Rotavirus vaccine impact assessment surveillance in India: protocol and methods

Introduction Rotavirus infection accounts for 39% of under-five diarrhoeal deaths globally and 22% of these deaths occur in India. Introduction of rotavirus vaccine in a national immunisation programme is considered to be the most effective intervention in preventing severe rotavirus disease. In 2016, India introduced an indigenous rotavirus vaccine (Rotavac) into the Universal Immunisation Programme in a phased manner. This paper describes the protocol for surveillance to monitor the performance of rotavirus vaccine following its introduction into the routine childhood immunisation programme. Methods An active surveillance system was established to identify acute gastroenteritis cases among children less than 5 years of age. For all children enrolled at sentinel sites, case reporting forms are completed and a copy of vaccination record and a stool specimen obtained. The forms and specimens are sent to the referral laboratory for data entry, analysis, testing and storage. Data from sentinel sites in states that have introduced rotavirus vaccine into their routine immunisation schedule will be used to determine rotavirus vaccine impact and effectiveness. Ethics and dissemination The Institutional Review Board of Christian Medical College, Vellore, and all the site institutional ethics committees approved the project. Results will be disseminated in peer-reviewed journals and with stakeholders of the universal immunisation programme in India.


GENERAL COMMENTS
This is a well written piece by Nair et al., which provides an easy to follow protocol for rotavirus vaccine impact evaluation via surveillance. It is clearly essential to conduct a thorough evaluation of a new vaccine programme particularly given the baseline burden of rotavirus disease in India, therefore this an important study.
In line with BMJ Open Protocol publishing guidance the manuscript contains no data or conclusions.
I do have some comments / questions below which I feel will enhance the paper and need clarification.
As per protocol guidance please could the authors add some timelines for the study and the data, how much pre-and post-data will be available for each site?
The authors mention that if there is low or high vaccine coverage (Pg 16 line 52 and Pg 17 line 3-4) there may need to be a change to analytical plan. Presumably in this scenario this would add limitations to the case-control study because of any of the following, for example; herd effect (lack of cases), different exposure risk between vaccinated and unvaccinated or larger sample size needed. In reference to this, can the authors explain why surveillance was not/is not being conducted in areas without vaccination as these states could potentially be used as natural population level geographical controls? Is it because there is state based heterogeneity in the baseline rate of hospitalised rotavirus disease?

GENERAL COMMENTS
This is clearly written and mostly straight-forward. There is a paragraph on ethics and dissemination between the discussion and conclusion which I would have thought should come before the discussion. I have provided some minor comments directly on the attached manuscript which should be fixed or clarified.
The reviewer provided a marked copy with additional comments. Please contact the publisher for full details.

GENERAL COMMENTS
Overall very well constructed protocol and well written manuscript. This protocol describes how post Rotavac (new rotavirus) vaccine introduction will be assessed with respect to impact on all-cause AGE, circulating rotavirus genotypes, and vaccine effectiveness. Some minor points of note: -Is it possible to give an approximate estimate of the number of children affected by rotavirus in India. The first two paragraphs contain a range of %, but an absolute estimate (range) would be good to include -bottom page 5 and later in manuscript authors allude to need to monitor safety, especially impact on intussusception rates. I realise this is not the topic of this study, but mention in Discussion on plan/s to monitor impact on IS would be of interest -pg 7 states all sentinel sites selected had at least 250 diarrhea admissions annually, but This is a well written piece by Nair et al., which provides an easy to follow protocol for rotavirus vaccine impact evaluation via surveillance. It is clearly essential to conduct a thorough evaluation of a new vaccine programme particularly given the baseline burden of rotavirus disease in India, therefore this an important study.
In line with BMJ Open Protocol publishing guidance the manuscript contains no data or conclusions.
I do have some comments / questions below which I feel will enhance the paper and need clarification.
Reviewers comment: 1. As per protocol guidance please could the authors add some timelines for the study and the data, how much pre-and post-data will be available for each site?
Response: Data collection will be for 4 years from each site. Five sites in the study (PGIMS, Rohtak; RPGMC, Tanda; Hitech hospital, BBSR; SVMC, Tirupati and CMC, Vellore) have 3-5 years data of prior to vaccine introduction. All the other sites have data after the introduction of vaccine.
The text now reads "Vaccine impact will be determined by monitoring trends in rotavirus hospitalizations pre-and post-rotavirus vaccine introduction at the five sites that have pre-introduction data [ Table.1]. The other sites will have four years of surveillance after vaccine introduction." In Pg 8 lines 13-14.
Reviewers comment: 2. The authors mention that if there is low or high vaccine coverage (Pg 16 line 52 and Pg 17 line 3-4) there may need to be a change to analytical plan. Presumably in this scenario this would add limitations to the case-control study because of any of the following, for example; herd effect (lack of cases), different exposure risk between vaccinated and unvaccinated or larger sample size needed. In reference to this, can the authors explain why surveillance was not/is not being conducted in areas without vaccination as these states could potentially be used as natural population level geographical controls? Is it because there is state based heterogeneity in the baseline rate of hospitalised rotavirus disease?
In this study, surveillance is not being conducted in areas without vaccination because of state-and site-based heterogeneity in the baseline rate of hospitalized rotavirus disease. This was seen in earlier surveillance and is also being seen in the data being generated in this study, in the postintroduction data from different states and at sites where there is more than one site per state. We have revised the statement which now reads "Rotavirus was also found to cause significant disease burden in children <5 years of age treated for diarrhea as outpatients [12], with a study in Kolkata reporting 48% positivity for rotavirus over 36 months" Page 4 lines 12, 13.
Reviewers comment: 5. Pg 5 line 53. Please make it clear that Rotavac is also an oral vaccine.

Methods
Reviewers comment: 6. The coverage estimates by WHO in India for 2017 for the first dose of OPV/ Pentavalent vaccine is more than 80%, and therefore, we expected similar coverage for oral rotavirus vaccine which is given along with the first dose of OPV/pentavalent. While oral rotavirus coverage has actually been reported as being lower, this is in the early stages of introduction, and we expect coverage rates to rise during the course of the study. Response: We intend to compare the rates of rotavirus positive gastroenteritis among all children enrolled in to surveillance with a collected stool sample. This comparison is for the five sentinel hospitals namely, Vellore, Rohtak, Tanda, Bhubaneswar and Tirupati where the pre-vaccine data is available from 2012 onwards.