Lanka [7] [24]	International versions in 9 languages [42]
Key Claim	Designed for large-scale studies [11]	Cost-effective for national surveys [24]	Clinically validated against biomarkers [42]

Analytical Framework and Evaluation Criteria

This comparison evaluates the three tools based on two primary sources of dietary assessment error:

• Systematic Underreporting of Energy and Nutrients: Assessed by comparing tool-reported intakes against objective recovery biomarkers and weighed food records.
• Gaps in Food Lists for Ethnic and Cultural Minorities: Assessed through usability studies, validation research, and documented tool adaptation processes.

Experimental data from validation studies are synthesized to provide a quantitative and qualitative comparison of tool performance.

Experimental Protocols for Tool Validation

To objectively compare the performance of ASA24, Intake24, and myfood24, researchers have employed rigorous experimental protocols. The following diagram outlines a generalized workflow for a validation study that incorporates multiple reference methods, including objective biomarkers.

Protocol 1: Biomarker-Based Validation

The most robust validation studies employ recovery biomarkers, which provide objective, non-self-reported measures of intake [22] [11].

• Population: Typically involves dozens to hundreds of healthy adults. For example, the IDATA study for ASA24 included 1,077 men and women aged 50-74 [11], while a myfood24 study included 71 Danish adults [22].
• Design: Participants complete multiple recalls on the web-based tool. Concurrently, energy expenditure is measured via doubly labeled water, and intakes of protein, sodium, and potassium are assessed via 24-hour urine collection [11]. Plasma or serum folate can also be measured as a biomarker for fruit and vegetable intake [22].
• Analysis: Reported energy and nutrient intakes from the tool are compared to biomarker values using correlation and Bland-Altman analyses to quantify the degree of under- or over-reporting.

Protocol 2: Comparative and Usability Studies

These studies evaluate a tool's performance against other dietary assessment methods and its practicality in target populations.

• Food List Coverage Analysis: Researchers collect dietary data from a diverse sample using a non-structured method (e.g., a free-text listing of all foods consumed). The items are then checked against the tool's food list to calculate the percentage of matches, identifying gaps in ethnic and traditional foods [43].
• Usability Testing: A mixed-methods approach where participants complete a recall while researchers record their screen and elicit verbal feedback ("think-aloud" technique). This is followed by a survey to identify challenges in navigation, food search, and portion size estimation [24].
• Cohort Application: The tool is deployed in a large-scale, diverse cohort study. Completion rates, times, and the quality of the collected data (e.g., number of food items reported, missing data rates) are analyzed across different demographic subgroups [7] [11].

Quantitative Performance Comparison

Data from independent validation studies reveal how each tool performs regarding accuracy and feasibility.

Underreporting of Energy and Nutrients

A critical challenge for all self-reported dietary tools is the systematic underreporting of energy intake. However, the magnitude of this error varies.

Table 2: Comparison of Energy and Nutrient Underreporting

Tool & Study	Reference Method	Energy Intake Difference	Nutrient Intake Correlation
ASA24 (IDATA Study) [11]	Doubly Labeled Water (Energy), Urinary Nitrogen (Protein)	Underreported vs. TEE	Protein intake closer to biomarker for women than men
myfood24 (Danish Validation) [22]	7-day Weighed Food Record & Biomarkers	87% classified as "acceptable reporters" via Goldberg cut-off	Strong correlation for folate (ρ=0.62); Acceptable for protein (ρ=0.45) & potassium (ρ=0.42)
myfood24 (Reproducibility) [22]	Repeated myfood24 administration	N/A	Strong correlations (ρ ≥ 0.50) for most nutrients; Highest for folate (ρ=0.84)

Operational Feasibility and User Engagement

The practicality of a tool for large-scale studies is determined by how easily participants can use it and the quality of data it generates.

Table 3: Comparison of Feasibility and User Experience Metrics

Metric	ASA24	Intake24	myfood24
Median Completion Time	~24 minutes (average) [26]	13 minutes (South Asia Biobank) [7]	Information Missing
Completion Rate	High (≥3 recalls by 91% men, 86% women) [11]	Information Missing	Information Missing
User Feedback	Designed for 5th-grade reading level [26]	84% reported "easy to use" but noted search/portion challenges [24]	Information Missing

Addressing the Pitfalls: A Detailed Analysis

Pitfall 1: Underreporting of Energy Intake

Underreporting is a well-documented issue across all dietary assessment methods. The IDATA study found that ASA24-estimated energy intakes were lower than energy expenditure measured by doubly labeled water for both men and women [11]. This indicates a systematic bias towards underreporting, a common phenomenon in self-reported data. The same study noted that the accuracy of protein reporting via ASA24 differed by sex, being closer to the urinary nitrogen biomarker for women than for men [11].

In contrast, a validation study for myfood24 that also used biomarkers reported that 87% of participants were classified as "acceptable reporters" of energy intake when assessed via the Goldberg cut-off [22]. Furthermore, this study demonstrated a strong Spearman’s rank correlation for total folate intake and serum folate (ρ=0.62), and acceptable correlations for protein and potassium, suggesting it is a useful tool for ranking individuals by intake within a study population [22].

Pitfall 2: Missing Ethnic and Cultural Foods

The failure of a dietary tool to include foods commonly consumed by ethnic and cultural minorities leads to item omission, misclassification, and ultimately, data that does not represent the true dietary habits of the population.

• Adaptation of Intake24: The South Asia Biobank project undertook a major adaptation of Intake24, developing a food database with 2,283 items to reflect the diverse food supply, preparations, and eating behaviors across Bangladesh, India, Pakistan, and Sri Lanka [7]. This demonstrates the tool's capacity for large-scale customization to ensure good coverage of foods consumed across diverse regions.

• Expansion of Foodbook24 (A Model for myfood24/ASA24): A study in Ireland expanded the Foodbook24 tool by adding 546 foods commonly consumed by Brazilian and Polish minorities. The updated list contained 86.5% (302 out of 349) of the foods listed by participants in a validation study, showing a high degree of comprehensiveness post-adaptation [43]. This process involved translating foods into Polish and Portuguese and using national nutrient databases from Brazil and Poland for culturally specific items [43].

• Usability Challenges in Intake24-NZ: A qualitative usability study of the New Zealand version of Intake24 found that despite overall positive feedback, participants faced challenges related to search terms and the types/order of foods displayed in search results [24]. This highlights that even after adaptation, continuous refinement of the user interface and food synonym list is necessary to capture the full diversity of dietary intake.

The Scientist's Toolkit: Research Reagent Solutions

The following table details key resources and methodologies used in the validation and adaptation of dietary assessment tools, as evidenced by the cited research.

Table 4: Essential Reagents and Methods for Dietary Tool Validation

Reagent / Method	Function in Validation/Adaptation	Exemplar Use Case
Doubly Labeled Water (DLW)	Objective biomarker for total energy expenditure; considered the gold standard for validating reported energy intake.	Used in the IDATA study to quantify the underreporting of energy in ASA24 [11].
24-Hour Urine Collection	Provides recovery biomarkers for nutrients like protein (via nitrogen), sodium, and potassium, against which self-reported intake can be compared.	Collected in the IDATA [11] and myfood24 Danish validation [22] studies.
National Food Consumption Surveys	Source of data on commonly consumed foods and typical portion sizes for specific national or ethnic sub-populations.	Used to identify 546 foods for addition to Foodbook24 for Brazilian and Polish groups in Ireland [43].
Food Composition Database (FCDB)	The nutrient lookup table underlying a dietary tool. Accuracy depends on using a comprehensive and context-appropriate FCDB.	The South Asia Intake24 project developed a regional FCDB [7]. Foodbook24 used Brazilian and Polish FCDBs for culturally unique foods [43].
Portion Size Image Sets	Visual aids to help users estimate the quantity of food consumed. Must be culturally appropriate (e.g., correct plateware, common brands).	Intake24 uses "as-served" and "guide" images. The NZ adaptation required updating these images to include local brands and servewares [24].

Accurate dietary assessment is a cornerstone of nutritional epidemiology, public health monitoring, and clinical nutrition research. Within this field, portion size estimation represents a critical source of measurement error that can significantly impact the validity of nutrient intake data [44]. For decades, researchers have sought to improve the accuracy of portion size reporting through various estimation aids (PSEAs), with food photographs and digital images emerging as prominent tools in both traditional and technologically-assisted dietary assessment methods [44] [45].

The evolution from text-based descriptions to visual aids reflects an ongoing effort to enhance precision while reducing participant burden. As dietary assessment has transitioned to digital platforms, understanding the comparative effectiveness of different visual estimation methods has become increasingly important for researchers selecting appropriate methodologies for large-scale surveys and clinical studies [36]. This guide objectively compares the performance of food photographs and digital images within the context of validating three prominent automated dietary assessment tools: ASA24, Intake24, and MyFood24.

Methodological Foundations of Portion Size Estimation

Core Concepts and Challenges

Portion size estimation accuracy is influenced by three fundamental cognitive processes: perception (relating an actual food amount to its representation in an aid), conceptualization (forming a mental picture of a consumed portion), and memory (accurately recalling the amount eaten) [45]. Each of these elements introduces potential error into dietary reporting.

Research has consistently identified several persistent challenges in portion size estimation:

• Flat-slope phenomenon: Large portions tend to be underestimated while small portions are overestimated [44] [45]
• Food form variability: Amorphous foods (e.g., mashed potatoes, scrambled eggs) and items without defined shapes are reported less accurately than single-unit foods (e.g., bread slices, fruits) [44] [45]
• Serving size effects: Foods consumed in small portions (e.g., spreads) are typically estimated more accurately than larger portions [44]

Evolution of Portion Size Estimation Aids

Traditional portion size estimation aids have included household measures, standard portion sizes, food models, and text-based descriptions. The development of image-based aids represents a significant advancement, beginning with photographic atlases and evolving into integrated digital image systems within automated dietary assessment platforms [44] [45].

Table 1: Evolution of Visual Portion Size Estimation Aids

Era	Primary Aid Types	Key Characteristics	Limitations
Pre-2000s	Food models, household utensils, text descriptions	Physical objects, verbal quantification	Inconvenient for large surveys, dependent on respondent familiarity with measures
2000-2010	Printed photographic booklets, atlases	Standardized 2D representations, multiple portion sizes	Limited food variety, static formats, production costs
2010-Present	Integrated digital images in automated systems	Dynamic presentation, extensive food databases, algorithmic support	Technology access barriers, varying image quality standards

Comparative Analysis of Automated Dietary Assessment Tools

Three automated dietary assessment tools have emerged as prominent platforms for large-scale dietary assessment: ASA24, Intake24, and MyFood24. Each incorporates visual portion size estimation aids differently, reflecting varying design philosophies and technical approaches.

