Reliability Issues of Mobile Nutrition Apps for Cardiovascular Disease Prevention: Comparative Study
Bibliographic record
Abstract
Background: Controlling saturated fat and cholesterol intake is important for the prevention of cardiovascular diseases. Although the use of mobile diet-tracking apps has been increasing, the reliability of nutrition apps in tracking saturated fats and cholesterol across different nations remains underexplored. Objective: This study aimed to examine the reliability and consistency of nutrition apps focusing on saturated fat and cholesterol intake across different national contexts. The study focused on 3 key concerns: data omission, inconsistency (variability) of saturated fat and cholesterol values within an app, and the reliability of commercial apps across different national contexts. Methods: Nutrient data from 4 consumer-grade apps (COFIT, MyFitnessPal-Chinese, MyFitnessPal-English, and LoseIt!) and an academic app (Formosa FoodApp) were compared against 2 national reference databases (US Department of Agriculture [USDA]-Food and Nutrient Database for Dietary Studies [FNDDS] and Taiwan Food Composition Database [FCD]). Percentages of missing nutrients were recorded, and coefficients of variation were used to compute data inconsistencies. One-way ANOVAs were used to examine differences among apps, and paired 2-tailed t tests were used to compare the apps to national reference data. The reliability across different national contexts was investigated by comparing the Chinese and English versions of MyFitnessPal with the USDA-FNDDS and Taiwan FCD. Results: Across the 5 apps, 836 food codes from 42 items were analyzed. Four apps, including COFIT, MyFitnessPal-Chinese, MyFitnessPal-English, and LoseIt!, significantly underestimated saturated fats, with errors ranging from -13.8% to -40.3% (all P<.05). All apps underestimated cholesterol, with errors ranging from -26.3% to -60.3% (all P<.05). COFIT omitted 47% of saturated fat data, and MyFitnessPal-Chinese missed 62% of cholesterol data. The coefficients of variation of beef, chicken, and seafood ranged from 78% to 145%, from 74% to 112%, and from 97% to 124% across MyFitnessPal-Chinese, MyFitnessPal-English, and LoseIt!, respectively, indicating a high variability in saturated fats across different food groups. Similarly, cholesterol variability was consistently high in dairy (71%-118%) and prepackaged foods (84%-118%) across all selected apps. When examining the reliability of MyFitnessPal across different national contexts, errors in MyFitnessPal were consistent across different national FCDs (USDA-FNDSS and Taiwan FCD). Regardless of the FCDs used as a reference, these errors persisted to be statistically significant, indicating that the app's core database is the source of the problems rather than just mismatches or variances in external FCDs. Conclusions: The findings reveal substantial inaccuracies and inconsistencies in diet-tracking apps' reporting of saturated fats and cholesterol. These issues raise concerns for the effectiveness of using consumer-grade nutrition apps in cardiovascular disease prevention across different national contexts and within the apps themselves.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.096 | 0.304 |
| Meta-epidemiology (narrow) | 0.001 | 0.001 |
| Meta-epidemiology (broad) | 0.001 | 0.002 |
| Bibliometrics | 0.003 | 0.003 |
| Science and technology studies | 0.001 | 0.002 |
| Scholarly communication | 0.003 | 0.003 |
| Open science | 0.002 | 0.003 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.002 | 0.001 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".