Review Reports - Association Between Food Security Status and Adherence to Mediterranean Diet Among Adults in Saudi Arabia: A Cross-Sectional Study

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

This manuscript addresses an important public health topic and provides useful data on the relationship between food insecurity and diet quality in Saudi Arabia. The study is generally well-structured, and the use of validated tools strengthens the analysis.

However, it currently needs major revision before it is suitable for publication. Please find below some comments to address:

Introduction

The introduction provides sufficient background and clearly explains the rationale for the study. It situates the research within both global and regional contexts and supports statements with appropriate references. The research gap is identified logically, and the study objective is clearly stated. However, the introduction could be slightly more concise by reducing repetition around socioeconomic determinants and dietary transitions;

Methods

The design is clearly identified as cross-sectional and the recruitment channels are described. However, this section needs more detail on recruitment and participation. Because the survey was distributed via social media and university email, the manuscript should explicitly discuss the likely selection bias toward younger, more educated, digitally connected participants. It would also help to report whether duplicate submissions were prevented, whether IP checks or one-response restrictions were used, and whether the survey was open nationwide or targeted through specific networks;

The authors should clarify how many people accessed the survey, how many consented, how many were excluded, and why. A flow diagram would improve transparency. It would also be useful to explain why adults older than 65 were excluded, especially since food insecurity and diet quality are highly relevant in older age groups;

The sample is heavily skewed toward participants from the Western region (81.6%), which limits national representativeness. This should already be signaled in the Results and later emphasized more strongly in the Discussion. Please expand the discussion of this limitation and clarify how it may affect the interpretation of the findings;

There is an editorial issue: the manuscript jumps from section 2.2 to 2.4, suggesting section 2.3 is missing;

The authors used a modified 13-item MEDAS after removing the wine item, which is understandable in an Islamic setting. However, once the tool is modified, the authors should provide stronger justification for the chosen cut-offs and discuss whether the modified score has been validated in Arabic-speaking Saudi adults. The sentence “The scale was translated into Arabic for use in this study [7]” also appears inadequately supported by the cited reference. At minimum, the authors should clarify whether they used an existing Arabic version or performed a new translation, and whether reliability or internal consistency was assessed in the present sample;

This subsection is too brief. Several variables later analyzed in the results are not listed here, including geographic region and marital status. The Methods should define all covariates exactly as used in analysis, including category construction for age, income, marital status, region, and education.

Statistical analysis

The authors state that a “generalized linear model” was used, but they do not report the family, link function, distributional assumptions, or why this model was preferred over ordinary linear regression. Since the MedDiet adherence score was treated as continuous, the authors should explain whether assumptions of linearity, normality of residuals, and homoscedasticity were examined. If the outcome was ordinal or bounded, that should also be discussed. In addition, collapsing the three food insecurity categories into one binary group for bivariate analyses is acceptable for descriptive purposes, but the rationale should be stated. The authors should also indicate how missing data were handled and whether any sensitivity analyses were conducted;

Using the FIES-SM is appropriate. However, the manuscript needs more precise reporting on the Arabic version and on the categorization. Most importantly, the references cited here appear problematic: the statement “validated for use in Arab countries [8,9]” does not seem to match the references listed, because those references concern Mediterranean diet adherence rather than FIES validation.

Author Response

Please see the attachment

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

The relationship between food insecurity and adherence to the Mediterranean diet is important, and the sample size is reasonably large for a cross-sectional study. The paper is also easy to follow at a general level. However, in its current form, several methodological and reporting issues limit confidence in the findings and should be addressed before the manuscript can be considered further.

-Participants were recruited through Google Forms distributed on social media and university email using convenience sampling. In addition, 81.6% of the sample came from the Western region, and the educational profile was skewed toward university-educated respondents. These features substantially limit the representativeness of the sample and make it difficult to generalize the findings to adults in Saudi Arabia more broadly. The manuscript should should be revised to avoid implying national representativeness.

-The manuscript states that the original 14-item MEDAS was reduced to 13 items by removing the wine item and that the scale was translated into Arabic for this study. However, no details are provided on forward-backward translation, cultural adaptation, pilot performance, internal consistency, or any psychometric evaluation of the modified version. Since the score was altered and then categorized into low, medium, and high adherence groups, the authors should justify the validity of these categories in the modified Arabic version or present sensitivity analyses.

-The manuscript states that the Arabic FIES was obtained from FAO and “validated for use in Arab countries,” but the citations attached to that sentence do not appear to match that claim. A careful audit of the references is needed to ensure that the cited studies truly support the statements made in the Methods and Introduction. This is important because instrument validity is central to the study.

