Euvichol-plus vaccine campaign coverage during the 2017/2018 cholera outbreak in Lusaka district, Zambia: a cross-sectional descriptive study

Objective To determine the coverage for the oral cholera vaccine (OCV) campaign conducted during the 2017/2018 cholera outbreak in Lusaka, Zambia. Study design A descriptive cross-sectional study employing survey method conducted among 1691 respondents from 369 households following the second round of the 2018 OCV campaign. Study setting Four primary healthcare facilities and their catchment areas in Lusaka city (Kanyama, Chawama, Chipata and Matero subdistricts). Participants A total of 1691 respondents 12 months and older sampled from 369 households where the campaign was conducted. A satellite map-based sampling technique was used to randomly select households. Data management and analysis A pretested electronic questionnaire uploaded on an electronic tablet (ODK V.1.12.2) was used for data collection. Descriptive statistics were computed to summarise respondents’ characteristics and OCV coverage per dose. Bivariate analysis (χ2 test) was conducted to stratify OCV coverage according to age and sex for each round (p<0.05). Results The overall coverage for the first, second and two doses were 81.3% (95% CI 79.24% to 83.36%), 72.1% (95% CI 69.58% to 74.62%) and 66% (95% CI 63.22% to 68.78%), respectively. The drop-out rate was 18.8% (95% CI 14.51% to 23.09%). Of the 81.3% who received the first dose, 58.8% were female. Among those who received the second dose, the majority (61.0%) were females aged between 5 and 14 years (42.6%) and 15 and 35 years (27.7%). Only 15.5% of the participants aged between 36 and 65 and 2.5% among those aged above 65 years received the second dose. Conclusion These findings confirm the 2018 OCV campaign coverage and highlight the need for follow-up surveys to validate administrative coverage estimates using population-based methods. Reliance on health facility data alone may mask low coverage and prevent measures to improve programming. Future public health interventions should consider sociodemographic factors in order to achieve optimal vaccine coverage.

The article is well written and the study appears to have been conducted with adequate quality.The results generated from the study are in line with earlier findings and as such they do bring limited new knowledge to the research field, apart from providing data on the implementation of the cholera vaccine campaign in the districts where the study was conducted.Our position this is supported by the authors´ own conclusion that "The generalizability of these findings to other settings within the country may be limited."Below follows some more detailed comments on the article.

Abstract
The methods did not specify the use of a logistic regression model, including the measures of effects (adjusted odds ratio with 95% confidence intervals).It is therefore pertinent that authors capture the information in the methods (which currently suggests a descriptive analysis).What is rationale for the age group classification?Is there any evidence to support the classification, which is presently too wide?

Introduction
The spelling of the 'world health organization' in lines 29-30 should be corrected: World Health Organization; the same applies to the spelling of GTFCC in lines 8-8 on page 7.There is information on when the first OCV campaign commenced, but nothing about the timeline for the second campaign.While the objective suggests that the study primarily focuses on OCV coverage, the last sentence (lines 36-37) suggests that the coverage of OCV campaigns and its acceptability among the Zambian population were the focus of the study.Authors need to clarify this throughout the manuscript.

Methods
Authors reported that face-to-face interview was conducted in the local language.While this is a good practice, it would help the reader if the authors clarify whether or not the interview questions were administered by the research assistants or interpreted by the survey participants.This information has an implication for information bias if not explicitly described-see details in line 17 on page 9.Under quality assurance, authors stated that householders who were not present during the field were revisited to avoid selection bias.Please specify the number of visits here.

Data processing and analysis
Authors provided a detailed descriptive analysis of the data collected, which is in line with the study's title and specific objectives.However, in the abstract, odds ratios for vaccine acceptance were presented, which I explained appeared to have deviated from the study's objective.Authors need to provide the analytical details in the methods to justify the inclusion of the results.

Ethical considerations
Authors stated that interviews were conducted in the house away from the public interference to ensure privacy for participants.For a household with about 4 persons, how was privacy issue addressed?This section of the methods seems too wordy.Please consider making it more precise.

Results
A total of 1,325 of participants was estimated as adequate to address the research question(s).However, authors stated in the result chapter (lines 29-3-of page 11) that a total of 2,000 participants were recruited into the study, suggesting that the sample size was probably estimated after data collection.If this was not the case, authors need to justify the reason for sampling more participants than necessary.E.g., authors stated in line 34 on page 11 that "a total final sample of 1,691 respondents (84.5%) were included in the final analysis." Again, consider my previous comments on the introduction of risk factors when the methods did not describe the methodological steps.

