Truace
Subscribe

Blinded professional ratings of artificial intelligence-generated and dietitian-designed meal plans

Objectives: Artificial intelligence-assisted systems increasingly produce personalized meal plans, but their value depends on whether clinicians judge those plans credible and usable. We evaluated complete plans from a constrained artificial intelligence-assisted pipeline under blinded professional review. Methods: In a cross-sectional exploratory web study, 54 eligible anonymous evaluator records from Brazilian dietitians contributed 130 evaluations of five matched plan pairs, with one pipeline-generated and on…

By the Truvace desk2 min read1 sourcePeer-reviewedTRV-2026-1886· cite & verify
Blinded professional ratings of artificial intelligence-generated and dietitian-designed meal plans
"Health professional reviews nutrition data while preparing healthy meal plans at a modern kitchen table" by nenadstojkovicart is licensed under CC BY 2.0. To view a copy of this license, visit https://creativecommons.org/licenses/by/2.0/.

In brief

Researchers asked Brazilian dietitians to rate meal plans without knowing whether each came from an AI-assisted pipeline or a dietitian. Five personas each had one pipeline-generated plan and one dietitian-designed plan, and 54 eligible evaluator records produced 130 evaluations across nine rated items.

The result suggests AI-assisted drafting could be credible to clinicians, supporting a structured first-draft workflow where dietitians edit the output. The study was exploratory and rated plans rather than measuring patient outcomes, so it does not show that the plans improve health or that clinicians would adopt them in practice.

Main points

  1. The study was a cross-sectional exploratory web study with anonymous evaluator records from Brazilian dietitians.
  2. Mean overall ratings were 3.67 for pipeline plans and 3.11 for dietitian-designed plans, with an effect size of 0.86.
  3. Domain differences favored the pipeline by 0.76 for nutritional adequacy, 0.69 for technical quality and 0.21 for practical applicability.
  4. In an optional pairwise task, 32 of 42 eligible choices picked the pipeline-generated plan as more appropriate.

The gain

In blinded review by Brazilian dietitians, meal plans from a constrained AI-assisted pipeline scored higher on average than matched dietitian-designed plans, mainly on nutritional adequacy and technical quality.

The rundown

The design covered five matched plan pairs, one pipeline-generated and one dietitian-designed per persona. Evaluators rated nine items across three domains, and the overall score required all nine items, leaving 124 evaluations with scores.

Among 38 evaluator records with complete scores for both origins, 31 favored the pipeline plans, 6 favored dietitian plans and 1 was tied. Pipeline plans scored higher in all five persona pairs, though the authors call this descriptive because evaluator records overlapped across pairs.

The authors conclude the findings support prospective evaluation of a structured first-draft workflow, with required clinician edits and patient-relevant outcomes as endpoints.

Sources

  1. Peer-reviewedNutrition2026-10-16

The debate