-The authors state that a generalized linear model was used, but they do not specify the distribution family, link function, or model diagnostics. Since the MedDiet adherence score is a bounded discrete score, the analytical choice should be justified more clearly. It is also unclear whether assumptions were checked for the independent t-test used in Table 4. At minimum, the manuscript should report why these tests were selected, whether assumptions were met, and whether effect sizes or model fit indices were examined.

-The statement that “65.8% of younger participants (18–34 years) experienced food insecurity” is not supported by Table 3 as written. The table shows age-specific proportions of food insecurity of 34.3% for ages 18–24 and 31.5% for ages 25–34. The current wording appears to confuse the distribution of food-insecure cases with within-group prevalence. This section should be revised carefully, and the authors should re-check all percentages against their denominators.

-Please check p values in the text and the tables. Are they matched?

-The adjusted association between food insecurity score and MedDiet adherence score is statistically significant, but the magnitude of the association is modest. Given the cross-sectional design, convenience sampling, and limited representativeness, the policy and intervention implications should be presented more cautiously. The paper should avoid language that sounds stronger than the evidence allows.

-The manuscript states that adherence to the MedDiet was “moderate,” yet the mean score was 5.45/13 and the majority of participants were classified as having low adherence. The wording should be made more precise and consistent across the Results and Discussion.

-Table 3 and its title should be edited for clarity. The phrasing “food security and food insecurity” is awkward, and the presentation of merged categories should be explained more cleanly.

-The reference list should be checked carefully for duplication and formatting problems. For example, references 12 and 13 appear to cite the same paper.

Author Response

Please see the attachment

Author Response File: Author Response.pdf

Reviewer 3 Report

Comments and Suggestions for Authors

I have carefully reviewed the manuscript entitled "Association between food security status and adherence to the Mediterranean diet among adults in Saudi Arabia". The study addresses a relevant and timely public health topic, particularly in the context of nutritional transition and socioeconomic disparities in non-Mediterranean settings. While the manuscript presents some strengths, several methodological, analytical, and interpretative issues need to be addressed before it can be considered for publication.

Major Comments

I am concerned about the use of convenience sampling via social media platforms (X, WhatsApp, Snapchat, Telegram). This approach introduces substantial self-selection bias and limits external validity. The sample is heavily skewed toward participants from the Western region (81.6%) and individuals with relatively high education levels, raising serious concerns about representativeness of the Saudi adult population. This issue is not sufficiently acknowledged across the manuscript, including in the Methods and Discussion sections. Moreover, no response rate is reported, and recruitment bias is not adequately discussed. The inclusion of a sample size calculation is misleading given the non-probabilistic sampling approach and should either be removed or clearly reframed. Additionally, the regional imbalance should be explicitly reported and discussed in both the Results and Discussion sections, rather than being only implicitly presented in descriptive tables.
The adaptation of the MEDAS tool from 14 to 13 items due to the exclusion of alcohol consumption is understandable in this cultural context. However, I consider the lack of validation of this modified version a major methodological limitation. The manuscript does not report: Internal consistency (e.g., Cronbach’s alpha); Construct validity of the adapted scale. This raises concerns regarding measurement validity and comparability across studies. At minimum, the authors should provide reliability metrics and justify maintaining the original cut-off points after modification.

Furthermore, the questionnaire validation process is insufficient. Face validity was assessed with only five individuals, and content validation relied on three experts without a clear description of the evaluation procedure (e.g., criteria used, agreement levels, revisions made). This does not meet acceptable standards, particularly given that key variables rely on self-reported data.

Regarding the Methods section, the use of FIES is appropriate; however, the authors should clarify whether the Arabic version has been specifically validated in the Saudi context and justify the recall period used.

Although the use of a generalized linear model is appropriate, the statistical methodology is insufficiently described. The manuscript does not report: Model assumptions (normality, linearity); Distribution and link function used; Multicollinearity diagnostics. In addition, the absence of key confounders such as BMI, physical activity, and health status weakens the analytical framework and may bias the reported associations. This lack of detail in the Methods section reduces reproducibility and limits confidence in the robustness of the findings.