Discussion
The discussion concludes that the findings from the survey are much in line with findings from other surveys in similar settings.It would be useful if the authors could highlight what new knowledge the study brings to the field of cholera control.
A key finding in the study is that the reported vaccine coverage is considerably higher than the one found in the household survey.There is little discussion on the reasons for the possibly overestimated coverage from the routine reports.Also, the authors just assumes that the data from the survey are valid and correct.We feel there should be an unbiased short discussion on the possible reasons for the discrepancy between the two sets of data.

References
References to websites should contain information on what date the website was accessed.

VERSION 1 -AUTHOR RESPONSE
Reviewer 1 Overall, the manuscript is satisfactorily written.However, 1) in the abstract, please improve the write up of the statistical analysis.Why is fishers exact test used to stratify OCV coverage, where there very few entries in some of the cells?2) the objective of the paper on page 6 could be shortened; 3) under sampling technique, the author could briefly state (range) of how many persons were sampled per household Response 1: We appreciate the reviewer's comments.We have now corrected abstract to read: descriptive statistics were computed to summarise respondents' characteristics and OCV coverage per dose.Bivariate analysis (chi square test) was conducted to stratify OCV coverage according to age and sex for each round.All analyses were conducted in STATA (p<0.05)(page 2) Response 2: We appreciate the reviewer's comment; we have now edited and shortened the objective of the study (page 5) Response 3: We have now included the missing statement under sampling to read: 1 to 6 participants were sampled per households (page 6) Reviewer 2 The article is well written and the study appears to have been conducted with adequate quality.The results generated from the study are in line with earlier findings and as such they do bring limited new knowledge to the research field, apart from providing data on the implementation of the cholera vaccine campaign in the districts where the study was conducted.Our position is supported by the authors´ own conclusion that "The generalizability of these findings to other settings within the country may be limited."Below follows some more detailed comments on the article.Abstract Query: The methods did not specify the use of a logistic regression model, including the measures of effects (adjusted odds ratio with 95% confidence intervals).It is therefore pertinent that authors capture the information in the methods (which currently suggests a descriptive analysis).What is rationale for the age group classification?Is there any evidence to support the classification, which is presently too wide?
Response 1: We appreciate the reviewer's comment.We have now edited the abstract to include use of logistic regression (page 2) Response 2: We appreciate the reviewer's comment on the rationale for the age group classification.We agree that the classification is too wide.Unfortunately, we are unable to edit the classification on the vaccine coverage; this is how the data was collected on using the pr-tested electronic questionnaire which was uploaded on the electronic tablet using ODK Version 1.12.2.We used the correct classification in table 1 on descriptive statistics (page 9).

Strengths and weaknesses
Query: Authors stated that conducting the survey shortly after the OCV campaign minimised recall bias; however, in the methods, they stated that the survey was conducted following the second round of the 2018 OCV campaign.The question is what is the interval between the second round of campaign and collection for this study?This needs to be carefully considered in order to justify the stated strength.

Introduction
Query: The spelling of the 'world health organization' in lines 29-30 should be corrected: World Health Organization; the same applies to the spelling of GTFCC in lines 8-8 on page 7. Response: We appreciate the reviewer's observation.We have made the suggested corrections (Page 4 and 5) Query: There is information on when the first OCV campaign commenced, but nothing about the timeline for the second campaign.Response: We appreciate the reviewer's comment.We have now provided information on the timeline for the second campaign (see page 5) While the objective suggests that the study primarily focuses on OCV coverage, the last sentence (lines 36-37) suggests that the coverage of OCV campaigns and its acceptability among the Zambian population were the focus of the study.Authors need to clarify this throughout the manuscript.Response: We appreciate this observation.The study focused on the coverage of the OCV campaigns; we have now corrected this throughout the manuscript Methods Query: Authors reported that face-to-face interview was conducted in the local language.While this is a good practice, it would help the reader if the authors clarify whether or not the interview questions were administered by the research assistants or interpreted by the survey participants.This information has an implication for information bias if not explicitly described-see details in line 17 on page 9. Response: We have edited this section; it now reads: "Research assistants administered the questionnaire and read out the interview questions during the face-to-face interview which were conducted in the local language".Query: Under quality assurance, authors stated that householders who were not present during the field were revisited to avoid selection bias.Please specify the number of visits here.Response: We appreciate the comment.In the data collection section we mentioned that two visits were made if the householders were not available during the first visits.The household was skipped if respondents were not found or refused to participate during the second visit (page 7)