Seeing Inequalities: Self-Reported Visual Difficulty and Eye Care Utilization Among Women in Malawi, a Secondary Analysis of the 2024 Demographic and Health Survey Data

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

General Comments

This manuscript presents a secondary analysis of the 2024 Malawi Demographic and Health Survey to investigate predictors of self-reported visual difficulty and the use of spectacles or contact lenses among women of reproductive age. The topic is relevant from both public health and global eye health perspectives, particularly given the limited epidemiological evidence available from low-income countries. The use of a nationally representative dataset and the large sample size constitute important strengths of the study.

However, although the dataset is valuable, the manuscript has several conceptual and methodological weaknesses that substantially limit the strength of its conclusions. Most importantly, the study evaluates self-reported visual difficulty rather than objectively measured visual impairment, yet much of the manuscript, including the Introduction, Discussion, and Conclusions, interprets the findings as if they reflected actual visual impairment. Furthermore, several potentially important confounding variables are unavailable or not considered, and the interpretation of spectacle use as a proxy for access to eye care is often overly simplistic. Finally, the Discussion requires a more critical interpretation of the findings and a clearer distinction between association and causation.

 

Introduction

-The Introduction provides an adequate overview of the global burden of visual impairment and emphasizes the importance of equitable access to refractive correction. The rationale for investigating visual health inequalities among women in Malawi is appropriate and timely.  Nevertheless, the Introduction would benefit from a clearer distinction between visual impairment, self-reported visual difficulty, and uncorrected refractive error. These concepts are used almost interchangeably throughout the manuscript despite representing different clinical entities. The DHS questionnaire does not diagnose visual impairment or refractive error, but rather captures participants' subjective perception of visual functioning. Consequently, several statements imply a level of clinical certainty that is not supported by the available data.

 

-In addition, the authors should better justify why reproductive-aged women were selected as the exclusive study population. While the DHS dataset focuses on women of reproductive age, the Introduction should explain why this specific subgroup deserves separate investigation beyond data availability.

 

Methods

-The use of a nationally representative survey and an established sampling framework represents a major strength of the study. The statistical analyses are generally appropriate for the study objectives, and the regression models are correctly described.  However, several methodological concerns should be addressed.

First, the study relies entirely on self-reported visual difficulty obtained from a single questionnaire item. This variable cannot distinguish between refractive error, cataract, retinal disease, neurological disorders, presbyopia, or temporary visual complaints. Therefore, the manuscript should avoid attributing the observed associations to specific ocular conditions unless supported by additional evidence.

 

-Second, several potentially important confounding variables are absent from the regression models. Household wealth index, employment status, health insurance, healthcare accessibility, previous eye examinations, and chronic systemic diseases are well-known determinants of both visual health and healthcare utilization. Their omission should be discussed more explicitly.

 

-Third, spectacle and contact lens use are analysed together as a single outcome. Given that contact lens wear is expected to be extremely uncommon in Malawi, combining both modalities provides little additional information and may even be misleading. Reporting spectacles alone would likely provide a more meaningful public health outcome.

 

-Finally, although the authors appropriately acknowledge unstable regression estimates for some outcome categories, the presence of extremely sparse categories ("cannot see at all") raises the question of whether collapsing outcome categories would have produced more robust and interpretable models.

 

Results

-The Results are clearly presented, and the tables provide a comprehensive description of the sociodemographic characteristics and regression analyses. The large sample size allows reasonably precise estimates for the main outcomes.  Nevertheless, several findings require more cautious interpretation.

The regression models identify statistically significant associations between age, education, residence, region, and self-reported visual difficulty. However, the observed odds ratios are interpreted as determinants of visual impairment rather than determinants of perceived visual difficulty. These are not necessarily equivalent.

 

-The district-level comparisons are interesting from a descriptive perspective but should be interpreted carefully. The explanation that higher spectacle use in some districts may simply reflect sunglasses or fashion eyewear is entirely speculative and is not supported by the available dataset. Alternative explanations, including differential access to eye care services, socioeconomic variation, or random sampling variability, should also be considered.

 

-Furthermore, the manuscript repeatedly discusses spectacle use as evidence of access to eye care. However, spectacle wear is influenced by numerous factors, including affordability, awareness, cultural acceptance, previous diagnosis, and individual visual needs. These complexities deserve greater consideration when interpreting the findings.

 

Discussion and Conclusions

-The Discussion appropriately places the findings within the broader context of visual health inequalities in sub-Saharan Africa and highlights several important public health implications. The authors also acknowledge some limitations, including the cross-sectional design and reliance on self-reported data.

However, the Discussion frequently overinterprets the results. Throughout the manuscript, self-reported visual difficulty is discussed as though it represents objectively confirmed visual impairment. This distinction should be maintained consistently, as subjective perception of vision may be influenced by cultural, educational, psychological, and socioeconomic factors in addition to ocular disease.

