Patient‑Reported Outcome Measures for Food Preference in Adults: A COSMIN‑Based Systematic Review Authors: Yun-Xia Wang, Ting Liu, Zhi-Shui Wu, Yin-Ying Zhang, Jian-Zheng Cai Published: 2026 ABSTRACT BackgroundFood preference is a key determinant of eating behavior and dietary choice in adults. Reliable and valid assessment tools are essential for tailoring dietary interventions, increasing adherence, and improving clinical and nutritional outcomes.ObjectiveTo systematically identify and evaluate the psychometric properties of patient-reported outcome measures (PROMs) assessing food preference in adults, specifically focusing on the evidence provided in their initial development and validation reports.DesignSystematic review of measurement properties conducted in accordance with the COSMIN guidelines.MethodsWe searched PubMed, Embase, Web of Science, and CINAHL for relevant literature published from database inception through September 2, 2025. Studies were included if they reported the initial development, validation, or cross-cultural adaptation of a self-reported food preference PROM in adults. Methodological quality and measurement properties were assessed using the COSMIN Risk of Bias checklist and COSMIN criteria for good measurement properties (version 2.0). Certainty of evidence was graded using a modified GRADE approach.ResultsThirty unique PROMs were identified. Five instruments—Preferences and Self-Efficacy of Diet and Physical Activity Behaviors Questionnaire for Latina Women (PSEDPALW), Brazilian version of the Food Choice Questionnaire (FCQ-BR), Food Liking Questionnaire (FLQ), the Food Liking Questionnaire in a French-Canadian Population (FLQ-FC), and Food Choice Determinants Questionnaire (FCDQ)—demonstrated sufficient content validity and internal consistency with moderate-to-high certainty of evidence and were classified as Class A recommendations. No included study reported evidence for measurement error, cross-cultural validity/measurement invariance, or responsiveness.ConclusionFCDQ, FCQ-BR, FLQ, FLQ-FC, and PSEDPALW were reliable tools for assessing food preference in adults. While initial development efforts often establish robust internal structure, properties such as responsiveness and qualitative transparency in item generation are frequently overlooked during a tool’s inception. Future research should ensure that the initial stages of PROM development adhere more strictly to COSMIN’s qualitative standards and longitudinal validation protocols. Introduction Food preference is a complex behavioral construct defined as an individual’s consistent inclination toward, or aversion to, specific dietary targets. Conceptually, it encompasses both the hedonic evaluation of a food item (“liking”) and the motivational drive to consume it (“wanting”).1 Unlike transient hunger, food preference functions as a stable, trait-like determinant of dietary intake.2,3 In clinical populations, particularly among adults managing obesity or metabolic disorders, measuring these preferences is critical. Consequently, it is essential to accurately profile an individual’s food preferences for designing personalized dietary interventions, improving adherence to nutritional therapy, and predicting long-term clinical outcomes. Given the subjective nature of food preference, patient-reported outcome measures (PROMs) are the primary mechanism for assessment. PROMs capture the patient’s direct perception of their eating behavior and offer insights that objective biomarkers or observational methods cannot provide, particularly in resource-limited clinical settings.4 These instruments allow clinicians to tailor dietary counseling to the patient’s intrinsic motivations, thus enhancing the quality of care. However, the current landscape of assessment tools is fragmented. There are various PROMs for food preference, including the food choice questionnaire (FCQ),5 the food preference questionnaire (FPQ),6 the food craving acceptance and action questionnaire (FAAQ),7 the food choice determinants questionnaire (FCDQ),8 etc. Despite the abundance, these tools vary significantly in theoretical underpinning, scope, and developmental rigor. While many of them have demonstrated utility in specific clinical or research scenarios, few have undergone systematic validation across diverse populations or cultural contexts. This heterogeneity presents a significant challenge: without a standardized evaluation of these tools, researchers and clinicians risk selecting instruments with unknown or poor psychometric properties, potentially compromising data integrity and the comparability of study results. A dedicated audit of existing instrumentations is required to address this knowledge gap. The consensus-based standards for the selection of health measurement instruments (COSMIN) provide a robust methodological framework and offer specific guidelines for rigorously evaluating the methodological quality and psychometric properties of PROMs.9 There is currently no comprehensive review that uses up-to-date standards to synthesize the psychometric quality of food preference PROMs. Therefore, in this work, we systematically reviewed the methodological quality and measurement properties of currently available PROMs for assessing food preference in adults. By utilizing the COSMIN methodology, we aim to provide evidence-based recommendations for research and clinical practice. Methods This systematic review was conducted in strict accordance with the PRISMA-COSMIN guideline for systematic reviews of outcome measurement instruments.9 Table S1 provides the PRISMA checklist. The review protocol was prospectively registered with PROSPERO (Registration Number: CRD420251138897). Search Strategy A comprehensive systematic search was executed across four major electronic databases: PubMed, Embase, Web of Science, and CINAHL. The time span was database inception through September 2, 2025. The search strategy utilized a combination of controlled vocabulary (MeSH terms/headings) and free-text keywords (Table S2). The search syntax combined constructs related to the domain (eg, “food preference*”, “dietary preference*”, “taste preference*”, “food liking”, “taste liking”) with filter terms for measurement tools (eg, “questionnaire*”, “scale*”, “instrument*”, “inventory”, “measure*”). Inclusion and Exclusion Criteria of Studies Eligibility was determined based on the following inclusion criteria: (1) the study population consisted of adults (≥18 years); (2) the primary aim was the development, validation, or cross-cultural adaptation of an instrument; (3) the study evaluated at least one psychometric property related to food preference; and (4) the instrument was a self-reported scale. Studies were excluded if (1) they were editorials, conference abstracts, or secondary literature such as systematic reviews; (2) they did not report on dietary or food preferences as a primary construct; and (3) the administered instruments were not self-reported (eg, clinician-administered scales, objective behavioral tasks, or biomarker-based assessments). To focus on unique instrumentation and its cross-cultural utility, we retained only the original development studies and subsequent cross-cultural adaptations (translations and cultural validations) and excluded subsequent psychometric evaluations of an existing instrument within the same cultural and linguistic context. In addition, studies were excluded if they utilized a PROM solely as an outcome measure without reporting new psychometric validation data. Literature Screening and Data Extraction Retrieved records were imported into EndNote X7 (Clarivate; London, UK) and duplicates were removed algorithmically. Literature screening and data extraction were performed independently by two researchers trained in evidence-based methodology and the COSMIN guidelines. The process began with a title and abstract screening to eliminate irrelevant records, followed by a full-text review of potentially eligible articles. To ensure coverage, the reference lists of all eligible studies were also manually screened. The data from the included studies were extracted using a standardized form to capture study characteristics (first author, publication year, country), PROM specifics (name, language, target population, sample size), and structural details (number of items, domains, scoring format, range, and completion time). Disagreements at any stage were resolved through consensus