Author's response to reviews Title: Sex education in Mumbai: exposure and opinions of adolescent students in Mumbai: A cross-sectional survey Authors:

1. You state that you will ‘explore the potential impact of school based sex education..’ however, given the design of your study this was not possible nor discussed in the paper itself. I suggest that you remove mention of ‘potential impact’ of such education from the abstract as this is misleading. Mention of ‘potential impact was removed from abstract’. Background 2. Para 3, last sentence: You suggest that countries with a long-history of school-based sex education have low teenage birth rates and low rates of STIs among, however, you fail to acknowledge that there are likely to be other cultural, familial, societal and health service factors that influence the rates of pregnancy and STIs in those countries. You should discuss the possibility of other causes of the low rates or remove the sentence. The following sentence was added to acknowledge the possibility of other factors ausing low rates of pregnancy and STIs: ‘However, it is important to acknowledge that other societal, familial, cultural and health service factors exist which may influence rates of teenage pregnancy and STIs in these countries’. Methods 3. Para 1, sentence 2: You state that the state’s government refused to take part in the federal government initiative but did they offer an alternative? Or were there any other groups e.g. NGOs offering sex education to these students? Please provide more details. Paragraph changed to state that it was a single government initiativethe Adolescence Education Programme specifically. 4. Para 1, sentence 3: More information needs to provided here on the selection of these 5 schools i.e. how representative were they of schools in the city? What is the total number of such schools in the city and how were these 5 selected? Also, it is essential that you should provide details of the grades that were involved in the survey and the age-ranges and sex-distributions within those grades. Explanation of how the colleges were selected has been added to methods, paragraph 2. Additionally, the age and sex distributions of students within grades has been added to the results in table 1. 5. Para 3, sentence 2: This explanation of how classes were selected is unclear. How many classes of the relevant grades were in each school and how many were chosen to participate? Any difference between the classes chosen and not chosen? e.g. were the classes streamed for academic ability, by sex etc.? A further explanation has been added to methods paragraph 4 detailing how classes were chosen (by subject studiedstudents were streamed to study an arts, science or commerce routenot by academic ability/sex) and how many classes participated in the study. 6. Para 4, sentence 2: The basis of the sample size calculation is unclear. Please provide the proportion of students exposed to sex education that was used in the calculation. Was the plan to analyse for each sex separately and if so were you aiming for 400 students of each sex? In fact, the need for such a sample size calculation is unclear as you do not present 95% CI for the proportion of students exposed to schoolbased sex education and do not carry out any statistical comparisons e.g. between sexes for this outcome. A reference has been added explaining the sample size chosen. 400 students were needed in total (which has been clarified in the paper). Statistical comparisons have since been added following reviewers’ comments. As no data exists on the proportion of students exposed to sex education in this community the prevalence was assumed to be 50% which provides the greatest sample size that would be needed (see reference for further explanation). 7. Para 4, sentence 4: You mention that you carried out logistic regression yet you do not provide sufficient details of the methods used to create your models e.g. was it forwards stepwise regression? What was the p-value that you used a cut-off when including/excluding variables from your model? In general, it is unclear to me why you wanted to look at factors associated with the students opinions on the importance of having sex education as part of the school curriculum. What is the public health importance of this? And if this was a key question (requiring such statistical analysis) then why was it not one of the aims of your study? (as described in the abstract) I suggest that you clarify the reasons for looking at this association and the statistical methods used or alternatively remove this analysis (and Table 2) from the paper. Taking your comment into consideration (and further comments from other reviewers) this analysis has been removed from the paper. 8. Para 5, last sentence: Please explain why individual written informed consent was not obtained from the students and also state whether students were given the chance to ask questions following receipt of the ‘information statement’ (Methods, para 3). Consent was in the form of informed consent. By completing the questionnaire informed consent was assumed from students; this was in keeping with the Population Sciences and Humanities Internal Ethics Review Committee, University of Birmingham. The following sentence was added to the paragraph: ‘Students were given the opportunity to ask questions about the study’. Results 9. Para 1, sentence 1: Why so few males? (see my earlier suggestion to include age and sex distributions within the relevant grades in the schools) Age and sex distributions within the relevant grades have been included now as table 1. 