 

-Moreover, several causal interpretations should be moderated. The study demonstrates associations rather than causal relationships, yet some sections imply that improving access to spectacles would directly reduce the observed inequalities. While this is plausible, the present data cannot establish causality.

 

-Finally, the Conclusions should better reflect the limitations of secondary survey data. The results identify demographic correlates of self-reported visual difficulty rather than determinants of objectively measured visual impairment.

 

Minor Comments

1. Throughout the manuscript, please replace "visual impairment" with "self-reported visual difficulty" whenever referring to the study outcome, unless objective clinical data are available.
2. The title could more clearly indicate that the study evaluates self-reported visual difficulty rather than clinically diagnosed visual impairment.
3. Please clarify why contact lenses and spectacles were analysed together despite the expected very low prevalence of contact lens wear in this population.
4. The Discussion would benefit from incorporating additional literature on the validity of self-reported vision compared with objectively measured visual acuity.
5. Several regression coefficients for the "cannot see at all" category are unstable due to sparse data. Consider simplifying the outcome categories or removing this category from the regression analyses.
6. Figure 1 is informative but would benefit from the inclusion of confidence intervals or sample sizes for each district, as some apparent differences may simply reflect small numbers.
7. The explanation that spectacle wear may reflect the use of sunglasses or fashion eyewear is speculative and should either be supported by evidence or removed.
8. Please discuss the absence of important socioeconomic variables (e.g., household wealth index, occupation, healthcare accessibility) that may confound the reported associations.
9. A final language revision is recommended to improve grammar and eliminate several repetitive expressions throughout the manuscript.
10. The Conclusions should be slightly more conservative and avoid implying causal relationships between demographic factors and visual difficulty. 

Author Response

Response to Reviewer Comments

Title: Seeing Inequalities: Self-reported Visual Difficulty and Eye Care Utilization Among Women in Malawi, a secondary analysis of the 2024 Demographic and Health Survey data

Reviewer 1

General Comments

This manuscript presents a secondary analysis of the 2024 Malawi Demographic and Health Survey to investigate predictors of self-reported visual difficulty and the use of spectacles or contact lenses among women of reproductive age. The topic is relevant from both public health and global eye health perspectives, particularly given the limited epidemiological evidence available from low-income countries. The use of a nationally representative dataset and the large sample size constitute important strengths of the study.

However, although the dataset is valuable, the manuscript has several conceptual and methodological weaknesses that substantially limit the strength of its conclusions. Most importantly, the study evaluates self-reported visual difficulty rather than objectively measured visual impairment, yet much of the manuscript, including the Introduction, Discussion, and Conclusions, interprets the findings as if they reflected actual visual impairment. Furthermore, several potentially important confounding variables are unavailable or not considered, and the interpretation of spectacle use as a proxy for access to eye care is often overly simplistic. Finally, the Discussion requires a more critical interpretation of the findings and a clearer distinction between association and causation.

Response: We thank the reviewer for their positive and constructive comments. We appreciate the recognition of the importance of the research topic, the use of a nationally representative dataset, and the strengths associated with the large sample size. We have carefully considered all of the reviewer's comments and have revised the manuscript accordingly. Specifically, we have clarified the methodology, strengthened the interpretation of the findings, expanded the discussion of the study's limitations and public health implications, and addressed each point raised in detail below. We have also edited the entire manuscript to ensure that we interpret the findings as self-reported findings and not actual visual impairment. We have also discussed potential confounding variables that have not been measured in the data and included a statement in the limitations section regarding the effect of these factors.

Line 312-316: Second, the data analysed lacked information on household wealth index, occupation, health insurance, eye-care accessibility, previous eye examinations and chronic systemic diseases with ocular manifestations like diabetes and hypertension. These unmeasured factors may have confounded the observed associations, particularly with BMI.

 

Introduction

-The Introduction provides an adequate overview of the global burden of visual impairment and emphasizes the importance of equitable access to refractive correction. The rationale for investigating visual health inequalities among women in Malawi is appropriate and timely. 

Response: We thank the reviewer for this favourable assessment.

Nevertheless, the Introduction would benefit from a clearer distinction between visual impairment, self-reported visual difficulty, and uncorrected refractive error. These concepts are used almost interchangeably throughout the manuscript despite representing different clinical entities. The DHS questionnaire does not diagnose visual impairment or refractive error, but rather captures participants' subjective perception of visual functioning. Consequently, several statements imply a level of clinical certainty that is not supported by the available data.

Response: Thank you for the comment. We revised the Introduction to distinguish clearly between clinical visual impairment, uncorrected refractive error, and self-reported visual difficulty. We now state that the DHS question captures perceived difficulty seeing and does not measure visual acuity or diagnose visual impairment, refractive error, or the cause of reported difficulty. We have therefore used the term “self-reported visual difficulty” consistently throughout the manuscript when referring to the study outcome and revised the study aim accordingly.

