Researchers have developed and conducted preliminary validation of NutriNumerics, a specialized instrument designed to measure nutrition-related quantitative literacy and numeracy among registered dietitian nutritionists and dietetics trainees. In an evaluation published in Frontiers in Nutrition, the new 29-item assessment demonstrated high overall content validity with a scale-level index of 0.96 and strong exploratory test-retest reliability with an intraclass correlation coefficient of 0.90, despite exhibiting modest internal consistency due to restricted score variability. The multi-phase instrument development study was conducted by Luc R. LaBonte, Sarah Elliott, and Jessica Faucette from the Department of Nutrition Science at East Carolina University. By bridging applied numerical calculation and real-world clinical reasoning, the tool aims to establish an objective standard for evaluating how nutritional professionals handle complex mathematical data, interpret clinical figures, and calculate dietary interventions across various professional environments.
Quantitative competencies play an essential role in daily healthcare delivery, where practitioners must interpret numerical evidence, analyze diagnostic data, and quantify clinical risks to make accurate decisions. Within nutritional practice, registered dietitian nutritionists routinely calculate personalized caloric and macronutrient needs, convert complex enteral feeding regimens, evaluate anthropometric charts, and interpret laboratory panels. Despite the centrality of these tasks, existing literature highlighted by the study authors indicates that quantitative literacy and numeracy are frequently inadequate across the healthcare workforce. Previous investigations cited by the researchers revealed that up to twenty-five percent of medication management calculations performed by nurses carried a high risk of error, while thirty percent of medical students and residents demonstrated unsatisfactory numeracy, which was linked to a seven-fold increase in the odds of misinterpreting clinical risk. Furthermore, clinicians often experience false confidence regarding their mathematical capabilities, underscoring the necessity of objective measurement tools.
Prior to the creation of NutriNumerics, existing numeracy metrics failed to capture the specialized cognitive and computational demands encountered in dietetics. Established instruments such as the General Health Numeracy Test, the Newest Vital Sign, and the Nutrition Literacy Assessment Instrument were constructed primarily for general adult, patient, or introductory student cohorts, focusing on basic comprehension rather than professional calculation. Other prominent assessments, including the Berlin Numeracy Test and higher education quantitative reasoning inventories, measure abstract mathematical or statistical concepts without addressing nutrition-specific workflows. Even clinician-oriented measures like the Medical Interpretation and Numeracy Test and the Numeracy Understanding in Medicine Instrument do not encompass practical nutritional formulas or contextual dietary decision-making. Consequently, dietitians faced a persistent gap in validated evaluative instruments, potentially leaving unaddressed deficiencies that could impair clinical nutrient prescriptions, resource allocation, and patient care outcomes.
To address these limitations, the research team designed a multi-phase investigation comprising expert-informed item generation, content and face validity appraisals, and preliminary psychometric evaluation. Content domains were established using the 2024 Academy of Nutrition and Dietetics Scope and Standards of Practice, the 2022 Commission on Dietetic Registration examination resources, and the Accreditation Council for Education in Nutrition and Dietetics standards. The development protocol covered six defined practice domains: Leadership and Entrepreneurship, Clinical Services, Practice Management, Professional Practice and Public Policy, Science and Evidence-Based Practice, and Client Services and Community Nutrition. Panelists underwent a structured orientation distinguishing calculation-focused numeracy from contextual quantitative literacy before drafting multiple-choice questions. Drafting procedures utilized artificial intelligence-assisted language editing through ChatGPT strictly to enhance readability and grammatical clarity, with all conceptual constructs, scenarios, and substantive items created entirely by the human experts.
The content validation phase engaged an expert panel of eight practicing registered dietitian nutritionists recruited from faculty and clinicians affiliated with East Carolina University, representing a forty percent acceptance rate among twenty solicitations. Panelists were required to possess an active credential, at least five years of specialized experience in their domain, and a screening score of at least 13 out of 20 on the Numeracy Understanding in Medicine Instrument to ensure baseline competency. The panel achieved a mean screening score of 19.4 with a standard deviation of 0.52. During the initial review of thirty-nine candidate items, thirty-four met the prespecified item-level content validity index threshold of 0.78. Following revisions to eight items based on qualitative and quantitative feedback, a second review achieved complete consensus, with all thirty-nine items meeting the threshold to produce an average scale-level content validity index of 0.96 with a standard deviation of 0.07.
Face validity was subsequently evaluated by eight participants comprising undergraduate Didactic Program in Dietetics seniors and graduate students who also achieved qualifying numeracy screening scores, averaging 17.6 with a standard deviation of 2.13. Reviewers evaluated item appropriateness using binary responses, with acceptable scale performance defined as an index of at least 0.83. Six items fell below the item-level face validity threshold, prompting the exclusion of one leadership item and the textual revision of another, yielding an unadjusted scale-level face validity index of 0.93 across thirty-eight items. Although an exploratory post-hoc adjustment that removed certain negative ratings attributed to reviewer unfamiliarity increased the scale index to 0.95, the authors prioritized the unadjusted metrics. Across the thirty-eight common items administered during initial testing and a follow-up survey completed at least one week later by eight participants, test-retest reliability demonstrated an intraclass correlation coefficient of 0.88.