ASA24 (Automated Self-Administered 24-hour Recall): Developed by the National Cancer Institute, ASA24 utilizes a comprehensive set of aerial food photographs based on the Food Intake Recording Software System [45]. The system employs a "multiple pass" method to enhance recall completeness and includes thousands of food images portraying varying portion sizes.

Intake24: This online 24-hour recall system, initially developed for use in the United Kingdom, utilizes food photographs specifically developed and validated against UK National Diet and Nutrition Survey data [10]. The system incorporates over 3,000 food photographs with portion sizes based on reported consumption patterns in the target population.

MyFood24: While also based on some principles of the Automated Multiple-Pass Method, MyFood24 adopts a more selective approach to visual aids, incorporating both images and text-based estimation methods [10].

Experimental Protocols and Validation Methodologies

Validation studies for these tools have employed rigorous methodologies to assess accuracy under controlled conditions:

Controlled Feeding Studies: The most robust validation approach involves controlled feeding studies where participants consume pre-weighed meals in laboratory settings, followed by dietary recall using the tested method. For example, one study involving 152 participants utilized a crossover design where subjects consumed breakfast, lunch, and dinner with unobtrusive weighing of foods and beverages, followed by 24-hour recalls using one of four technology-assisted methods [46].

Observational Feeding Studies: These studies allow participants to self-serve foods in buffet-style settings, with researchers unobtrusively weighing servings and plate waste to establish "true" intake. Participants then return the following day to estimate portion sizes using the digital tools being tested [45]. This approach balances controlled measurement with more natural eating behaviors.

Comparison Studies: Large-scale field comparisons involve participants completing both the test method (e.g., online recall) and an established reference method (e.g., interviewer-led recall). For example, the INTAKE24 validation involved 180 participants aged 11-24 years completing both INTAKE24 and interviewer-led recalls on four separate occasions [10].

Performance Comparison: Quantitative Findings

Accuracy of Portion Size Estimation Methods

Recent controlled studies provide direct comparisons of estimation accuracy between different visual approaches and tools:

Table 2: Accuracy of Portion Size Estimation Methods Based on Controlled Studies

Estimation Method	Study Characteristics	Energy Intake Accuracy (% of true intake)	Key Findings on Portion Size Estimation
Text-Based (TB-PSE)	40 participants, 2-24h recall delay [44]	--	31% of estimates within 10% of true intake; 50% within 25% of true intake
Image-Based (IB-PSE)	Same study population as above [44]	--	13% of estimates within 10% of true intake; 35% within 25% of true intake
ASA24	152 participants, controlled feeding [46]	+5.4% (overestimation)	Reasonable validity for average intake estimates
Intake24	Same study population as above [46]	+1.7% (slight overestimation)	Accurate estimation of intake distributions for energy and protein
Image-Assisted Interviewer-Administered 24HR	Same study population as above [46]	+15.0% (substantial overestimation)	Highest estimation error among tested methods

Impact of Food Characteristics on Estimation Accuracy

The effectiveness of visual estimation methods varies significantly by food type and characteristics:

Table 3: Estimation Accuracy by Food Form and Type

Food Category	Examples	Relative Estimation Accuracy	Noteworthy Findings
Single-unit foods	Bread slices, bread rolls, fruits	Highest accuracy	More easily conceptualized and matched to reference images [44]
Amorphous foods	Scrambled eggs, yogurt, muesli	Lowest accuracy	Lack defined shapes, increasing estimation error [44] [45]
Liquids	Milk, juice, water	Moderate accuracy	Varies by container transparency and measurement unit
Spreads	Margarine, jam	Moderate to high accuracy for small portions	Small absolute errors due to small typical portion sizes [44]

Figure 1: Factors Influencing Portion Size Estimation Accuracy. This diagram illustrates the cognitive processes, food characteristics, and methodological approaches that collectively determine estimation accuracy, with color-coding indicating the nature of each factor's influence.

Research Reagent Solutions

Implementing robust portion size estimation research requires specific materials and methodological components:

Table 4: Essential Research Materials and Their Functions

Research Material	Function/Purpose	Implementation Examples
Standardized Food Image Sets	Provide consistent visual references for portion estimation	ASA24 picture book (9,000+ aerial photographs) [45]; INTAKE24's 3,000+ UK-specific food images [10]
Calibrated Weighing Scales	Establish "true" intake values for validation studies	UltraShip digital scales (accurate to 2g) used in observational feeding studies [45]
Digital Image Presentation Platforms	Standardize image display and response collection	Custom computer applications; web-based dietary assessment tools (ASA24, INTAKE24, MyFood24) [44] [10]
Portion Size Range Standards	Define appropriate minimum and maximum portions for images	5th to 95th percentiles of consumption from national surveys (e.g., NHANES) [45]
Reference Measurement Protocols	Standardize the quantification of true intake	Weighed serving containers pre- and post-selection; plate waste measurements [44] [45]

Experimental Workflow for Validation Studies

Figure 2: Experimental Workflow for Portion Size Estimation Validation. This standardized methodology illustrates the sequential phases for validating portion size estimation methods, from initial preparation through data collection to statistical analysis.

Implications for Research and Practice

The comparative evidence on portion size estimation methods yields several important considerations for researchers and practitioners:

Tool Selection Criteria: When choosing between automated dietary assessment tools, researchers should consider:

• Population characteristics: Age, tech literacy, and cultural background influence estimation accuracy
• Food focus: Studies emphasizing amorphous foods may require supplemental estimation methods
• Resource constraints: Balance between implementation cost and data precision requirements
• Analysis needs: Nutrient-level analysis versus food group pattern identification

Methodological Recommendations: Based on current evidence:

• Text-based estimation (household measures, standard portions) may provide superior accuracy to image-based approaches for many applications [44]
• When using images, presenting multiple portion options (e.g., 8 images) significantly outperforms single "average" portion images [45] [47]
• Food-specific estimation strategies accounting for food form characteristics can enhance overall accuracy
• Method consistency may be more important than absolute accuracy for longitudinal measures

Future Directions: Emerging technologies including artificial intelligence, machine learning, and virtual reality present opportunities for next-generation portion size estimation [2]. These approaches may help address persistent challenges with amorphous foods and the flat-slope phenomenon while potentially reducing participant burden through automated image analysis and immersive estimation environments.

This guide objectively compares the performance of three major online 24-hour dietary assessment tools—ASA24, Intake24, and myfood24—focusing on strategies to enhance participant compliance through usability testing, reminders, and incentives. The content is framed within the broader thesis of comparatively validating these tools for research applications.

Online, self-administered 24-hour dietary recall tools have revolutionized data collection in nutritional epidemiology by reducing costs and administrative burden. However, their effectiveness hinges on participant compliance and their ability to generate accurate data. ASA24 (US), Intake24 (UK), and myfood24 (UK) are three leading tools that have undergone significant development and validation. Evidence indicates that the strategies of usability testing, participant reminders, and judicious use of incentives are critical success factors, though their implementation and effectiveness vary across platforms. The following comparison synthesizes experimental data to guide researchers in selecting and deploying these tools.

A primary strategy for improving compliance is to ensure the tool is intuitive and easy to use. Usability testing directly identifies challenges that lead to user errors and abandonment.

Table 1: Tool Overview and Key Usability Metrics

Feature	ASA24	Intake24	myfood24
Developer/Origin	National Cancer Institute (USA) [3]	Newcastle University (UK), open-source [24] [4]	University of Leeds (UK) [48] [49]
Core Methodology	Automated Multiple-Pass Method (AMPM) [3]	Multiple-pass 24-hour recall [24] [23]	Adapts aspects of AMPM [49]
Reported Completion Time	~40 minutes [50]	Information missing	31 minutes (beta), improved to 16 minutes (live) [48]
System Usability Scale (SUS) Score	Information missing	Information missing	Beta: 66/100, Live: 74/100 (considered "good") [48]
Key Usability Findings	Only 1 in 39 low-income users completed recall unassisted; 71% knowingly entered incorrect data [50]	Users experienced challenges with search terms, portion sizes, and food prompts [24]	Navigation and presentation errors in beta version significantly improved in live version [48]

Experimental Protocols for Usability Testing

The methodologies for obtaining the usability data in Table 1 provide a blueprint for researchers to evaluate tools for their own populations.

• ASA24 Usability Study (2019): This mixed-methods study involved 39 low-income adults (85% female, mean age 48.2). Participants were asked to complete a single 24-hour recall using ASA24. Sessions were recorded using audio and screen capture software. Researchers analyzed these recordings to calculate quantitative metrics like task success rates and time, and to qualitatively code for specific usability issues. The "think-aloud" technique, where participants verbalize their thought process, was used to gain insight into cognitive challenges [50].

• Intake24-NZ Usability Study (2024): This study also used a mixed-methods approach with 37 participants aged ≥11 years. The protocol consisted of two components: 1) completion of a dietary recall with screen observation recordings, and 2) a follow-up usability survey. The combination of observed behavior and self-reported feedback allowed researchers to identify discrepancies between what users do and what they say, uncovering challenges like incorrect search term usage and portion size estimation [24].

• myfood24 Formative Evaluation (2015): This two-stage evaluation first tested a beta-version with 14 adolescents, using screen capture software and questionnaires. After amendments were made to create a live version, 70 adolescents were recruited to use the tool and complete the same questionnaire. This iterative process allowed researchers to quantify improvements in usability via the System Usability Scale (SUS) and reduction in completion time [48].

Comparative Validation and Data Accuracy

Ultimately, compliance strategies must support the goal of collecting valid data. The following table compares the performance of these tools against objective and interviewer-led benchmarks.

Table 2: Comparative Validity and Reliability Data

Validation Metric	ASA24	Intake24	myfood24
Criterion Validity (vs. Doubly Labeled Water)	Under-reported EI vs. biomarkers [19]	Under-reported EI by ~25% [23]	Attenuation factors of 0.2-0.3 against biomarkers [20]
Comparison to Interviewer-Led Recalls	Found to be comparable [19] [51]	Yields comparable estimates [23]	Nutrient estimates ~10-20% lower, with moderate agreement (ICC ~0.4-0.5) [20]
Reliability (Intra-class Correlation for Single Recall)	Information missing	0.35 for Energy Intake [23]	Information missing

Experimental Protocols for Validation Studies

• Intake24 Criterion Validity Protocol (2019): To validate energy intake (EI), 98 UK adults (40-65 years) were recruited. Total Energy Expenditure (TEE) was measured using the doubly labeled water (DLW) method, the gold standard. Participants consumed a weight-specific dose of DLW and provided urine samples over 9-10 days. During this period, they completed the Intake24 recall at least twice. Reported EI was then compared to TEE using Bland-Altman analysis and correlation tests [23].