Line 84-86: In this study, we sought to determine the prevalence of self-reported visual disability among women of reproductive age in Malawi using data from the 2024 Malawi DHS. We also explored factors associated with their uptake of spectacle and contact lens correction.

Line 102-106: Self-reported difficulty with vision, classified into five categories: “no difficulty,” “some difficulty,” “a lot of difficulty,” “cannot see at all,” and “don’t know” [25]. These items only capture self-reported visual functioning and do not constitute or distinguish between clinical measures of visual impairment, visual acuity, refractive error or ocular disease.

 

-In addition, the authors should better justify why reproductive-aged women were selected as the exclusive study population. While the DHS dataset focuses on women of reproductive age, the Introduction should explain why this specific subgroup deserves separate investigation beyond data availability.

Response: We thank the reviewer for this comment. We have added a brief justification for focusing on women aged 15–49 years, noting the potential relevance of visual functioning to education, employment, caregiving, and health-service use during this life stage, while clarifying that the findings are not generalisable to men or older women.

Line 73-76: In women of reproductive age, visual difficulty may affect education, employment, caregiving and health service use during this critical life stage. In Malawi, there is limited literature on the uptake of optical correction in general [22,23] and among women of reproductive age.

 

Methods

-The use of a nationally representative survey and an established sampling framework represents a major strength of the study. The statistical analyses are generally appropriate for the study objectives, and the regression models are correctly described. 

Response: We thank the reviewer for this comment.

However, several methodological concerns should be addressed.

First, the study relies entirely on self-reported visual difficulty obtained from a single questionnaire item. This variable cannot distinguish between refractive error, cataract, retinal disease, neurological disorders, presbyopia, or temporary visual complaints. Therefore, the manuscript should avoid attributing the observed associations to specific ocular conditions unless supported by additional evidence.

Response: Thank you. We have clarified in the Methods and Discussion that the DHS item measures self-reported visual difficulty only and does not establish the presence, severity, or cause of clinical visual impairment. We have revised the entire manuscript to use appropriate language to avoid attributing observed associations to specific ocular conditions or causes.

Line 102-106: Self-reported difficulty with vision, classified into five categories: “no difficulty,” “some difficulty,” “a lot of difficulty,” “cannot see at all,” and “don’t know” [25]. These items only capture self-reported visual functioning and do not constitute or distinguish between clinical measures of visual impairment, visual acuity, refractive error or ocular disease.

 

-Second, several potentially important confounding variables are absent from the regression models. Household wealth index, employment status, health insurance, healthcare accessibility, previous eye examinations, and chronic systemic diseases are well-known determinants of both visual health and healthcare utilization. Their omission should be discussed more explicitly.

Response: Thank you for highlighting this important limitation. Indeed, the DHS variables available for our analysis did not include measures of household wealth index, employment status, health insurance, healthcare accessibility, previous eye examinations, or chronic systemic diseases with potential ocular manifestations. We were therefore unable to include these variables in the regression models. We have now stated this limitation more explicitly in the Discussion and acknowledge that their absence may have resulted in residual confounding. Consequently, the observed associations are interpreted as associations adjusted only for the sociodemographic variables available in the dataset, rather than as causal effects.

Line 282-289: Although a multi-country population-based study of adults aged 18-49 years in Ghana, Gambia and Togo found that marital status remained significantly associated with self-reported visual difficulties even after adjusting for sociodemographic factors [36], we note that the relationship between marital status and self-reported visual difficulty may be context specific and influenced by individual, social, cultural or healthcare related factors [23,35], that were not fully captured in this study. Hence, the unexpectedly lower odds associated with widowed and divorced/separated women may reflect residual confounding by these unmeasured factors.

Line 312-316: Second, the data analysed lacked information on household wealth index, occupation, health insurance, eye-care accessibility, previous eye examinations and chronic systemic diseases with ocular manifestations like diabetes and hypertension. These unmeasured factors may have confounded the observed associations, particularly with BMI.

 

-Third, spectacle and contact lens use are analysed together as a single outcome. Given that contact lens wear is expected to be extremely uncommon in Malawi, combining both modalities provides little additional information and may even be misleading. Reporting spectacles alone would likely provide a more meaningful public health outcome.

Response: We thank Reviewer 1 for this observation. In the 2024 Malawi DHS questionnaire, the relevant item collected spectacle and contact-lens use as a single combined response and did not distinguish between the two modalities. Separate data for spectacle use and contact-lens use were therefore not available, and we could not analyse spectacle use alone without misrepresenting the source data. We have clarified this in the Methods section and revised the terminology throughout the manuscript to refer consistently to “reported use of glasses or contact lenses” rather than to “glasses use” alone. We also acknowledge in the limitations that the combined measure does not allow us to determine the relative contribution of glasses and contact-lens use or to assess modality-specific determinants of correction use.

