Fifteen Years of Patient Experience with Hospital Food in a Spanish Long-Term Care Hospital

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

Fifteen Years of Patient Experience with Hospital Food in a Spanish Long-Term Care Hospital

Title: Appropriate.

Abstract: Well written, the most relevant information is included.

Introduction: Short but with good information.

Materials and Methods: the strongest part. Very well developed, explained, especially the instrument/questionnaire [PREM] development/adaptation and psychometric validation. 

Results: Also, very strong. The authors selected the most important findings. 

Discussion: Although time trends (or trend studies) might be considered a type of longitudinal research, authors do not track the same participants over time and do not do 'causal' analysis, then, I suggest changing the word 'longitudinal' used in lines 318, 332, and 413 to 'trend series’ or 'time series’.

Author Response

We are grateful to the editor and reviewers for their valuable comments and suggestions. These observations have provided an excellent opportunity to strengthen and refine both the style and the scientific content of the manuscript.

All changes made in response to the reviewers’ comments have been clearly highlighted in the revised text.

Reviewer 1.

Title: Appropriate.

Abstract: Well written, the most relevant information is included.

Introduction: Short but with good information.

Materials and Methods: the strongest part. Very well developed, explained, especially the instrument/questionnaire [PREM] development/adaptation and psychometric validation. 

Results: Also, very strong. The authors selected the most important findings.

Discussion: Although time trends (or trend studies) might be considered a type of longitudinal research, authors do not track the same participants over time and do not do 'causal' analysis, then, I suggest changing the word 'longitudinal' used in lines 318, 332, and 413 to 'trend series’ or 'time series’.

Thank you for this important comment. We agree with this observation and have revised the manuscript accordingly, replacing the term longitudinal research with more appropriate wording throughout the text.

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

Dear Authors,

This manuscript addresses a clinically relevant but often underestimated aspect of care in long-term hospitalization, namely the patient experience with hospital food and food service. The long observation window is a clear strength, and the dataset has practical value because it reflects routine care rather than an artificial research setting. I also appreciate the authors’ attempt to update and psychometrically evaluate a food-related PREM, which could become useful for quality improvement in similar institutions. At the same time, several parts of the methodological framing remain less clear than they should be, especially regarding comparability across questionnaire versions, the source of responses, and the interpretation of temporal changes. In my opinion, the paper has good potential, but it would benefit from a more cautious and transparent presentation before it is ready for publication.

• Please clarify the unit of analysis and the issue of independence of observations. Because the survey was administered annually to inpatients in a long-term care setting, it seems possible that some patients may have contributed more than one response across years. This point is important for both the psychometric analyses and the temporal trend analyses, and it should be explained explicitly.
• Please clarify who actually completed the questionnaire. The supplementary forms indicate that the respondent could be the patient, a paid caregiver, or a family member, whereas the Methods section reads more as if the analysis concerned patient responses only and excluded cognitively impaired patients. Please report the proportion of proxy responses, if available, and consider a sensitivity analysis or at least a discussion of the possible influence of respondent type.
• The harmonization of the older 4-point version and the newer 5-point version needs a stronger justification. A simple transformation of the old scale into the new metric may affect comparability across years, especially when the wording and number of response options changed. I would strongly recommend a sensitivity analysis restricted to the 2024–2025 version, or at minimum a separate presentation showing that the main conclusions do not depend on the scale conversion strategy.
• The interpretation of the organizational milestones should be more cautious. The observed worsening after the catering company change and the improvement after the introduction of new carts are plausible and interesting, but with annual cross-sectional data from a single center these findings should be framed as associations rather than as causal effects. This is particularly relevant for the 2025 cart intervention, where the post-intervention window is very limited.
• The psychometric discussion should be strengthened and slightly toned down. The overall scale performs reasonably well, but Factor 2 shows modest reliability, and some items such as temperature, presentation, and timing appear weaker in the factor models. I think the manuscript would be stronger if the authors acknowledged more directly that the total score currently seems more robust than the subscale interpretation.
• There is a methods/results inconsistency in the comparison across diet types: the Methods section refers to Kruskal–Wallis with Dunn’s post hoc testing, whereas the Results section reports ANOVA with an F statistic. Please align the statistical description throughout the manuscript.
• In the subsection title, “Confirmatory Factor Analysis (EFA)” should be corrected to “Confirmatory Factor Analysis (CFA)”.
• The English is generally understandable, but several sentences still sound slightly awkward and would benefit from polishing. Examples include “soft diet without food that pose a choking risk”, “hospital unis”, and a few discussion sentences with uneven syntax.
• Please explain the handling of missing data in greater detail. It would be helpful to know whether missingness differed by year, hospital unit, or diet type, and whether complete-case analysis may have introduced bias.
• Because the global satisfaction item is conceptually very close to the PREM content, the manuscript would benefit from a more nuanced description of criterion validity. I would avoid presenting this as strong external validation and instead describe it as supportive evidence using a closely related anchor item.
• To enrich the clinical discussion, the authors may also consider citing DOI:10.3390/jcm14051494, because it discusses biochemical, hematological, and immunological indicators of malnutrition in hospitalized older inpatients, and DOI:10.3390/antiox12030569, because it addresses oxidative stress and adipokine markers in older hospitalized patients with diverse nutritional status. .

