Determinants of Influenza Vaccination Coverage among Greek Health Care Workers amid COVID-19 Pandemic

Round 1

Reviewer 1 Report

This is an online questionnaire survey study. However, the items assessing the respondents’ attitude toward, effect and safety of influenza vaccination were imperfect. Table 1 described the items used for assessing influenza vaccination contained the word “vaccines” but not “influenza vaccines.” Given that this study analyzed the secondary questionnaire-based data on the COVID-19 vaccination acceptance, it seemed impossible for the respondents to understand what “vaccines” on the items indicate, vaccines for COVID-19, influenza, or both? Influenza vaccines had a long history of development compared with COVID-19 vaccines. People may know and trust on influenza vaccines well but feel unfamiliar with COVID-19 vaccines. Therefore, the items with the incomplete definition limited the value of this study.

Author Response

Reviewer 1

This is an online questionnaire survey study. However, the items assessing the respondents’ attitude toward, effect and safety of influenza vaccination were imperfect. Table 1 described the items used for assessing influenza vaccination contained the word “vaccines” but not “influenza vaccines.” Given that this study analyzed the secondary questionnaire-based data on the COVID-19 vaccination acceptance, it seemed impossible for the respondents to understand what “vaccines” on the items indicate, vaccines for COVID-19, influenza, or both? Influenza vaccines had a long history of development compared with COVID-19 vaccines. People may know and trust on influenza vaccines well but feel unfamiliar with COVID-19 vaccines. Therefore, the items with the incomplete definition limited the value of this study.

Response: We would like to thank the reviewer for constructive comments. The paper is based on secondary analysis of COVID-19 vaccination coverage (dependent variable) data and flu vaccination coverage was one of the dependent variables. Regarding the questions related to the attitudes towards the vaccines these questions refers to the vaccines in general. By the use of these questions we attempted to explore the attitude of the participants towards vaccination in general. However, the attitudes towards COVID-19 vaccines were investigated by the use of a separate question. Since our study is based on secondary data and the primary outcome of interest was COVID 19 vaccination coverage, we didn’t assess attitudes of the participants towards influenza vaccination. Nevertheless, as the reviewer pointed out, influenza vaccines had a long history of development and application and could be considered as “old” vaccines. Consequently, we may consider that our questions on the attitudes of participants towards vaccines in general, indirectly may include influenza vaccines too.

Reviewer 2 Report

Review of Manuscript; R0; Determinants of Influenza vaccination coverage among Greek 2  health care workers amid COVID-19 pandemic Authors: Rachiotis et al. Reviewer: Michael B. Steinborn         Background     This study reports aspects of attitude/opinion on vaccination effectiveness in a Greek medical staff sample (mostly academics, physicians) of more than 300 participants. In general, it was aimed to explore predictors of attitude using logistic-regression analyses. To this end, patients were administered with a set of questions, asking for the relevant aspects.      Evaluation     The study is relevant, and the manuscript is well written and concise. Some information (mostly method and design) is lacking and should be provided before publication. The results seem to add meaningfully to existing knowledge of what determines attitudes towards vaccine. My comments are given below.     Line comments     line 22-34 abstract the abstract could benefit from rewriting. Basically, I would recommend providing a brief statement of the goals of the study, the main hypotheses, method and procedure, and give a understandable overview of the main and most important results. Abbreviations should be avoided and statistical details should not be included in the abstract.     line 53-60, methods & design description some important information is lacking concerning aspects of study design (list of questions, counterbalancing, contextual variables, etc.) which is authors refer to is explained elsewhere in more detail (in a previous publication), however, I would recommend to provide some basic information needed to understand what was exactly done here.      line 49-52 (hypotheses) "....Consequently, using the same dataset, we attempted to study possible factors associated with influenza vaccination  coverage among Greek physicians, dentists and 51 pharmacists during the flu season 2020-2021......"-------comment: the hypotheses are too vague and should be formulated more clearly as they determine the implied analyses. It is important to know at least what the outcome variable is?     line 91-93, sample the sample is very specific as it contains mostly academic medical personal. this might be addressed in the discussion later on.      line 95-112, tables tables should be prepared according to APA standards     line 116-221 (statistics) I would suggest providing a table with population statistics for the most relevant variables of interest (including M, SD, skewness, range, etc.), in order to evaluate the basic preconditions of variables such as aspects of sample distribution, etc. In the next step, I would suggest including an inter-correlation table so that one can see and evaluate the correlation between variable at one glance. Finally, the more specific regression-based analyses should be presented. I would suggest my own work (at the same time apologising for it) to make my point more clear at this point (doi:10.1037/pas0000482).      line 116-221 (statistics, score computation) I suggest providing more information about the exact questions and the scoring methods used to aggregate the questions. How many questions were used and how are they scored under one label or concept? Note that the predictor value of likert-type self-report scores depend on their reliability and thus on the number of items underneath each score, which is which important because regression-model outputs largely depend on them (see doi:10.1037/pas0000482, for a methodological tutorial).     line 113-end, discussion The discussion could be improved, I would recommend to present a much deeper elaboration of the findings. I would suggest, (a) presenting a brief overview of the findings, then (b) relate the results to previous empirical results, (c) elaborate on methodological aspects, (d) discuss limitations, (e) provide some recommendations or guidelines that could be derived from the present study.