Line 107-108: Use of visual correction was defined as one single item (wearing glasses or contact lenses) and scored as “yes” or “no” [25].

Line 317-318: Third, the DHS item combined glasses and contact lens use; hence, the study is unable to estimate the use of each modality separately.

 

-Finally, although the authors appropriately acknowledge unstable regression estimates for some outcome categories, the presence of extremely sparse categories ("cannot see at all") raises the question of whether collapsing outcome categories would have produced more robust and interpretable models.

Response: We thank the reviewer for this thoughtful suggestion. We acknowledge that the small number of participants reporting "Cannot see at all" resulted in unstable estimates and wide confidence intervals. However, we retained this category because it represents a distinct level of visual difficulty as defined in the Malawi DHS questionnaire and preserves comparability with previous and future studies using DHS visual difficulty measures. We thought that the instability of the estimates in itself reflects the rarity of this outcome in the concerned population. We have highlighted the instability of these estimates in the table footnote. We agree that future MDHS studies with larger numbers of severely visually impaired participants may benefit from combining categories to improve the stability of the estimate. We have explained and included notes on this in the footnotes for Table 4.

Line 190-193: *ORs were extremely large or small with extremely wide or undefined 95% CIs; these estimates are likely unreliable. Sex was also excluded because it has only 1 category (Females). The estimates for the level of education were unreliable and thus excluded. Collapsing the cells with small numbers into one may lead to loss of information, as each visual difficulty category provides unique information, so this was not done. No observations were available for BMI for those who cannot see at all; hence, these cells are empty.

Results

-The Results are clearly presented, and the tables provide a comprehensive description of the sociodemographic characteristics and regression analyses. The large sample size allows reasonably precise estimates for the main outcomes.  Nevertheless, several findings require more cautious interpretation.

Response: We thank the reviewer for the favourable assessment of the manuscript. We address the ensuing issues below.

The regression models identify statistically significant associations between age, education, residence, region, and self-reported visual difficulty. However, the observed odds ratios are interpreted as determinants of visual impairment rather than determinants of perceived visual difficulty. These are not necessarily equivalent.

Response: Thank you. We have revised the Results and Discussion to refer consistently to self-reported visual difficulty rather than visual impairment. We now interpret the regression findings as associations between sociodemographic characteristics and “self-reported difficulty” seeing, not as determinants of clinical visual impairment or of any specific ocular condition. We have also replaced causal language, including “determinants” and “predictors,” with “factors associated with.”

Line 172-178: Using “no difficulty seeing” as the reference, the Multinomial logistic regression model was significant (likelihood ratio χ² = 303.4, df = 28, p < 0.001) and the goodness of fit test did not indicate lack of fit (χ² = 1988.8, df = 2472, p = 1.00) (Table 4). Increasing age was associated with higher odds of reporting “some difficulty” with seeing (OR 1.07, 95% CI 1.06 - 1.09, p < 0.001) and a “don’t know” response (OR: 1.09, 95% CI 1.07 - 1.10, p < 0.001) but was not significantly associated with “cannot see at all.” Compared with women in the Southern region, women in the Northern (OR 2.1, 95% CI 1.5 - 2.9, p < 0.001) and Central regions (OR: 1.6, 95% CI 1.2 - 2.1, p = 0.003) had higher odds of reporting “some difficulty” seeing. Women in the central region also had higher odds of reporting “a lot of difficulty” (OR: 4.3, 95% CI 2.1 - 8.6, p < 0.001) in seeing. Urban residence was associated with 40% odds of reporting “some difficulty” (OR: 1.4, 95% CI 1.1 - 1.9, p = 0.020). Compared with the reference marital-status group, women who were married/living with a partner (OR: 0.3, 95% CI 0.2 - 0.5, p < 0.001), divorced/separated (OR: 0.3, 95% CI 0.2 - 0.5, p < 0.001) or widowed (OR: 0.3, 95% CI 0.1 - 0.5, p < 0.001) had lower odds of reporting “some difficulty”. BMI showed no consistent association with self-reported visual difficulty. Estimates for the “cannot see at all” category were unstable due to sparse data and are likely unreliable.

 

-The district-level comparisons are interesting from a descriptive perspective but should be interpreted carefully. The explanation that higher spectacle use in some districts may simply reflect sunglasses or fashion eyewear is entirely speculative and is not supported by the available dataset. Alternative explanations, including differential access to eye care services, socioeconomic variation, or random sampling variability, should also be considered.

Response: We agree. We have removed the speculative explanation concerning sunglasses or fashion eyewear because the dataset does not provide information on the purpose, type, or clinical appropriateness of spectacles or contact lenses. We now present district-level differences as descriptive findings only. We also note that these differences may reflect unmeasured variation in resource disparities and service delivery gaps, which are typical for many African countries.

