Correction of Refractive Error in the Adult Population of Bangladesh: Meeting the Unmet Need

Abstract

purpose. To assess the extent of uncorrected refractive error and associated factors in Bangladesh and to suggest ways in which this need can be met.

methods. A nationally representative sample of 12,782 adults (≥30 years of age) was selected. Of them, 11,624 subjects underwent a demographic interview, visual acuity (logarithm of the minimum angle of resolution [logMAR]) measurement, automated refraction, and optic disc examination. Subjects with visual acuity less than 6/12 in either eye also had a corrected refraction measurement, cataract grading, and dilated retinal examination.

results. Of the 11,624 subjects examined, 2,469 (22.1%) were myopes (less than −0.5 D) and 2,308 (20.6%) hyperopes (more than +0.5 D). The spectacle coverage percentage, calculated as [met need/(met need + unmet need) × 100%] was 25.2% and 40.5%, using 6/12 and 6/18 visual acuity cutoffs, respectively, and was higher in men and urban inhabitants. Older subjects and the literate and more highly educated were more likely to wear spectacles; however, most spectacle wearers (81%) had inadequate correction. Of the 1142 subjects who would benefit from spectacles, 827 (72.4%) would be suitable for off-the-shelf spectacles. Subjects without spectacles with less than 6/12 in the better eye (n = 835), would achieve 6/12 or better with correction (unmet need). Extrapolation to the national population yields an estimate that 1.5 million (6.7%) adult men and 1.8 million (9.2%) women have an unmet need for refractive correction.

conclusions. In Bangladesh, there is low spectacle coverage with a large unmet need. This survey identified risk groups, in particular women and those living in rural areas. This description of the availability of refractive services suggests areas for improvement (e.g., off-the-shelf spectacles) that may enable Bangladesh to achieve the goals of the World Health Organization’s Vision 2020 initiative.

Refractive error is a remediable cause of visual impairment, with correction of significant refractive error being a priority of Vision 2020: The Right to Sight, the World Health Organization’s (WHO) Global Initiative for the Elimination of Avoidable Blindness. ¹ The scale of the problem of refractive error has perhaps received less attention in the past, given that the WHO categories of visual impairment refer to best corrected visual acuity in the better eye rather than presenting visual acuity. ² Refractive error has severe social and economic effects on individuals and communities, restricting educational and employment opportunities of otherwise healthy individuals. The earlier onset of refractive error compared with cataract can account for twice as many blind-person years. ³  

Recently, several population-based studies of refractive error and associated visual impairment have been performed, with the use of a common protocol in school age children in five geographic regions (i.e., Eastern Nepal, ⁴ rural China, ⁵ urban Chile, ⁶ rural southern India, ⁷ and urban northern India ⁸ ). Refractive error was the cause of visual impairment (<6/12 in the better eye) in 81.7% of children in the urban Indian population and in 61% in the rural Indian population. These studies are an important contribution to the understanding of refractive error prevalence in this age group. However, in adults the available information derived from population-based studies of nonwhite populations is more limited. Surveys in East and Southeast Asia ^{9 10} have shown a substantial level of refractive error, in particular a high prevalence of myopia (39% in Singapore Chinese adults aged 40 years or older ⁹ ) that appears to be increasing. ^{11 12}  

Fewer such studies have been conducted in South Asia. The Andhra Pradesh Eye Disease Study ¹³ reported a prevalence of myopia of 29% among adults aged 30 years and older. The Bangladesh National Blindness and Low Vision Survey ¹⁴ (1999–2000) recently reported the prevalence and causes of blindness and visual impairment in 11,624 Bangladeshi adults (aged 30 years or more). The main cause of a presenting visual acuity of less than 6/12 (0.3 logarithm of the minimum angle of resolution [logMAR]) in the better eye was cataract (73.4%), followed by refractive error (18.9%) and uncorrected aphakia (1.2%). In relation to the causes of bilateral blindness (defined as <3/60, or 1.3 logMAR, in the better eye), cataract was also the principal disorder (in 129 [79.6%] of the 162 blind subjects) followed by uncorrected aphakia (6.2%). The prevalence of refractive errors in Bangladesh (in subjects with phakic eyes and for whom refractive error data were recorded [n = 11,189]) has been reported recently. ¹⁵ Twenty-two percent of this adult study population (aged 30 years or older) had myopia (less than −0.5 D) and 20.6% hyperopia (more than +0.5 D). A large increase in prevalence of myopia from the age of 50 years was noted, which was ascribed to cataract. The prevalence of hyperopia also increased steadily from the age of 30 years, reaching a peak in the 50- to 59-year age group, and decreasing in older persons. 