10. Para 1, sentence 2: Please state the % that returned a completed questionnaire. This has been added. Results paragraph 1. 11. Para3, sentence 1: Was there any information available from the students or their teachers/another source on the content of the sex education lessons that had taken place at the study schools? No information was available from all students as to the content of sex education lesson. This has been discussed as part of the study’s limitations. 12. Para 4, last 2 sentences: See my earlier comments re the usefulness of this logistic regression analysis especially as almost 90% of the students thought that sex education was important. 13. Para 7, last sentence: Such statements where proportions are compared should be accompanied with some indication of whether this was a real difference between sexes, e.g. provide a p-value from a chi-squared analysis. P-value added. Conclusions 14. Para 2, sentence 2: Please provide some indication of whether the study in Dehli took place in the same time period and among a similar age/sex group of participants? Comment added with regards to when the study took place (2007) and that participants were of a similar age. 15. Para 2, sentence 3: A major limitation of this study is the absence of detail on what sex education lessons the students received and so any comparison with the level of ‘sex education’ in other studies and/or the ‘expected/unexpected’ level of ‘sex education’ are futile. For example, the ‘sex education’ that the students are referring to could simply have been one lesson where the teachers told them not to have sex (i.e. not comprehensive sex education). If this were the case then it might not be unexpected that ~62% of students had received such a message. Comment noted and this point has been added to the study’s limitations. 16. Para 3, last sentence: Please see my earlier comment re: this research questionif you are going to retain this analysis in the paper and suggest that further work be carried out on this question then you need to explain why is this important in terms of public health? Analysis removed as has this point in keeping with your former suggestion. 17. Para 4: Please discuss the possibility of biased reporting given that the interviews took place in the school? i.e. Is it possible that students overemphasised the importance of knowledge obtained in school. This has been discussed under limitations of the study, paragraph 1. 18. Para 6: Please discuss the possibility of social-desirability bias e.g. if the survey was led by a doctor then would students be more likely to say that they would like to obtain sex education from a doctor? Issue of social desirability bias discussed under limitations, paragraph 1. Survey was not led by a doctor but by a medical student as part of their dissertation in a degree in International Health. The interviewer introduced herself as a university student studying international health, but not as a medical student. 19. Para12, first sentence: Please provide more details of the ‘robust evidence which suggests that curriculum-based sex education programmes are beneficial in preventing HIV, STIs and early pregnancy in adolescents’. To my knowledge the only study (at least in SSA) that has shown this has been the Stepping Stones study in South Africa which found a decrease in HSV2 (not HIV). See the update to the Steady, Ready, Go review (http://www.evidence4action.org/images/stories/documents/srgreview.pdf and associated article that is due to be published soon (Mavdezenge SN, Doyle AM, Ross DA. HIV prevention in young people in sub-Saharan Africa: A systematic review Adolescent Health In press.) This statement is referenced and evidence to support it can be found in the reference provided: ‘Global perspectives on the sexual and reproductive health of adolescents: patterns, prevention, and potential’a series of articles on adolescent health from the Lancet. General comments 20. A major limiation of this study is the lack of depth of information that was obtained e.g. on previous exposure to sex education and also on what kind of sex education programme students would like to happen in the future. A further limitation in terms of usefulness of the findings stems from the fact that even if students responded to questions such as ‘Do you feel you have good access to the advice you need?’ honestly, they may in fact not know what ‘good and comlete advice’ consists of as they have never received it. These limiations, their implications for interpretation of the findings and the subsequent usefulness of the findings for public health should be throughly discussed in the paper. Further discussion of the potential policy implications of these findings should be provided as it is not clear that the findings from this study provide enough information to inform the design or implementation of sex education programmes in schools. These points have been taken into consideration and have been discussed as part of limitations of the study. • Minor Essential Revisions The author can be trusted to make these. For example, missing labels on figures, the wrong use of a term, spelling mistakes. Done 21. The first sentence of the abstract is currently a fragment and should be made into a proper sentence e.g. The aim of this study was to determine.... This has been amended: ‘The aim of this study was to determine students’ exposure to sex education and identify students’ perceptions of accessibility to sexual health advice and their preferences in implementing sex education.’ 