Line 210-221: There was substantial variation in the proportions of self-reported visual difficulty across districts, ranging from approximately 0.7% in Chikwawa to 22.5% in Dowa (camps)(Figure 1). Relatively higher percentages were also observed in Mzuzu City (13.7%) and Zomba City (12.0%), whereas Mangochi (1.4%), Zomba (1.9%), and Chiradzulu (1.9%) had lower proportions. Reported use of glasses or contact lenses likewise varied by districts, from 0.6% in Phalombe to 13.2% in Dowa (camps), with medium levels observed in Mzuzu City (11.6%) and Zomba City (10.8%). The district-level patterns of self-reported visual difficulty and glasses/contact lens use did not correspond consistently. For example, Blantyre (2.2% visual difficulty vs 5.0% glasses use) and Zomba (1.9% vs 4.7%) had relatively higher reported glasses/contact lens use despite low self-reported visual difficulty, whereas most of the other regions, including Phalombe (5.0% vs 0.6%), showed the reverse pattern.

Line 304-307: District-level analysis revealed substantial geographic variation in visual difficulty and self-reported use of glasses/contact lenses, reflecting resource disparities and service delivery gaps documented across Africa, and highlighting the need for locally tailored interventions [7,18,19,39].

-Furthermore, the manuscript repeatedly discusses spectacle use as evidence of access to eye care. However, spectacle wear is influenced by numerous factors, including affordability, awareness, cultural acceptance, previous diagnosis, and individual visual needs. These complexities deserve greater consideration when interpreting the findings.

Response: Thank you. We have revised the interpretation of reported use of spectacles or contact lenses to avoid treating it as a direct proxy for access to eye care. We now state that this outcome may be influenced by affordability, awareness, cultural acceptance, previous eye examination or diagnosis, perceived visual need, and availability of services and correction. We have indeed addressed these factors in the discussion.

Line 296-302: Overall, reported use of glasses/contact lenses reflected visual needs and varied by age, education, residence and district. Older women and women reporting greater visual difficulty were more likely to report using glasses or contact lenses. There was also higher reported use in rural areas but lower odds of reported use in urban areas after adjustment. Such discrepancies may relate to variations in factors such as awareness, affordability, cultural acceptance, previous eye examination and perceived individual needs, rather than access alone [38].

 

Discussion and Conclusions

-The Discussion appropriately places the findings within the broader context of visual health inequalities in sub-Saharan Africa and highlights several important public health implications. The authors also acknowledge some limitations, including the cross-sectional design and reliance on self-reported data.

Response: We thank the reviewer for the acknowledgement.

However, the Discussion frequently overinterprets the results. Throughout the manuscript, self-reported visual difficulty is discussed as though it represents objectively confirmed visual impairment. This distinction should be maintained consistently, as subjective perception of vision may be influenced by cultural, educational, psychological, and socioeconomic factors in addition to ocular disease.

Response: Thank you. We have revised the language in the entire manuscript including the Discussion to distinguish consistently between self-reported visual difficulty and clinically measured visual impairment. We now state specifically in the methods that the DHS item reflects respondents' perceived difficulty seeing and does not provide an objective measure of visual acuity, ocular disease, refractive error, or the cause of the reported difficulty. We also acknowledge that responses may be influenced by social, educational, psychological, cultural, and socioeconomic factors, in addition to underlying eye conditions. Accordingly, we no longer interpret reported visual difficulty as evidence of objectively confirmed visual impairment.

Line 102-106: Self-reported difficulty with vision, classified into five categories: “no difficulty,” “some difficulty,” “a lot of difficulty,” “cannot see at all,” and “don’t know” [25]. These items only capture self-reported visual functioning and do not constitute or distinguish between clinical measures of visual impairment, visual acuity, refractive error or ocular disease.

 

-Moreover, several causal interpretations should be moderated. The study demonstrates associations rather than causal relationships, yet some sections imply that improving access to spectacles would directly reduce the observed inequalities. While this is plausible, the present data cannot establish causality.

Response: We agree. We have removed or revised causal language in the Discussion and Conclusions. The findings are now described as sociodemographic factors associated with self-reported visual difficulty and reported use of spectacles or contact lenses. We no longer state or imply that improving access to spectacles or contact lenses would directly reduce the observed inequalities. Instead, we conclude that the findings may inform factors that influence access to eye care and identify vulnerable groups that need targeted strategies to prevent avoidable causes of visual impairment.

Line 320-333:

5. Conclusions

This study shows that self-reported visual difficulty and the reported use of spectacles or contact lenses among women in Malawi are influenced by a complex interplay of age, education, geographic location, and other social determinants. Older and less-educated women appear particularly vulnerable. Reported access to corrective lenses is strongly influenced by educational attainment and varies significantly across regions and districts. Given that a large proportion of visual difficulty is preventable or treatable with relatively simple interventions such as glasses or contact lenses, information on self-reported visual difficulty and use of spectacles will guide public health policies aimed at improving access to affordable eye care and eye health education, particularly in underserved populations. Targeted strategies, including community-based screening, health education, and expansion of subsidised spectacle provision initiatives, are essential to reduce inequalities and address the unmet need for vision care among women of reproductive age and other vulnerable groups.