Please respond point by point to all of the comments above and indicate clearly how each issue has been addressed in the revised manuscript.

Best regards,

The reviewer.

Author Response

Reviewer 2

This manuscript addresses a clinically relevant but often underestimated aspect of care in long-term hospitalization, namely the patient experience with hospital food and food service. The long observation window is a clear strength, and the dataset has practical value because it reflects routine care rather than an artificial research setting. I also appreciate the authors’ attempt to update and psychometrically evaluate a food-related PREM, which could become useful for quality improvement in similar institutions. At the same time, several parts of the methodological framing remain less clear than they should be, especially regarding comparability across questionnaire versions, the source of responses, and the interpretation of temporal changes. In my opinion, the paper has good potential, but it would benefit from a more cautious and transparent presentation before it is ready for publication.

Thank you for this overall assessment, which we greatly appreciate. We have used the comments and suggestions to improve the manuscript.

• Please clarify the unit of analysis and the issue of independence of observations. Because the survey was administered annually to inpatients in a long-term care setting, it seems possible that some patients may have contributed more than one response across years. This point is important for both the psychometric analyses and the temporal trend analyses, and it should be explained explicitly.

We thank the reviewer for this comment. The PREM survey was administered once per year and collected anonymously; therefore, observations were analyzed as independent repeated cross-sectional data. Although it cannot be confirmed, repeated responses are unlikely given the average length of stay (approximately 90 days for long-term care and 120 days for mental health). However, this possibility cannot be fully excluded and has been acknowledged as a limitation.

• Please clarify who actually completed the questionnaire. The supplementary forms indicate that the respondent could be the patient, a paid caregiver, or a family member, whereas the Methods section reads more as if the analysis concerned patient responses only and excluded cognitively impaired patients. Please report the proportion of proxy responses, if available, and consider a sensitivity analysis or at least a discussion of the possible influence of respondent type.

We thank the reviewer for this helpful observation. Whenever possible, the questionnaire was completed directly by the patient. However, when the patient was unable to respond due to frailty, clinical condition, or cognitive impairment, a family member or paid caregiver completed the PREM as a proxy respondent, following routine practice in our long‑term care setting. We have now clarified this point in the Methods section, and we have also included a table reporting the distribution of respondent (Table S1). Regarding the potential influence of respondent type, we conducted a sensitivity analysis comparing PREM scores across patients, relatives, and paid caregivers (Table S2). Patients completed the majority of questionnaires (80%), while proxy respondents accounted for the remaining 20%. Although proxy respondents tended to provide slightly higher ratings, the differences were small and did not alter the overall pattern of results. These findings suggest that proxy completion introduces minimal variation and does not materially affect the interpretation of the PREM.

• The harmonization of the older 4-point version and the newer 5-point version needs a stronger justification. A simple transformation of the old scale into the new metric may affect comparability across years, especially when the wording and number of response options changed. I would strongly recommend a sensitivity analysis restricted to the 2024–2025 version, or at minimum a separate presentation showing that the main conclusions do not depend on the scale conversion strategy.

We thank the reviewer for this valuable suggestion. To assess whether the scale change could affect the interpretation of the results, we performed a sensitivity analysis comparing mean PREM scores between the period using the 4‑point scale (2011-2023) and the period using the 5‑point scale (2024-2025) (Table S3). As expected, absolute scores were higher in the 5‑point period; however, the relative pattern across dimensions remained stable. The dimensions with the highest and lowest ratings were consistent across periods, and the overall interpretation of the results did not change. This indicates that the scale modification affects only the absolute level of the scores, not the substantive conclusions. We have added this justification and the results of the sensitivity analysis to the manuscript.