or adjudication by a third reviewer. Quality Assessment The quality assessment followed the three-step COSMIN methodology (http://www.cosmin.nl), which comprised the evaluation of methodological quality, the assessment of measurement properties, and the grading of evidence quality. Two evaluators independently conducted the appraisal, and discrepancies were resolved through discussion or third-party consultation. Assessment of Methodological Quality The methodological rigor of each study was evaluated using the COSMIN Risk of Bias checklist.10 This tool assesses ten measurement properties: (1) PROM development, (2) content validity, (3) structural validity, (4) internal consistency, (5) cross-cultural validity/measurement invariance, (6) reliability, (7) measurement error, (8) criterion validity, (9) hypotheses testing for construct validity, and (10) responsiveness. Each property was rated on a four-point scale: “very good”, “adequate”, “doubtful”, or “inadequate”. According to the “worst score counts” principle, the overall methodological quality rating for each property was determined by the lowest score assigned to any item within that box. Assessment of Measurement Properties The psychometric performance of each PROM was evaluated against the COSMIN criteria for good measurement properties (version 2.0).11 Each property was rated as “sufficient (+)”, “insufficient (−)”, or “indeterminate (?)” based on the reported statistical results. Ratings were first assigned at the individual study level and then qualitatively synthesized for each PROM. If findings were consistent across studies, the overall rating reflected the individual ratings. In cases of inconsistency, potential sources of heterogeneity were explored. If heterogeneity could be explained, the results were stratified; otherwise, the overall rating was determined by the majority finding (≥75% agreement) or labeled as inconsistent (±). Assessment of Evidence Quality The certainty of evidence for each measurement property was graded using a modified grading of recommendations assessment, development and evaluation (GRADE) approach.12 Two reviewers independently rated the evidence quality as “high”, “moderate”, “low”, or “very low”. The starting point for evidence quality was “high”, which was subsequently downgraded based on the presence of risk of bias, inconsistency, imprecision, or indirectness. For internal consistency, the initial evidence level was anchored to the structural validity evidence. Measurement properties with an “indeterminate” rating were not graded. Final recommendations for PROMs were classified into three categories: Class A: PROMs with “sufficient (+)” content validity and at least “low” evidence quality for internal consistency.Class C: PROMs with high quality evidence for an “insufficient (−)” measurement property.Class B: PROMs not meeting the criteria for Class A or Class C and thus requiring further validation. Class A: PROMs with “sufficient (+)” content validity and at least “low” evidence quality for internal consistency. Class A: PROMs with “sufficient (+)” content validity and at least “low” evidence quality for internal consistency. Class C: PROMs with high quality evidence for an “insufficient (−)” measurement property. Class C: PROMs with high quality evidence for an “insufficient (−)” measurement property. Class B: PROMs not meeting the criteria for Class A or Class C and thus requiring further validation. Class B: PROMs not meeting the criteria for Class A or Class C and thus requiring further validation. Results Search Results The initial database search yielded 1314 records, and 1262 unique articles were screened after duplicate removal. After a rigorous preliminary review of titles and abstracts, followed by a full-text assessment against the eligibility criteria, 30 studies5–8,13–38 were ultimately included. The PRISMA flow diagram illustrates the selection process and specific reasons for exclusion (Figure 1). Figure 1Flow Chart of Article Retrieval.The flowchart details the identification and screening process of studies from databases and registers. In the identification phase, 1314 records were identified from PubMed (49), EMBASE (1253), Web of Science (12) and CINAHL (0). Duplicate records removed before screening totaled 52. In the screening phase, 1262 records were screened based on title and abstract, with 1123 records filtered out. Reports sought for retrieval numbered 139, with none having unavailable full text. Reports assessed for eligibility also numbered 139, with 109 reports excluded for reasons such as not adults (66), wrong publication type (5), no reported dietary preferences (35) and instrument not self-reported (3). Finally, 30 studies were included in the review.A flowchart of study identification and screening process from databases and registers. Flow Chart of Article Retrieval. Study Description Table 1 summarizes the characteristics of the included studies. The datasets span the period from 1992 to 2024, and most PROMs (29/30) were published in English. All included studies have been approved by an ethics review committee. Nineteen studies5–8,13,14,19–21,24,26,27,29,33–38 focused on the development and validation of new scales, and 11 studies15–18,22,23,25,28,30–32 focused on cross-cultural adaptation. Geographically, research output was concentrated in the Americas (16 studies)6,7,14–17,20–22,24,26,29,31,33,35,36 and Europe (11 studies),5,13,18,19,27,28,30,32,34,37,38 and only three studies were conducted in Asia.8,23,25 Regarding response formats, seven studies13,17–19,27,34,37 employed a 0–100 scoring system and the others (23/30) employed Likert-type response formats with 4–9 levels. Table 1Characteristics of the Included StudiesStudy†CountryPopulation†Sample Size and Subject Age§Instrument†Items†Response OptionsPsychometric Properties Measured†Ammann et al (2018)13SwitzerlandAdultsExploratory: n = 57 (31.89 ± 12.53);Confirmatory: n = 538 (45.44 ± 16.15);Test-retest: n = 226 (62.22 ± 12.40)FDPS80–100①②③⑤⑦Barnes et al (2014)14USAWomen from a large universityn = 585 (22.29 ± 6.36)FTSI15Likert 5 (1–5)①②③⑥⑦Brisbois-Clarkson et al (2009)15CanadaAdultsExploratory: n = 160 (three age groups);Confirmatory: n = 239 (three age groups)MPC-NA320–8 for each of the four categories①②③⑤⑥Carbonneau et al (2017)16CanadaAdults (18–65 years)Development: n = 31 (45.6 ± 13.9);Validation: n = 143 (47.4 ± 13.3)FLQ-FC32Likert 9 (1–9)①②③⑤⑥Carvalho-Ferreira et al (2019)17BrazilAdults (18–60 years)n = 48 (32.8 ± 1.4)LFPQ-BR160–100①②③Charlot et al (2019)18FranceAdultsn = 69 (35.8 ± 10.8)LFPQ-Short160–100①②③⑦Dalton et al (2015)19UKAdultsn = 215 (29.7 ± 10.3)CoEQ210–100①②③Deglaire et al (2012)38FranceAdultsn = 198 (39.4 ± 12.3)FLQ147Likert 5 (1–5)①②③⑤Finlayson et al (2007)37UKAdults (18–30 years)n = 60 (21.4 ± 3.3)LFPQ200–100①②Foerde et al (2018)21USAAdults (18–45 years)n = 36 (26.5 ± 6.0)FCT43Likert 5 (1–5)①②③⑤Geiselman et al (1998)6USAAdults (22–34 years)n = 9 (26.8 ± 4.1)FPQ72Likert 9 (1–9)①②③⑤Heitor et al (2019)22BrazilAdultsn = 502 (46.4 ± 19.3)FCQ-BR36Likert 5 (1–5)①②③⑤Juarascio et al (2011)7USAAdultsn = 463 (24.28 ± 6.03)FAAQ36Likert 6 (1–6)①②③⑤⑦Kamei et al (2024)23JapanAdults (20–69 years)Preliminary 1: n = 202 (20–69);Preliminary 2: n = 207 (20–69);Main: n = 1079 (20–69);Follow-up: n = 500 (20–69)J-FNS-A8Likert 7 (1–7)①②③⑤⑥Koch et al (2023)24USAHispanic breast cancer survivors, ≥21 years oldValidation: n = 118 (56.4 ± 10.5);Test-retest: n = 30 (53.3 ± 8.8)PSEDPALW104Likert 5 (1–5) andLikert 6 (1–6)①②③⑤Komatsu et al (2008)25JapanJapanese femalesn = 180 (18.6 ± 0.7)FCI-J47Likert 5 (0–4)①②③⑤⑥Ledikwe et al (2007)20USAAdults (18–45 years)Laboratory-based: n = 63 (24 ± 1);Free-living, cross-sectional n = 148 (37 ± 5);Free-living, longitudinal: n = 71 (45 ± 1)Fat PQ19TASTE/FREQ scores on Likert scale①②③⑤⑥⑦Lewis et al (2021)26USAAdults preparing to have bariatric surgeryn = 999 (43.1 ± 11.3)BSATED37Likert 5 (1–5)①②③Müller et al (2008)27GermanySample 1: Healthy adultsSample 2: StudentsSample 1: n = 231 (24.2 ± 4.7)Sample 2: n = 106 (23.0 ± 3.7)ACQ240–100①②③⑤Paupério et al (2014)28PortugalPregnant women in their final trimester and mothers of newborns in the first week of lifen = 219 (three age categories)P-FNS10Likert 7 (1–7)①②③Pliner and Hobden (1992)29CanadaAdultsItem generation: n = 27 (18–74);Initial item testing: n = 55 (18–74);Psychometric analysis: n = 210 (18–74);Validation: n = 156 (18–49)FNS10Likert 7 (1–7)①②③⑤⑥Roudsari et al (2020)8IranAdultsn = 70 (42.3 ± 10.2)FCDQ60Likert 5 (1–5)①②③Rodríguez et al (2007)30UK and SpainAdultsUK: n = 293 (20.57 ± 4.93);Spain: n = 373 (20.3 ± 2.4)FCCQ39Likert 6 (1–6)①②③Salas-García et al (2024)31MexicoAdultsn = 788 (24.7 ± 8.5)FCQ-M72Likert 4 (1–4)①②③⑤Steptoe et al (1995)5UKAdultsn = 358 (18–74)FCQ36Likert 5 (1–5)①②③⑤⑦Suzuki et al (2017)35USAAdults meeting income eligibilityn = 160 (56.9 ± 19.8)FBC22Likert 4 (1–4)①②③⑤Szakály et al (2021)32HungaryAdultsn = 500 (three age groups)H-FNS10Likert 7 (1–7)①②③Thivel et al (2022)34FranceHealthy weight adultsn = 430 (42.2 ± 12.7)LFPQ-fr360–100①②⑤Toll et al (2009)36USAAdults smoking a daily average of ≥20 cigarettes for at least one yearn = 385 (45.95 ± 11.17)QCSRF14Likert 7 (1–7)①②③⑥⑦White et al (2002)33USAAdultsPreliminary: n = 474 (20.30 ± 2.98);Primary: n = 379 (37.85 ± 13.80)FCI28Likert 7 (1–7)①②③⑤⑥⑦Notes: Instrument: ACQ, the attitudes to chocolate questionnaire; BSATED, bariatric surgical alterations in tolerability, enjoyment and cravings in the diet; CoEQ, control of eating questionnaire; FAAQ, food craving acceptance and action questionnaire; Fat PQ, fat preference questionnaire; FBC, food behavior checklist; FCCQ, the food chocolate-craving questionnaire; FCDQ, the food choice determinants questionnaire; FCI, the food craving inventory; FCI-J, food craving inventory for Japanese; FCQ, food choice questionnaire; FCQ-BR, Brazilian version scale of the food choice questionnaire; FCQ-H, the food choice questionnaire for Hungary; FCQ-M, food choice questionnaire for Mexico; FCT, food choice task; FDPS, food disgust picture scale; FLQ, the food liking questionnaire; FLQ-FC, the food liking questionnaire in a French-Canadian population; FNS, the food neophobia scale; FPQ, food preference questionnaire; FTSI, food thought suppression inventory; H-FNS, food neophobia scale for Hungary; J-FNS-A, Japanese version of alternative food neophobia scale; LFPQ, the Leeds food preference; LFPQ-BR, the Leeds food preference questionnaire in Brazil; LFPQ-fr, French version of the Leeds food preference questionnaire; LFPQ-Short, shortened version of the Leeds food preference questionnaire; MPC-NA, macronutrient preference checklist use in North America; P-FNS, the food neophobia scale for pregnant women; PSEDPALW, preferences and self-efficacy of diet and physical activity behaviors questionnaire for Latina women; QCSRF, the questionnaire on craving for sweet or rich foods. † All included articles are cross-sectional studies. By default, adults mean ≥18 years old. “Items” denote the number of items in the corresponding instrument. Measures: ① Content validity; ② Structural validity; ③ Internal consistency; ④ Cross-Cultural validity/measurement invariance; ⑤ Reliability; ⑥ Criterion validity; ⑦ Hypothesis testing. § Age is expressed as a range or mean ± standard deviation (unit: years). Table 1Characteristics of the Included StudiesStudy†CountryPopulation†Sample Size and Subject Age§Instrument†Items†Response OptionsPsychometric Properties Measured†Ammann et al (2018)13SwitzerlandAdultsExploratory: n = 57 (31.89 ± 12.53);Confirmatory: n = 538 (45.44 ± 16.15);Test-retest: n = 226 (62.22 ± 12.40)FDPS80–100①②③⑤⑦Barnes et al (2014)14USAWomen from a large universityn = 585 (22.29 ± 6.36)FTSI15Likert 5 (1–5)①②③⑥⑦Brisbois-Clarkson et al (2009)15CanadaAdultsExploratory: n = 160 (three age groups);Confirmatory: n = 239 (three age groups)MPC-NA320–8 for each of the four categories①②③⑤⑥Carbonneau et al (2017)16CanadaAdults (18–65 years)Development: n = 31 (45.6 ± 13.9);Validation: n = 143 (47.4 ± 13.3)FLQ-FC32Likert 9 (1–9)①②③⑤⑥Carvalho-Ferreira et al (2019)17BrazilAdults (18–60 years)n = 48 (32.8 ± 1.4)LFPQ-BR160–100①②③Charlot et al (2019)18FranceAdultsn = 69 (35.8 ± 10.8)LFPQ-Short160–100①②③⑦Dalton et al (2015)19UKAdultsn = 215 (29.7 ± 10.3)CoEQ210–100①②③Deglaire et al (2012)38FranceAdultsn = 198 (39.4 ± 12.3)FLQ147Likert 5 (1–5)①②③⑤Finlayson et al (2007)37UKAdults (18–30 years)n = 60 (21.4 ± 3.3)LFPQ200–100①②Foerde et al (2018)21USAAdults (18–45 years)n = 36 (26.5 ± 6.0)FCT43Likert 5 (1–5)①②③⑤Geiselman et al (1998)6USAAdults (22–34 years)n = 9 (26.8 ± 4.1)FPQ72Likert 9 (1–9)①②③⑤Heitor et al (2019)22BrazilAdultsn = 502 (46.4 ± 19.3)FCQ-BR36Likert 5 (1–5)①②③⑤Juarascio et al (2011)7USAAdultsn = 463 (24.28 ± 6.03)FAAQ36Likert 6 (1–6)①②③⑤⑦Kamei et al (2024)23JapanAdults (20–69 years)Preliminary 1: n = 202 (20–69);Preliminary 2: n = 207 (20–69);Main: n = 1079 (20–69);Follow-up: n = 500 (20–69)J-FNS-A8Likert 7 (1–7)①②③⑤⑥Koch et al (2023)24USAHispanic breast cancer survivors, ≥21 years oldValidation: n = 118 (56.4 ± 10.5);Test-retest: n = 30 (53.3 ± 8.8)PSEDPALW104Likert 5 (1–5) andLikert 6 (1–6)①②③⑤Komatsu et al (2008)25JapanJapanese femalesn = 180 (18.6 ± 0.7)FCI-J47Likert 5 (0–4)①②③⑤⑥Ledikwe et al (2007)20USAAdults (18–45 years)Laboratory-based: n = 63 (24 ± 1);Free-living, cross-sectional n = 148 (37 ± 5);Free-living, longitudinal: n = 71 (45 ± 1)Fat PQ19TASTE/FREQ scores on Likert scale①②③⑤⑥⑦Lewis et al (2021)26USAAdults preparing to have bariatric surgeryn = 999 (43.1 ± 11.3)BSATED37Likert 5 (1–5)①②③Müller et al (2008)27GermanySample 1: Healthy adultsSample 2: StudentsSample 1: n = 231 (24.2 ± 4.7)Sample 2: n = 106 (23.0 ± 3.7)ACQ240–100①②③⑤Paupério et al (2014)28PortugalPregnant women in their final trimester and mothers of newborns in the first week of lifen = 219 (three age categories)P-FNS10Likert 7 (1–7)①②③Pliner and Hobden (1992)29CanadaAdultsItem generation: n = 27 (18–74);Initial item testing: n = 55 (18–74);Psychometric analysis: n = 210 (18–74);Validation: n = 156 (18–49)FNS10Likert 7 (1–7)①②③⑤⑥Roudsari et al (2020)8IranAdultsn = 70 (42.3 ± 10.2)FCDQ60Likert 5 (1–5)①②③Rodríguez et al (2007)30UK and SpainAdultsUK: n = 293 (20.57 ± 4.93);Spain: n = 373 (20.3 ± 2.4)FCCQ39Likert 6 (1–6)①②③Salas-García et al (2024)31MexicoAdultsn = 788 (24.7 ± 8.5)FCQ-M72Likert 4 (1–4)①②③⑤Steptoe et al (1995)5UKAdultsn = 358 (18–74)FCQ36Likert 5 (1–5)①②③⑤⑦Suzuki et al (2017)35USAAdults meeting income eligibilityn = 160 (56.9 ± 19.8)FBC22Likert 4 (1–4)①②③⑤Szakály et al (2021)32HungaryAdultsn = 500 (three age groups)H-FNS10Likert 7 (1–7)①②③Thivel et al (2022)34FranceHealthy weight adultsn = 430 (42.2 ± 12.7)LFPQ-fr360–100①②⑤Toll et al (2009)36USAAdults smoking a daily average of ≥20 cigarettes for at least one yearn = 385 (45.95 ± 11.17)QCSRF14Likert 7 (1–7)①②③⑥⑦White et al (2002)33USAAdultsPreliminary: n = 474 (20.30 ± 2.98);Primary: n = 379 (37.85 ± 13.80)FCI28Likert 7 (1–7)①②③⑤⑥⑦Notes: Instrument: ACQ, the attitudes to chocolate questionnaire; BSATED, bariatric surgical alterations in tolerability, enjoyment and cravings in the diet; CoEQ, control of eating questionnaire; FAAQ, food craving acceptance and action questionnaire; Fat PQ, fat preference questionnaire; FBC, food behavior checklist; FCCQ, the food chocolate-craving questionnaire; FCDQ, the food choice determinants questionnaire; FCI, the food craving inventory; FCI-J, food craving inventory for Japanese; FCQ, food choice questionnaire; FCQ-BR, Brazilian version scale of the food choice questionnaire; FCQ-H, the food choice questionnaire for Hungary; FCQ-M, food choice questionnaire for Mexico; FCT, food choice task; FDPS, food disgust picture scale; FLQ, the food liking questionnaire; FLQ-FC, the food liking questionnaire in a French-Canadian population; FNS, the food neophobia scale; FPQ, food preference questionnaire; FTSI, food thought suppression inventory; H-FNS, food neophobia scale for Hungary; J-FNS-A, Japanese version of alternative food neophobia scale; LFPQ, the Leeds food preference; LFPQ-BR, the Leeds food preference questionnaire in Brazil; LFPQ-fr, French version of the Leeds food preference questionnaire; LFPQ-Short, shortened version of the Leeds food preference questionnaire; MPC-NA, macronutrient preference checklist use in North America; P-FNS, the food neophobia scale for pregnant women; PSEDPALW, preferences and self-efficacy of diet and physical activity behaviors questionnaire for Latina women; QCSRF, the questionnaire on craving for sweet or rich foods. † All included articles are cross-sectional studies. By default, adults mean ≥18 years old. “Items” denote the number of items in the corresponding instrument. Measures: ① Content validity; ② Structural validity; ③ Internal consistency; ④ Cross-Cultural validity/measurement invariance; ⑤ Reliability; ⑥ Criterion validity; ⑦ Hypothesis testing. § Age is expressed as a range or mean ± standard deviation (unit: