Smartphone App for Real-Time Dietary Tracking and Nutrient Analytics

Thesis Abstract

In an era of rising non-communicable disease prevalence and fragmented dietary reporting, there remains a critical gap between real-time dietary intake data and actionable nutrient analytics that can inform individualized nutrition interventions. This study investigates the development and evaluation of a smartphone application designed to enable real-time dietary tracking and immediate nutrient analytics to support informed dietary decisions among adults. The aim is to assess the app’s usability, accuracy of nutrient estimation, and its effects on short-term dietary behavior and knowledge. Specific objectives include (1) to evaluate the concordance between user-recorded intake captured by the app and corresponding 24-hour dietary recalls; (2) to assess the app’s accuracy in calculating macro- and micronutrient intakes against standardized food composition databases; (3) to determine the impact of real-time feedback on portion-size estimation and adherence to individualized dietary goals; (4) to explore user experience, engagement, and perceived usefulness through qualitative insights; and (5) to examine potential moderating effects of health literacy and prior nutrition knowledge on app effectiveness. The methodology adopts a mixed-methods sequential explanatory design. A pilot phase will develop and refine the app’s user interface and nutrient database in collaboration with a multidisciplinary team including dietitians, computer scientists, and behavioral scientists. The main study will recruit 320 adults aged 18–65 from urban and suburban communities, sampled through stratified random sampling to ensure representation across gender, SES, and BMI categories. Quantitative data will be collected over a four-week period, including app-recorded dietary data, three 24-hour recalls at baseline and end of study, anthropometric measures, and validated questionnaires on health literacy (Newest Vital Sign) and nutrition knowledge (General Nutrition Knowledge Questionnaire). A subset of 40 participants will participate in semi-structured interviews to elucidate user experience, barriers, and facilitators of engagement. Data collection instruments include the smartphone app with integrated barcode scanning, portion-size visuals, a nutrient analytics dashboard, a push-notification system for real-time feedback, and data extraction templates for recalls and questionnaires. Validity and reliability will be established through cross-validation of app-estimated nutrient intakes against multiple-pass 24-hour recalls (n=960 recalls) and a test-retest assessment for the nutrient database’s estimates. Descriptive statistics will summarize engagement metrics, while Bland-Altman analysis and intraclass correlation coefficients will evaluate agreement between app data and recalls. Regression analyses will examine predictors of adherence to dietary goals and changes in nutrient adequacy. A linear mixed-effects model will assess longitudinal changes in macronutrient and micronutrient intakes, adjusting for covariates such as age, sex, BMI, health literacy, and nutrition knowledge. Thematic analysis will be applied to the qualitative interview data to identify recurring themes related to usability, perceived accuracy, and behavioral impact. Expected findings include (i) moderate to strong concordance between app-recorded nutrient estimates and recalls for energy intake and macronutrients (ICC > 0.70; Bland-Altman within acceptable limits), with variable accuracy for micronutrients dependent on database completeness; (ii) significant reductions in energy from ultra-processed foods and improvements in fruit and vegetable intake among participants receiving real-time feedback, particularly among those with higher baseline nutrition knowledge; (iii) high usability scores (SUS > 70) and positive engagement metrics, with identified barriers including initial setup complexity and data entry burden; (iv) evidence that health literacy moderates the effectiveness of feedback, with greater behavioral change observed in participants with adequate health literacy. The study contributes to knowledge by integrating real-time nutrient analytics within a user-facing mobile platform and evaluating its efficacy as a behavioral intervention tool in diverse populations. It provides empirical evidence on the feasibility of leveraging ICT-enabled dietary monitoring to improve nutrient adequacy and informs design considerations for scalable digital nutrition interventions. Practical implications include informing guidelines for clinical nutrition practice, public health nutrition programs, and the development of standardized benchmarks for mobile dietary assessment tools. The study may be limited by self-report biases in recalls and variable device adoption rates. Recommendations include iterative refinement of nutrient databases, enhancement of automatic portion-size estimation, and exploration of integration with wearable sensors to further augment accuracy and personalization.