Subsequent psychometric evaluation of the preliminary instrument revealed critical nuances regarding response variation and internal consistency. Ten individuals completed the initial administration, including six practicing dietitians, three graduate students, and one senior, recording a mean screening score of 19.4. Initial analysis of the twenty quantitative literacy items revealed a Cronbach’s alpha of 0.40, heavily suppressed because eleven items exhibited zero response variance and four negatively influenced reliability. An exploratory subset analysis of the five items with response variance yielded an alpha of 0.69. Similarly, the initial eighteen numeracy items produced an alpha of 0.52, where fourteen items demonstrated no variance. Table 3 documented the items showing response variation, including calculations of enteral free water, annual software savings, and longitudinal weight trajectory interpretations. When restricted to the three numeracy items exhibiting variance, exploratory alpha reached 0.89, highlighting how score homogeneity in high-performing cohorts deflates traditional reliability coefficients.
Through systematic refinement balancing statistical performance against broad professional domain representation, the authors finalized NutriNumerics as a 29-item instrument comprising 12 quantitative literacy items and 17 numeracy items. Retrospective analysis of the retained items showed an average scale content validity index of 0.96 and an unadjusted scale face validity index of 0.92, with twenty-nine items meeting individual content criteria and twenty-five meeting face validity standards. The final quantitative literacy component demonstrated an internal consistency alpha of 0.60, while the numeracy component recorded an alpha of 0.63. On retrospective test-retest evaluation among the eight follow-up completers, the retained 29-item total score yielded an intraclass correlation coefficient of 0.90 with a 95 percent confidence interval from 0.56 to 0.98, indicating favorable preliminary temporal stability despite the substantial statistical uncertainty reflected in the interval width.
Participant performance on the finalized 29-item scale demonstrated high overall accuracy, with mean scores reaching 11.10 out of 12 for quantitative literacy and 16.60 out of 17 for numeracy. However, the researchers emphasized substantial methodological limitations that contextualize these preliminary metrics. The small participant sample and recruitment from a single university network constrain the broader generalizability of the findings across diverse professional backgrounds. The screening prerequisite requiring a high score on the general medicine numeracy instrument likely induced a pronounced ceiling effect by selecting individuals with superior baseline mathematical aptitude. Combined with potentially modest item difficulty, this restricted score variance directly depressed Cronbach’s alpha coefficients. Furthermore, the inclusion of the study authors in item writing and refinement introduced possible bias, and the investigation did not gather data to evaluate convergent validity against outside objective measures.
Looking ahead, the authors emphasized that NutriNumerics is not currently validated or intended for high-stakes professional decisions, such as employee hiring, credentialing, administrative promotion, or formal clinical remediation. Instead, its immediate utility lies in academic program evaluation, curriculum development, and formative continuing education to identify areas where trainees require additional mathematical support. Future empirical investigations must administer the 29-item assessment to substantially larger, geographically diverse, and academically heterogeneous cohorts to rigorously establish its dimensional structure, examine item discrimination, and verify construct validity. Longitudinal research is also recommended to determine whether NutriNumerics scores are sensitive to curricular coursework, supervised practice rotations, or targeted didactic interventions. Once its measurement properties are comprehensively verified, the tool could provide a vital framework for elevating quantitative precision, nutritional calculation safety, and evidence-based decision-making throughout clinical dietetics.
Subject of Research: Development and preliminary psychometric evaluation of the NutriNumerics assessment instrument in dietetics.
Article Title: Development and preliminary validation of NutriNumerics: assessment of nutrition-related quantitative literacy and numeracy in dietetics
Article References: LaBonte, L. R., Elliott, S., & Faucette, J. (2026). Development and preliminary validation of NutriNumerics: assessment of nutrition-related quantitative literacy and numeracy in dietetics. Frontiers in Nutrition, 13, Article 2010455. https://doi.org/10.3389/fnut.2026.2010455
Image Credits: AI Generated
DOI: 10.3389/fnut.2026.2010455
Keywords: NutriNumerics, quantitative literacy, numeracy assessment, registered dietitian nutritionists, dietetics education, content validity index, face validity index, test-retest reliability, Cronbach's alpha, intraclass correlation coefficient, Numeracy Understanding in Medicine Instrument, psychometric evaluation, enteral nutrition calculations, dietetics practice management, clinical dietetics competencies, Didactic Program in Dietetics, item-level content validity
News Source: Ophelia Keating. (November 12, 2026). NutriNumerics Tool Demonstrates Strong Preliminary Validity for Assessing Dietetic Numeracy Skills. Scienmag.
