This paper is a secondary analysis of a Malawian household survey exploring associations of patients who self-reported as having used formal eye care services. It is a useful idea to use this survey data for this purpose, but the author needs to check that they are using the correct source numbers for their statistical analysis and only report the numbers actually surveyed—not the national estimated numbers derived from these.

6. Most of the key elements are covered, but it is poorly written. The following paper might be another useful reference: Tafida A, Kyari F, Abdull MM, et al. Poverty and blindness in Nigeria: findings from the national survey of visual impairment and blindness. Ophthalmic Epidemiol. 2015;22(5):333-41 [doi: 10.3109/09286586.2015.1077259] [Medline: 26395660]

Response: We have rewritten the introduction and included the reference suggested above.

7. Please clarify the aims and objectives of the study. For example, results on reasons for not accessing services are presented, but this is not mentioned beforehand.

8. To better understand the source of data used in the analysis, please explain the sample size calculation and the sampling strategy for IHS. Presumably, households are the last sampling unit; are all eligible individuals normally living in the selected households then interviewed?

Response: We have explained in the text.

9. The paragraph on sampling weights is not clear. Presumably, this reflects the sampling strategy for the IHS so that the findings can be extrapolated to the whole population aged ≥15 years? If so, please explain.

Response: We have included this explanation in the text.

10. Please outline what confusion matrix techniques are for, what the method entails, and the outputs of the analysis.

Response: We have removed this from the article.

11. What symptoms were used in the analysis?

Response: We have indicated symptoms related to eye problems.

12. “Data” is plural: should say “data were....”

Response: We have edited.

13. Define young adults and older adults.

Response: We have edited.

14. I assume Table 1 shows findings extrapolated to the whole country. The results of the IHS should be presented first, followed by extrapolations to the whole country, specifying the weights used, so that the analysis is more transparent.

Response: We have presented the results of IHS without the weighting before extrapolations under Multimedia Appendix 1.

15. Table 1: Put the data into a proper table, as the results are difficult to see at the moment.

Response: We have modified Table 1 and included a write up for the the rest.

16. The results need to be better ordered. At the moment regional differences in prevalence are included in the paragraph on where participants sought care.

Response: We have rearranged.

17. I do not understand the findings in Table 3. Presumably, these are the findings from the confusion matrix technique?

Response: We have removed Table 3.

18. It would be interesting to know whether the nature of the symptoms reported influenced health-seeking behavior. For example, were those who reported loss of vision more likely to access services?

Response: This was beyond the scope of this study, as the data did not cover aspects of vision loss.

19. Please put negligence in inverted commas, as this is a judgmental term that means failure to give enough care or attention to someone or something that you are responsible for.

Response: We have modified.

20. An important reason why chronic conditions were more common than other conditions, which are often short-lived, is because of the cross-sectional nature of the study and the short 2-week period over which participants were to report eye conditions.

Response: We have included this suggestion in the text.

21. Follow instructions to authors. For example, how many authors should be quoted (ref 1 in particular)?

Response: We have corrected it.