Prevalence of blindness and its determinants in Bangladeshi adult population: results from a national cross-sectional survey

Objective The objective of this study was to determine the prevalence of blindness and its determinants in Bangladeshi adult population. Study design A cross-sectional population-based survey conducted at household level with national representation. Samples were drawn from the 2011 national census frame using a multistage stratified cluster sampling method. Setting and participants The survey was done in urban and rural areas in 2013 using a probability proportionate to size sampling approach to locate participants from 72 primary sampling units. One man or one woman aged ≥40 years was randomly selected from their households to recruit 7200. In addition to sociodemographic data, information on medication for hypertension and diabetes was obtained. Blood pressure and capillary blood glucose were measured. Eyelids, cornea, lens, and retina were examined in addition to visual acuity and refraction testing. Primary outcome measures The following definition was used to categorise subjects having (1) blindness: visual acuity <3/60, (2) low vision: ≥3/60 to <6/60 and (3) normal vision: ≥6/12 after best correction. Results We could recruit 6391 (88.8%) people among whom 2955 (46.2%) were men. Among them, 1922 (30.1%) were from urban and 4469 (69.9%) were from rural areas. The mean age was 54.3 (SD 11.2) years. The age-standardised prevalence, after best correction, of blindness and low vision was 1.0% (95% CI 0.5% to 1.4%) and 12.1% (95% CI 10.5% to 13.8%), respectively. Multivariable logistic regression indicated that cataract, age-related macular degeneration and diabetic retinopathy were significantly associated with low vision and blindness after adjustment for age and sex. Population attributable risk of cataract for low vision and blindness was 79.6%. Conclusions Low vision and blindness are common problems in those aged 40 years or older. Extensive screening and eye care services are necessary for wider coverage engaging all tiers of the healthcare system especially focusing on cataract.

The authors say that is a national representative study, how was this represented? Are all the people in Bangladesh in the population pool listed for the randomly sampling? The article did not articulate.
Background: P5, line 8-11 Please give the time estimated. P6, line 33-34, what were the specific national strategies for preventing avoidable blindness? P6 -7: sampling: what the census used? Are all people from the country over 40 years old in the sample framework? How many clusters/size of clusters were selected for the 7,200 people.
P9, line 36-40 and P13, table 1, p-value should be given for each of the characteristics to tell the significance.
P9, line 47-53, these percentage should have confidence interval P10, discussion: as the age specific prevalence of BL/LV were not presented, so the comparison of prevalence due to ageing is not clear.
P10, line 27-28, your prevalence was 12.1%, how could this be the same as the countries listed? With less than 1%? P10, line 35 -41 , Cataract is the major cause of blindness, the reasons of low prevalence in other countries that the authors given could not convince readers.
In general, the discussion is not robust.

REVIEWER: 1
This large-scale population-based study addressed the prevalence and associations of visual impairment and blindness in Bangladesh. There is a paucity of ophthalmic epidemiological data from Bangladesh. Therefore, it may be worth publishing. However, there are obvious limitations in study methods, which currently are not fully addressed in the manuscript. >>Thank you very much for your encouraging comments.
1. Is refraction performed in the study? If not, the vision won't be best corrected. This needs to be discussed. If yes, how was the refraction performed? Please clarify.
>> Yes. Auto refraction was done, and then subjective refraction was performed. This has been shown in Figure 1, and relevant texts added (page 6, line 23) 2. It is also unclear how the authors defined presenting visual acuity. Please clarify. >> Distance visual acuity was measured on unaided participants with Snellen 'E' chart and a hand-held tally counter, if necessary, at three meters by ophthalmic nurses. Depending on acuity, finger count, hand movement and light projections were used. Medical technologists have done autorefraction. Thereafter, subjective refractions were done by the ophthalmologists. (page 6, line 19-23) 3. The methods to detect glaucoma and AMD, DR is not standardized. The current methods (examined by clinicians using direct or indirect biomicroscopy) may lead to the underdetection of these diseases, hence under-estimate the prevalence. These methods are also not verifiable.
>> We have added a sub-section to address this issue as below: (page 5, line 10-17)

Training of the survey team
The survey team was comprised of experienced enumerators, ophthalmic nurses, and medical technologists and ophthalmologists. They were trained in the National Institute of Ophthalmology by the investigators. Upon completion of their training, a dry-run was given in two nearby rural and urban areas. They were trained (as a team) using a using a study manual before launching the survey to reduce inter-observer variations and improve diagnostic accuracy done by the ophthalmologists. Their findings were randomly checked by the investigators at least once in a primary sampling unit.
In a large-scale survey in low resource settings such approach is being used. However, we acknowledge the possibility of an under-estimation in the Limitation sub-section (page 10, line 10).
4. The limitations in disease diagnosis methods caused marked difficulty in interpreting the results.
>>We acknowledged the limitation of diagnosis AMD and DR in addition to the non-availability of color photograph of the fundus. Possibility of an under-estimation has been added (page 10, line 10). Otherwise, our diagnoses were accurate and reliable.

REVIEWER: 2
The study reported valuable findings from a national blindness and low vision survey from the country and it is very important to inform future work across the country. >>Thank you very much for your encouraging comments.
There are some uncertainties as follows: >> The characteristics were given just to provide the background of the participants. There is no hypothesis behind this. Therefore, we humbly submit that adding p-value to any variable is not necessary. Such approach will invite a problem of multiple testing. As you know, with a set alpha of 0.05, 1 in 20 such tests may appear statistically significant due to chance alone*.
Better we avoid such p values here. 10. P9, line 47-53, these percentage should have confidence interval >> 95% confidence intervals added 11. P10, discussion: as the age specific prevalence of BL/LV were not presented, so the comparison of prevalence due to ageing is not clear.
>>Age-specific prevalence was presented in the Results section (page 8, lines 9-13, and in Table 2) but for two major groups. We have added Figure 2 to show age-specific prevalence to comment that blindness increases with age as reported by others (24,27). (page 8, line 2-4).
In addition, we have added the following sentences for low vision also. (page 9, line 9-13) Apparently, we observed a higher prevalence of low vision (12.1%) compared to studies in India (9.3%) 24 ,Pakistan (3.3) 21 , Iran (4.0%) 27 , but it was somewhat similar to that reported from South American countries (5.9 ˗-18.7%). x These differences should be cautiously interpreted because variation in age composition of the respondents, and some other factors, is an important determinant of low vision.
. 12. P10, line 27-28, your prevalence was 12.1%, how could this be the same as the countries listed? With less than 1%? >>Our claim of similarity is based on blindness prevalence, which is 1.0% in our study. (page 8, line 33-34) The 12.1% rate is for low vision. Kindly see clarifications above.
13. P10, line 35 -41, Cataract is the major cause of blindness, the reasons of low prevalence in other countries that the authors given could not convince readers.
>> Please see the response above: the prevalence in other countries is almost same as ours.
14. In general, the discussion is not robust. >>Because cataract is the major cause of blindness and low vision, we have added a paragraph on cataract: (page 9, line 30-35 continued to page 10, line 1-2) Cataracts attribution to blindness in our sample (79.6%) is a little higher than that reported in an India population (62.1%). Therefore, addressing cataract will be bring most benefit to prevent blindness. In addition to promotion of healthy ageing, a few other factors such as ultraviolet ray exposures, diabetes, hypertension, use of certain drugs, and smoking can be considered. Accessibility to socioeconomically deprived people especially in remote areas should be enhanced. Blindness prevention programe's success will largely depend on the