Editorial
Research Methodology
How we design surveys, run cohort analyses, and evaluate nutrition apps. Methodology choices are restated on every research piece.
Practitioner surveys
Survey research uses a recruitment frame of credentialed Registered Dietitians sourced through professional networks and dietetic practice groups (DPGs). Instruments are pilot-tested with a small panel of dietitians representative of the survey's target specialty. We report sample size, response rate, region distribution, and years-in-practice distribution on every survey.
Where surveys include free-text responses, two reviewers independently code responses; we report inter-rater agreement (Cohen's kappa) and disclose any disagreements that required adjudication.
Cohort research
Cohort analyses are observational and follow self-selected populations who consent to anonymized data use. We disclose the sampling frame, the period of observation, the inclusion and exclusion criteria, and analytic decisions including how we handle missing data, dropout, and confounding.
Where we draw on third-party datasets (e.g., the Dietary Assessment Initiative cohort), the dataset is cited with version and access date. We do not claim causal inference from observational data.
App evaluation criteria
Nutrition apps are evaluated on six clinically relevant domains:
- Database accuracy — proportion of generic and branded items resolving to validated entries; presence of laboratory-tested values.
- Macronutrient transparency — visibility of macronutrient breakdown including fiber, fat subtypes where claimed, and sodium.
- Photo-AI validation — vendor's published mean absolute percentage error (MAPE) on portion estimation, with whether the figure is independently replicated.
- Clinical workflow fit — exportability to common EHRs, support for renal/diabetes/CKD presets, ability for a clinician to view patient data.
- Privacy posture — published privacy policy, HIPAA-readiness for clinical use, third-party data sharing.
- Replicability of vendor claims — whether accuracy and outcome claims published by the vendor are reproducible by independent testers.
We test on current iOS and Android builds at time of review. App version numbers and test dates are listed on every review.
Citation policy
Primary sources (peer-reviewed nutrition, endocrinology, and sports-medicine journals) are cited first. Where we draw on industry-published methodology (e.g., Foodvision-Bench, vendor white papers), we label it as such. Position papers from the Academy of Nutrition and Dietetics and other professional societies are cited as professional society guidance.