• The interpretation of the organizational milestones should be more cautious. The observed worsening after the catering company change and the improvement after the introduction of new carts are plausible and interesting, but with annual cross-sectional data from a single center these findings should be framed as associations rather than as causal effects. This is particularly relevant for the 2025 cart intervention, where the post-intervention window is very limited.

We agree that our wording may have suggested causal interpretations that are not warranted with annual cross‑sectional data from a single center. We have revised the text to consistently describe the changes in PREM scores as temporal associations with the catering company change and the introduction of new carts, rather than as causal effects. We now explicitly state that these findings should be interpreted as exploratory and hypothesis‑generating, and this clarification has been added to the Limitations section of the manuscript.

• The psychometric discussion should be strengthened and slightly toned down. The overall scale performs reasonably well, but Factor 2 shows modest reliability, and some items such as temperature, presentation, and timing appear weaker in the factor models. I think the manuscript would be stronger if the authors acknowledged more directly that the total score currently seems more robust than the subscale interpretation.

We appreciate this thoughtful observation. In response, we have strengthened and moderated the psychometric discussion to more clearly acknowledge that, although the overall scale performs well, Factor 2 shows somewhat lower reliability and weaker item loadings. We now explicitly state that the total score currently appears more robust than the subscale interpretation and is therefore the most reliable indicator for routine use. These clarifications have been incorporated into the Discussion.

• There is a methods/results inconsistency in the comparison across diet types: the Methods section refers to Kruskal–Wallis with Dunn’s post hoc testing, whereas the Results section reports ANOVA with an F statistic. Please align the statistical description throughout the manuscript.

We thank the reviewer for detecting this inconsistency. The planned analysis for comparisons across diet types was non‑parametric (Kruskal-Wallis with Dunn’s post hoc tests), and this is what we actually used. The reference to ANOVA and the F statistic in the Results was an oversight and has now been corrected to report the Kruskal–Wallis statistic and p‑value, in line with the Methods section.

• In the subsection title, “Confirmatory Factor Analysis (EFA)” should be corrected to “Confirmatory Factor Analysis (CFA)”.

This typo has been corrected.

• The English is generally understandable, but several sentences still sound slightly awkward and would benefit from polishing. Examples include “soft diet without food that pose a choking risk”, “hospital unis”, and a few discussion sentences with uneven syntax.

We have reviewed the English and identified a few sentences that could be improved for better clarity and fluency. The sentences have been refined and now sound more natural.

• Please explain the handling of missing data in greater detail. It would be helpful to know whether missingness differed by year, hospital unit, or diet type, and whether complete-case analysis may have introduced bias.

We appreciate the reviewer’s request for further clarification regarding the handling of missing data. As noted in the manuscript, missingness was low overall. To address this comment, we conducted a structured analysis of missing data by year, hospital unit, and diet type. The results (now included in the Supplementary Material, Tables S4-S6) show no meaningful differences across these dimensions and no systematic patterns of missingness. All analyses were performed using complete cases, which is stated explicitly in the Methods section. Given the low and relatively stable level of missingness across subgroups, the risk of substantial bias due to complete‑case analysis is likely limited. Nonetheless, we now acknowledge in the Limitations section that complete‑case analysis may still introduce some degree of bias.

• Because the global satisfaction item is conceptually very close to the PREM content, the manuscript would benefit from a more nuanced description of criterion validity. I would avoid presenting this as strong external validation and instead describe it as supportive evidence using a closely related anchor item.

Because no gold standard measure of patient experience with hospital food was available, a single global satisfaction item was used as a closely related anchor to provide supportive evidence regarding the PREM’s performance. Given the conceptual proximity between this item and the PREM content, the results were interpreted cautiously and not as evidence of strong external validation.

• To enrich the clinical discussion, the authors may also consider citing DOI:10.3390/jcm14051494, because it discusses biochemical, hematological, and immunological indicators of malnutrition in hospitalized older inpatients, and DOI:10.3390/antiox12030569, because it addresses oxidative stress and adipokine markers in older hospitalized patients with diverse nutritional status. .

Thank you for this valuable suggestion. It has helped to reinforce the message regarding the importance of addressing malnutrition in this setting, and both references have been included in the revised manuscript.

 

 

Author Response File: Author Response.pdf

Reviewer 3 Report

Comments and Suggestions for Authors

The manuscript entitled “Fifteen Years of Patient Experience With Hospital Food in a Spanish Long-Term Care Hospital” addresses a highly relevant issue, namely nutrition and the consequences of its inadequacy in long-term care settings.

