Health and Comorbidities in Minority Ethnic Adults Living with Visual Impairment in the UK

Round 1

Reviewer 1 Report

This is a really well written study. Visual impairment impacts a lot of lives and for people who are minorities and may face barriers in seeking care and support, the effect may even be greater. It is a huge credit to the authors for bringing to light how VI is affecting such groups in the UK.

Comment 1

Page 1 line 44-45 The authors provide the report on comorbidity found in Glaucoma patients in Saudi Arabia, but did not indicate if these patients were visually impaired. Unless all the patients with Glaucoma in the study were visually impaired the study may not qualify to be used in this context.

Comment 2

The authors make comparisons between Asia and Black and go ahead to make a separate comparison for MEC and White populations. And sometimes compares all the 4 groups when one group is a composite of two other groups. The study recruited disproportionate number of Asian and Black participants compared to the other minority ethic communities. Since most of the results and conclusions are based on Asian, black and white communities. I suggest the authors leave out MEC as a variable and make comparisons between the 3 groups. It will make it easier for readers to follow the analysis and the conclusions that will be reached.

Comment 3

It was noted that the authors try to satisfy assumptions of the various test by mixing up parametric and non parametric tests. If there is a significant number of data that are are not eligible for parametric statistics, the authors should consider using non parametric statistics throughout especially in table 3, 7 and 9.

Comment 4

In Tables 3,4,7 and 8, it is observed that the authors calculated chi-square with counts for only one of the variables and the other variable have zero count. This virtually invalidates the test. The authors may provide reasons why this should be acceptable.

Comment 5

In table 2 the authors present range as a variable, It is recommended that they remove this from the table and report it in the main text

Comment 6

One of the major selling points of this study was the fact that it covers younger people. However there is very limited focus on the younger population throughout the study. There was no comparison on the impact of VI and comorbidity among different age groups etc

Comment 7

A little suggestion; it will have been interesting to see how VI severity and/ or vision changes impacted self rated health, physical exercise etc.

Author Response

First of all, we would like to thank you for taking the time to review our manuscript and your very much constructive comments. We have reviewed these and made changes in the revised Word document. Any revisions have been made in track changes to help you locate them. We further provide our responses to each comment below. Please note that we have also made changes to the manuscripts in its entirety to improve its readability. For instance, we have revised the sentence structures in the Conclusion. We hope this resolves issues around English language.

Comment 1

Page 1 line 44-45 The authors provide the report on comorbidity found in Glaucoma patients in Saudi Arabia, but did not indicate if these patients were visually impaired. Unless all the patients with Glaucoma in the study were visually impaired the study may not qualify to be used in this context.

 Instead of removing this reference, we have added the following information about the sample in this article to our manuscript for the information or readers:

 It should be noted that not all patients in this study had V.I.: 16.0% had normal vision, while 19.2% had near normal vision and 64.8% had moderate V.I. to legal blindness. 

Comment 2

The authors make comparisons between Asia and Black and go ahead to make a separate comparison for MEC and White populations. And sometimes compares all the 4 groups when one group is a composite of two other groups. The study recruited disproportionate number of Asian and Black participants compared to the other minority ethic communities. Since most of the results and conclusions are based on Asian, black and white communities. I suggest the authors leave out MEC as a variable and make comparisons between the 3 groups. It will make it easier for readers to follow the analysis and the conclusions that will be reached.

This article forms part of a series of articles which use a similar analysis but focus on different topics. We believe that the comparison of the combined MEC group to matched control sample of White adults is useful to retain. For one, it provides a comparison of two equally sized groups who have been matched on age, gender and geographical location, thus identifying possible ethnic group differences. A similar comparison of Asian, Black and White participants, although interesting, would be less robust due to the marked differences in subgroup sample sizes and the lack of matching on potential confounders. We do appreciate that having four groups and two comparisons impacts on the narrative flow of the paper, and we have restructured the text to separate out as much as possible the results presented for the MEC and White groups from the results presented for the Asian and Black groups.

