TY - JOUR
T1 - Is this food healthy? Reframing nutrition evidence through counterfactual comparisons
AU - López-Moreno, Miguel
AU - López-Gil, José Francisco
N1 - Publisher Copyright:
© 2026 The Author(s). Published by Elsevier Ltd. This is an open access article under the CC BY license. http://creativecommons.org/licenses/by/4.0/
PY - 2026/6
Y1 - 2026/6
N2 - SummaryNutrition science frequently produces conclusions that appear inconsistent or context dependent, contributing to uncertainty in dietary recommendations. Many debates framed around questions such as whether a specific food is “healthy” implicitly assume that foods exert intrinsic effects independent of dietary context. However, because diet is inherently compositional, increasing intake of one food necessarily implies decreasing another, meaning that dietary effects represent substitution contrasts rather than isolated exposures. This article aims to critically evaluate how inadequate specification of counterfactual contrasts and comparators in nutrition research limits causal interpretation in evidence synthesis, and to outline methodological approaches based on causal inference and network meta-analysis (NMA) to improve interpretability. Drawing on the potential outcomes framework, we propose that meaningful causal interpretation requires explicitly defining the comparator (“in comparison with that?”) and ensuring that exposures correspond to well-defined interventions consistent with the consistency assumption. In this context, NMA offers a methodological framework that preserves comparator structure by jointly modeling multiple competing alternatives, thereby aligning more closely with the relational nature of dietary interventions and causal inference logic. Practical recommendations are provided to improve future research, including explicit specification of counterfactual contrasts, clearer definition of dietary exposures, transparent reporting of substitution context and energy balance, and appropriate use of NMA when multiple alternatives exist. We conclude that improving nutrition evidence synthesis does not require abandoning meta-analysis but rather reframing research questions and analytical strategies around clearly defined causal contrasts.
AB - SummaryNutrition science frequently produces conclusions that appear inconsistent or context dependent, contributing to uncertainty in dietary recommendations. Many debates framed around questions such as whether a specific food is “healthy” implicitly assume that foods exert intrinsic effects independent of dietary context. However, because diet is inherently compositional, increasing intake of one food necessarily implies decreasing another, meaning that dietary effects represent substitution contrasts rather than isolated exposures. This article aims to critically evaluate how inadequate specification of counterfactual contrasts and comparators in nutrition research limits causal interpretation in evidence synthesis, and to outline methodological approaches based on causal inference and network meta-analysis (NMA) to improve interpretability. Drawing on the potential outcomes framework, we propose that meaningful causal interpretation requires explicitly defining the comparator (“in comparison with that?”) and ensuring that exposures correspond to well-defined interventions consistent with the consistency assumption. In this context, NMA offers a methodological framework that preserves comparator structure by jointly modeling multiple competing alternatives, thereby aligning more closely with the relational nature of dietary interventions and causal inference logic. Practical recommendations are provided to improve future research, including explicit specification of counterfactual contrasts, clearer definition of dietary exposures, transparent reporting of substitution context and energy balance, and appropriate use of NMA when multiple alternatives exist. We conclude that improving nutrition evidence synthesis does not require abandoning meta-analysis but rather reframing research questions and analytical strategies around clearly defined causal contrasts.
KW - Causal inference
KW - Counterfactual framework
KW - Dietary substitution
KW - Food comparisons
KW - Network meta-analysis
KW - Randomized controlled trials
UR - https://www.scopus.com/pages/publications/105035495527
U2 - 10.1016/j.clnu.2026.106655
DO - 10.1016/j.clnu.2026.106655
M3 - Artículo
AN - SCOPUS:105035495527
SN - 0261-5614
VL - 61
JO - Clinical Nutrition
JF - Clinical Nutrition
M1 - 106655
ER -