22. In the methods section of the abstract you should mention the age and sex of the students interviewed. The sex and age of participants has been added to the methods section of the abstract. 23. The phrase ‘majority desire’ in the conclusions section of the abstract is unclear and you should edit e.g. the majority of those interviewed indicated their desire for.... These phrase has been changed and now reads: ‘The majority of those interviewed indicated a desire for more widespread implementation of school-based sex education , particularly amongst female respondents.’ Background 24. Para 1: Reference 3I’m surprised that you quote HIV statistics from a 2007 online article. I suggest that you use (and reference) the data from the most recent UNAIDS HIV/AIDS epidemic update that is available online. The reference used was from the World Health Organisation, rather than UNAIDS report, as UNAIDS most recent report was from 2006 and the WHO article from 2007. Regardless, both references indicate the exact same prevalence of HIV. 25. Para1: Reference 4Again I suggest that you update this incidence figure using recent UNAIDS datait is unlikely that the HIV incidence in 1996 is relevant for a study carried out in 2010. This reference quotes STIs other than HIV statistic and does not refer to HIV incidence. This was the most current figure available following a comprehensive literature review on the incidence of STIs. I have kept this reference as the UNAIDS statistics only refer to incidence for HIV. 26. Para 2, sentence 3: Please clarify if you are referring to ‘new’ (incident) or ‘existing’ (prevalent) HIV cases. This has been clarifiedcases are now described as ‘existing cases’. 27. Para 4, sentence 4: Please explain ‘co-curricular’. Co-curricular has been defined as: ‘complementing but not part of the regular curriculum’. 28. Para 4, last sentence: I suggest that you also/instead reference the more recent UNAIDS report http://data.unaids.org/pub/Outlook/2010/20100713_outlook_youngpeople_en.pdf Done. 29. Para 5, last sentence: I suggest that you replace ‘lacunae’ (a less commonly used word) with ‘gap’. The word ‘lacunae’ has been changed to ‘gap’. 30. Para 6, first sentence: I suggest that you update references 3 and 4 (see my earlier comment) Please see response to comments 24 and 25 above. Methods 31. Para 3, 2nd last sentence: I suggest that you replace the word ‘ensure’ with ‘encourage’ as there is no way of ensuring honest responses. The word ‘ensure’ has been changed to ‘encourage’. 32. Para 5, 2nd last sentence: Reference 20 does not seem appropriate for Indian policy on issues of informed consentis there not a government document that states the rules re parental consent? A comprehensive search for a government document that states this was not found. This was stated in the paper by McManus et al published in the BMC Women’s Health 2008 without a reference. Results 33. Para 1, sentence 3: Was religion missing for the other 10 participants or did they state other religions? The religion of the other 10 participants has now been noted in paragraph 1 of the results. Conclusions 34. Para 5, last sentence: Please see my suggestion re a replacement/addition to reference 18. Reference changedsee point 28. 35. Para 7, last sentence: Please check that this is the latest literature on the effectiveness of peer-educators. It might be worth referencing the Steady, Ready, Go! Review carried out by WHO and others which reviews studies in all developing countries: (http://www.who.int/child_adolescent_health/documents/trs_938/en/index.html and the recent update to this review covering studies in SSA only: http://www.evidence4action.org/images/stories/documents/srgreview.pdf) Reference 28 was more recent that above mentioned reviews, except the Steady, Ready Go! Review which does not specifically refer to peer-led interventions, but rather simply school based interventions. 36. Para 9, sentence 2: Insert the word ‘reported’ in this sentence as behaviours were reported not observed and reported behaviour is prone to many potential biases. The word ‘found’ was changed to ‘reported’. 37. Paragraph 13, second sentence: this should be reference 12 not 13 (I think) Reference 13 changed to 12. Tables 38. Table 2: If you are going to retain this table (see my earlier comments) then (i) Please insert the number of students who were in each category (e.g. number who reported prior formal sex education), and within each group the number who thought that sex education was/ was not important. (ii) Please provide details of the model e.g. in a footnote describe what you adjusted for. (iii) Minor suggestion: 2 decimal places is sufficient