Comment 3

It was noted that the authors try to satisfy assumptions of the various test by mixing up parametric and non parametric tests. If there is a significant number of data that are are not eligible for parametric statistics, the authors should consider using non parametric statistics throughout especially in table 3, 7 and 9.

Considering the small subsample sizes of Asian and Black participants, we have revised the results presented in Tables 3, 7 and 9 to show results of Fisher’s exact tests only. We took fewer than 3 conditions meeting the test assumptions for chi-square test and we have added the following explanation in the Statistical analysis section:

Where test assumptions for chi-square tests relating to expected frequencies were violated, Fisher’s exact tests were calculated in R. Fisher’s exact results only were calculated to compare prevalence of eye and physical health conditions among Asian and Black participants because fewer than 3 of the individual conditions met test assumptions for chi-square tests.

As a result, we did not update the results for the group comparisons between MEC and White participants, for whom larger sample sizes were available. For MEC and White participants chi-square results are shown unless test assumptions relating to expected cell frequencies are violated.

Comment 4:

In Tables 3,4,7 and 8, it is observed that the authors calculated chi-square with counts for only one of the variables and the other variable have zero count. This virtually invalidates the test. The authors may provide reasons why this should be acceptable.

Just to clarify that these variables were part of multi-code questions, which allowed participants to select all that apply. They were subsequently transformed into individual derived variables with ‘selected’/’not selected’ answer codes. For instance, Table 2 includes the variable ‘Dry AMD’. Seven Asian participants (15.2%) but none of the Black participants had selected Dry AMD in the multi-code question. In other words, 100% (n=22) of Black participants did not select Migraines compared to 84.8% (n=39) of Asian participants. As such, there is a difference between the groups, which, compared to other variables, trends towards statistical significance (p=.087). In all cases where one cell is equal to zero, the subgroups are compared using Fisher’s exact test. We cannot see a reason not to report the Fisher’s exact p-value for these cases (chi-square was not used for any of these cases), but we are happy to remove them from the tables if there is a compelling statistical reason.

Comment 5

In table 2 the authors present range as a variable, It is recommended that they remove this from the table and report it in the main text

We have removed range from Table 2 as suggested.

Comment 6

One of the major selling points of this study was the fact that it covers younger people. However there is very limited focus on the younger population throughout the study. There was no comparison on the impact of VI and comorbidity among different age groups etc

Unfortunately, due to the small sample sizes, it was not possible to compare different age groups. We have incorporated the following comment (highlighted in bold in the paragraph below) into the Discussion and the comment listed under Comment 7 to highlight the fact that this should be explored in future research.

Overall, this study identified few statistically significant differences between MEC and White participants, and even fewer between Asian and Black participants. Despite the relatively young age, comorbidity, particularly of physical health conditions, was high: nine in ten Black participants, over two thirds of Asian and over three quarters of White participants reported a comorbid condition. Due to small sample sizes, it was not possible to compare health status across different age groups and this should be addressed in future research.

Comment 7

A little suggestion; it will have been interesting to see how VI severity and/ or vision changes impacted self rated health, physical exercise etc.

We completely agree that this would be interesting to explore within the different communities. However, we believe that this would be better explored in future research, partly due to the subjective measures used to collect V.I. severity, but mainly due to the already small subsample sizes for our ethnic communities. To encourage future research in this area we have added the following comment in the Limitations section:

For instance, it was not possible to explore the impact of age, V.I. severity or changes in vision on aspects such as self-rated health or exercise participation within each community. This will need to be explored in future research with a larger sample.

Reviewer 2 Report

This is a good paper on a very interesting and current topic. While the paper is very important, it could be more concise (especially in the results section) and some areas of the methods should be explained more clearly. Below are some suggestions and comments.