years). Characteristics of the Included Studies Notes: Instrument: ACQ, the attitudes to chocolate questionnaire; BSATED, bariatric surgical alterations in tolerability, enjoyment and cravings in the diet; CoEQ, control of eating questionnaire; FAAQ, food craving acceptance and action questionnaire; Fat PQ, fat preference questionnaire; FBC, food behavior checklist; FCCQ, the food chocolate-craving questionnaire; FCDQ, the food choice determinants questionnaire; FCI, the food craving inventory; FCI-J, food craving inventory for Japanese; FCQ, food choice questionnaire; FCQ-BR, Brazilian version scale of the food choice questionnaire; FCQ-H, the food choice questionnaire for Hungary; FCQ-M, food choice questionnaire for Mexico; FCT, food choice task; FDPS, food disgust picture scale; FLQ, the food liking questionnaire; FLQ-FC, the food liking questionnaire in a French-Canadian population; FNS, the food neophobia scale; FPQ, food preference questionnaire; FTSI, food thought suppression inventory; H-FNS, food neophobia scale for Hungary; J-FNS-A, Japanese version of alternative food neophobia scale; LFPQ, the Leeds food preference; LFPQ-BR, the Leeds food preference questionnaire in Brazil; LFPQ-fr, French version of the Leeds food preference questionnaire; LFPQ-Short, shortened version of the Leeds food preference questionnaire; MPC-NA, macronutrient preference checklist use in North America; P-FNS, the food neophobia scale for pregnant women; PSEDPALW, preferences and self-efficacy of diet and physical activity behaviors questionnaire for Latina women; QCSRF, the questionnaire on craving for sweet or rich foods. † All included articles are cross-sectional studies. By default, adults mean ≥18 years old. “Items” denote the number of items in the corresponding instrument. Measures: ① Content validity; ② Structural validity; ③ Internal consistency; ④ Cross-Cultural validity/measurement invariance; ⑤ Reliability; ⑥ Criterion validity; ⑦ Hypothesis testing. § Age is expressed as a range or mean ± standard deviation (unit: years). Evaluation of Methodological Quality and Measurement Properties The methodological rigor and psychometric outcome of the 30 included studies were assessed via the COSMIN Risk of Bias checklist (Table 2 and Figure 2). Table S3 provides the detailed psychometric properties for each instrument. The scope of psychometric evaluation varied significantly, ranging from two to six properties per study, and 12 studies5,7,13–16,20,23,25,29,33,36 examined five or more properties. Content validity (n = 30) and internal consistency (n = 28) were the most frequently reported metrics. Reliability, criterion validity, and convergent validity were reported in 19, 9, and 8 studies, respectively. No data were available regarding cross-cultural validity, measurement error, or responsiveness. Table 2Methodological Quality Assessment and Psychometric Properties Rating of the Included StudiesInstrument§StudyDVLM†Content Validity†Structural Validity†Internal Consistency†CCV/MI†Reliability†Criterion Validity†Hypothesis Testing†CMPHNRLVCCMPHBMQQRMQQRMQQRMQQRMQQRMQQRACQ27Muller et al (2008)DD/+D/?D/?V−V+NANAD−NANANANABSATED26Lewis et al (2021)DD/+D/?D/?A−V+NANAV−NANANANACoEQ19Dalton et al (2015)DD/+D/?D/?D+V+NANANANAV?NANAFAAQ7Juarascio et al (2011)DD/+D/?D/?A?V+NANAV+NANAV+Fat PQ20Ledikwe et al (2007)DD/+D/?D/?I?D?NANAD+V−V+FBC35Suzuki et al (2017)DD/?D/?D/?D?V+NANAV+NANANANAFCCQ30Rodríguez et al (2007)DD/+D/?D/?V+V+NANANANANANANANAFCDQ8Roudsari et al (2020)AA e/+D p/+A e/+D p/+A e/+D p/+A−V+NANANANANANANANAFCI33White et al (2002)DD/+D/?D/?V+V+NANAV+V+V+FCI-J25Komatsu et al (2008)DD/+D/?D/?V+V+NANAD+V−NANAFCQ5Steptoe et al (1995)DD/+D/?D/?V+V+NANAV+NANAV+FCQ-BR22Heitor et al (2019)AA/+A/+A/+V+V+NANAV+NANANANAFCQ-M31Salas-García et al (2024)DD/+D/?D/?V+V+NANAV−NANANANAFCT21Foerde et al (2018)DD/+D/?D/?I?I?NANAV+NANANANAFDPS13Ammann et al (2018)DD/+D/+D/?V+V+NANAD+NANAV+FLQ38Deglaire et al (2012)DA e,p/+A e,p/+A e,p/+V+V+NANAV+NANANANAFLQ-FC16Carbonneau et al (2017)DA e,p/+A e,p/+A e,p/+A?V+NANAV+V−NANAFNS29Pliner and Hobden (1992)DD/+D/?D/?I?V+NANAD+V−NANAFPQ6Geiselman et al (1998)DD/+D/?D/?I?I?NANAD+NANANANAFTSI14Barnes et al (2014)DD/+D/?D/?V+V+NANANANAV?V+H-FNS32Szakály et al (2021)DD/+D/?D/?V+V+NANANANANANANANAJ-FNS-A23Kamei et al (2024)DD/+D/?D/?V+V+NANAD+V+NANALFPQ37Finlayson et al (2007)DD/?D/?D/?I?I?NANANANANANANANALFPQ-BR17Carvalho-Ferreira et al (2019)DD/+D/?D/?I?I?NANANANANANANANALFPQ-fr34Thivel et al (2022)DD/+D/?D/?I?I?NANAV?NANANANALFPQ-Short18Charlot et al (2019)DD/+D/?D/?I?I?NANANANANANAV+MPC-NA15Brisbois-Clarkson et al (2009)DD/+D/?D/?I?V+NANAD+V−NANAP-FNS28Paupério et al (2014)DD/+D/?D/?V+V+NANANANANANANANAPSEDPALW24Koch et al (2023)DA e,p/+A e,p/+A e,p/+I?V+NANAV+NANANANAQCSRF36Toll et al (2009)DD/?D/?D/?D?V+NANANANAV?V+Notes: § Instrument: ACQ, the attitudes to chocolate questionnaire; BSATED, bariatric surgical alterations in tolerability, enjoyment and cravings in the diet; CoEQ, control of eating questionnaire; FAAQ, food craving acceptance and action questionnaire; Fat PQ, fat preference questionnaire; FBC, food behavior checklist; FCCQ, the food chocolate-craving questionnaire; FCDQ, the food choice determinants questionnaire; FCI, the food craving inventory; FCI-J, food craving inventory for Japanese; FCQ, food choice questionnaire; FCQ-BR, Brazilian version scale of the food choice questionnaire; FCQ-H, the food choice questionnaire for Hungary; FCQ-M, food choice questionnaire for Mexico; FCT, food choice task; FDPS, food disgust picture scale; FLQ, the food liking questionnaire; FLQ-FC, the food liking questionnaire in a French-Canadian population; FNS, the food neophobia scale; FPQ, food preference questionnaire; FTSI, food thought suppression inventory; H-FNS, food neophobia scale for Hungary; J-FNS-A, Japanese version of alternative food neophobia scale; LFPQ, the Leeds food preference; LFPQ-BR, the Leeds food preference questionnaire in Brazil; LFPQ-fr, French version of the Leeds food preference questionnaire; LFPQ-Short, shortened version of the Leeds food preference questionnaire; MPC-NA, macronutrient preference checklist use in North America; P-FNS, the food neophobia scale for pregnant women; PSEDPALW, preferences and self-efficacy of diet and physical activity behaviors questionnaire for Latina women; QCSRF, the questionnaire on craving for sweet or rich foods. † DVLM, PROM development; CCV/MI, cross-cultural validity/measurement invariance; CPHSN, comprehensiveness, superscripts e and p denote asking expert and patients, respectively; RLVC, relevance; CPHSB, comprehensibility; MQ, methodological quality, rated as V (very good); A (adequate); D (doubtful), or I (inadequate); QR, quality of the results rated, given as + (sufficient), − (insufficient), ? (indeterminate), or NA (not available). Figure 2COSMIN Classification of the Criteria for Good Measurement Properties.A stacked horizontal bar graph with categories on the vertical axis: PROM development, Content validity, Structural validity, Internal consistency, Cross-Cultural validity, Measurement invariance, Reliability, Criterion validity, Hypothesis testing, Overall. Horizontal axis label, Percentage left parenthesis percent right parenthesis, range 0 to 100, with tick labels 0, 20, 40, 60, 80, 100. Legend lists five rating levels: Very Good, Adequate, Doubtful, Inadequate, Not Available. PROM development: small Adequate segment then Doubtful to 100. Content validity: Adequate to about 15, then Doubtful to 100. Structural validity: Very Good to about 45, Adequate to about 58, Doubtful to about 68, Inadequate to 100. Internal consistency: Very Good to about 80, Doubtful to about 83, Inadequate to 100. Cross-Cultural validity: Not Available to 100. Measurement invariance: Not Available to 100. Reliability: Very Good to about 40, Doubtful to about 66, Not Available to 100. Criterion validity: Very Good to about 33, Not Available to 100. Hypothesis testing: Very Good to about 26, Not Available to 100. Overall: Very Good to about 24, Adequate to about 28, Doubtful to about 52, Inadequate to about 58, Not Available to 100.A stacked horizontal bar graph showing COSMIN ratings across measurement properties and overall. Table 2Methodological Quality Assessment and Psychometric Properties Rating of the Included StudiesInstrument§StudyDVLM†Content Validity†Structural Validity†Internal Consistency†CCV/MI†Reliability†Criterion Validity†Hypothesis Testing†CMPHNRLVCCMPHBMQQRMQQRMQQRMQQRMQQRMQQRACQ27Muller