• myfood24 Biomarker Comparison Protocol (2018): This study recruited 212 metabolically stable adults. Each participant completed both the myfood24 online recall and an interviewer-administered multiple-pass recall on three separate occasions. Estimated intakes of protein, potassium, and sodium were compared against urinary biomarker concentrations, while sugar intake was compared to a predictive biomarker. This design allowed for a direct comparison of the online tool against both biomarker standards and a traditional method [20].

The Scientist's Toolkit: Research Reagent Solutions

The following table details key materials and methodological "reagents" essential for conducting the types of experiments cited in this guide.

Table 3: Essential Research Reagents for Dietary Assessment Validation

Reagent / Tool	Function in Research	Example Application
Doubly Labeled Water (DLW)	Criterion method for measuring total energy expenditure in free-living individuals [23].	Serves as an objective benchmark to validate self-reported energy intake from tools like Intake24 [23].
24-Hour Urinary Biomarkers	Objective measures of nutrient intake (e.g., protein via nitrogen, sodium, potassium) [20].	Used to validate reported intake of specific nutrients against their urinary recovery, as in the myfood24 validation [20].
Screen & Audio Recording Software	Captures real-time user interaction and verbal feedback during tool use [24] [50].	Critical for in-depth usability testing to identify interface problems and user errors, as performed for ASA24 and Intake24-NZ [24] [50].
System Usability Scale (SUS)	A standardized, reliable questionnaire for measuring subjective perceptions of usability [48].	Provides a quick, quantitative score to compare usability across different tool versions or populations, as used in myfood24 development [48].
Automated Multiple-Pass Method (AMPM)	A structured interview protocol to enhance completeness and accuracy of dietary recalls [3].	Forms the foundational methodology for ASA24 and influences the design of other tools like myfood24 [3] [49].

Workflow Visualization for Compliance Strategy Testing

The following diagram synthesizes the key steps in a comprehensive strategy for developing and validating a compliant-friendly dietary assessment tool, as illustrated by the experimental protocols.

Compliance Strategy Testing Workflow

Synthesis and Recommendations

The evidence demonstrates that while all three tools are viable for research, a one-size-fits-all approach to compliance is ineffective. Usability is not a binary feature but a continuous target that requires iterative testing with specific populations, as starkly shown by the struggles of low-income adults with ASA24 despite its sophisticated design [50]. Furthermore, validation is context-dependent; all tools exhibit under-reporting compared to biomarkers, but this is a known issue across dietary assessment methods and their performance is generally comparable to more costly interviewer-led recalls [23] [20].

For researchers, the following evidence-based recommendations are made:

• Prioritize Population-Specific Usability: Do not assume a tool validated in one population will perform equally in another. Conduct pilot usability tests mimicking the protocols described for Intake24-NZ and ASA24 [24] [50].
• Provide Proactive Support: The finding that 71% of users knowingly entered incorrect data on ASA24 suggests that providing on-demand, proactive technical support is crucial, not optional [50].
• Use Multi-Day Recalls: Reliability data for Intake24 shows that a single recall is insufficient to rank individuals accurately (ICC=0.35 for energy); using the mean of two or more recalls significantly improves reliability (ICC=0.52) [23]. Study design should plan for multiple recalls per participant.
• Select Tools Based on Research Context: The choice between ASA24, Intake24, and myfood24 should be guided by the target population's location, the need for customization (favoring the open-source Intake24), and the specific nutrient databases and features required by the research question [3] [24] [49].

In nutritional epidemiology and clinical research, the quality of dietary intake data is fundamentally dependent on the underlying food and nutrient databases that support assessment tools. Automated, self-administered 24-hour dietary recall systems have emerged as viable alternatives to traditional interviewer-led methods, offering advantages in scalability, cost-effectiveness, and reduced participant burden [10]. Among these, ASA24, INTAKE24, and myfood24 represent three prominent tools built upon the Automated Multiple-Pass Method (AMPM) methodology. While these systems share similar methodological foundations, their database management approaches—including comprehensiveness of food items, nutrient coding procedures, and portion size estimation methods—vary significantly, creating important distinctions for researchers to consider when selecting appropriate assessment tools. This guide provides a detailed comparison of these systems, with particular emphasis on their database management infrastructures and the experimental evidence supporting their performance.

The foundation of any dietary assessment tool is its underlying database architecture, which dictates the comprehensiveness, accuracy, and usability of the system. The three tools examined employ distinct approaches to database management, reflecting their different developmental origins and target populations.

ASA24 leverages the United States Department of Agriculture's (USDA) Food and Nutrient Database for Dietary Studies (FNDDS) and Food Pattern Equivalent Database (FPED) to derive nutrient and food group estimates [52]. This connection to nationally representative databases ensures consistency with major governmental surveys. The system incorporates a highly standardized multi-pass web-based recall that adapts the interviewer-administered AMPM method used in "What We Eat in America," the dietary interview component of the National Health and Nutrition Examination Survey (NHANES) [13]. This methodological alignment facilitates comparability with national benchmark data.

INTAKE24 utilizes a database linked to the UK National Diet and Nutrition Survey (NDNS) Nutrient Databank, containing over 2,500 foods [10] [23]. The system was specifically developed for use with 11-24-year-olds but has since been extended for the general adult population. A distinctive feature of INTAKE24 is its portion size estimation system, which employs a series of over 3,000 food photographs developed based on portion sizes reported in UK National Diet and Nutrition Surveys and validated in both feeding studies and relative validations against 4-day weighed intakes [10].

myfood24, mentioned briefly in comparative literature, adopts only some aspects of the AMPM method and has been developed for use with the UK population, validated against face-to-face interviewer-led recalls in 11-18-year-olds [10]. While detailed database specifications for myfood24 are not fully elaborated in the available search results, it represents another UK-focused alternative in the landscape of automated dietary assessment tools.

For researchers requiring exceptionally detailed nutrient coverage, the Nutrition Data System for Research (NDSR) software maintains a comprehensive database of approximately 19,500 foods, including 8,100 brand name products, with 181 nutrients, nutrient ratios, and other food components [53]. Though not one of the three primary tools examined, it represents the upper range of database comprehensiveness available in dietary assessment software.

Table 1: Database Architecture and Coverage Comparison

Tool	Primary Nutrient Database	Number of Foods	Nutrients Tracked	Portion Size Estimation	Regional Focus
ASA24	USDA FNDDS & FPED	Not specified	Comprehensive set aligned with USDA databases	Food photographs, dimensional estimates	United States
INTAKE24	UK NDNS Nutrient Databank	2,500+	Comprehensive set aligned with UK databases	3,000+ validated food photographs	United Kingdom, with international adaptations
myfood24	Not specified in available literature	Not specified	Not specified	Not specified	United Kingdom
NDSR	Proprietary database	~19,500 (including 8,100 brand products)	181 nutrients, ratios, and components	Multiple methods	United States

Experimental Validation and Comparative Performance

Validation studies provide critical insights into how these tools perform in real-world research settings, particularly regarding data accuracy and reliability. The available experimental evidence reveals important differences in validation approaches and outcomes.

INTAKE24 Validation Evidence

INTAKE24 has undergone extensive validation across multiple age groups. A comparison study with interviewer-led multiple pass 24-hour recalls in 180 participants aged 11-24 years found that mean intakes reported using INTAKE24 were similar to those reported in interviewer-led recalls for energy and most nutrients [10]. Specifically, INTAKE24 underestimated energy intake by just 1% on average compared to interviewer-led recalls, with limits of agreement ranging from -49% to +93% [10]. In the 11-16-year-old subgroup, mean energy intakes were underestimated by only 3%, with all INTAKE24 energy intakes falling within 50% of interviewer-led values [54].

A particularly rigorous validation study compared INTAKE24-reported energy intake against total energy expenditure measured using doubly labeled water (DLW) in 98 UK adults aged 40-65 years [23]. This criterion validation revealed that participants under-reported energy intake by 25% in the first recall, 22% for the average of first two recalls, and 25% for the first three recalls. The correlations between reported energy intake and measured energy expenditure were 0.31 (first recall), 0.47 (first two recalls), and 0.39 (first three recalls), indicating moderate ability to rank individuals by energy intake [23].

Test-retest reliability assessment for INTAKE24 demonstrated an intra-class correlation coefficient (ICC) of 0.35 for energy intake from a single recall, with values ranging from 0.31 for iron to 0.43 for non-milk extrinsic sugars. When considering pairs of recalls, the ICC improved to 0.52 for energy intake, ranging from 0.37 for fat to 0.63 for non-milk extrinsic sugars [23].

ASA24 Validation Evidence

ASA24 has been evaluated through the Food Reporting Comparison Study (FORCS), which compared self-reported intakes using ASA24-2011 with data collected using the interviewer-administered AMPM in 1,076 participants [13]. For dietary supplement use—a particularly challenging aspect of dietary assessment—the proportions reporting use via ASA24 and AMPM were 46% and 43% respectively, with these proportions judged equivalent with a small effect size of less than 20% [13].

However, usability studies have revealed important challenges with ASA24, particularly in certain populations. A study focusing on low-income adults found that of 39 participants, only one was able to complete a dietary recall unassisted [55]. Researchers identified 286 usability issues within 22 general usability categories, including difficulties using the search function, misunderstanding questions, and uncertainty regarding how to proceed to the next step. Notably, 71.4% of participants knowingly misentered dietary information at least once [55].

Table 2: Experimental Validation Metrics Across Tools

Validation Metric	INTAKE24	ASA24	myfood24
Energy Intake Comparison vs. Interviewer-Led Recall	-1% to -3% mean difference [10] [54]	Not specified	Not specified
Criterion Validity (DLW) Correlation	0.31-0.47 (energy intake vs. TEE) [23]	Not specified	Not specified
Under-reporting vs. DLW	22-25% [23]	Not specified	Not specified
Test-Retest Reliability (ICC)	0.35-0.52 (energy) [23]	Not specified	Not specified
Supplement Reporting Equivalence	Not specified	Equivalent to AMPM (46% vs. 43%) [13]	Not specified
Usability Challenges	Not specified	Significant issues in low-income populations [55]	Not specified

Methodology of Key Validation Experiments

Understanding the experimental protocols used to validate these tools provides critical context for interpreting their performance data and applying them appropriately in research settings.