How reliable is it to classify these people in this way with self-reported measures? They could just not be wearing the correct glasses for reading or other co-morbidities could be taking hold and causing issues with completing daily life tasks. In my experience, many people don’t know if they are registered as SSI/SI and they often don’t know what you’re talking about when you say near, distance and peripheral. I understand you have put this in the limitations.

Can you explain how they were classified more clearly? What made you put them in each section?

I found the sample paragraph from line 170-177 very difficult to understand with so many acronyms etc. Perhaps this could be set out more clearly?

Line 222 - were slightly but not significantly more likely to not know what their eye condition was. This is a difficult sentence to understand. Perhaps you mean – MEC participants were slightly but not significantly less likely to know what their eye condition was?

Changes in vision – Is it not difficult to classify as different conditions will progress faster, if they are waiting on injections etc. for AMD or other treatment for eye conditions? What was the purpose of this question?

Figure 1 - Did you classify them into mild, moderate, and severe according to their responses or was this their responses?

279 – grammar – Although, none of the Asian and black participants reported very bad health – there would need to be a second part of this sentence to use Although 

You will need to look at grammar and sentence structure throughout the paper. 

Author Response

First of all, we would like to thank you for taking the time to review our manuscript and your very much constructive comments. We have reviewed these and made changes in the revised Word document. Any revisions have been made in track changes to help you locate them. We further provide our responses to each comment below. Please note that we have also made changes to the manuscripts in its entirety to improve its readability. For instance, we have revised the sentence structures in the Conclusion. We hope this resolves issues around English language.

How reliable is it to classify these people in this way with self-reported measures? They could just not be wearing the correct glasses for reading or other co-morbidities could be taking hold and causing issues with completing daily life tasks. In my experience, many people don’t know if they are registered as SSI/SI and they often don’t know what you’re talking about when you say near, distance and peripheral. I understand you have put this in the limitations.

Can you explain how they were classified more clearly? What made you put them in each section?

We agree with the reviewer that the use of self-report measures to assess V.I. has limitations. As noted, we do raise this as a limitation in the Limitations section. We have provided more detail about the process of classifying participants, which we hope helps clarity. We also highlight the issue around self-report questions in the Methods and Materials section:

As part of this, potential participants were first asked if they had ‘any difficulty seeing, or wear glasses or contact lenses?’. V.I. severity was then determined by the interviewer via coding of responses to a set of questions that asked about the participant’s registration status (whether they were registered as severely sight impaired, partially sight impaired, unsure of the category in which they were registered, or not registered), whether they were legally allowed to drive, and their near, distance and peripheral vision. The latter asked participants about difficulties in reading newsprint, seeing a person from a certain distance, and difficulties seeing people or things in the periphery of their vision. Participants who reported wearing glasses or contact lenses were prompted to rate difficulties when wearing their glasses or contact lenses to assess best-corrected vision. Response options for these were: ‘cannot do’, ‘severe difficulties’, ‘moderate difficulties’, ‘mild difficulties’ and ‘no difficulties’. Participants were categorised as having mild, moderate or severe V.I. based on their responses to these three questions (Supplementary Figure 1).  Individual classifications assigned to participants were discussed with members of charities that commissioned the survey and amended to reflect the participant’s experience. The pre-coded severity variable was included in the deidentified data set delivered by the data controllers. As with any self-report question, self-reported V.I. relies on a respondent’s honest and accurate assessment of their vision. Two comparisons of subjective and objective measures of V.I. found some over-identification of V.I. when using subjective vs objective measures [47, 48], but both concluded that self-report measures were a valid and suitable indicator of V.I., which continues to be used in V.I. research [47, 49-55], including in general population surveys in the UK [56].

I found the sample paragraph from line 170-177 very difficult to understand with so many acronyms etc. Perhaps this could be set out more clearly?

We have simplified this paragraph as follows. Please note that, as a result, some detail has been removed. Please let us know if this is acceptable.