Author Response

Reviewer 2

Review of Manuscript; R0; Determinants of Influenza vaccination coverage among Greek health care workers amid COVID-19 pandemic Authors: Rachiotis et al. Reviewer: Michael B. Steinborn         Background     This study reports aspects of attitude/opinion on vaccination effectiveness in a Greek medical staff sample (mostly academics, physicians) of more than 300 participants. In general, it was aimed to explore predictors of attitude using logistic-regression analyses. To this end, patients were administered with a set of questions, asking for the relevant aspects.      Evaluation     The study is relevant, and the manuscript is well written and concise. Some information (mostly method and design) is lacking and should be provided before publication. The results seem to add meaningfully to existing knowledge of what determines attitudes towards vaccine. My comments are given below.

 Response: We would like to thank the reviewer for thoughtful and valuable comments which helped us to substantially improve the quality of our manuscript.

    Line comments     line 22-34 abstract the abstract could benefit from rewriting. Basically, I would recommend providing a brief statement of the goals of the study, the main hypotheses, method and procedure, and give a understandable overview of the main and most important results. Abbreviations should be avoided and statistical details should not be included in the abstract.

Response: We have revised the abstract, accordingly. We have added a sentence on research hypothesis and we also omitted statistical details and abbreviations. 

    line 53-60, methods & design description some important information is lacking concerning aspects of study design (list of questions, counterbalancing, contextual variables, etc.) which is authors refer to is explained elsewhere in more detail (in a previous publication), however, I would recommend to provide some basic information needed to understand what was exactly done here.

Response: The secondary nature of our data is a considerable restriction for us to replicate additional methodological details from our previously published manuscript. Please note that we have consulted with the editorial office on this issue. Thank you in advance, for your understanding.

    line 49-52 (hypotheses) "....Consequently, using the same dataset, we attempted to study possible factors associated with influenza vaccination  coverage among Greek physicians, dentists and 51 pharmacists during the flu season 2020-2021......"-------comment: the hypotheses are too vague and should be formulated more clearly as they determine the implied analyses. It is important to know at least what the outcome variable is?  

Response: We would like to thank the reviewer for this very constructive comment. Both in abstract and main manuscript, we have now included a brief description of our research hypothesis as follows: “We formulated the hypothesis that COVID 19 pandemic exerted a positive impact on the influenza vaccination coverage of HCWs”. In addition, we clearly stated that influenza vaccination coverage is the outcome variable (last lines of introduction section).

 

   line 91-93, sample the sample is very specific as it contains mostly academic medical personal. this might be addressed in the discussion later on.

Response: We have addressed this issue in the discussion of the revised form. In particular, the following sentence was added on to discussion section (page 5) of the revised manuscript: “Apart from its regional nature, our sample was very specific since it comprised of physicians, dentists and pharmacists. Both dentists and pharmacists are health care professionals and as role models for their patients they could influence their vaccine uptake. This is very important for pharmacists particularly who are actively engaged in the implementation of national vaccination program.”

      line 95-112, tables tables should be prepared according to APA standards     line 116-221 (statistics) I would suggest providing a table with population statistics for the most relevant variables of interest (including M, SD, skewness, range, etc.), in order to evaluate the basic preconditions of variables such as aspects of sample distribution, etc.