INTAKE24 Comparison Study Protocol

The INTAKE24 validation study employed a rigorous comparison design with the following methodology [10]:

• Participants: 180 people aged 11-24 years (60 aged 11-16 years and 120 aged 17-24 years) recruited through schools and various community methods.
• Study Design: Each participant completed both an INTAKE24 recall and an interviewer-led 24-hour recall on the same day on four separate occasions over a one-month period, including at least one weekend day.
• Randomization: A weighted randomization was used whereby 75% of participants completed INTAKE24 first on each occasion, and 25% completed the interviewer-led recall first to control for order effects.
• Interviewer-Led Recall Protocol: The comparison method followed the interview protocol used in the Low Income Diet and Nutrition Survey (LIDNS), which is based on the USDA AMPM method, with portion size assessment assisted by the Young Person's Food Atlas [10].
• Analysis: Daily energy and nutrient intakes reported in INTAKE24 were compared to those from interviewer-led recalls using ratio analysis and limits of agreement.

INTAKE24 Doubly Labeled Water Validation Protocol

The criterion validity study employed the following methodology [23]:

• Participants: 98 UK adults (50 men and 48 women) aged 40-65 years across a wide BMI range recruited from the Fenland Study.
• DLW Administration: Participants received a body weight-specific dose of DLW (deuterium oxide-18) and collected daily urine samples for 9-10 days. Total energy expenditure was calculated using the method of Schoeller with a fixed space ratio of 1.0316.
• Intake24 Administration: Participants completed Intake24 at least twice during the DLW measurement period.
• Laboratory Analysis: Urine samples were analyzed in duplicate for 18O enrichment using the CO2 equilibration method of Roether and for 2H enrichment using isotope ratio mass spectrometry.
• Statistical Analysis: The method of Bland and Altman was used to examine accuracy and precision of reported energy intake against measured TEE. Correlation analysis quantified the ability of the instrument to rank individuals.

Diagram 1: INTAKE24 DLW Validation Workflow

ASA24 Supplement Reporting Comparison Protocol

The FORCS study methodology for comparing supplement reporting included [13]:

• Participants: 1,076 men and women from three integrated health care systems in the U.S., with quota sampling to ensure balance of sex, age, and race/ethnicity.
• Study Design: Participants were randomly assigned to one of four groups completing two 24-hour recalls: two ASA24s; two AMPM interviews; ASA24 first then AMPM; or AMPM first then ASA24.
• Supplement Coding: Reported dietary supplements were coded using the 2007-2008 NHANES Dietary Supplement Database (NHANES-DSD). For ASA24, "unfound supplements" were reviewed and matched to NHANES-DSD codes where possible.
• Statistical Analysis: The two one-sided test was used to assess equivalence of reported supplement use between methods, with standard errors estimated using delete-one jackknife procedure.

Database Management and Technical Considerations

Effective database management in dietary assessment tools requires addressing several technical challenges that impact data quality and usability.

Food Matching and Search Functionality: Both ASA24 and INTAKE24 utilize free-text search for food entry, but usability studies indicate this presents significant challenges. ASA24 demonstrated particular difficulties in low-income populations, where participants struggled with the search function and frequently misentered dietary information [55]. This suggests database search algorithms and user interfaces require careful optimization for different population groups.

Portion Size Estimation: The approach to portion size estimation represents a critical database management consideration. INTAKE24's extensive library of over 3,000 food photographs developed from UK National Diet and Nutrition Survey data represents a significant investment in this aspect of database management [10]. The system's validation against weighed food diaries suggests this approach enhances accuracy [23].

International Adaptation: Database management must accommodate regional food patterns and nutrient composition differences. INTAKE24 has been adapted for use in Portugal, Denmark, New Zealand, and the United Arab Emirates, with versions for India and Australia under development [23]. This requires substantial database modification to include local foods and reflect regional nutrient composition differences.

Dietary Supplement Integration: Comprehensive dietary assessment requires effective integration of supplement databases. ASA24's linkage to the NHANES Dietary Supplement Database demonstrates one approach to this challenge [13]. The equivalence in supplement reporting between ASA24 and interviewer-administered recalls (46% vs. 43%) suggests this integration can be effective [13].

Diagram 2: Database Management Architecture

Table 3: Research Reagent Solutions for Dietary Assessment Studies

Tool/Resource	Function	Application Context
Doubly Labeled Water (DLW)	Criterion method for measuring total energy expenditure in free-living individuals [23]	Validation studies to assess energy intake under-reporting
Young Person's Food Atlas	Standardized portion size estimation using food photographs [10]	Interviewer-led 24-hour recalls, particularly with younger populations
NHANES Dietary Supplement Database (NHANES-DSD)	Comprehensive coding system for dietary supplements [13]	Standardized coding of supplement data in 24-hour recalls
USDA Automated Multiple-Pass Method (AMPM)	Interviewer-administered 24-hour recall protocol [10] [13]	Gold standard comparison method for validation studies
Food and Nutrient Database for Dietary Studies (FNDDS)	USDA nutrient composition database [52]	Primary nutrient database for ASA24 and NHANES
NDNS Nutrient Databank	UK nutrient composition database [10]	Primary nutrient database for INTAKE24

The comparative analysis of ASA24, INTAKE24, and myfood24 reveals that while all three tools leverage technology to improve dietary assessment efficiency, their database management approaches yield important differences in performance, validity, and appropriate application contexts. INTAKE24 demonstrates strong convergent validity against interviewer-led recalls and moderate criterion validity against doubly labeled water, though with significant under-reporting similar to traditional methods. ASA24 shows equivalence to interviewer-administered recalls for supplement reporting but faces substantial usability challenges in vulnerable populations. The comprehensiveness of each tool's underlying nutrient database, portion size estimation methods, and search functionality significantly influence their effectiveness in research settings. Researchers should select tools based on alignment with their target population's characteristics, regional dietary patterns, and specific nutrient focus, while continuing to account for the systematic under-reporting inherent in all self-reported dietary assessment methods.

Evidence-Based Performance Review Against Biomarkers and Gold Standards

Validating dietary assessment tools is fundamental to ensuring the accuracy of nutritional research. The validation hierarchy progresses from relative comparisons against established methods like interviewer-led recalls to criterion validation against objective, non-invasive biomarkers such as doubly labelled water (DLW). This guide objectively compares the validation evidence for three prominent online 24-hour dietary recall tools: ASA24, Intake24, and myfood24. These systems were developed to reduce the cost and participant burden associated with traditional dietary assessment while maintaining data quality. Their validation strategies provide a framework for understanding their relative strengths and limitations in capturing energy and nutrient intake. The convergence of evidence from different validation tiers is essential for researchers to select the most appropriate tool for their specific study objectives and population.

Comparative Validity Data

The validity of online tools is assessed through two primary approaches: relative validity against interviewer-led recalls and criterion validity against objective biomarkers. The data below summarize key performance metrics from multiple studies.

Table 1: Relative Validity Compared to Interviewer-Led 24-Hour Recalls

Tool	Study Population	Mean Energy Intake Difference	Correlation/Agreement	Key Findings
Intake24	180 participants, aged 11-24 [10] [28]	-1% (underestimate)	Limits of agreement: -49% to +92%	Mean intakes for energy and most nutrients were within 4% of interviewer-led recall.
myfood24	UK adults [12]	~10-20% lower for most nutrients	Intraclass Correlation Coefficient (ICC): ~0.4-0.5	Produced lower estimates than interviewer-led recalls but with consistent moderate agreement.

Table 2: Criterion Validity Against Doubly Labelled Water (DLW) and Other Biomarkers

Tool	Biomarker	Study Population	Mean Energy/Nutrient Intake Difference	Correlation with Biomarker
Intake24	DLW (Energy)	98 UK adults, 40-65 years [23]	-25% (underestimate vs. TEE*)	0.31 (single recall), 0.47 (mean of two recalls)
ASA24	DLW (Water)	1082 US adults, 50-74 years [56]	-18% to -31% (water intake)	0.46 (single recall), 0.58 (mean of six recalls)
myfood24	Urinary Nitrogen (Protein) & Potassium	212 UK adults [12]	Protein: -10% vs. biomarker [12]	Protein: ~0.3-0.4 (partial correlation) [12]
myfood24	Urinary Potassium	German adults [57]	Potassium: No significant difference vs. biomarker [57]	Potassium: Moderate agreement (pc=0.44) [57]
myfood24	Serum Folate	Danish adults [22]	N/A	Strong correlation (ρ=0.62) with total folate intake [22]

*TEE: Total Energy Expenditure

Experimental Protocols for Key Validation Studies

The quantitative data presented above are derived from rigorous experimental protocols. Understanding these methodologies is critical for interpreting the results and assessing the quality of the validation evidence.

Criterion Validation with Doubly Labelled Water

The doubly labelled water (DLW) method is considered the reference standard for estimating free-living total energy expenditure (TEE) and is used to validate reported energy intake (EI) under the assumption of energy balance [23]. The protocol for validating Intake24 is representative of this high-standard approach [23]:

• Participant Cohort: Ninety-eight UK adults (40-65 years) across a range of BMI categories were recruited from the Fenland Study.
• DLW Administration: Participants ingested a weight-specific dose of DLW (D₂¹⁸O). Urine samples were collected for 9-10 days post-dose. Samples were analyzed for ¹⁸O and ²H enrichment using isotope ratio mass spectrometry, and TEE was calculated.
• Dietary Assessment: Participants completed the Intake24 online recall at least twice during the DLW measurement period.
• Data Analysis: The accuracy of reported EI was assessed against TEE using Bland-Altman analysis for mean bias (accuracy) and limits of agreement (precision). Correlation analyses quantified the tool's ability to rank individuals correctly.

Relative Validation against Interviewer-Led Recalls

Relative validation studies compare the test method against an established, albeit still subjective, dietary assessment method. The protocol for the Intake24 comparison study illustrates this design [10] [28]:

• Participant Cohort: One hundred and eighty participants aged 11-24 years.
• Study Design: A repeated-measures cross-sectional study where each participant completed both an Intake24 recall and an interviewer-led multiple-pass 24-hour recall on the same day. This was repeated on four separate occasions over one month.
• Methodology Control: A weighted randomization ensured 75% of participants completed Intake24 first to avoid testing the online tool in an unnatural context (after an interview). This also provided a methodological check on order effects.
• Reference Method: The interviewer-led recalls followed a standardized protocol based on the UK Low Income Diet and Nutrition Survey, which itself is derived from the USDA Automated Multiple-Pass Method.
• Data Analysis: Mean intakes of energy and nutrients from both methods were compared. Paired t-tests, correlation coefficients, and Bland-Altman limits of agreement were used to assess agreement.

Validation against Urinary and Blood Biomarkers

Biomarkers like urinary nitrogen (protein) and potassium offer objective measures for validating specific nutrient intakes. The validation of myfood24-Germany exemplifies this protocol [57]:

• Participant Cohort: Ninety-seven German adults.
• Study Design: Participants completed a 3-day weighed dietary record (WDR) with a 24-hour urine collection on the third day. This was followed by at least one myfood24 recall for the same day as the WDR/day 3.
• Biomarker Collection and Analysis: A 24-hour urine sample was collected. Completeness was verified via volume and self-reported collection time.
  • Urinary Nitrogen: Measured in duplicate by the Dumas method and used to calculate protein intake (using a conversion factor of 6.25 and assuming 80% of dietary nitrogen is excreted).
  • Urinary Potassium: Measured by atomic absorption spectroscopy to estimate potassium intake (assuming 80% excretion).
• Data Analysis: Protein and potassium intakes from myfood24 and the WDR were compared to the biomarker estimates using correlation and concordance statistics.