 

-Finally, the Conclusions should better reflect the limitations of secondary survey data. The results identify demographic correlates of self-reported visual difficulty rather than determinants of objectively measured visual impairment.

Response: We have revised the Conclusions accordingly. They now describe the findings as demographic correlates of self-reported visual difficulty among women aged 15–49 years in Malawi, based on secondary DHS data. The results will inform public health policies that will be aimed at reducing visual impairment.

Line 320-333:

5. Conclusions

This study shows that self-reported visual difficulty and the reported use of spectacles or contact lenses among women in Malawi are influenced by a complex interplay of age, education, geographic location, and other social determinants. Older and less-educated women appear particularly vulnerable. Reported access to corrective lenses is strongly influenced by educational attainment and varies significantly across regions and districts. Given that a large proportion of visual difficulty is preventable or treatable with relatively simple interventions such as glasses or contact lenses, information on self-reported visual difficulty and use of spectacles will guide public health policies aimed at improving access to affordable eye care and eye health education, particularly in underserved populations. Targeted strategies, including community-based screening, health education, and expansion of subsidised spectacle provision initiatives, are essential to reduce inequalities and address the unmet need for vision care among women of reproductive age and other vulnerable groups.

 

Minor Comments

Throughout the manuscript, please replace "visual impairment" with "self-reported visual difficulty" whenever referring to the study outcome, unless objective clinical data are available.

Response: Thank you for these constructive suggestions. We have revised the manuscript as follows:

 We replaced “visual impairment” with “self-reported visual difficulty” whenever referring to the DHS study outcome. “Visual impairment” is now used only when discussing clinically measured vision loss in the broader literature. We have also revised the language of the entire manuscript to ensure that it reflects accurately what was done.

 

The title could more clearly indicate that the study evaluates self-reported visual difficulty rather than clinically diagnosed visual impairment.

Response:  Thank you. We have already included the word “self-reported” in the title. But we have further revised the title to make clear that the study evaluates self-reported visual difficulty and it is a secondary data analysis rather than an assessment and analysis of a clinically diagnosed visual impairment.

Our revised title: “Seeing Inequalities: Self-reported Visual Difficulty and Eye Care Utilization Among Women in Malawi, a secondary analysis of the 2024 Demographic and Health Survey data”

Line 1-4: Seeing Inequalities: Self-reported Visual Difficulty and Eye Care Utilization Among Women in Malawi, a secondary analysis of the 2024 Demographic and Health Survey data

Please clarify why contact lenses and spectacles were analysed together despite the expected very low prevalence of contact lens wear in this population.

Response: We thank the Reviewer for this observation. In the 2024 Malawi DHS questionnaire, the relevant item collected spectacle and contact-lens use as a single combined response and did not distinguish between the two modalities. Separate data for spectacle use and contact-lens use were therefore not available, and we could not analyse spectacle use alone without misrepresenting the source data. We have clarified this in the Methods section and revised the terminology throughout the manuscript to refer consistently to “reported use of glasses or contact lenses” rather than to “glasses use” alone. We also acknowledge in the limitations that the combined measure does not allow us to determine the relative contribution of glasses and contact-lens use or to assess modality-specific determinants of correction use.

Line 107-108: Use of visual correction was defined as one single item (wearing glasses or contact lenses) and scored as “yes” or “no” [25].

Line 317-318: Third, the DHS item combined glasses and contact lens use; hence, the study is unable to estimate the use of each modality separately.

 

The Discussion would benefit from incorporating additional literature on the validity of self-reported vision compared with objectively measured visual acuity.

Response: We have revised the manuscript to clearly distinguish between visual impairment, uncorrected refractive error, and self-reported visual difficulty, emphasising that these are distinct but related concepts. We have clarified that self-reported visual difficulty is a subjective measure of functional vision and may not correspond directly to objectively measured visual acuity or visual impairment.

Line 77-83: Visual impairment, URE and self-reported visual difficulty are related aspects, with uncorrected refractive errors contributing to both visual impairment and difficulties in performing visual tasks. In Demographic Health Survey (DHS) questions, what is captured is the perceived difficulty of seeing. Although such surveys do not measure visual acuity to objectively diagnose and determine the cause of visual impairment, they provide information that can be used for identifying inequalities in access to care for specific health conditions, including vision.

 

Several regression coefficients for the "cannot see at all" category are unstable due to sparse data. Consider simplifying the outcome categories or removing this category from the regression analyses.