et al (2008)DD/+D/?D/?V−V+NANAD−NANANANABSATED26Lewis et al (2021)DD/+D/?D/?A−V+NANAV−NANANANACoEQ19Dalton et al (2015)DD/+D/?D/?D+V+NANANANAV?NANAFAAQ7Juarascio et al (2011)DD/+D/?D/?A?V+NANAV+NANAV+Fat PQ20Ledikwe et al (2007)DD/+D/?D/?I?D?NANAD+V−V+FBC35Suzuki et al (2017)DD/?D/?D/?D?V+NANAV+NANANANAFCCQ30Rodríguez et al (2007)DD/+D/?D/?V+V+NANANANANANANANAFCDQ8Roudsari et al (2020)AA e/+D p/+A e/+D p/+A e/+D p/+A−V+NANANANANANANANAFCI33White et al (2002)DD/+D/?D/?V+V+NANAV+V+V+FCI-J25Komatsu et al (2008)DD/+D/?D/?V+V+NANAD+V−NANAFCQ5Steptoe et al (1995)DD/+D/?D/?V+V+NANAV+NANAV+FCQ-BR22Heitor et al (2019)AA/+A/+A/+V+V+NANAV+NANANANAFCQ-M31Salas-García et al (2024)DD/+D/?D/?V+V+NANAV−NANANANAFCT21Foerde et al (2018)DD/+D/?D/?I?I?NANAV+NANANANAFDPS13Ammann et al (2018)DD/+D/+D/?V+V+NANAD+NANAV+FLQ38Deglaire et al (2012)DA e,p/+A e,p/+A e,p/+V+V+NANAV+NANANANAFLQ-FC16Carbonneau et al (2017)DA e,p/+A e,p/+A e,p/+A?V+NANAV+V−NANAFNS29Pliner and Hobden (1992)DD/+D/?D/?I?V+NANAD+V−NANAFPQ6Geiselman et al (1998)DD/+D/?D/?I?I?NANAD+NANANANAFTSI14Barnes et al (2014)DD/+D/?D/?V+V+NANANANAV?V+H-FNS32Szakály et al (2021)DD/+D/?D/?V+V+NANANANANANANANAJ-FNS-A23Kamei et al (2024)DD/+D/?D/?V+V+NANAD+V+NANALFPQ37Finlayson et al (2007)DD/?D/?D/?I?I?NANANANANANANANALFPQ-BR17Carvalho-Ferreira et al (2019)DD/+D/?D/?I?I?NANANANANANANANALFPQ-fr34Thivel et al (2022)DD/+D/?D/?I?I?NANAV?NANANANALFPQ-Short18Charlot et al (2019)DD/+D/?D/?I?I?NANANANANANAV+MPC-NA15Brisbois-Clarkson et al (2009)DD/+D/?D/?I?V+NANAD+V−NANAP-FNS28Paupério et al (2014)DD/+D/?D/?V+V+NANANANANANANANAPSEDPALW24Koch et al (2023)DA e,p/+A e,p/+A e,p/+I?V+NANAV+NANANANAQCSRF36Toll et al (2009)DD/?D/?D/?D?V+NANANANAV?V+Notes: § Instrument: ACQ, the attitudes to chocolate questionnaire; BSATED, bariatric surgical alterations in tolerability, enjoyment and cravings in the diet; CoEQ, control of eating questionnaire; FAAQ, food craving acceptance and action questionnaire; Fat PQ, fat preference questionnaire; FBC, food behavior checklist; FCCQ, the food chocolate-craving questionnaire; FCDQ, the food choice determinants questionnaire; FCI, the food craving inventory; FCI-J, food craving inventory for Japanese; FCQ, food choice questionnaire; FCQ-BR, Brazilian version scale of the food choice questionnaire; FCQ-H, the food choice questionnaire for Hungary; FCQ-M, food choice questionnaire for Mexico; FCT, food choice task; FDPS, food disgust picture scale; FLQ, the food liking questionnaire; FLQ-FC, the food liking questionnaire in a French-Canadian population; FNS, the food neophobia scale; FPQ, food preference questionnaire; FTSI, food thought suppression inventory; H-FNS, food neophobia scale for Hungary; J-FNS-A, Japanese version of alternative food neophobia scale; LFPQ, the Leeds food preference; LFPQ-BR, the Leeds food preference questionnaire in Brazil; LFPQ-fr, French version of the Leeds food preference questionnaire; LFPQ-Short, shortened version of the Leeds food preference questionnaire; MPC-NA, macronutrient preference checklist use in North America; P-FNS, the food neophobia scale for pregnant women; PSEDPALW, preferences and self-efficacy of diet and physical activity behaviors questionnaire for Latina women; QCSRF, the questionnaire on craving for sweet or rich foods. † DVLM, PROM development; CCV/MI, cross-cultural validity/measurement invariance; CPHSN, comprehensiveness, superscripts e and p denote asking expert and patients, respectively; RLVC, relevance; CPHSB, comprehensibility; MQ, methodological quality, rated as V (very good); A (adequate); D (doubtful), or I (inadequate); QR, quality of the results rated, given as + (sufficient), − (insufficient), ? (indeterminate), or NA (not available). Methodological Quality Assessment and Psychometric Properties Rating of the Included Studies Notes: § Instrument: ACQ, the attitudes to chocolate questionnaire; BSATED, bariatric surgical alterations in tolerability, enjoyment and cravings in the diet; CoEQ, control of eating questionnaire; FAAQ, food craving acceptance and action questionnaire; Fat PQ, fat preference questionnaire; FBC, food behavior checklist; FCCQ, the food chocolate-craving questionnaire; FCDQ, the food choice determinants questionnaire; FCI, the food craving inventory; FCI-J, food craving inventory for Japanese; FCQ, food choice questionnaire; FCQ-BR, Brazilian version scale of the food choice questionnaire; FCQ-H, the food choice questionnaire for Hungary; FCQ-M, food choice questionnaire for Mexico; FCT, food choice task; FDPS, food disgust picture scale; FLQ, the food liking questionnaire; FLQ-FC, the food liking questionnaire in a French-Canadian population; FNS, the food neophobia scale; FPQ, food preference questionnaire; FTSI, food thought suppression inventory; H-FNS, food neophobia scale for Hungary; J-FNS-A, Japanese version of alternative food neophobia scale; LFPQ, the Leeds food preference; LFPQ-BR, the Leeds food preference questionnaire in Brazil; LFPQ-fr, French version of the Leeds food preference questionnaire; LFPQ-Short, shortened version of the Leeds food preference questionnaire; MPC-NA, macronutrient preference checklist use in North America; P-FNS, the food neophobia scale for pregnant women; PSEDPALW, preferences and self-efficacy of diet and physical activity behaviors questionnaire for Latina women; QCSRF, the questionnaire on craving for sweet or rich foods. † DVLM, PROM development; CCV/MI, cross-cultural validity/measurement invariance; CPHSN, comprehensiveness, superscripts e and p denote asking expert and patients, respectively; RLVC, relevance; CPHSB, comprehensibility; MQ, methodological quality, rated as V (very good); A (adequate); D (doubtful), or I (inadequate); QR, quality of the results rated, given as + (sufficient), − (insufficient), ? (indeterminate), or NA (not available). COSMIN Classification of the Criteria for Good Measurement Properties. Quality Assessment: PROM Development and Content Validity Table 3 details the assessment for PROM development and content validity. In the conceptual elicitation phase, literature reviews were the predominant theoretical foundation, often supplemented by expert panels. While item generation generally relied on theory or existing instruments, the reporting of qualitative methodologies was frequently deficient. Consequently, 28 studies were deemed methodologically “doubtful” in PROM development due to poor transparency about the qualitative process. The Food Choice Determinants Questionnaire (FCDQ)8 was an exception (considered “adequate”) as it was developed using in-depth interviews and grounded theory, which ensured that constructs were clearly defined and items were generated through systematic, inductive data analysis. The Brazilian version scale of the food choice questionnaire (FCQ-BR)22 was also deemed “adequate” because it applied a rigorous cross‑cultural adaptation protocol ahead of the quantitative validation. Table 3Evaluation of Evidence Level of the Included Scales and the Formation of RecommendationsInstrument†Content Validity§Structural Validity§Internal Consistency§CCV/MI§Reliability§Criterion Validity§Hypothesis Testing§ClassORQEORQEORQEORQEORQEORQEORQEACQ27?VLa−M+HNANA−LNANANANABBSATED26?VLa−M+HNANANANANANANANABCoEQ19?VLa+M+HNANANANA−MNANABFAAQ7?VLa?L+HNANA+HNANANANABFat PQ20?VLa?L?LcNANA+MNANA+HBFBC35?VLa?L+HNANA+HNANANANABFCCQ30?VLa+H+HNANANANANANANANABFCDQ8+M−M+MbNANA−MNANANANAAFCI33?VLa+H+HNANA+H+H+HBFCI-J25?VLa+H+HNANA+H−MNANABFCQ5?VLa+H+HNANA+HNANA+HBFCQ-BR22+H+H+HNANA+HNANANANAAFCQ-M31?VLa+H+HNANA−MNANANANABFCT21?VLa?L?LcNANA+LcNANANANABFDPS13?VLa+H+HNANA+HNANA+HBFLQ38+H+H+HNANA+HNANANANAAFLQ-FC16+H?L+HNANA+Mb−MNANAAFNS29?VLa−L−MNANA+M−MNANABFPQ6?VLa?L?LcNANA+LcNANANANABFTSI14?VLa+H+HNANANANA?L+HBH-FNS32?VLa+H+HNANANANANANANANABJ-FNS-A23?VLa+H+HNANA+M+HNANABLFPQ37?VLa?L?VLNANANANANANANANABLFPQ-BR17?VLa?L?LcNANANANANANANANABLFPQ-fr34?VLa?L?VLNANA?LNANANANABLFPQ-Short18?VLa?L?LcNANANANANANA+HBMPC-NA15?VLa?L+HNANA+M−MNANABP-FNS28?VLa+H+HNANANANANANANANABPSEDPALW24+H?VL+HNANA+HNANANANAAQCSRF36?VLa?L+HNANANANA?L+HBNotes: † Instrument: ACQ, the attitudes to chocolate questionnaire; BSATED, bariatric surgical alterations in tolerability, enjoyment and cravings in the diet; CoEQ, control of eating questionnaire; FAAQ, food craving acceptance and action questionnaire; Fat PQ, fat preference questionnaire; FBC, food behavior checklist; FCCQ, the food chocolate-craving questionnaire; FCDQ, the food