I believe the authors have valuable and interesting data; however, the results are presented by addressing topics that would be more appropriately divided into at least two separate manuscripts. What is particularly confusing is the aim of the study. It is unclear whether the objective is the validation of the questionnaire, the analysis of hospital diets, or an organizational evaluation comparing the effectiveness of two different food providers.

I would suggest that the authors consider splitting the work into two distinct publications.

The first manuscript, more clinically oriented, should:
a) Provide a stronger and more comprehensive introduction. This should include discussion of the nutritional consequences of long-term hospitalization, other contributing factors to worsening nutritional status beyond hospital diet (e.g., depression, fasting for diagnostic procedures), information on how regularly patients are assessed at admission and during hospitalization, differences in food provision between countries such as the USA and Spain, risks of iatrogenic malnutrition in different patient categories (e.g., dysphagic vs. surgical patients), and known causes of dissatisfaction with hospital diets reported in the literature.

1. b) Clarify participant selection. The current description is somewhat unclear, as it suggests that all hospitalized patients were enrolled, yet after 15 years the authors questioned the reliability of responses from cognitively impaired patients. Moreover, the questionnaires indicate that, for patients unable to respond, caregivers answered on their behalf. This aspect needs to be better explained and justified.
2. c) Justify the use of a 5-point Likert scale. The use of an odd-numbered scale introduces the risk that respondents may prefer neutral answers, potentially leading to overall neutral results. The authors should explain their choice and discuss its implications.

Results: In this type of study, only questionnaire results should be presented, with reference to the different items. Given the change in food provider, the authors should interpret their findings by highlighting the overall result of the 15-year analysis—for example, whether no significant changes occurred over time, whether patients consistently expressed dissatisfaction, or whether specific factors (such as portion size) remained the main source of dissatisfaction (these are illustrative examples, not references to actual data).

In my opinion, the current title appears to have been defined in a way that allows inclusion of all results, but these should instead be separated: questionnaire validation on one side, and diet acceptability and its consequences on the other.

In this clinical context, the discussion should also include data on mealtime assistance, barriers to food consumption, and information on how patients’ weight and BMI evolve during hospitalization.

Statistical considerations related to the questionnaire should be presented in a separate manuscript, with appropriate restructuring of the content.

 

Author Response

The manuscript entitled “Fifteen Years of Patient Experience With Hospital Food in a Spanish Long-Term Care Hospital” addresses a highly relevant issue, namely nutrition and the consequences of its inadequacy in long-term care settings.

I believe the authors have valuable and interesting data; however, the results are presented by addressing topics that would be more appropriately divided into at least two separate manuscripts. What is particularly confusing is the aim of the study. It is unclear whether the objective is the validation of the questionnaire, the analysis of hospital diets, or an organizational evaluation comparing the effectiveness of two different food providers.

I would suggest that the authors consider splitting the work into two distinct publications.

The first manuscript, more clinically oriented, should:

a) Provide a stronger and more comprehensive introduction. This should include discussion of the nutritional consequences of long-term hospitalization, other contributing factors to worsening nutritional status beyond hospital diet (e.g., depression, fasting for diagnostic procedures), information on how regularly patients are assessed at admission and during hospitalization, differences in food provision between countries such as the USA and Spain, risks of iatrogenic malnutrition in different patient categories (e.g., dysphagic vs. surgical patients), and known causes of dissatisfaction with hospital diets reported in the literature.

We thank the reviewer for this valuable suggestion. The introduction has been enriched by incorporating a broader discussion on the nutritional consequences of prolonged hospitalization and additional factors contributing to the deterioration of nutritional status.

b) Clarify participant selection. The current description is somewhat unclear, as it suggests that all hospitalized patients were enrolled, yet after 15 years the authors questioned the reliability of responses from cognitively impaired patients. Moreover, the questionnaires indicate that, for patients unable to respond, caregivers answered on their behalf. This aspect needs to be better explained and justified.

We have clarified that approximately 80% of questionnaires were completed directly by patients (Table S1). When patients were unable to respond due to clinical condition, frailty, or cognitive impairment, caregivers or relatives completed the PREM on their behalf, reflecting routine practice in our hospital. We have also added a sensitivity analysis comparing patient and proxy responses, which shows that respondent type had only minimal influence on PREM scores.

c) Justify the use of a 5-point Likert scale. The use of an odd-numbered scale introduces the risk that respondents may prefer neutral answers, potentially leading to overall neutral results. The authors should explain their choice and discuss its implications.

