Review Reports

Reviewer  comments

Changes 

In section Methods in subsection `2.3. Sampling method` / in `2.3.2. Sample size calculation` stated that `target group included 1200 persons` (on Line 105), and then described in detail `2.3.2. Sampling process`. However, the specified target group of 1,200 people was not reached in this work, because it was stated in the Results section that this study included 906 people.
Additionally, nowhere in the paper is it explained why the estimated size of the target group was not reached.

Also, if this study was conducted as stated in the text (on `2.3.2. Sampling process`, on Lines 114-126), the question is why the estimated size of the target group was not reached.
Does the stated description of the sampling methodology in this study correspond to what was realized during the study?
Why was the study not carried out until the estimated size of the target group was reached?
Neither the `Participation Rate` nor the `Response Rate` are mentioned in the paper. How was the issue of missing answers resolved? Do the people who refused to participate in the study differ in characteristics from the participants in the study?
Limitations of this study do not mention the issue of the size of the target group at all.

Our target population was 1200, with a minimum number of included participants of 900. We included 906 participants.  We have added the following “As the expected size of each group’s total target population in the PPHCs was between 10,000 and 100,000, our target group included 1200 persons (with a minimum number of 900 individuals and an assumption of 35% non-response rate).”

In the Abstract of the paper, in the Results, Discussion and Conclusions sections, some of the results are stated, described and interpreted completely incorrectly. For example, I quote `Working in rural compared to urban areas was associated with good knowledge (aOR=1.7, 95%CI=1.0-2.8) and positive attitude (aOR=4.6, 95%CI=1.1-19.2) to COVID-19 vaccination.`, on Lines 25-27, is not a correct wording, etc.

We have revised and corrected as follow :”Working in urban compared to rural areas was associated with good knowledge (aOR=1.57, 95%CI=1.12-2.21) and positive attitude (aOR=4.94, 95%CI=1.19-20.44) to COVID-19 vaccination”

In the Abstract of the paper, in the Results, Discussion and Conclusions sections, some of the results are stated, described and interpreted completely incorrectly. For example, I quote `Working in rural compared to urban areas was associated with good knowledge (aOR=1.7, 95%CI=1.0-2.8) and positive attitude (aOR=4.6, 95%CI=1.1-19.2) to COVID-19 vaccination.`, on Lines 25-27, is not a correct wording, etc.

Working in urban compared to rural areas was associated with good knowledge (aOR=1.57, 95%CI=1.12-2.21) and positive attitude (aOR=4.94, 95%CI=1.19-20.44) to COVID-19 vaccination

The discussion of the paper is very scarce, without a detailed analysis of the correlation between the described characteristics of subjects and KAP scores.  

We have included the standard sections of a manuscript discussion, including the summary of main results, interpretation and comparison with the literature, strengths and limitations, recommendations and conclusions. We have added some sentences to the discussion as requested by other reviewers. We have also provided a discussion of the factors associated with KAP such as the region, years of professional experience etc, and compared with findings from the literature.

Reviewer 

Changes 

Line 56 Reference 9 did not open the file.

Thank you for noting this, we have corrected it.

Parisi A, Regazzi L, Spanaus ES, Valz Gris A, Cadeddu C, Moore M, et al. Impact of COVID-19 pandemic on vaccine hesitancy and sentiment changes: A survey of healthcare workers in 12 countries. Public Health. 2025 Jan 1;238:188–96.

Line 78 Present the complete interview form in the appendix.

The complete form of the interview is in appendix A

Line 84 Explain why other professional categories that provide care to people with COVID-19 were not included in the study.

We aimed to include HCWs with direct contact with patients, because they are directly included in recommending the vaccines for patients, and therefore could influence COVID-19 vaccine update in the population directly.  However, we agree with the reviewer that this group could potentially be of interest, for example if such caregivers should advise on vaccines after recovery, to help prevent individuals from being infected again in the future. This could be of interest in a future study but was not within the scope of the current paper.