There are several inconsistencies in how results are presented and interpreted. The descriptive results indicate low adherence to the Mediterranean diet, yet this is described as “moderate” in the Discussion, which is not consistent with the reported data. The merging of food insecurity categories into a binary variable reduces analytical detail and should be clearly justified, as it may obscure important gradients. Additionally, some subgroup analyses involve small sample sizes, which may compromise statistical validity and should be acknowledged.
While the association between food insecurity and MedDiet adherence is statistically significant (B ≈ −0.10; p ≈ 0.02), the effect size is very small and likely of limited practical relevance. The difference in adherence scores between food-secure and food-insecure individuals is minimal (5.54 vs 5.19), and the categorical distribution is not statistically significant (p = 0.089). I find that the manuscript overinterprets these findings. The Discussion and Conclusion sections should be revised to reflect the modest magnitude of the association and avoid overstating its implications.
The manuscript largely confirms well-established evidence that food insecurity is associated with poorer diet quality. While the regional focus is valuable, the conceptual contribution remains limited. The Introduction does not sufficiently justify the relevance of studying the Mediterranean diet in a non-Mediterranean context, and cross-country comparisons are presented without adequate contextualisation. The Discussion would benefit from deeper engagement with: Food environment and structural determinants; Cultural applicability of the Mediterranean diet in Saudi Arabia; Broader frameworks such as the nutrition transition and food systems perspective. Additionally, some interpretations (e.g., food cost and affordability) are speculative, as these variables were not directly measured in the study.

The strengths of the study are somewhat overstated, and key limitations are underdeveloped. In particular, the manuscript should more clearly acknowledge: Sampling bias and limited representativeness; Lack of validation of the adapted MEDAS instrument; Absence of relevant behavioural and health-related confounders. The Implications section is too generic and not sufficiently supported by the presented data. Policy recommendations should be more cautious and directly aligned with the study’s findings.
The title should specify the cross-sectional design for clarity. The abstract reports p-values but lacks effect sizes, which limits interpretability. Statements regarding socioeconomic factors are overly broad and should reflect the specific variables analysed. The conclusions are generally aligned with the results but should be more conservative. Any causal language should be avoided due to the cross-sectional design.

Author Response

Please see the attachment

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

I have no further comments.

Author Response

Comment 1: I have no further comments.

Respond: We thank the reviewer for the constructive feedback and are pleased that the revisions have addressed the reviewer’s concerns.

Reviewer 3 Report

Comments and Suggestions for Authors

I would like to thank the authors for their thorough revisions and detailed responses to the previous comments. The manuscript has improved substantially in terms of methodological transparency, statistical reporting, and interpretation of findings. In particular, the authors have strengthened the description of the analytical approach, expanded the limitations section, and adopted a more cautious interpretation of the observed associations. The discussion is now better contextualized within the broader literature on food insecurity, dietary patterns, and nutrition transition, and several previously raised concerns have been adequately addressed. However, one important issue still requires further clarification before the manuscript can be considered for acceptance.

Although the authors now report the internal consistency of the modified 13-item MEDAS instrument, the reported Cronbach’s alpha remains very low (α = 0.389). This raises important concerns regarding the reliability and internal consistency of the adapted instrument within the studied population. While the authors acknowledge this limitation, I believe the implications of such low reliability should be discussed more explicitly and critically, particularly regarding: potential measurement error; attenuation of observed associations; comparability with studies using the original validated MEDAS; interpretation of adherence categories and scores. The authors should also clarify whether any item-level analysis was performed to identify problematic items contributing to the low internal consistency.

Overall, the manuscript is now much improved and close to being suitable for publication. I therefore recommend Minor Revision.

Author Response

Comment 1:

Overall, the manuscript is now much improved and close to being suitable for publication. I therefore recommend Minor Revision.

Response 1:

We thank the reviewer for the feedback and positive evaluation of the revised manuscript. We acknowledge the concern regarding the low internal consistency of the modified 13-item MEDAS in the current sample (Cronbach’s α = 0.389). The relatively low internal consistency may reflect the multidimensional nature of the instrument, as the MEDAS assesses multiple distinct dietary behaviors rather than a single underlying construct. Similar findings were reported by Hou et al. (2023), who evaluated the same adapted MEDAS in Taiwan and identified several items with low item-total correlations, including butter/cream, red meat, and commercial sweets. Despite the low internal consistency, the instrument was retained to preserve comparability with the original MEDAS, which was also the approach adopted in the current study.

Additional discussion has been added to the limitations section (page 12, lines 415–425) to address the implications of this finding, including the possibility of measurement error, attenuation of the observed associations, and limited comparability with studies using the original validated MEDAS instrument. Therefore, adherence scores and categories should be interpreted with caution.

Hou, Y.-C., Li, J.-Y., Chen, J.-H., Hsiao, J.-K., & Wu, J.-H. (2023). Short Mediterranean diet screener detects risk of prediabetes in Taiwan. Scientific Reports, 13, 1220. https://doi.org/10.1038/s41598-023-28573-5

Author Response File: Author Response.pdf