Response: In fact, age was the only numerical variable in our study. Apart from Mean and Standard Deviation, we have included data on skewness and kurtosis (results section; revised version of the manuscript).

In the next step, I would suggest including an inter-correlation table so that one can see and evaluate the correlation between variable at one glance.

Response: To our knowledge, the study of intercorrelations among the independent variable/ predictors are useful for identifying multicollinearity in the regression analysis. However, all but one variables (age) employed in our analysis were qualitative, dichotomous variables.

 

Finally, the more specific regression-based analyses should be presented. I would suggest my own work (at the same time apologizing for it) to make my point more clear at this point (doi:10.1037/pas0000482).      line 116-221 (statistics, score computation) I suggest providing more information about the exact questions and the scoring methods used to aggregate the questions. How many questions were used and how are they scored under one label or concept? Note that the predictor value of likert-type self-report scores depend on their reliability and thus on the number of items underneath each score, which is which important because regression-model outputs largely depend on them (see doi:10.1037/pas0000482, for a methodological tutorial).  

Response: We would like the reviewer for these commends. As we have stated in the manuscript (methods section; page 2): “Participants were asked to rate on a four-point Likert scale the importance, safety and effectiveness of vaccinations (Fully agree/agree/disagree/fully disagree). For the purposes of the logistic regression analysis these four response categories were reduced (Fully agree/agree vs. disagree/fully disagree).” Moreover, three questions were included in the logistic regression model (revised version; results section; last lines): Three questions were found to be significant in univariate analysis (Vaccines, in general, are important for public health; Vaccines, in general, are effective; have you been vaccinated against COVID 19?) and were included in the logistic regression. In the limitations of our study we have included the following sentence with a new reference: “ Furthermore, our questionnaire was short and this may impact on reliability [13]” .

  line 113-end, discussion The discussion could be improved, I would recommend to present a much deeper elaboration of the findings. I would suggest, (a) presenting a brief overview of the findings, then (b) relate the results to previous empirical results, (c) elaborate on methodological aspects, (d) discuss limitations, (e) provide some recommendations or guidelines that could be derived from the present study.

Response: Τhe discussion section has been modified, accordingly.

Reviewer 3 Report

Dear Editor,

Rachiotis and colleagues investigated the predictors of influenza vaccination acceptance in a cohort of about 400 Greek heath care professionals in central Greece. 74% of the cohort was vaccinated against influenza, which was significantly higher than previous reports from Greece. Willingness to be vaccinated against COVID-19 (the study was conducted prior to availability of COVID-19 vaccines) was found to be the only predictors associated with influenza vaccination after adjusting for a few other variables.

I would like to commend the authors for this work and would also like to thank the editor for giving me the opportunity to review this study.

Overall, this is a well-conducted study.

Please see my comments below:

1. Although this is a well-written study, it needs significant English language revision in terms of grammar and syntax so that the authors can present their nice work in a more presentable way
2. It seems that vaccination coverage assessment was central part of this work. Therefore, this may need to be presented as one of the objectives in the last paragraph of introduction.
3. The tables need improvement in terms of presentation
4. Would prefer to see the ORs in the univariate analysis if possible
5. Would be interested to see the role of age as a predictor. Would divide the cohort in to equal groups (e.g. above median age and below median age) and would put this in the univariate analysis (and if significant, in the multivariate analysis).
6. Why dentists and pharmacists were pooled together? Is it because of the sample size?
7. If the sample size is not a problem and if the data are available, would further group physicians in medical vs surgical vs lab specialties
8. Ideally the first paragraph of the discussion need to summarize the main findings, which, per my understanding, are 1) vaccination rate compared to the past and 2) significant predictors of coverage

 

 

Author Response

Rachiotis and colleagues investigated the predictors of influenza vaccination acceptance in a cohort of about 400 Greek heath care professionals in central Greece. 74% of the cohort was vaccinated against influenza, which was significantly higher than previous reports from Greece. Willingness to be vaccinated against COVID-19 (the study was conducted prior to availability of COVID-19 vaccines) was found to be the only predictors associated with influenza vaccination after adjusting for a few other variables.