A total of 745 adults aged 18 and over participated in the survey, including 78 adults from MEC and 667 from White communities. There were no statistically significant differences between these two groups in terms of gender, Χ² (1, 742)=0.63, p=.427, nor V.I. severity, U=28365.5, p=.092. But MEC adults were statistically significantly younger (M = 40.91 years) than their White counterparts (M = 58.72 years), t(743) = 9.37, p < 0.001. In addition, there were statistically significant differences between MEC and White adults in their location in the UK, Χ2 (11, 741) = 71.69, p <.001, Cramer’s V =.311, and whether they lived in a rural vs town setting, Χ2 (2, 741)=26.38, p <.001, Cramer’s V=.189.

Line 222 - were slightly but not significantly more likely to not know what their eye condition was. This is a difficult sentence to understand. Perhaps you mean – MEC participants were slightly but not significantly less likely to know what their eye condition was?

We have reworded this sentence as follows. We prefer this wording as it allows for percentages to be reported in the sentence and allows for some uncertainty around whether participants selected Don’t know for other reasons. We hope the revision improves the readability of this sentence.

MEC participants were slightly but not significantly more likely to select ‘Don’t know’ when asked what eye condition(s) they had (11.7% vs 7.8%).

Changes in vision – Is it not difficult to classify as different conditions will progress faster, if they are waiting on injections etc. for AMD or other treatment for eye conditions? What was the purpose of this question?

This question formed part of a set of questions which explored participants’ V.I. In the Discussion we comment on the possible impact of treatment on reported improvements in vision. We have now added a comment (highlighted in bold below) about the impact of treatment delays vs natural progression of eye diseases as possible reasons for deteriorations in vision:

Interestingly, eight participants (two Asian, two Black and four White participants) reported that their vision had improved recently. These participants reported having congenital cataract, detached retina, glaucoma, uveitis, optic nerve condition and/or keratoconus. Some of these conditions may indeed be treatable (e.g. with cataract or retinal detachment surgery) [52, 53], while treatments for others may slow but not reverse sight loss (e.g. eye drops for glaucoma) [54]. Participants were not asked whether they had been receiving any treatments for their eye conditions, so it is not clear if these contributed to these self-reported improvements. Furthermore, it is not clear which aspects of vision improved. The same applies to self-reported deteriorations in vision. Around half of Asian, White and MEC participants had experienced a deterioration in their vision compared to just over a third of Black participants. Deterioration in vision may be caused by a range of factors, such as delays to receiving treatment, ineffective treatment, or progression of disease, the rates of which will differ between individuals and eye conditions.

Figure 1 - Did you classify them into mild, moderate, and severe according to their responses or was this their responses?

As indicated in our previous response, we have provided more detail about how V.I. status was assessed. In addition to added detail, we have replaced Table 1 with a flowchart which is to be submitted as supplementary materials. This figure illustrates how participants’ V.I. status was categorised. We hope that the reviewers (and readers) will find this helpful. It should be noted that, as is indicated in the manuscript, inconsistent responses were discussed by the data controllers and collectors and categorisation adjusted accordingly.

279 – grammar – Although, none of the Asian and black participants reported very bad health – there would need to be a second part of this sentence to use Although 

‘Although’ was used as a synonym of ‘however’ rather than ‘in spite of’ in this sentence. To make this clearer, we have replaced it with ‘however’.

Round 2

Reviewer 2 Report

Thank you for editing the paper and providing more information on some of the methods. I think it is much improved. 

Very minor English editing 

Author Response

Thank you for your positive response to our changes. We note that you have requested further edits to the language. We have  asked a  colleague who is a native speaker to review the language for any grammar or wording issues, although all authors went through tertiary education in the UK and the first author, Syeda Hussain, is British and therefore a native speaker.  We suspect that some of the language issues you raised by may be stylistic rather than mistakes. We hope that the changes we have made to simplify the text will improve the readability of the text.