Line 100 Explain the choice of Slovin's formula for sample size calculation, considering that there is information on the number of professionals working in primary care.

In KAPP studies, we lacked robust prior estimates of the KAPP metrics (outcome), therefore we used Slovins formula  to obtain the minimum number of participants.

Line 267 Table 3. Explain why experience < 13 years and ≥ 13 years was used and why those with continuing education and training in vaccines did not opt ​​for this.

Continuing education and training were not part of the Influenza and COVID-19 vaccination strategy.  Our sample included nurses, physicians and other HCWs. Published health-workforce literature frequently divides careers into early (0–5 or up to 10 years), mid-career (10–20 years), and late-career (>20 years). Based on these commonly used ranges, our selected cutoff of 13 years corresponds to the transition between early/intermediate and mid-career stages. To clarify this we have added “A threshold of 13 years was used to indicate the transition between early career and mid-career transition in HCWs (16,17).

Line 267 Table 3. Explain why the age of the professionals was categorized into age groups < 40 years and ≥ 40 years, considering that older individuals have a higher risk of having a non-communicable disease. Having a non-communicable disease was one of the variables analysed.

At 40 years of age, many individuals have passed early-career / early-adult challenges; they may have established family, stable employment, or long-term commitments. Thus 40 can mark a transition between “young / early adulthood” and “full adulthood / middle adulthood.” This type of social-age rationale is used sometimes in life-course studies.

Line 288 Table 4. Explain why good knowledge and perception, positive attitude, and good practices of health professionals regarding COVID-19 vaccines, and occupation were presented as doctors and others, when other professional categories were interviewed.

All HCWs were included in the study but physicians were the only category to prescribe treatment, and most likely to recommend the vaccines (also decide the contraindication for vaccinations).  For this part of the analysis, we compared doctors with other professional categories.

Reviewer  comments

Changes 

Line 16-18: This is just the study’s objective. Briefly talk about the background.

Thank you for the suggestion. Considering the 250 word limit we have added one sentence to the background. To keep to the word limit we have edited other parts of the abstract.

“HCWs in primary care settings play a significant role in recommending vaccines to patient. Therefore, it is relevant to explore the factors influencing the attitudes, perceptions and practices regarding COVID-19 vaccination.”

Line 19: February 2024, please include the specific date.

We have added the dates “29^th  January to 3^rd February 2024”

Please express all the percentage as follows: e.g.: 37.7% (342/906)

Changes have been made for all the percentages in the manuscript as requested (in red).

Please round off all the results at two decimal places

Changes have been made for all the percentages in the manuscript (in red).

Line 35-37: “Tailored interventions……”, this is a general recommendation. Please be specific based on the study results, what type of evidence-based intervention do you recommend?

We agree that we need to be more specific, we have added “Targeting interventions, particularly in HCWs with less professional experience working in rural settings to increase good practices among HCWs are needed to help improve COVID-19 vaccination coverage in Tunisia. Importantly, future research should explore reasons behind vaccine hesitancy in these subgroups to help tailor interventions that address these.”

For example, a Belgian study ( PMID: 34066476 ) revealed that concerns about potential side effects and the impression that the vaccine was developed too quickly were the main reasons against COVID-19 vaccination in hospital staff and that key factors in the decision to receive a vaccine were data on vaccine efficacy and safety, and knowing that vaccination went well in others. Randomized control trials are needed to determine whether interventions tailored to subgroups that target specific reasons for vaccine hesitancy are effective in increasing uptake.

Please include the epidemiological profile of COVID-19 in Tunisia, with main focus on healthcare workers.

Thank you for this good suggestion. We have added some relevant information. “As of October 2023, the ONMNE reported 1,156,613 COVID-19 cases and 29,494 deaths in Tunisia (10). Five major outbreaks were reported between September 2020 and November 2025, with a Crude Mortality Rate (CMR) of 247,9 per 100,000 (11) . Between March and May 2020, 14.8% of COVID-19 national cases were among HCWs (12).“

Please include references regarding COVID-19 vaccines safety and effectiveness.