I would like to commend the authors for this work and would also like to thank the editor for giving me the opportunity to review this study.

 

Overall, this is a well-conducted study.

Response: We would like to thank the reviewer for helpful and valuable comments.

 

Please see my comments below:

 

Although this is a well-written study, it needs significant English language revision in terms of grammar and syntax so that the authors can present their nice work in a more presentable way

It seems that vaccination coverage assessment was central part of this work. Therefore, this may need to be presented as one of the objectives in the last paragraph of introduction.

The tables need improvement in terms of presentation

Response: We have modified table 1 and we have presented ORs and 95% C.I.

Would prefer to see the ORs in the univariate analysis if possible

Response: See previous response, please.

Would be interested to see the role of age as a predictor. Would divide the cohort in to equal groups (e.g. above median age and below median age) and would put this in the univariate analysis (and if significant, in the multivariate analysis).

Response: The mean age of vaccinated is almost identical to the mean age of non- vaccinated. We also transformed age into a dichotomous variable (</45 ;> 45 years), and the vaccination coverage was similar between the two groups (p=0.46).

Why dentists and pharmacists were pooled together? Is it because of the sample size?

Response: They were pooled together because of their similar vaccination rates.

If the sample size is not a problem and if the data are available, would further group physicians in medical vs surgical vs lab specialties

Response: Unfortunately, these data are not available.

Ideally the first paragraph of the discussion need to summarize the main findings, which, per my understanding, are 1) vaccination rate compared to the past and 2) significant predictors of coverage

Response: We have modified the first paragraph of the discussion, accordingly.

 

Reviewer 4 Report

The study of Rachiotis et al. entitled “Determinants of Influenza vaccination coverage among Greek health care workers amid COVID-19 pandemic” the authors evaluated possible factors associated with influenza vaccination coverage among Greek physicians, dentists and pharmacists during the flu season 2020-2021. For this, an online survey was conducted among the members of the Larissa (Thessaly, Central Greece) Medical, Dentists and Pharmacists Associations. The influenza vaccination coverage was 74% (251/340). Acceptance of COVID-19 vaccination coverage was the only factor independently associated with the likelihood of influenza vaccination coverage (OR= 2.06; 95% C.I. = 1.15- 3.67). Finally, the authors reinforce the need to increase influenza vaccine coverage among healthcare professionals and especially among physicians. Theme is especially important given the need for these health professionals to be vaccinated and protected against infectious diseases. But the way it was conducted, as only a questionnaire applied once, leaves the data weakened to draw any kind of conclusion. For these reasons, the manuscript should be rejected.

1) Authors must read and proofread the manuscript carefully, there are double spacing and loose letters in the text.

2) Manuscript has serious flaws, additional experiments needed, research not conducted correctly, since:

2.1) The authors only did one interview.

2.2) This interview was not repeated during the pandemic.

2.3) If the data had been collected at the time the professionals were being vaccinated, they would have been more robust.

3) For these reasons the article should not be accepted.

Author Response

Reviewer 4.

The study of Rachiotis et al. entitled “Determinants of Influenza vaccination coverage among Greek health care workers amid COVID-19 pandemic” the authors evaluated possible factors associated with influenza vaccination coverage among Greek physicians, dentists and pharmacists during the flu season 2020-2021. For this, an online survey was conducted among the members of the Larissa (Thessaly, Central Greece) Medical, Dentists and Pharmacists Associations. The influenza vaccination coverage was 74% (251/340). Acceptance of COVID-19 vaccination coverage was the only factor independently associated with the likelihood of influenza vaccination coverage (OR= 2.06; 95% C.I. = 1.15- 3.67). Finally, the authors reinforce the need to increase influenza vaccine coverage among healthcare professionals and especially among physicians. Theme is especially important given the need for these health professionals to be vaccinated and protected against infectious diseases. But the way it was conducted, as only a questionnaire applied once, leaves the data weakened to draw any kind of conclusion. For these reasons, the manuscript should be rejected.

