Methodology Brief

IRB / ethics approval considerations for weighed-food and image-based dietary assessment studies

A methodology brief

Sofia Rivera, MS, RD; Helena Weiss, PhD, MPH, RD
Published September 1, 2025

Background

A weighed-food or image-based dietary assessment validation study typically involves healthy adult participants, non-invasive measurements, and foods the participant would otherwise consume. Many investigators treat such studies as low-risk or exempt, and in some jurisdictions they qualify for expedited review. Nevertheless, the combination of photographic data that can contain faces, hands, or identifying surroundings, together with linked dietary and potentially health data, raises human-subjects protections beyond those of a traditional weighed-food diary.

The Initiative’s position is that ethical review is always required, that image-data consent must be explicit and separable from overall study consent, and that retention and sharing plans must be made at protocol-writing time.

The Method

Initiative studies are required to satisfy the following before data collection begins:

Ethics approval or exemption letter. An approval or formal exemption from an appropriate ethics board (IRB, REC, or equivalent) must be in file, and the protocol number must be reported in every Initiative publication drawing on the data.

Separable image consent. The consent form includes a distinct checkbox for (a) participation in the study, (b) collection and internal use of images, and (c) release of de-identified images in a public dataset. A participant may consent to (a) and (b) but decline (c); such participants’ data remain in the study but their images are not released.

Incidental-finding policy. Weighed-food and image-based studies do not generate clinical data in the usual sense, but the diary or photographic record may reveal information of clinical interest (for example, signs of disordered eating, alcohol use patterns). The consent form states explicitly whether and under what circumstances the investigator will contact the participant or refer to a clinician. The Initiative’s default is that no unsolicited clinical contact is made, and that any participant question is directed to their usual care provider.

Data retention and deletion. The protocol specifies retention duration for identifiable data (default: 7 years from publication), de-identification procedures, and conditions under which raw images may be destroyed earlier at participant request.

Data-use agreement for third parties. If images or diary data will be shared with any external party (for example, a technology vendor whose tool is being evaluated), the consent form states this, the data-use agreement is in file, and the agreement is cited in the publication.

Special populations. Studies including minors, pregnant participants, patients with eating disorders, or other populations identified as vulnerable by the reviewing board are required to include a formal protocol addendum addressing additional protections.

Worked example

A checklist for a representative Initiative validation study might read as follows.

Of 92 enrolled participants, 84 consented to public image release; the remaining 8 are included in the study but their images will not appear in the released dataset.

Common pitfalls

• Treating image data as “just photos of food.” Many food photographs contain hands, faces reflected in surfaces, or identifying tattoos; image-data consent requires its own language.
• Bundling consent for study participation and public dataset release into a single checkbox. This practice disadvantages participants who would otherwise contribute.
• Relying on retrospective de-identification to release a dataset whose consent form did not anticipate public release.
• Failing to state a retention duration. “As long as necessary” is not an adequate statement for most ethics boards.
• Sharing images with a vendor for model training under a research consent that did not contemplate commercial use.
• Omitting the IRB approval number and ethics-board name from the published methods section.

Recommended reporting

• Name of the reviewing ethics board and approval (or exemption) number.
• Date of approval and any amendments.
• Consent-form language for image data (paraphrased or included as a supplement).
• Incidental-finding policy summary.
• Retention and deletion timeline.
• Third-party data sharing, if any, with data-use agreement summary.
• Proportion of participants consenting to each separable data-use tier.

References

1. Rivera M, Weiss R. Ethics practice in dietary assessment validation: a survey of recent studies. J Nutr. 2024;154(6):1788-1796.
2. Klausen H, Morales F. Image data in nutrition research: consent practices and gaps. Public Health Nutr. 2023;26(8):1660-1669.
3. Weiss R. Separable consent in digital health research: why bundled consent fails participants. JMIR mHealth Uhealth. 2023;11(10):e48220.
4. Ahlgren P, Rivera M. Retention and deletion rules for food-image datasets. Nutrients. 2024;16(5):705.
5. Henriksen D. Data-use agreements with technology vendors in nutrition research. Am J Clin Nutr. 2022;116(9):2410-2418.
6. Kessler F. Incidental findings in non-clinical dietary research: a policy framework. Br J Nutr. 2023;130(4):550-559.
7. Yoshida H. De-identification of food photography: a practical protocol. Public Health Nutr. 2024;27(11):2402-2411.
8. Rivera M. When is a dietary study minimal risk? J Nutr. 2025;155(2):315-319.