   In the introduction we have added “ Several studies have proven, however,  the safety and effectiveness of COVID-19 vaccines, especially among immunocompromised patients (16,17). In Tunisia, COVID-19 vaccines were also found to be effective in preventing SARS-CoV-2 infections and severe forms of COVID-19 (18).”

Line 127: Data collection, this title should be “Data collection and management.

We have changed as suggested.

Line 177: “aas” should be written “as”

We have corrected this.

How did you deal with data confidentiality?

Personal data was not collected and HCWs were given an ID unique identifier. To clarify this we have added “and we did not collect any personally identifiable information such as name, family name, or date of birth.”

In the data analysis section, please state how did you check multicollinearity between variables?

We have added this information in the methods “Multicollinearity was assessed using Variance Inflation Factor (VIF) (values >5 indicates collinearity), Tolerance and Condition indices (values >30 indicates multicollinearity).”

Line 192: using T-Student test for differences in means. This is for normal distribution. What about skewed distribution? Which parametric test was used?

We have clarified this by adding “Student’s t-test for normally distributed continuous variables, and Mann–Whitney U test for skewed distributions.”

In the data analysis section, how to you think that the multivariable logistic regression model fits to the data?

We have added “Hosmer-Lemeshow Goodness-of-Fit Test was used to assess the model fit. A non-significant result (p >0.05) indicated good model fit.”

Please provide the ethical approval number in the section of ethical considerations

We have added the relevant information “Approval number 01_KAPP and date of approval 23 January 2024”

Please express all the percentage as follows: e.g.: 37.7% (342/906)

Changes have been made for all the percentages across the manuscript (in red)

The way of presenting table 3 and 4 is not effective. It is much better to combine table 3 and 4, including two sections: univariable and multivariable logistic regression (column). Each of these sections should report the odd ratios (crude or adjusted), 95%CI, and P-values.  In the rows, include Good knowledge and perception, Positive attitude, and Good practice

Thank you for this excellent suggestion, we have merged the tables as requested.

Please improve the way of reporting the results by including figure such as coeplot (reporting aOR and 95%CI) for each statistically variable in the multivariable logistic regression.

Thank you for this suggestion. However, the odds ratios are already presented in Table 3 and it is usually not necessary to provide the same data in both  a table and a figure. We have produced a figure and ask the reviewers and editors to decide if it should be added. However, we feel it is important to present the exact numbers for all odds ratios, so we prefer to keep the table. This allows for the reader to see the exact numbers and, for example, enables meta-analyses to be conducted etc.

Line 315-320: Please look for pre-and post-COVID-19 studies and compare the study's findings to those studies. E.g: Gabriel ANA, Wang XY, Jamil L, Shimbre MS, Bikorimana G, Zhao L, Cao WC. Knowledge, attitude, and perception towards COVID-19 vaccinations among the adults in Rwanda: a cross-sectional study. BMC Public Health. 2024 Jul 17;24(1):1919. doi: 10.1186/s12889-024-19082-9. PMID: 39020329; PMCID: PMC11256467.

Thank you for the reference. However, we are not sure how to address this point, as COVID-19 vaccines were not available pre-COVID-19.

Please highlight all the limitations related to observational studies.

Although observational studies provide valuable descriptive results, causal inferences cannot be made and there may be unmeasured or residual confounding,  selection bias and information bias .

Please included evidence-based recommendations based on the study’s results.

In light of these findings, we highlight the importance of strengthening HCWs’ knowledge through supportive policies and clear communication, and enabling good practice through adequate resources and leadership. Continuous education and training programs may be particularly beneficial for less experienced HCWs to improve their  knowledge. Targeting regions and areas with lower KAP metrics is crucial while implementing vaccination strategies and campaigns. We have added references in the manuscript supporting our recommendations.

The quality of English language should be improved.

One of the coauthors is a native English speaker with extensive scientific writing and editing experience. She has reviewed the manuscript thoroughly and made adjustments where needed.