choice determinants questionnaire; FCI, the food craving inventory; FCI-J, food craving inventory for Japanese; FCQ, food choice questionnaire; FCQ-BR, Brazilian version scale of the food choice questionnaire; FCQ-H, the food choice questionnaire for Hungary; FCQ-M, food choice questionnaire for Mexico; FCT, food choice task; FDPS, food disgust picture scale; FLQ, the food liking questionnaire; FLQ-FC, the food liking questionnaire in a French-Canadian population; FNS, the food neophobia scale; FPQ, food preference questionnaire; FTSI, food thought suppression inventory; H-FNS, food neophobia scale for Hungary; J-FNS-A, Japanese version of alternative food neophobia scale; LFPQ, the Leeds food preference; LFPQ-BR, the Leeds food preference questionnaire in Brazil; LFPQ-fr, French version of the Leeds food preference questionnaire; LFPQ-Short, shortened version of the Leeds food preference questionnaire; MPC-NA, macronutrient preference checklist use in North America; P-FNS, the food neophobia scale for pregnant women; PSEDPALW, preferences and self-efficacy of diet and physical activity behaviors questionnaire for Latina women; QCSRF, the questionnaire on craving for sweet or rich foods. § CCV/MI, Cross-cultural validity/measurement invariance. OR, overall rating, including + (sufficient), − (insufficient), ? (indeterminate), ± (inconsistent), and NA (not applicable); QE, quality of evidence, including VL (very low), L (low), M (moderate), H (high), and NA (not available). The superscripts in the assessment of evidence quality represent the following: a, evidence downgraded three levels due to serious risk of bias (multiple studies of doubtful quality); b, evidence downgraded one level due to imprecision (sample size only 50–100 in one study); c, evidence downgraded two levels due to imprecision (sample size <50 in one study). Table 3Evaluation of Evidence Level of the Included Scales and the Formation of RecommendationsInstrument†Content Validity§Structural Validity§Internal Consistency§CCV/MI§Reliability§Criterion Validity§Hypothesis Testing§ClassORQEORQEORQEORQEORQEORQEORQEACQ27?VLa−M+HNANA−LNANANANABBSATED26?VLa−M+HNANANANANANANANABCoEQ19?VLa+M+HNANANANA−MNANABFAAQ7?VLa?L+HNANA+HNANANANABFat PQ20?VLa?L?LcNANA+MNANA+HBFBC35?VLa?L+HNANA+HNANANANABFCCQ30?VLa+H+HNANANANANANANANABFCDQ8+M−M+MbNANA−MNANANANAAFCI33?VLa+H+HNANA+H+H+HBFCI-J25?VLa+H+HNANA+H−MNANABFCQ5?VLa+H+HNANA+HNANA+HBFCQ-BR22+H+H+HNANA+HNANANANAAFCQ-M31?VLa+H+HNANA−MNANANANABFCT21?VLa?L?LcNANA+LcNANANANABFDPS13?VLa+H+HNANA+HNANA+HBFLQ38+H+H+HNANA+HNANANANAAFLQ-FC16+H?L+HNANA+Mb−MNANAAFNS29?VLa−L−MNANA+M−MNANABFPQ6?VLa?L?LcNANA+LcNANANANABFTSI14?VLa+H+HNANANANA?L+HBH-FNS32?VLa+H+HNANANANANANANANABJ-FNS-A23?VLa+H+HNANA+M+HNANABLFPQ37?VLa?L?VLNANANANANANANANABLFPQ-BR17?VLa?L?LcNANANANANANANANABLFPQ-fr34?VLa?L?VLNANA?LNANANANABLFPQ-Short18?VLa?L?LcNANANANANANA+HBMPC-NA15?VLa?L+HNANA+M−MNANABP-FNS28?VLa+H+HNANANANANANANANABPSEDPALW24+H?VL+HNANA+HNANANANAAQCSRF36?VLa?L+HNANANANA?L+HBNotes: † Instrument: ACQ, the attitudes to chocolate questionnaire; BSATED, bariatric surgical alterations in tolerability, enjoyment and cravings in the diet; CoEQ, control of eating questionnaire; FAAQ, food craving acceptance and action questionnaire; Fat PQ, fat preference questionnaire; FBC, food behavior checklist; FCCQ, the food chocolate-craving questionnaire; FCDQ, the food choice determinants questionnaire; FCI, the food craving inventory; FCI-J, food craving inventory for Japanese; FCQ, food choice questionnaire; FCQ-BR, Brazilian version scale of the food choice questionnaire; FCQ-H, the food choice questionnaire for Hungary; FCQ-M, food choice questionnaire for Mexico; FCT, food choice task; FDPS, food disgust picture scale; FLQ, the food liking questionnaire; FLQ-FC, the food liking questionnaire in a French-Canadian population; FNS, the food neophobia scale; FPQ, food preference questionnaire; FTSI, food thought suppression inventory; H-FNS, food neophobia scale for Hungary; J-FNS-A, Japanese version of alternative food neophobia scale; LFPQ, the Leeds food preference; LFPQ-BR, the Leeds food preference questionnaire in Brazil; LFPQ-fr, French version of the Leeds food preference questionnaire; LFPQ-Short, shortened version of the Leeds food preference questionnaire; MPC-NA, macronutrient preference checklist use in North America; P-FNS, the food neophobia scale for pregnant women; PSEDPALW, preferences and self-efficacy of diet and physical activity behaviors questionnaire for Latina women; QCSRF, the questionnaire on craving for sweet or rich foods. § CCV/MI, Cross-cultural validity/measurement invariance. OR, overall rating, including + (sufficient), − (insufficient), ? (indeterminate), ± (inconsistent), and NA (not applicable); QE, quality of evidence, including VL (very low), L (low), M (moderate), H (high), and NA (not available). The superscripts in the assessment of evidence quality represent the following: a, evidence downgraded three levels due to serious risk of bias (multiple studies of doubtful quality); b, evidence downgraded one level due to imprecision (sample size only 50–100 in one study); c, evidence downgraded two levels due to imprecision (sample size <50 in one study). Evaluation of Evidence Level of the Included Scales and the Formation of Recommendations Notes: † Instrument: ACQ, the attitudes to chocolate questionnaire; BSATED, bariatric surgical alterations in tolerability, enjoyment and cravings in the diet; CoEQ, control of eating questionnaire; FAAQ, food craving acceptance and action questionnaire; Fat PQ, fat preference questionnaire; FBC, food behavior checklist; FCCQ, the food chocolate-craving questionnaire; FCDQ, the food choice determinants questionnaire; FCI, the food craving inventory; FCI-J, food craving inventory for Japanese; FCQ, food choice questionnaire; FCQ-BR, Brazilian version scale of the food choice questionnaire; FCQ-H, the food choice questionnaire for Hungary; FCQ-M, food choice questionnaire for Mexico; FCT, food choice task; FDPS, food disgust picture scale; FLQ, the food liking questionnaire; FLQ-FC, the food liking questionnaire in a French-Canadian population; FNS, the food neophobia scale; FPQ, food preference questionnaire; FTSI, food thought suppression inventory; H-FNS, food neophobia scale for Hungary; J-FNS-A, Japanese version of alternative food neophobia scale; LFPQ, the Leeds food preference; LFPQ-BR, the Leeds food preference questionnaire in Brazil; LFPQ-fr, French version of the Leeds food preference questionnaire; LFPQ-Short, shortened version of the Leeds food preference questionnaire; MPC-NA, macronutrient preference checklist use in North America; P-FNS, the food neophobia scale for pregnant women; PSEDPALW, preferences and self-efficacy of diet and physical activity behaviors questionnaire for Latina women; QCSRF, the questionnaire on craving for sweet or rich foods. § CCV/MI, Cross-cultural validity/measurement invariance. OR, overall rating, including + (sufficient), − (insufficient), ? (indeterminate), ± (inconsistent), and NA (not applicable); QE, quality of evidence, including VL (very low), L (low), M (moderate), H (high), and NA (not available). The superscripts in the assessment of evidence quality represent the following: a, evidence downgraded three levels due to serious risk of bias (multiple studies of doubtful quality); b, evidence downgraded one level due to imprecision (sample size only 50–100 in one study); c, evidence downgraded two levels due to imprecision (sample size <50 in one study). The assessment of content validity focused on the comprehensiveness, relevance, and comprehensibility of the instruments as evaluated by patients and experts. Three studies, which validated the FLQ,38 FLQ-FC,16 and PSEDPALW,24 had “adequate” methodological rigor. These studies consulted both patients and experts and had “sufficient” psychometric outcome in all content domains. In the case of FCDQ,8 the method was considered “adequate” by experts but “doubtful” by patients, and the psychometric outcome was consistently “sufficient (+)” across all domains. For FCQ-BR,22 the methodological rigor for assessing content validity was “adequate”, although it was unclear if the appraisal involved experts or patients. All remaining studies were “doubtful” in methodological rigor due to insufficient reporting of cognitive interview protocols or pilot testing procedures. Internal Structure Since data on cross-cultural validity and measurement