Response: We thank the reviewer for this suggestion. We acknowledge that the small number of participants reporting "Cannot see at all" resulted in unstable estimates and wide confidence intervals. However, we retained this category because we thought it represents a distinct level of visual difficulty as defined in the Malawi DHS questionnaire and preserves comparability with previous and future studies using DHS visual difficulty measures. We thought that the instability of the estimates in itself reflects the rarity of this outcome in the concerned population. We have, however highlighted the instability of these estimates in the table footnote and excluded the ones that are extremely unstable, like “Education level”. We agree that future MDHS studies with larger numbers of severely visually impaired participants may benefit from combining categories at the survey design level to improve the stability of the estimate. We have explained and included notes on this in the footnotes for Table 4.

Line 190-193: *ORs were extremely large or small with extremely wide or undefined 95% CIs; these estimates are likely unreliable. Sex was also excluded because it has only 1 category (Females). The estimates for the level of education were unreliable and thus excluded. Collapsing the cells with small numbers into one may lead to loss of information, as each visual difficulty category provides unique information, so this was not done. No observations were available for BMI for those who cannot see at all; hence, these cells are empty.

 

Figure 1 is informative but would benefit from the inclusion of confidence intervals or sample sizes for each district, as some apparent differences may simply reflect small numbers.

Response: We revised Figure 1 to include sample size and 95% confidence intervals for each district as suggested by the reviewer, allowing readers to assess the precision of district-level estimates.

 

The explanation that spectacle wear may reflect the use of sunglasses or fashion eyewear is speculative and should either be supported by evidence or removed.

Response: Thank you. We have removed the speculative statement that reported use of spectacles or contact lenses may reflect sunglasses or fashion eyewear.

Please discuss the absence of important socioeconomic variables (e.g., household wealth index, occupation, healthcare accessibility) that may confound the reported associations.

Response: We expanded the Discussion to acknowledge that the available dataset did not include household wealth index, occupation or employment status, health insurance, healthcare accessibility, previous eye examinations, or chronic systemic diseases with potential ocular manifestations. These unmeasured factors may have confounded the observed associations; therefore, findings are interpreted as adjusted associations based only on the variables available in the dataset, not as causal effects.

Line 286-290…we note that the relationship between marital status and self-reported visual difficulty may be context-specific and influenced by individual, social, cultural or healthcare related factors [23,35], that were not fully captured in this study. Hence, the unexpectedly lower odds associated with widowed and divorced/separated women may reflect residual confounding by these unmeasured factors.

 

A final language revision is recommended to improve grammar and eliminate several repetitive expressions throughout the manuscript.

Response: Thank you. We have conducted a full language revision to improve grammar, reduce repetition, and ensure consistent terminology throughout the manuscript.

The Conclusions should be slightly more conservative and avoid implying causal relationships between demographic factors and visual difficulty.

Response: We revised the Conclusions to use more cautious language. They now describe demographic correlates of self-reported visual difficulty and reported use of spectacles or contact lenses, rather than determinants of objectively measured visual impairment. The Conclusions also state that the cross-sectional, self-reported data do not establish causal relationships.

Line 320-333: 5. Conclusions

This study shows that self-reported visual difficulty and the reported use of spectacles or contact lenses among women in Malawi are influenced by a complex interplay of age, education, geographic location, and other social determinants. Older and less-educated women appear particularly vulnerable. Reported access to corrective lenses is strongly influenced by educational attainment and varies significantly across regions and districts. Given that a large proportion of visual difficulty is preventable or treatable with relatively simple interventions such as glasses or contact lenses, information on self-reported visual difficulty and use of spectacles will guide public health policies aimed at improving access to affordable eye care and eye health education, particularly in underserved populations. Targeted strategies, including community-based screening, health education, and expansion of subsidised spectacle provision initiatives, are essential to reduce inequalities and address the unmet need for vision care among women of reproductive age and other vulnerable groups.

Reviewer 2 Report

Comments and Suggestions for Authors

Major points

1. Many readers may know little about Malawi. The authors should provide an overview of the country (such as its economic situation and educational standards).
2. Among Malawian women, there is a high proportion of individuals who are divorced or separated. This likely has implications for factors such as economic circumstances and stress levels. Have the reasons for this high rate of divorce or separation been identified?
3. The authors state that a strong association was observed between aging and visual difficulty. Since the study participants are aged 49 or younger, presbyopia is more likely to be a significant issue than conditions such as cataracts among age-related ophthalmic disorders. Since presbyopia can often be corrected with glasses or similar aids, appropriate optical correction can alleviate visual difficulties. Are the authors investigating whether the glasses they are using are appropriate?
4. In this study, no clear association was observed between BMI and visual difficulty. Obesity increases the risk of developing diabetes and hypertension, which in turn can lead to eye diseases such as diabetic retinopathy and retinal vein occlusion. These diseases are thought to be involved in obesity-related visual difficulty. Since medical treatment is effective in preventing diabetic retinopathy and retinal vein occlusion, information regarding the history of medical treatment among participants classified as overweight or obese should be added.
5. In the Discussion section, the authors should discuss the number of participants unable to use glasses due to financial constraints, the appropriateness of the glasses currently in use, and the medical treatment status of participants classified as overweight or obese.