• Authors must read and proofread the manuscript carefully, there are double spacing and loose letters in the text.

Response: We have corrected and proofread the manuscript.

2) Manuscript has serious flaws, additional experiments needed, research not conducted correctly, since:

2.1) The authors only did one interview.

Response: Actually, this is an online questionnaire study and not an interview one. Questionnaire studies are frequently used for the evaluation of vaccination coverage.

2.2) This interview was not repeated during the pandemic.

 Response: This study has been conducted during December 2020, two weeks prior to the start of the COVID 19 national vaccination campaign. The vaccination of HCWs starts on early October and it is expected that during December has been completed. Consequently, a repetition of the study would not yield significant additional cases of HCWs vaccinated against influenza.   

 

2.3) If the data had been collected at the time the professionals were being vaccinated, they would have been more robust.

3) For these reasons the article should not be accepted.

Round 2

Reviewer 1 Report

The item assessing the attitude toward influenza vaccines was not specific to influenza vaccines. It is impossible for the respondents to understand what “vaccines” on the items indicate. Although the authors explained that this item assessed the general attitude toward all vaccines, it is not correct to consider all the vaccines as the same.

Author Response

The item assessing the attitude toward influenza vaccines was not specific to influenza vaccines. It is impossible for the respondents to understand what “vaccines” on the items indicate. Although the authors explained that this item assessed the general attitude toward all vaccines, it is not correct to consider all the vaccines as the same.

 

Response: The following paragraph has been included in the limitations (discussion section; page 5; revised manuscript): “Last we acknowledge that our work is based on secondary analysis of COVID-19 vaccination coverage, and flu vaccination coverage was one of the dependent variables. Regarding the questions related to the attitudes towards the vaccines these questions refers to the vaccines in general. By the use of these questions we attempted to explore the attitude of the participants towards vaccination in general. However, the attitudes towards COVID-19 vaccines were investigated by the use of a separate question. Since our study is based on secondary data and the primary outcome of interest was COVID 19 vaccination coverage, we didn’t directly assess attitudes of the participants towards influenza vaccination. Nevertheless, influenza vaccines had a long history of development and application and could be considered as “old” vaccines. Consequently, we may consider that our questions on the attitudes of participants towards vaccines in general, may indirectly include influenza vaccines too.

Author Response File: Author Response.pdf

Reviewer 2 Report

This is the revised version of a manuscript I was asked to review earlier. I read the manuscript again and found it improved further (it was already in quite good shape in the first submission). In my view, this work can be accepted as it is, and I have only one very minor comment regarding the presentation of tables that will not take much time to consider (and can be done after acceptance, during the preparation for final publication). Namely, I would suggest presenting tables in classic APA style (e.g., avoiding horizontal lines between data points). My congratulations to this fine work.

Michael B. Steinborn, Universität of Würzburg

Author Response

This is the revised version of a manuscript I was asked to review earlier. I read the manuscript again and found it improved further (it was already in quite good shape in the first submission). In my view, this work can be accepted as it is, and I have only one very minor comment regarding the presentation of tables that will not take much time to consider (and can be done after acceptance, during the preparation for final publication). Namely, I would suggest presenting tables in classic APA style (e.g., avoiding horizontal lines between data points). My congratulations to this fine work.

Response: We would like to thank the author for persisting efforts to improve the quality of our manuscript.  We have revised the tables, accordingly.

Author Response File: Author Response.pdf

Reviewer 4 Report

I still think the results of the online questionnaire study are too fragile to publish, which makes the hypotheses fragile as well. For example, there is no comparative or control group, such as health professionals and professionals in other areas. For these reasons I maintain my decision not to accept the article.

Author Response

I still think the results of the online questionnaire study are too fragile to publish, which makes the hypotheses fragile as well. For example, there is no comparative or control group, such as health professionals and professionals in other areas. For these reasons I maintain my decision not to accept the article.

 

Response: We disagree with the reviewer on this issue. In fact, we have included a control group. Dentists and pharmacists were used as control group vs. physicians. However, occupational group was not proved to be an independent predictor of flu influenza coverage.

Author Response File: Author Response.pdf