invariance were absent, the evaluation of internal structure was limited to structural validity and internal consistency (Tables 2 and S3). Regarding structural validity, factor analysis was the primary method employed. According to the COSMIN 2.0 guidelines,11 confirmatory factor analysis (CFA) is preferred for instruments with established theoretical frameworks. Thirteen studies had “very good” methodological rigor. Specifically, eight studies14,25,27,28,31–33,38 utilized both CFA and exploratory factor analysis (EFA), while five studies5,13,22,23,30 relied solely on CFA. Four studies7,8,16,26 were “adequate” as they used EFA alone, and 3 studies19,35,36 were “doubtful” as they only conducted principal component analysis (PCA). The remaining ten studies were considered inadequate. The psychometric outcome in structural validity was “sufficient (+)” for 13 PROMs, “insufficient (−)” for 3, and indeterminate (?) for 14. Regarding internal consistency, most studies (23/30)5,7,8,13–16,19,22–33,35,36,38 calculated Cronbach’s alpha coefficients and were considered having “very good” methodological rigor. All these studies had “sufficient (+)” psychometric outcome because they reported Cronbach’s alpha values ≥0.70 along with at least low-quality evidence for structural validity. The other 7 studies6,17,18,20,21,34,37 had “indeterminate (?)” psychometric outcome due to missing information or inconsistent findings. Among them, the methodological rigor was “inadequate” for 6 studies6,17,18,21,34,37 because they failed to report Cronbach’s alpha and “doubtful” for one20 because it only reported inter-item correlations. Other Measurement Properties Reliability was assessed in 20 studies. Eleven studies5,7,16,21,24,26,31,33–35,38 utilized intraclass correlation coefficients (ICC) and one22 calculated Kappa values; all 12 had “very good” methodological rigor. The other 8 studies6,13,15,20,23,25,27,29 relied on Pearson’s/Spearman correlations; they lacked clarity regarding the consistency of administration methods, and their methodological rigor was deemed “doubtful”. The psychometric outcome was “sufficient (+)” for 16 studies,5–7,13,15,16,20–25,29,33,35,38 as they reported an ICC, Kappa, or Pearson’s/Spearman correlation coefficient greater than 0.7. Three studies26,27,32 had “insufficient (−)” psychometric outcome because they reported an ICC less than 0.7. Criterion validity was evaluated in ten studies.14–16,19,20,23,25,29,33,36 These studies all had “very good” methodological rigor, as they compared the PROM scores against an established comparator (eg, BMI or existing validated scales) frequently treated as a proxy for a gold standard. Among the 10 studies, two studies23,33 had “sufficient (+)” psychometric outcome as they attained ≥0.70 correlation with the comparator. Five studies15,16,19,20,25,29 had “insufficient (−)” psychometric outcome as the reported correlations of <0.7, and the psychometric outcome was indeterminate for the remaining studies. Hypothesis testing for construct validity was conducted in 8 studies.5,7,13,14,18,20,33,36 Two studies13,36 reported on both convergent validity and known-groups validity, five5,7,14,18,20 on convergent validity only, and one33 on known-groups validity only. They all had “very good” methodological rigor, as the convergent validity was always assessed through comparisons with other widely used measurement tools, and the known-groups validity was always examined by subgroup comparisons. For all eight studies, the psychometric outcome was “sufficient (+)”, as they all confirmed that ≥75% of their results aligned with predefined hypotheses. Evidence Certainty and Recommendations Across all PROMs, the certainty of evidence ranged from “high” to “very low” (Table 3). Reasons for downgrading included risk of bias, inconsistency, and imprecision. Five instruments demonstrated robust psychometric profiles and fell under Class A: the Food Choice Determinants Questionnaire (FCDQ),8 the Brazilian version of the Food Choice Questionnaire (FCQ-BR),22 the Food Liking Questionnaire (FLQ),5 the Food Liking Questionnaire in a French-Canadian Population (FLQ-FC),16 and the Preferences and Self-Efficacy of Diet and Physical Activity Behaviors Questionnaire for Latina Women (PSEDPALW).24 All five instruments had synthesized ratings of “sufficient (+)” in both content validity and internal consistency, and the certainty of evidence was high for all PROMs except FCDQ. For FCDQ, the evidence certainty in internal consistency was downgraded one level to moderate due to sample size (only 50–100 participants). For FLQ-FC, reliability was rated “sufficient (+)”, but the evidence certainty was downgraded one level to moderate also due to sample size (only 50–100 participants). All other 25 PROMs had a synthesized rating of “doubtful” in content validity after a three-level downgrade in the evidence certainty due to risk of bias. In addition, the synthesized rating of internal consistency was “doubtful” for FCT, FPQ, LFPQ-short, Fat PQ, and LFPQ-BR, all of which had a two-level downgrade in the evidence certainty due to sample size (<50 participants). Discussion This systematic review provides a rigorous COSMIN-based evaluation of 30 PROMs designed to assess food preference in adults. Five of the evaluated tools, namely FCQ,5 FCQ-BR,22 FCDQ,8 FLQ-FC16 and PSEDPALW,24 demonstrated robust psychometric profiles and received a Class A recommendation. These tools were supported by moderate-to-high evidence certainty for their synthesized performance in content validity and internal consistency. These instruments serve distinct roles. The FCQ,5 FCDQ,8 and FLQ-FC16 offer generalized assessments of food choice, the FCQ-BR22 provides a culturally validated framework for Brazil, and the PSEDPALW24 addresses the specific clinical needs of breast cancer survivors. These results provide clear evidence to help improve patient care and guide nutrition policy. Development and Content Validity High-quality PROM development is the foundation for valid content. COSMIN requires transparent qualitative methods (target population involvement, interview guides, transcription, analysis, saturation) during item generation and cognitive testing. However, in this review, only FCQ-BR22 and FCDQ8 reported adequate PROM development procedures. All other development studies lacked sufficient detail and were therefore deemed “doubtful” for PROM development. This shortfall largely reflects incomplete reporting of qualitative methods and unclear evidence that target users were consulted during item refinement. The 28 studies with “doubtful” PROM development mostly relied only on literature reviews (sometimes with expert panels) rather than full qualitative elicitation with the target population. When content validity was explicitly assessed (relevance, comprehensiveness, comprehensibility), five instruments exhibited adequate procedures and favorable results: FCDQ,8 FCQ-BR,22 FLQ,38 FLQ-FC,16 and PSEDPALW.24 These studies reported consultation with experts and/or patients and produced “sufficient (+)” psychometric outcomes for content domains. Other studies either did not describe cognitive interview protocols or did not state whether patients participated in content appraisal, which creates a large reporting gap. To improve future instruments, developers should document qualitative protocols, include cognitive interviews or pilot testing with target users, and report how feedback shaped final items. Internal Structure and Internal Consistency No included study reported cross-cultural validity/measurement invariance. To evaluate structural validity, eight studies utilized both CFA and EFA, five studies used CFA only, and four studies used EFA only. Among those using CFA, the psychometric outcome was sufficient for all except ACQ (which was insufficient). To assess internal consistency, 23 studies reported Cronbach’s α, all attaining α ≥0.70, thus having “sufficient (+)” psychometric outcome. However, Cronbach’s α alone only indicates internal homogeneity and does not prove that a scale measures the intended construct.39 Future validation work should pair internal consistency with robust structural validity and consider additional indices (eg, item-total correlations, omega coefficients) to better characterize scale reliability.40 Moving beyond classical metrics and employing item response theory (IRT) or structural equation modeling (SEM) will help provide a more