Signaling Pathways and Workflows

The following diagram illustrates the hierarchical relationship between different types of validation studies, from relative to criterion, and the objective measures at the apex.

The Scientist's Toolkit: Key Reagents and Materials

Successful execution of a dietary validation study requires specific reagents and materials. The table below details essential items used in the protocols cited above.

Table 3: Essential Research Reagents and Materials for Dietary Validation Studies

Item Name	Function in Validation	Specific Examples from Research
Doubly Labelled Water (DLW)	A stable isotope tracer (²H₂¹⁸O) used to measure total energy expenditure (TEE) in free-living individuals, serving as a criterion measure for validating reported energy intake.	Deuterium oxide-18 (H₂¹⁸O) and ²H₂O [23].
Isotope Ratio Mass Spectrometer (IRMS)	The analytical instrument used to measure the enrichment of ¹⁸O and ²H in urine samples after DLW administration. It provides the precise data needed to calculate TEE.	AP2003 (Analytical Precision Ltd) and Isoprime (GV Instruments) [23].
Validated Food Photograph Atlas	A set of standardized food portion images used within online tools to assist users in estimating the volume or weight of consumed foods, improving portion size accuracy.	The Young Persons Food Atlas [10]; over 3000 food photographs in Intake24 [10].
24-Hour Urine Collection Kit	Materials provided to participants for the complete collection of all urine over a 24-hour period, which is necessary for biomarker analysis (e.g., nitrogen, potassium).	Includes collection containers, storage bottles, cooling elements, and a detailed protocol form [57] [22].
Online Dietary Recall System	The self-administered tool being validated. It includes a searchable food database, portion size estimation aids, and automated nutrient coding.	ASA24, Intake24, myfood24 [23] [56] [12].
Web-Based Kitchen Scale	Provided to participants to weigh all foods and beverages consumed during a weighed food record, which serves as a high-quality reference method.	Scales provided to participants in the myfood24 Denmark validation [22].

Accurate dietary assessment is fundamental for epidemiological research, clinical studies, and public health monitoring. For decades, researcher-administered 24-hour recalls, particularly the USDA's Automated Multiple-Pass Method (AMPM), have served as a reference standard despite their resource-intensive nature. The National Cancer Institute's Automated Self-Administered 24-hour Dietary Assessment Tool (ASA24) was developed to overcome these limitations by providing a freely available, web-based system for collecting high-quality dietary data [3]. This guide objectively evaluates ASA24's performance against interviewer-administered AMPM and biomarker-based validation studies, with comparative data against other automated systems like myfood24, to inform researchers and health professionals about its appropriate applications and limitations.

Criterion Validity Against True Intake

Table 1: ASA24 Performance Compared to True Intake in a Controlled Feeding Study [25]

Performance Metric	ASA24	Interviewer AMPM	P-value
Match Rate (Items consumed that were reported)	80%	83%	0.07
Intrusion Rate (Items reported but not consumed)	Significantly higher	Lower	<0.01
Energy intake accuracy	No significant difference from true intake	No significant difference from true intake	NS
Nutrient intake accuracy	No significant difference from true intake	No significant difference from true intake	NS
Portion size accuracy	No significant difference from true intake	No significant difference from true intake	NS

Comparative Validity Against Biomarkers

Table 2: ASA24 Correlation with Recovery Biomarkers [58] [56]

Biomarker	ASA24 Correlation (Single)	ASA24 Correlation (Multiple)	Comparison Method	Correlation with Biomarker
Protein (Urinary Nitrogen)	-	0.46 [58]	FFQ	0.46 [58]
Potassium (Urinary)	-	0.42-0.44 [58]	FFQ	0.42-0.48 [58]
Total Water (Doubly Labeled Water)	0.46 [56]	0.58 (6 recalls) [56]	FFQ	0.48 [56]
Energy (Doubly Labeled Water)	Attenuation factor: 0.28 [56]	Attenuation factor: 0.43 (6 recalls) [56]	4-day Food Record	Attenuation factor: 0.32-0.39 [56]

Cross-Platform Comparison: ASA24 vs. myfood24

Table 3: Automated Dietary Tool Performance Comparison [59] [60]

Performance Metric	ASA24	myfood24
Correlation with biomarkers (protein, potassium, sodium)	0.46, 0.42-0.44, 0.42-0.48 [58]	0.3-0.4 [59]
Energy intake attenuation	18-31% underestimation vs. DLW [56]	"Broadly similar" to interviewer-administered tool [59]
Usability in older adults	Rated less easy than voice-based tool (6.7/10) [61]	-
Folate validity correlation	-	0.62 (Danish population) [60]

Key Experimental Protocols

Feeding Study Design (True Intake Validation)

The most rigorous validation of ASA24 against known true intake was conducted through a feeding study with the following protocol [25]:

• Participants: 81 adults
• True Intake Measurement: Foods and beverages offered at a buffet were inconspicuously weighed before and after each participant served themselves. Plate waste was measured to determine exact consumption.
• Recall Methods: Participants were randomly assigned to complete either an ASA24 or an interviewer-administered AMPM recall the following day.
• Analysis: Linear and Poisson regression analyses examined associations between recall mode and match rates, exclusions, intrusions, and differences between true and reported intakes for energy, nutrients, food groups, and portion sizes.

Biomarker Validation Protocol

The Interactive Diet and Activity Tracking in AARP (IDATA) study validated ASA24 against objective biomarkers with this methodology [62] [56]:

• Participants: 1,082 men and women aged 50-74
• ASA24 Administration: Participants completed 6 unannounced ASA24 recalls (ASA24-2011 version) over 12 months.
• Biomarker Collection:
  • Doubly labeled water (DLW) to assess total energy expenditure and water intake
  • 24-hour urine collections to measure protein (via urinary nitrogen), potassium, and sodium excretion
• Comparison Methods: Food Frequency Questionnaires (DHQ II) and 4-day food records were also collected for comparison.
• Analysis: Geometric means of self-reported intakes were compared with biomarker values. Attenuation factors and correlation coefficients were calculated to assess measurement error.

Mode Comparison Study

A large-scale study compared ASA24-2011 directly with interviewer-administered AMPM [62]:

• Design: Healthy individuals from diverse geographic regions completed both ASA24 and interviewer-administered recalls.
• Measures: Nutrient, food group, and supplement intake estimates were compared between methods.
• Participant Experience: Response rates and preferences were assessed.
• Analysis: Comparability of reported intakes and participant preferences between the two modes.

Visualization of Validation Pathways

ASA24 Validation Pathways Diagram. This flowchart illustrates the three primary methodological approaches used to validate ASA24 against criterion standards (true intake), objective biomarkers, and the established interviewer-administered AMPM.

The Researcher's Toolkit

Table 4: Essential Research Reagents and Tools for Dietary Validation Studies

Tool/Reagent	Function in Validation	Example Application
Doubly Labeled Water (DLW)	Objective measure of total energy expenditure and water intake [56]	Criterion validation for reported energy and water intake [56]
24-Hour Urine Collection	Recovery biomarkers for protein (nitrogen), potassium, and sodium [58]	Reference method for validating self-reported nutrient intakes [58]
Standardized Buffet Protocol	Controlled setting to establish true intake [25]	Feeding studies to assess reporting accuracy against known consumption [25]
ASA24-2024	Latest version with current food/nutrient databases [63]	Most up-to-date dietary assessment with mobile capability [63]
Biomarker Database	Biological sample repository for validation	Longitudinal assessment of diet-disease relationships [58]

ASA24 demonstrates substantial criterion validity against true intake measures, performs comparably to interviewer-administered AMPM for most nutrients, and shows similar correlation patterns with recovery biomarkers as other self-report instruments. While some limitations exist—including slightly lower match rates compared to AMPM and systematic underestimation of energy and water intake—the tool offers researchers a scientifically valid, cost-effective alternative to traditional dietary assessment methods. The substantial feasibility advantages of ASA24, including its automated coding, 24/7 availability, and scalability, make it particularly valuable for large-scale studies where multiple dietary assessments are needed to estimate usual intake. Future developments should focus on improving portion size estimation and addressing systematic underreporting common to all self-reported dietary assessment tools.

For researchers and drug development professionals, selecting a robust dietary assessment tool is critical for obtaining reliable nutrient intake data in clinical and population studies. Automated 24-hour dietary recall tools have emerged as cost-effective alternatives to resource-intensive interviewer-led methods. Among the leading tools are ASA24, Intake24, and MyFood24, which are often evaluated for their validity and practicality. This guide focuses on the accuracy of Intake24, particularly its performance in estimating macronutrient intake compared to traditional interviewer-led recalls, and situates these findings within a comparative framework of the three major platforms.

Quantitative Comparison of Validation Data

The following table summarizes key validation findings for Intake24 and comparable data for ASA24 and MyFood24, providing a direct comparison of their performance against established methods.

Table 1: Summary of Validation Studies for Automated Dietary Recall Tools

Tool	Comparison Method	Sample Size & Population	Key Finding: Energy Intake	Key Finding: Macronutrients
Intake24	Interviewer-led multiple-pass 24-hr recall [35] [28]	180 participants, aged 11-24 years [35]	Underestimated by 1% on average [35] [28]	Mean intakes of all macronutrients were within 4% of the interviewer-led recall [35] [28]
ASA24	Controlled feeding study (True intake known) [36]	81 adults [36]	Underestimated true intake by approximately 6% [36]	Performance data against a true measure is available [36]
MyFood24	Interviewer-led 24-hr recall [36]	111 participants, aged 11-18 years [36]	Not specified in available data	Not specified in available data

Detailed Experimental Protocols

Intake24 Validation Protocol

The validation study for Intake24 was designed as a relative comparison against an established interviewer-led method. The detailed protocol underscores the rigor of the evaluation [35].