Author Response

Response to Reviewer Comments

Title: Seeing Inequalities: Self-reported Visual Difficulty and Eye Care Utilization Among Women in Malawi, a secondary analysis of the 2024 Demographic and Health Survey data

Reviewer 2

Major points

1. Many readers may know little about Malawi. The authors should provide an overview of the country (such as its economic situation and educational standards).

Response: Thank you for this suggestion. We added a brief contextual description of Malawi in the Methods section.

Line 88-90: Malawi is a landlocked country in southern Africa with a predominantly rural population of variable socioeconomic attributes. Its population was approximately 17.5 million at the time of the 2018 census [24] and is projected to reach 22.8 million by 2026.

1. Among Malawian women, there is a high proportion of individuals who are divorced or separated. This likely has implications for factors such as economic circumstances and stress levels. Have the reasons for this high rate of divorce or separation been identified?

Response: Thank you for this observation. In our analytic sample, marital status was recorded as a current-status variable, with divorced and separated women grouped in one category. The DHS data used for this analysis did not contain information on the timing or reasons for marital dissolution, or on its economic and psychosocial consequences. We were therefore unable to determine why the proportion of women classified as divorced/separated was high or to evaluate whether economic circumstances or stress explained the observed association with self-reported visual difficulty.

1. The authors state that a strong association was observed between aging and visual difficulty. Since the study participants are aged 49 or younger, presbyopia is more likely to be a significant issue than conditions such as cataracts among age-related ophthalmic disorders. Since presbyopia can often be corrected with glasses or similar aids, appropriate optical correction can alleviate visual difficulties. Are the authors investigating whether the glasses they are using are appropriate?

Response: We agree that, among women aged 15–49 years, presbyopia is a more plausible age-related explanation for increasing difficulty with near vision than cataract or other later-life ocular disorders. However, the DHS question does not assess near versus distance vision, refractive status, ocular disease, prescription history, or whether the spectacles or contact lenses are clinically appropriate. We have therefore removed causal attribution to presbyopia, cataract, or other specific conditions and now describe age as being associated with self-reported visual difficulty.

1. In this study, no clear association was observed between BMI and visual difficulty. Obesity increases the risk of developing diabetes and hypertension, which in turn can lead to eye diseases such as diabetic retinopathy and retinal vein occlusion. These diseases are thought to be involved in obesity-related visual difficulty. Since medical treatment is effective in preventing diabetic retinopathy and retinal vein occlusion, information regarding the history of medical treatment among participants classified as overweight or obese should be added.

Response: Thank you. We agree that obesity can be associated with systemic conditions that may affect vision. However, the available DHS data used in this analysis did not contain information on diabetes, hypertension, retinal disease, or participants’ medical-treatment histories. We were therefore unable to assess whether these conditions explained the BMI findings or to add treatment information for women classified as overweight or obese. We have mentioned in the limitation section that the above unmeasured chronic disease may have confounded the results and avoid interpreting the absence of an association between BMI and self-reported visual difficulty as evidence that BMI or obesity-related systemic disease is unrelated to visual health.

Line 313-317: Second, the data analysed lacked information on household wealth index, occupation, health insurance, eye-care accessibility, previous eye examinations and chronic systemic diseases with ocular manifestations like diabetes and hypertension. These unmeasured factors may have confounded the observed associations, particularly with BMI.

1. In the Discussion section, the authors should discuss the number of participants unable to use glasses due to financial constraints, the appropriateness of the glasses currently in use, and the medical treatment status of participants classified as overweight or obese.

Response: We agree that financial barriers, the appropriateness of correction, and medical treatment status are important for interpreting reported use of spectacles or contact lenses. Although the DHS variables available for this analysis did not specifically measure inability to obtain correction because of cost, the purpose or prescription, appropriateness of spectacles or contact lenses, prior eye examination, or treatment for systemic and ocular conditions and other confounding factors, we have mentioned in the discussion that these factors can influence the results and include a statement regarding this in the limitation section.

Line 313-317: Second, the data analysed lacked information on household wealth index, occupation, health insurance, eye-care accessibility, previous eye examinations, appropriateness of glasses being worn and chronic systemic diseases with ocular manifestations like diabetes and hypertension. These unmeasured factors may have confounded the observed associations, particularly with BMI.

We have also made minor changes throughout the entire manuscript to ensure consistent use of terminology,  improve the language and grammar to enhance readability and keep the language associative rather than causal.

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

I would like to congratulate the authors of this research for having adequately addressed all the suggested comments. This research has reached a sufficient standard to be accepted.

Reviewer 2 Report

Comments and Suggestions for Authors

Major points

1. O.K.
2. O.K.
3. O.K.
4. O.K.
5. O.K.