granular analysis of measurement precision across different levels of the preference construct. Other Measurement Properties Several measurement properties were unevenly evaluated: Reliability (test–retest): Twenty instruments reported some form of test–retest assessment. Twelve studies used ICC or Kappa (considered methodologically stronger), thus having stronger methodological rigor than the 8 other studies that relied on Pearson/Spearman correlations. However, none of these studies justified the chosen retest interval. This is a significant oversight because food preference possesses both trait-like stability and state-like fluctuations. Without a justified interval, it is difficult to distinguish between measurement error and true biological variation.Criterion validity: Ten studies attempted to establish criterion validity, though the field currently lacks a universally accepted biological or behavioral “gold standard” for food preference. Consequently, researchers frequently relied on surrogate measures or older scales as comparators.41 In the 10 studies, such comparisons yielded modest results, as only two studies achieved correlations ≥0.70. It is possible that existing PROMs are capturing unique dimensions of food preference not fully represented by current proxies. Future work should utilize multidimensional benchmarks to triangulate validity more effectively.Hypothesis testing (construct validity): Eight PROMs conducted hypothesis testing (convergent or known-groups validity), with the majority confirming ≥75% of predefined hypotheses. This high rate of success suggests that while gold standards are absent, the instruments possess strong theoretical foundations and are effectively capturing constructs that correlate predictably with established dietary behaviors and demographic traits.Measurement error and responsiveness: None of the 30 studies reported measurement error or responsiveness, which leaves an important gap for longitudinal and intervention research. This is a critical omission for longitudinal trials where monitoring dietary shifts is the primary objective. Reliability (test–retest): Twenty instruments reported some form of test–retest assessment. Twelve studies used ICC or Kappa (considered methodologically stronger), thus having stronger methodological rigor than the 8 other studies that relied on Pearson/Spearman correlations. However, none of these studies justified the chosen retest interval. This is a significant oversight because food preference possesses both trait-like stability and state-like fluctuations. Without a justified interval, it is difficult to distinguish between measurement error and true biological variation. Reliability (test–retest): Twenty instruments reported some form of test–retest assessment. Twelve studies used ICC or Kappa (considered methodologically stronger), thus having stronger methodological rigor than the 8 other studies that relied on Pearson/Spearman correlations. However, none of these studies justified the chosen retest interval. This is a significant oversight because food preference possesses both trait-like stability and state-like fluctuations. Without a justified interval, it is difficult to distinguish between measurement error and true biological variation. Criterion validity: Ten studies attempted to establish criterion validity, though the field currently lacks a universally accepted biological or behavioral “gold standard” for food preference. Consequently, researchers frequently relied on surrogate measures or older scales as comparators.41 In the 10 studies, such comparisons yielded modest results, as only two studies achieved correlations ≥0.70. It is possible that existing PROMs are capturing unique dimensions of food preference not fully represented by current proxies. Future work should utilize multidimensional benchmarks to triangulate validity more effectively. Criterion validity: Ten studies attempted to establish criterion validity, though the field currently lacks a universally accepted biological or behavioral “gold standard” for food preference. Consequently, researchers frequently relied on surrogate measures or older scales as comparators.41 In the 10 studies, such comparisons yielded modest results, as only two studies achieved correlations ≥0.70. It is possible that existing PROMs are capturing unique dimensions of food preference not fully represented by current proxies. Future work should utilize multidimensional benchmarks to triangulate validity more effectively. Hypothesis testing (construct validity): Eight PROMs conducted hypothesis testing (convergent or known-groups validity), with the majority confirming ≥75% of predefined hypotheses. This high rate of success suggests that while gold standards are absent, the instruments possess strong theoretical foundations and are effectively capturing constructs that correlate predictably with established dietary behaviors and demographic traits. Hypothesis testing (construct validity): Eight PROMs conducted hypothesis testing (convergent or known-groups validity), with the majority confirming ≥75% of predefined hypotheses. This high rate of success suggests that while gold standards are absent, the instruments possess strong theoretical foundations and are effectively capturing constructs that correlate predictably with established dietary behaviors and demographic traits. Measurement error and responsiveness: None of the 30 studies reported measurement error or responsiveness, which leaves an important gap for longitudinal and intervention research. This is a critical omission for longitudinal trials where monitoring dietary shifts is the primary objective. Measurement error and responsiveness: None of the 30 studies reported measurement error or responsiveness, which leaves an important gap for longitudinal and intervention research. This is a critical omission for longitudinal trials where monitoring dietary shifts is the primary objective. Strengths and Limitations This systematic review has several notable strengths. It was conducted in strict adherence to the PRISMA-COSMIN reporting guidelines for Outcome Measurement Instruments9 and utilized the updated COSMIN methodology to ensure methodological rigor. To maintain objectivity and minimize variance, all steps (including literature screening, data extraction, and quality appraisal) were performed independently by two researchers, with disagreements resolved through consensus or third-party adjudication. The use of the COSMIN Risk of Bias checklist and the GRADE approach further ensured a standardized, transparent synthesis of evidence across multiple psychometric domains. However, several limitations warrant consideration. First, some relevant instruments may have been missed if they were published in gray literature or not indexed in the searched databases. Second, many primary studies provided incomplete reporting of measurement properties, notably measurement error and responsiveness, which limited the overall evidence certainty and necessitated downgrading for risk of bias. Third, although independent dual-review was enforced, evidence grading is still inherently subjective. Finally, the evidence quality for several PROMs was constrained by small validation sample sizes (<50 or <100 participants) in the original studies, which may impact the generalizability of our findings. Conclusions Significant methodological gaps existed in the reviewed food preference PROMs. Most tools lacked transparency in their developmental phases, and no study addressed responsiveness or cross-cultural applicability. Of the 30 examined instruments, FCDQ, FCQ-BR, FLQ, FLQ-FC, and PSEDPALW demonstrated sufficient internal structure, construct validity, and content validity, whereas all others were inadequate. The “Grade A” recommendations provided in this review offer an evidence-based standard for instrument selection in research and clinical practice. However, the findings highlight a clear need for higher methodological standards. Future research should prioritize conducting rigorous, longitudinal validations of existing tools. In developing new scales, strict adherence to COSMIN guidelines is needed from the initial conceptualization phase. Greater emphasis must be placed on patient involvement during content validation and the assessment of measurement error to ensure that the assessments of food preference in adults are scientifically reliable and clinically meaningful.