Table 2: Key Components of the Intake24 Validation Study Protocol

Component	Description
Study Design	Cross-sectional comparison where each participant completed both an Intake24 recall and an interviewer-led recall on the same day. This was repeated on four separate days over one month [35].
Participant Recruitment	180 participants aged 11-24 years were recruited, with quotas to ensure an even distribution of age, gender, economic status, and ethnicity [35].
Comparison Method	The interviewer-led recalls followed the multiple-pass protocol used in the UK Low Income Diet and Nutrition Survey (LIDNS), a method itself validated against direct observation and doubly labeled water [35].
Portion Size Estimation	In Intake24, participants estimated portion sizes using over 3,000 validated food photographs based on UK National Diet and Nutrition Survey data. In the interviewer-led recall, the Young Persons Food Atlas was used [35].
Order Randomization	A weighted randomization was applied: 75% of participants completed Intake24 first, while 25% completed the interviewer-led recall first. This design tested Intake24 under realistic conditions (completed first) and controlled for the potential learning effect of completing one method on the other [35] [28].

Tool Selection and Evaluation Framework

A 2024 review outlined a formal process for selecting a 24-hour recall tool for a national nutrition survey, providing a structured framework for comparing tools like ASA24, Intake24, and MyFood24. The evaluation was based on pre-selection criteria, with each tool scored as follows [36]:

Table 3: Tool Evaluation Scores from a 2024 Systematic Review

Criterion	ASA24	Intake24	MyFood24
Evidence of Validation	2 points	2 points	1 point
Use in National Surveys	2 points	2 points	1 point
Availability	1 point	1 point	1 point
English Language	1 point	1 point	1 point
Portion Size Images	1 point	1 point	1 point
Data Coding Capabilities	2 points	2 points	2 points
Adaptability	0 points	2 points	2 points
Automatic Link to Food Data	2 points	2 points	2 points
Total Score	9/10	10/10	9/10

Visualizing the Validation and Selection Workflow

The diagram below illustrates the logical pathway and key decision points in the tool validation and selection process, as derived from the studies on Intake24 and the comparative review.

The Scientist's Toolkit: Essential Research Reagents and Materials

The following table details key components used in the validation and adaptation of automated dietary recall tools like Intake24.

Table 4: Essential Materials for Dietary Recall Validation Studies

Item	Function in Research
Validated Portion Size Photographs	Aids participants in estimating the amount of food consumed. The images in Intake24 are based on national survey data and are critical for converting food selections into quantitative data [35] [64].
Food Composition Database	A nutrient lookup table that automatically converts reported foods and portions into energy and nutrient intakes. Tools must be adapted to use country-specific databases (e.g., from the UK or New Zealand) [35] [64].
Multiple-Pass Recall Protocol	A structured interview technique designed to enhance memory and reduce forgetting. It forms the methodological backbone of both interviewer-led and automated recalls like Intake24 and ASA24 [35].
Localized Food List	A comprehensive database of foods, including local and traditional items, brand names, and common synonyms. This is essential for accurate food matching and must be adapted for different countries or cultures [4] [64].
Usability Testing Framework	A mixed-methods approach (e.g., screen recordings, think-aloud protocols, surveys) to identify user challenges and improve the tool's interface and functionality before full-scale deployment [64].

The evidence demonstrates that Intake24 provides a highly accurate estimate for macronutrient intake, showing close agreement with traditional interviewer-led recalls. This validation, combined with its high adaptability and open-source nature, makes it a robust choice for large-scale nutritional research and national surveys. When compared directly with ASA24 and MyFood24 in a structured evaluation framework, Intake24 consistently ranks highly, particularly for its proven use in national surveys and its flexibility for localization. For researchers requiring a validated, cost-effective dietary assessment tool, Intake24 represents a compelling solution with documented accuracy comparable to more resource-intensive methods.

Accurate dietary assessment is fundamental for exploring the relationship between nutrition and health. Traditional methods like weighed food records (WFR) are considered a reference but are administratively burdensome. The emergence of web-based, self-administered 24-hour dietary recalls (24HDRs) like myfood24 offers a promising alternative for large-scale studies, promising reduced cost, time, and participant burden [65] [66]. However, the validity of these tools must be rigorously tested. This guide objectively evaluates the performance of myfood24 by examining its agreement with traditional weighed food records and objective urinary biomarkers, situating its validity within the broader context of automated dietary assessment tools.

Experimental Protocols: How Validity is Assessed

Validation studies for dietary assessment tools typically employ a cross-sectional or repeated-measures design, comparing the test method (e.g., myfood24) against a reference method. The key protocols are outlined below.

Method Comparison: Versus Weighed Food Records

This approach compares the intake of energy and nutrients estimated by myfood24 with those from a WFR, which is traditionally considered a "gold standard" in dietary assessment.

• Procedure: Participants are instructed to complete both methods for the same intake day. In the German validation study, participants kept a 3-day paper-based WDR, meticulously weighing and recording all consumed foods and beverages. On the third day, they also completed a myfood24 24HDR for the same 24-hour period [65] [67]. The WDRs were then manually coded by trained staff using a specialized nutrient database.

Biomarker Comparison: Versus Objective Biological Measures

This method provides an objective, non-self-report measure of true intake for specific nutrients. It helps overcome biases inherent in all self-reported dietary data.

• Procedure: Participants collect a complete 24-hour urine sample, which is analyzed for biomarkers such as nitrogen (a proxy for protein intake) and potassium. In the German and Danish studies, participants collected urine on the final day of their food recording [65] [22]. Protein intake from the biomarker was calculated from urinary nitrogen using a standard conversion factor (assuming 80% of dietary nitrogen is excreted in urine and a nitrogen-to-protein conversion factor of 6.25) [67]. Potassium intake was estimated directly from urinary potassium excretion, also assuming an 80% excretion rate [67].

Table 1: Key Research Reagents and Materials for Validation Studies

Item	Function in Validation Research
Kitchen Scale	Used by participants in a Weighed Food Record (WFR) to accurately measure the weight of all consumed food and drink items [22].
24-Hour Urine Collection Containers & Protocols	Essential for the complete collection of all urine over a 24-hour period for subsequent biomarker analysis (e.g., nitrogen, potassium, creatinine) [67].
Atomic Absorption Spectrometer	Analytical instrument used to measure the concentration of potassium in urine samples with high precision [67].
Food Composition Database (FCDB)	The nutrient calculation engine behind any dietary tool. myfood24 adaptations use localized FCDBs (e.g., German BLS, UK Composition of Foods) to convert reported foods into nutrient estimates [67] [66].

Validity and Reproducibility Metrics

Statistical analyses are used to quantify the agreement between methods.

• Correlation Analysis: Spearman's rank correlation coefficients (ρ) determine the strength of the relationship for ranking individuals by intake.
• Concordance Analysis: Concordance correlation coefficients (pc) and weighted Kappa coefficients (κ) assess the overall agreement, accounting for both precision and bias.
• Mean Difference Tests: Paired t-tests check for systematic over- or under-reporting by the test method.

The following diagram illustrates the standard workflow for a myfood24 validation study integrating both method and biomarker comparisons.

Results: Quantitative Performance of MyFood24

The validity of myfood24 has been evaluated in several populations, with results demonstrating its utility as a dietary assessment tool for research.

Agreement with Weighed Food Records

The method comparison shows that myfood24 produces nutrient estimates that are broadly comparable to those from traditional WFRs.

• German Adult Study: Significant correlations were found for energy and all 32 tested nutrients, with coefficients ranging from 0.45 to 0.87 [65] [67]. There was no significant difference in the mean intake of energy and macronutrients. However, myfood24 underestimated the mean intake of 15 out of 32 nutrients, primarily micronutrients [65].

• Danish Adult Study (2025): The reproducibility of myfood24 was strong for most nutrients and food groups when repeated after 4 weeks. The highest correlations were observed for folate (ρ = 0.84) and total vegetable intake (ρ = 0.78) [22] [68].

Table 2: myfood24 Performance Against Weighed Food Records and Biomarkers

Comparison & Population	Key Metric	Findings for myfood24
vs. Weighed Food Record (German Adults) [65] [67]	Nutrient Correlation Range	0.45 - 0.87 for energy and 32 nutrients
	Mean Energy & Macronutrient Intake	No significant difference from WFR
	Systematic Error	Underestimated mean intake of 15 nutrients
vs. Urinary Biomarkers (German Adults) [65] [67]	Protein Intake (vs. Urinary Nitrogen)	10% lower than biomarker; Good agreement (pc=0.58, κ=0.51)
	Potassium Intake (vs. Urinary Potassium)	No significant mean difference; Moderate agreement (pc=0.44, κ=0.30)
vs. Urinary Biomarkers (UK Adults) [66]	Protein, Potassium, Sodium	Attenuation factors of ~0.2-0.3; Ranking correlations of ~0.3-0.4
Reproducibility (Danish Adults) [22] [68]	Correlation after 4 weeks (ρ)	Folate: 0.84; Total Vegetables: 0.78; Most nutrients ≥0.50

Agreement with Urinary Biomarkers

When compared against objective biomarkers, myfood24 demonstrates a level of validity that is comparable to, and sometimes better than, other self-report methods.

• Protein Validation: In the German study, protein intake reported by myfood24 was on average 10% lower than the intake estimated from the urinary nitrogen biomarker. Despite this mean underestimation, the statistical agreement was classified as "good" (pc=0.58) [65] [67]. A UK study similarly found that myfood24 performed on par with an interviewer-administered 24HDR when compared to protein biomarkers [66].

• Potassium Validation: The German study found no significant difference in the mean potassium intake assessed by myfood24 and the urinary potassium biomarker, with agreement classified as "moderate" (pc=0.44) [65]. Both the WFR and myfood24 showed a similar shared bias in assessing potassium, suggesting a limitation possibly rooted in the food composition database rather than the tool itself [67].

• Overall Biomarker Performance: A UK validation study concluded that while both myfood24 and interviewer-led recalls resulted in attenuation (underestimation of true effect sizes) compared to biomarkers, the online tool was comparable to the more costly and burdensome traditional method across a range of measures [66].

Comparative Context: MyFood24 in the Landscape of Automated Tools

Framing myfood24's validity within a broader comparison to other major automated tools like ASA24 and Intake24 provides crucial context for researchers.

• Validity Benchmark: The ASA24 tool has been validated in feeding studies where true intake is known. One such study found that ASA24 respondents reported 80% of items truly consumed, slightly lower than the 83% reported with an interviewer-administered recall, but with little difference in the accuracy of energy and nutrient estimates [25]. This demonstrates a level of criterion validity similar to what has been established for myfood24 against biomarkers and WFRs.

• Distinguishing Features: While both are web-based 24HDRs, myfood24 often highlights its validation against nutrient biomarkers in multiple countries [42] [66]. In contrast, a key feature of the ASA24 is that it is freely available to researchers and is built on the USDA's Automated Multiple-Pass Method [3]. Intake24, used in the UK National Diet and Nutrition Survey, is another tool validated for use in adolescents and adults.

The following diagram summarizes the logical relationship of conclusions derived from the validation evidence for myfood24.