We thank the reviewer for this valuable comment. The change in the response scale was introduced to improve the sensitivity and discriminative capacity of the instrument. Specifically, the use of a 5-point Likert scale allows for greater variability in patient responses, reducing the likelihood of clustering at extreme categories. In addition, the inclusion of a neutral midpoint was considered important, as it enables respondents to express opinions that are neither clearly positive nor negative. This helps to minimize forced-choice bias and enhances the validity of the measurements. In terms of implications, this modification supports a more nuanced assessment of patient experience, which may facilitate the identification of areas for improvement in hospital food services with greater precision. Furthermore, a sensitivity analysis was conducted comparing results before and after the scale modification. The analysis showed similar response patterns, supporting the robustness of the findings and suggesting that the change does not materially affect the study conclusions.

Results: In this type of study, only questionnaire results should be presented, with reference to the different items. Given the change in food provider, the authors should interpret their findings by highlighting the overall result of the 15-year analysis—for example, whether no significant changes occurred over time, whether patients consistently expressed dissatisfaction, or whether specific factors (such as portion size) remained the main source of dissatisfaction (these are illustrative examples, not references to actual data).

We took into account the suggestion to separate the validation of the questionnaire and the evaluation of diet acceptability into two separate publications. However, as mentioned earlier, we chose to maintain a single publication to avoid unnecessary fragmentation of the findings. We adjusted the approach to clearly separate the methodological and clinical aspects within the same manuscript. Therefore, Study 1 focuses on questionnaire validation, while Study 2 presents the clinical findings derived from its use.

In my opinion, the current title appears to have been defined in a way that allows inclusion of all results, but these should instead be separated: questionnaire validation on one side, and diet acceptability and its consequences on the other.

We greatly appreciate the reviewer’s overall assessment of the work and the suggestions for improvement. In preparing this manuscript, the dilemma accurately identified by the reviewer was indeed considered. However, a single publication was preferred in order to avoid salami slicing or any impression that the intention was to generate multiple publications from the same study. As an alternative, and in line with the reviewer’s concern, methodological aspects of the study that support the measurement approach have been included in study 1.

This structure was chosen to ensure methodological clarity and analytical rigor. Study 1 establishes the soundness of the measurement instrument, whereas Study 2 focuses on the substantive findings derived from its use. This approach avoids conflating evidence on the quality of the tool with evidence on temporal patterns in patient experience.

In this clinical context, the discussion should also include data on mealtime assistance, barriers to food consumption, and information on how patients’ weight and BMI evolve during hospitalization.

In the discussion, we expanded on data regarding mealtime assistance, barriers to food consumption, and information on how patients' weight and BMI evolve during hospitalization. This information, along with the review of the findings, allows for a more comprehensive and contextualized interpretation of the clinical results. We also addressed how these aspects can influence patient experience and how changes in food service impacted overall satisfaction.

Statistical considerations related to the questionnaire should be presented in a separate manuscript, with appropriate restructuring of the content.

We greatly appreciate the reviewer’s overall assessment of the work and the suggestions for improvement. In preparing this manuscript, the dilemma accurately identified by the reviewer was indeed considered. However, a single publication was preferred in order to avoid salami slicing or any impression that the intention was to generate multiple publications from the same study. As an alternative, and in line with the reviewer’s concern, methodological aspects of the study that support the measurement approach have been included in study 1.

This structure was chosen to ensure methodological clarity and analytical rigor. Study 1 establishes the soundness of the measurement instrument, whereas Study 2 focuses on the substantive findings derived from its use. This approach avoids conflating evidence on the quality of the tool with evidence on temporal patterns in patient experience.

 

 

Author Response File: Author Response.pdf

Round 2

Reviewer 2 Report

Comments and Suggestions for Authors

Dear Authors,

Thank you very much for the effort you put into improving your manuscript. Well-revised!

Best regards,

The reviewer.

Author Response

We thank the reviewer for the positive feedback and for the valuable comments provided throughout the review process, which have helped to improve the manuscript.

Reviewer 3 Report

Comments and Suggestions for Authors

The manuscript may be published now!

Author Response

We sincerely thank the reviewer for the positive evaluation of our manuscript. We are very grateful for the time and effort dedicated to reviewing our work, as well as for the constructive comments provided during the process, which have helped us to improve the manuscript and bring it to its current form. We truly appreciate your support in considering the manuscript for publication in this journal.