Reviewer  comments

Changes 

Slovin's formula is used to determine sample size. But this formula is ambiguous (at best). It is not clear to what the parameter e refers: a level of significance, or the confidence interval, or something else? In any case this parameter has meaning only with reference to a particular variable, of which the variance (or for binary variables, the proportion) is known or can be predicted. See e.g. Pourhoseingholi et al. (2013) Sample size calculation in medical studies, Gastroenterol Hepatol Bed Bench. 2013 Winter;6(1):14–17.  In fact Slovin's formula lacks credibility (no reference is given), see Slovin's Formula: What is it and When do I use it? - Statistics How To . Given this, reference to this formula detracts from the scientific worth of the article, rather than supporting this. It might therefore be better to omit it, and write that the sample size was determined by practical feasibility or convenience (see also below).

As there are different outcomes in our study (KAPP metrics) and we do not know the populations’ behavior, we opted for the Solvins formula for the sample size calculations.

In any case, such sample size calculations (with valid formulas) typically assume simple random sampling, while the sample used here is a (clustered) convenience sample. This also means that statistical inference is in principle not appropriate. I am aware that many  published studies use statistical inference with non-random samples. Nevertheless, the reader should be warned that for this reason the reported confidence intervals should be taken with some grains of salt. Also, the fact that the study uses a convenience sample should be mentioned as a limitation.

The convenience in the sampling did not include all the sampling method, but only the last stage, ie the choice of the centre in case we did not reach the sample size needed in that governorate.

The first step of the sampling was the random choice of the governorate (9/24). Because of logistical constraints (we had 10 investigators who conducted data collection over  6 days). The second stage was the random  selection of  health district  from the largest health districts in the governorate. All the PPHC from the health district were included. The convenience choice of the PPHC was made when the number of HCWs was not reached and then we needed to choose the next biggest PPHC – at this level we opted for the closest  PPHC. We have added this limitation to the discussion “Another limitation of our study that the sampling method, as we chose the most convention PPHC when the expected number was not reached due to logistical constraints.”

The finding that few health workers have a positive attitude towards Covid-19 vaccination is worrying and also surprising, seeing the high vaccination rate in Tunisia (compared to other North-African countries). This could perhaps be discussed a bit more extensively. For instance, Al-Kassim Hassan M, Adam Bala A, Jatau AI. Low rate of COVID-19 vaccination in Africa: a cause for concern. Ther Adv Vaccines Immunother. 2022 Mar 24, suggest that the low rate of vaccination in Africa is due to the feeling that other health issues are more important and to misinformation about Covid-19 vaccination. Are these factors relevant for Tunisia also? It struck me that more health care workers would accept an influenza vaccination (Table 2) so the bad attitude towards the Covid-19 vaccination may be specific to that vaccin. Please comment on this.

We agree with the reviewer on this interesting point – the positive attitude is very low and reasons behind it should be explored. We plan to investigate reasons behind the negative attitudes toward COVID-19 vaccines in a future study but it was beyond the scope of the current study where are objectives were to examine characteristics associated with different KAPP scores. We have added some reflection on this point. The reference by  Al-Kassim Hassan is an opinion letter so it unfortunately does not include any data to back up the opinion that “the low rate of vaccination in Africa is due to the feeling that other health issues are more important and to misinformation about COVID-19 vaccination” so we have not added it to our manuscript. We added this section to the discussion:

“Interesting, our study showed that vaccine uptake declined in HCWs after the pandemic; 96.47% received a COVID-19 vaccine when they were introduced in the country but only 35.21% took a booster dose and less than 5% took a booster vaccine in the 2023-24 season. This raises concerned for willingness to receive COVID-19 vaccines in future seasons. The reasons behind vaccine hesitancy should be explored in more depth in future studies. Information is currently sparse in Tunisia  and other countries in the region, although one study reported that COVID-19 vaccine hesitancy was substantial among Africans and was based on perceived risk of coronavirus infection and past experiences, including knowing someone who had experience a series side effect from a vaccine in the past (REF - PMCID: PMC8635331  PMID: 34851998)”