The body of evidence supports the conclusion that myfood24 is a valid tool for dietary assessment in research settings. Its performance in method comparisons shows it can generate nutrient data that is consistent with that from weighed food records, particularly for energy and macronutrients. Its agreement with urinary biomarkers for protein and potassium confirms that it can objectively rank individuals' intake with good to moderate accuracy, which is sufficient for many epidemiological studies focused on relative comparisons rather than absolute intake [22] [68].

A key strength of myfood24 is its reproducibility and demonstrated feasibility for repeated administration, which is crucial for estimating usual intake in large-scale studies [22]. Furthermore, its performance has been shown to be comparable to interviewer-administered 24-hour recalls [66], offering a substantial advantage in terms of cost and administrative burden.

However, researchers must be aware of its limitations. Like all self-reported dietary tools, myfood24 is subject to systematic error, including the underestimation of certain nutrient intakes and a general attenuation versus biomarkers [65] [66]. The validity of each nationally adapted version must be assessed independently due to changes in the underlying food composition database [22].

In the context of comparative validation with ASA24 and Intake24, myfood24 holds a strong position, with multiple robust biomarker-validated studies supporting its use. The choice between these tools will ultimately depend on the specific research needs, including target population, required nutrients, and budgetary constraints.

Accurate dietary assessment is fundamental to advancing our understanding of the links between nutrition, health, and disease. For researchers and professionals in drug development and public health, selecting the most appropriate dietary intake tool is a critical decision that directly impacts data quality, validity, and resource allocation. Traditional methods, such as interviewer-administered 24-hour recalls, are robust but often prohibitively costly and time-consuming for large-scale studies. The development of automated, self-administered 24-hour dietary recall (24HR) tools presents a promising solution, offering scalability and reduced administrative burden [12] [69].

Among the most prominent tools in this space are ASA24 (from the US National Cancer Institute), Intake24 (originally from Newcastle University, UK), and MyFood24 (from the University of Leeds, UK). While all three are web-based and designed to automate the 24-hour recall process, they differ in their development history, validation evidence, technical features, and suitability for specific populations. This guide provides an objective, evidence-based comparison of these three tools, synthesizing data from validation studies, usability research, and expert evaluations to empower researchers in making an informed selection for their specific scientific context.

The validity and reliability of any dietary assessment tool are paramount. The following table summarizes key quantitative findings and validation evidence for ASA24, Intake24, and MyFood24 from peer-reviewed studies.

Table 1: Summary of Validation and Performance Evidence for ASA24, Intake24, and MyFood24

Tool & Origin	Comparison Method	Key Findings on Nutrient Intake	Usability & Target Population
ASA24 (US, NIH/NCI) [3] [16]	Interviewer-administered AMPM [16]	Equivalent reported energy intake; proportions of supplement use equivalent (46% vs 43%) to AMPM [16]. Attenuation factors of ~0.2-0.3 for protein, potassium, sodium vs biomarkers, similar to interviewer-based tool [12].	Used in >1,000 studies as of 2025 [3]. Appropriate for those with ≥5th grade reading level; parents can report for children [3].
Intake24 (UK, Newcastle University) [35]	Interviewer-led 24HR (LIDNS method) [35]	Underestimated energy intake by 1% on average vs interviewer-led recall; limits of agreement: -49% to +93%. Most macronutrient and micronutrient intakes within 4% of interviewer-led recall [35].	Developed and tested for ages 11-24 [35]. An iterative, user-centered design process focused on usability for younger populations [35].
MyFood24 (UK, University of Leeds) [6] [12]	Biomarkers & Interviewer-led 24HR [12]	Attenuation factors ~0.2-0.3 and partial correlations ~0.3-0.4 against biomarkers, broadly similar to an interviewer-based tool [12]. Nutrient estimates ~10-20% lower than interviewer-based tool [12].	Mean System Usability Scale (SUS) score of 55.5 (indicating "less than satisfactory" usability) in older adults (60-74 years) without guidance; underreporting was observed [6].

Detailed Experimental Protocols and Methodologies

To critically appraise the evidence, it is essential to understand the methodologies of the key studies cited.

Protocol: Validation of ASA24 against the Interviewer-Administered AMPM

A primary study comparing ASA24 to the well-established Automated Multiple-Pass Method (AMPM) was the Food Reporting Comparison Study (FORCS) [16].

• Objective: To compare reported dietary supplement use and nutrient intakes between the self-administered ASA24-2011 and the interviewer-administered AMPM.
• Design: A randomized study where participants were assigned to one of four groups, each completing two 24-hour recalls: two ASA24s, two AMPMs, ASA24 then AMPM, or AMPM then ASA24.
• Participants: 1,076 men and women from three integrated healthcare systems in the US, with quota sampling to ensure diversity in age, sex, and race/ethnicity.
• Analysis: Equivalence of reported supplement use was assessed using the two one-sided tests (TOST) procedure. Nutrient intakes from foods and beverages were also compared.

This design allowed for a rigorous, head-to-head comparison of the two modes of administration while controlling for potential order effects [16].

Protocol: Validation of MyFood24 against Biomarkers and Interviewer Recalls

A comprehensive validation of MyFood24 was conducted against objective biomarkers and a traditional interviewer-led recall [12].

• Objective: To validate the MyFood24 online 24-h recall tool against biomarkers and compare its performance with an interviewer-administered multiple-pass 24-h recall (MPR).
• Design: Participants completed the MyFood24 recall, an interviewer-based MPR, and a suite of reference measures (biomarkers). These assessments were repeated on three separate occasions, approximately two weeks apart, to approximate longer-term usual intake. The order of the two recall methods was randomized.
• Participants: 212 metabolically stable adults aged 18-65.
• Biomarkers: Estimated intakes of protein, potassium, and sodium were compared with urinary biomarker concentrations. Total sugar intake was compared with a predictive biomarker, and energy intake was compared with energy expenditure measured by accelerometry and calorimetry.

This multi-faceted protocol provides a high level of evidence by benchmarking the self-report tool against objective, non-self-report measures [12].

Protocol: Usability Assessment of MyFood24 in an Older Population

A study focused specifically on the usability of the Norwegian version of MyFood24 highlights the importance of considering the target population [6].

• Objective: To examine the usability of a web-based dietary assessment tool (MyFood24) in older adults.
• Design: A cross-sectional study where participants completed a single 24-hour dietary recall using MyFood24 and then evaluated its usability.
• Participants: 60 Norwegian women and men aged 60-74 years, recruited via convenience and snowball sampling.
• Usability Measure: The System Usability Scale (SUS), a 10-item questionnaire with a 5-point Likert scale. An SUS score of ≥68 is considered satisfactory.
• Key Condition: No guidance or support was provided to complete the recall, testing the tool's intuitiveness.

This study underscores that tool performance can vary significantly across demographic groups and that pre-testing with a target population is crucial [6].

Tool Comparison: Technical and Practical Features

Beyond performance in validation studies, selection depends on practical features and adaptability.

Table 2: Comparison of Technical, Functional, and Adaptability Features

Feature	ASA24 [3] [36]	Intake24 [35] [36]	MyFood24 [12] [36]
Primary Method	Multiple-pass 24HR recall/Food record [3]	Multiple-pass 24HR recall [35]	24HR recall/Food record [12]
Food Composition Database	US Department of Agriculture (USDA) databases; Canadian and Australian versions also exist [3]	Linked to UK National Diet and Nutrition Survey (NDNS) Nutrient Databank [35]	UK Composition of Foods; extensive branded food database (>50,000 items) [12]
Portion Size Estimation	Images, household measures, etc. [3]	>3000 food photographs developed from UK NDNS data [35]	Photographic images, natural measures, household measures; >6000 common items have images [12]
Adaptability & Use in National Surveys	Used in many studies; new US versions released biennially [3] [36]	Used in UK national surveys; highly rated for adaptability in a 2024 review [36]	Multiple international versions (e.g., Germany, Norway); rated highly for adaptability [36]
Cost	Free [3]	Free or low-cost for academic use [36]	Commercial licensing model [36]

Workflow of a Typical Automated Self-Administered 24-Hour Recall

The following diagram illustrates the generalized multi-pass methodology common to tools like ASA24, Intake24, and MyFood24, which is adapted from the traditional interviewer-led Automated Multiple-Pass Method [3] [12] [35].

The Researcher's Toolkit: Essential Components for Dietary Assessment Studies

Planning a study that incorporates these tools requires consideration of several key "research reagents" or essential components.

Table 3: Essential Materials and Resources for Dietary Assessment Studies

Item	Function & Description	Considerations for Tool Selection
Food Composition Database	The underlying dataset linking consumed foods to their nutrient profiles. It is the foundation for all nutrient intake calculations [12].	Check if the tool's database aligns with your country's food supply (e.g., USDA vs. UK Composition of Foods) and if it includes branded products relevant to your population [12] [36].
Portion Size Estimation Aids	Visual or quantitative guides (e.g., photographs, household measures) that help participants estimate the volume of food consumed [12] [35].	Evaluate the quality, quantity, and cultural relevance of the image library. Assess if the tool allows for alternative estimation methods (e.g., weight entry) [35].
Participant Training & Support Materials	Resources (e.g., help videos, instruction sheets) to guide participants through the recall process, improving data accuracy and completeness [6] [12].	Determine the level of support your population needs. The absence of guidance can lead to poor usability and underreporting, especially in older adults [6].
Data Output & Management System	The backend researcher portal that provides access to cleaned, coded, and analyzed nutrient and food group data files [3].	Assess the format of the output data (e.g., raw files, pre-aggregated nutrients), ease of access, and tools for managing participant cohorts [3] [36].
Validation Evidence	Peer-reviewed studies comparing the tool to a reference method (e.g., biomarkers, interviewer recall) within a population similar to your target group [12] [16].	Do not assume validity for all populations. Scrutinize existing validation literature for your demographic of interest (e.g., children, elderly, specific ethnic groups) [6] [35].

The evidence indicates that ASA24, Intake24, and MyFood24 are all viable alternatives to costly interviewer-based recalls, yet each has distinct strengths.

• For researchers prioritizing a well-established, free tool with strong institutional backing from the NCI and extensive use in diverse studies, ASA24 is a robust choice, particularly for US-based populations. Its strong equivalence to the AMPM for supplement reporting and nutrient intake makes it a reliable option [3] [16].
• For studies focusing on younger populations (ages 11-24) or requiring a tool with a strong user-centered design ethos, Intake24 is highly suitable. Its validation against interviewer-led recalls shows good agreement for most nutrients, and its design prioritizes engagement for younger users [35] [36].
• For researchers needing a tool with a extensive branded food database and flexibility for international adaptation, MyFood24 is an excellent candidate. While its usability may require more support for older adults, its performance against biomarkers is comparable to an interviewer-led method [12] [36].

Ultimately, the final selection should be guided by a pilot test within the specific study population. This practice helps identify potential issues with usability, data completeness, and the relevance of the food list before launching the full-scale study [6].

Conclusion

The comparative analysis confirms that ASA24, Intake24, and MyFood24 are all valid alternatives to traditional, costly dietary assessment methods, each with distinct strengths. ASA24 is a robust, freely available tool with extensive validation against biomarkers. Intake24 demonstrates high usability and accuracy, particularly in younger populations. MyFood24 offers a comprehensive food database, validated in multiple countries. The choice of tool depends on specific research needs, including target population, budget, and required nutrient detail. Future directions should focus on enhancing portion size estimation, expanding databases for ethnic foods, and integrating these tools into large-scale clinical and biomedical research to better understand diet-disease relationships.

References

• 1. Relative Validation of an Artificial Intelligence–Enhanced ... [https://pmc.ncbi.nlm.nih.gov/articles/PMC9723975/]
• 2. Emerging Technologies for Investigating Food Consumer ... [https://pmc.ncbi.nlm.nih.gov/articles/PMC12620339/]
• 3. ASA24® Dietary Assessment Tool | EGRP/DCCPS/NCI/NIH [https://epi.grants.cancer.gov/asa24/]
• 4. Intake24 [https://intake24.org/]
• 5. Research [https://www.myfood24.org/research/]
• 6. Evaluation of a Web-Based Dietary Assessment Tool ... [https://pmc.ncbi.nlm.nih.gov/articles/PMC8956983/]
• 7. The Adaptation, Implementation, and Performance ... [https://www.sciencedirect.com/science/article/pii/S2475299125000022]
• 8. Comparing the Usability of the Web-Based 24-h Dietary ... [https://www.mdpi.com/2072-6643/14/21/4543]
• 9. Feasibility and usability of a voice-based dietary recall tool ... [https://pmc.ncbi.nlm.nih.gov/articles/PMC12612543/]
• 10. Comparison of INTAKE24 (an Online 24-h Dietary Recall ... [https://pmc.ncbi.nlm.nih.gov/articles/PMC4924199/]
• 11. Performance and feasibility of recalls completed using the ... [https://pmc.ncbi.nlm.nih.gov/articles/PMC7606702/]
• 12. Validity of an online 24-h recall tool (myfood24) for dietary ... [https://bmcmedicine.biomedcentral.com/articles/10.1186/s12916-018-1113-8]
• 13. Comparing reported dietary supplement intakes between ... [https://pmc.ncbi.nlm.nih.gov/articles/PMC6594012/]
• 14. Comparing self-reported energy intake using an online ... [https://www.sciencedirect.com/science/article/abs/pii/S0899900723002861]
• 15. Themes 2024- Research | Humanities 2025 Institute Research [https://www.ncl.ac.uk/nuhri/research/research-themes-2024-2025/]
• 16. Comparing Reported Dietary Supplement Intakes between ... [https://www.sciencedirect.com/science/article/abs/pii/S2212267218302545]
• 17. Expansion and Assessment of a Web-Based 24-Hour Dietary ... [https://ojphi.jmir.org/2025/1/e52380/]
• 18. NCI updates ASA24 online dietary assessment tool [https://aasm.org/nci-updates-asa24-online-dietary-assessment-tool/]
• 19. ASA24® Publications | EGRP/DCCPS/NCI/NIH [https://epi.grants.cancer.gov/asa24/resources/publications.html]
• 20. Validity of an online 24-h recall tool ( myfood ) for dietary assessment... 24 [https://biomedcentral.eu/articles/10.1186/s12916-018-1113-8]
• 21. How Accurate Is Multiple Imputation for Nutrient Intake ... [https://www.mdpi.com/2072-6643/17/15/2510]
• 22. Validity and reproducibility of a web-based dietary ... [https://pmc.ncbi.nlm.nih.gov/articles/PMC12187487/]
• 23. Validity and reliability of an online self-report 24-h dietary recall method... [https://pmc.ncbi.nlm.nih.gov/articles/PMC6722486/]
• 24. Evaluating User Experiences and Preferred Features of a ... [https://pmc.ncbi.nlm.nih.gov/articles/PMC11530719/]
• 25. Performance of the Automated Self-Administered 24-hour ... [https://www.sciencedirect.com/science/article/pii/S0002916523046919]
• 26. ASA24® Frequently Asked Questions (FAQs) [https://epi.grants.cancer.gov/asa24/resources/faq.html]
• 27. Development of a UK Online 24-h Dietary Assessment Tool [https://pmc.ncbi.nlm.nih.gov/articles/PMC4488770/]
• 28. Intake24 | Validation [https://www.intake24.com/info/validation]
• 29. Accuracy and Cost-effectiveness of Technology-Assisted ... [https://www.researchprotocols.org/2021/12/e32891/]
• 30. Validity and reliability of an online self-report 24-h dietary ... [https://www.cambridge.org/core/journals/journal-of-nutritional-science/article/validity-and-reliability-of-an-online-selfreport-24h-dietary-recall-method-intake24-a-doubly-labelled-water-study-and-repeatedmeasures-analysis/3CB8BB3B415198218F7E6B9D5AB3DA84]
• 31. Field Testing of the Use of Intake24—An Online 24-Hour ... [https://www.mdpi.com/2072-6643/10/11/1690]
• 32. The state of food composition databases: data attributes ... [https://www.frontiersin.org/journals/nutrition/articles/10.3389/fnut.2025.1552367/full]
• 33. The most comprehensive nutrient database available [https://www.myfood24.org/work-with-us/dataset/]
• 34. Data Analysis for the 2025 Guidelines Committee Report [https://www.dietaryguidelines.gov/2025-advisory-committee-report/data-analysis]
• 35. Comparison of INTAKE24 (an Online 24-h Dietary Recall ... [https://www.mdpi.com/2072-6643/8/6/358]
• 36. a review and evaluation of online 24-h dietary recall tools [https://pmc.ncbi.nlm.nih.gov/articles/PMC11705012/]
• 37. May 2025 Database Updates [https://www.myfood24.org/news/may-2025-database-updates/]
• 38. Development of a New Branded UK Food Composition ... [https://pubmed.ncbi.nlm.nih.gov/27527214/]
• 39. Methods matter for dietary supplement exposure ... [https://www.sciencedirect.com/science/article/pii/S0002916525001844]
• 40. Integrating ASA24 Into a Data-Driven Nutrition Education ... [https://www.sciencedirect.com/science/article/abs/pii/S1499404625004063]
• 41. Development [https://www.myfood24.org/research/academic-papers/development/]
• 42. myfood24 | Quick and Easy to Use Nutritional Analysis Software [https://www.myfood24.org/]
• 43. Online Journal of Public Health Informatics - Expansion and... [https://ojphi.jmir.org/2025/1/e52380]
• 44. The accuracy of portion size estimation using food images and ... [https://pmc.ncbi.nlm.nih.gov/articles/PMC9291996/]
• 45. Assessment of the Accuracy of Portion Size Reports Using ... [https://pmc.ncbi.nlm.nih.gov/articles/PMC3773715/]
• 46. Accuracy of energy and nutrient intake estimation versus ... [https://www.sciencedirect.com/science/article/pii/S0002916524004568]
• 47. Food Photography I: the perception of food portion size ... [https://www.cambridge.org/core/journals/british-journal-of-nutrition/article/food-photography-i-the-perception-of-food-portion-size-from-photographs/E6915CA4E3897F27470DB2091B8AFED6]
• 48. a UK online dietary assessments tool | BMC Nutrition [https://bmcnutr.biomedcentral.com/articles/10.1186/s40795-015-0016-8]
• 49. Development of a UK Online 24-h Dietary Assessment Tool [https://www.mdpi.com/resolver?pii=nu7064016]
• 50. Assessing the Usability of the Automated Self-Administered ... [https://www.mdpi.com/2072-6643/11/1/132]
• 51. How Accurate Is Multiple Imputation for Nutrient Intake ... [https://pmc.ncbi.nlm.nih.gov/articles/PMC12348990/]
• 52. Analysis Software for 24 Hour Dietary Recalls and Food ... [https://dietassessmentprimer.cancer.gov/learn/software-24.html]
• 53. NDSR Software Food and Nutrient Database [https://www.ncc.umn.edu/ndsr-database-page/]
• 54. (PDF) Comparison of INTAKE (an online 24hr dietary recall tool)... 24 [https://www.academia.edu/95666392/Comparison_of_INTAKE24_an_online_24hr_dietary_recall_tool_with_an_interviewer_led_24hr_recall_method_in_11_16_year_olds]
• 55. Assessing the Usability of the Automated Self-Administered ... [https://pubmed.ncbi.nlm.nih.gov/30634581/]
• 56. Validation of Total Water Intake from the Automated Self ... [https://pmc.ncbi.nlm.nih.gov/articles/PMC10613756/]
• 57. Validation of the web-based self-administered 24-h dietary ... [https://pmc.ncbi.nlm.nih.gov/articles/PMC8437850/]
• 58. Relative Validity of Nutrient Intakes Assessed by ... [https://pmc.ncbi.nlm.nih.gov/articles/PMC5928456/]
• 59. Validity of an online 24-h recall tool (myfood24) for dietary ... [https://pmc.ncbi.nlm.nih.gov/articles/PMC6083628/]
• 60. Validity and reproducibility of a web-based dietary ... [https://www.cambridge.org/core/product/0ECCD31EC8D8F53B62FBE64F45A1E906]
• 61. Feasibility and usability of a voice-based dietary recall tool ... [https://pubmed.ncbi.nlm.nih.gov/41246199/]
• 62. ASA24® Evaluation & Validation | EGRP/DCCPS/NCI/NIH [https://epi.grants.cancer.gov/asa24/respondent/validation.html]
• 63. Comparison of ASA24® Versions | EGRP/DCCPS/NCI/NIH [https://epi.grants.cancer.gov/asa24/comparison.html]
• 64. Evaluating User Experiences and Preferred Features of a Web ... [https://formative.jmir.org/2024/1/e63823]
• 65. comparison with a weighed dietary record and biomarkers [https://pubmed.ncbi.nlm.nih.gov/33974128/]
• 66. Validity of an online 24-h recall tool ( myfood )... | springermedizin.de 24 [https://www.springermedizin.de/validity-of-an-online-24-h-recall-tool-myfood24-for-dietary-asse/16023928]
• 67. comparison with a weighed dietary record and biomarkers [https://link.springer.com/article/10.1007/s00394-021-02547-7]
• 68. a cross-sectional study in an adult Danish population [https://pubmed.ncbi.nlm.nih.gov/40568623/]
• 69. Improving dietary data collection tools for better nutritional ... [https://www.sciencedirect.com/science/article/pii/S2666990022000180]