We set out to investigate the need for optical services among Bangladeshi adults and how the need might be met. Data from the National Blindness and Low Vision Survey were analyzed to

1.  
   Calculate the met and unmet need for refractive error correction and thereby advance a definition for spectacle coverage in the adult (≥30 years of age) population of Bangladesh.
2.  
   Investigate factors associated with lack of spectacle correction and the accuracy of habitual correction.
3.  
   Suggest ways in which the refractive need of Bangladesh can be met.

Methods

The methodology used in the survey has been described in detail elsewhere. ¹⁶ Based on the target age population of 30 years and older (44 million adults), a blindness prevalence estimate of 2.5%, a random sampling error of 0.35% with a design effect of 1.5 (a factor by which the sample size is multiplied to adjust for the cluster sampling strategy), a sample size of 12,900 subjects was determined. Multistage stratified cluster random sampling, with probability proportional-to-size procedures, was adopted as the strategy for the selection of a cross-sectional, nationally representative sample. Stratification of the sample according to rural and urban residence (corresponding to official municipality ordinance status) ¹⁷ was incorporated in the selection process. Within each of the six regional administrative divisions, a proportional number of clusters in relation to the overall national population were identified based on official census data. A total of 154 cluster sample sites were selected, of which 104 were rural villages and the remaining 50 were urban block areas. For logistic purposes, rural clusters consisted of 100 subjects, and urban clusters consisted of 50 subjects. 

The survey teams underwent specialized training during a 1-month period. Both inter- and intraobserver agreement were determined among the groups of ophthalmic nurses and ophthalmologists, with respect to visual acuity testing, intraocular pressure measurement, optic disc assessment, and lens opacity grading. To test the methods, two pilot surveys ¹⁶ were performed, one in a rural area and the other in an urban setting. Adjustment was made to the examination protocol in advance of the main survey which was undertaken between October 1999 and June 2000. The examination process began with an interview in which the interviewer confirmed that the individual was an enumerated subject and obtained his or her consent to participate. Demographic data such as age and sex were collected, in addition to specific information regarding socioeconomic status, employment, literacy, and the wearing of spectacles or contact lenses. Literacy was recorded as reads easily, reads with difficulty, and illiterate. 

Distance visual acuity was measured with a reduced logMAR-based tumbling-E chart. ¹⁸ Presenting visual acuity was measured with the subject’s current distance refractive correction, if worn, in each eye in turn. If a subject correctly identified 30 optotypes (0.1 logMAR), the acuity testing session for that eye was then terminated. All subjects underwent automated refraction without cycloplegia (model RM-8000B; Topcon Corp., Tokyo, Japan), performed by trained medical technicians. The average refractive error (based on three consecutive readings) was obtained. If the autorefractor did not yield a measurement (especially because of media opacity) in a subject with less than 6/12 (0.3 logMAR) presenting visual acuity (with habitual correction if worn), the ophthalmologist attempted a manual objective and subjective refraction. Subjects with less than 6/12 (0.3 logMAR) presenting visual acuity in either eye (with habitual correction if worn) were then retested with the autorefraction result placed in a trial frame, using spherical and cylindrical trial lenses. This was performed to estimate the contribution of refractive error to these subjects’ visual impairment. 

The ophthalmologists asked these subjects whether they had been treated for ocular disorders. The presence or absence of an intraocular lens and the use of an aphakic correction were also noted. The ophthalmologist attempted an examination of the optic discs of all subjects using direct ophthalmoscopy. Intraocular pressure was measured by Schiotz tonometry (Biro Gerhard, Hausen, Germany) if either or both optic discs had a vertical cup-to-disc ratio of 0.7 or more, or if the optic disc could not be visualized despite pupil dilation. All subjects with less than 6/12 (0.3 logMAR) presenting visual acuity in either eye were assessed for cataract. It was decided that the most appropriate cataract grading system for this field survey would be the Mehra/Minassian system ¹⁹ which uses six categories based on obscuration of the red reflex of an undilated normal pupil due to the presence of a lens opacity, as assessed using direct ophthalmoscopy. The grades were: 0, clear red reflex/no opacity; 1, a few small-dot opacities that occupy less than a 1-mm² maximum area; 2A, lens opacity that obscures the red reflex, with the area obscured smaller than the area of clear reflex; 2B, obscured area equal to or greater than area of clear red reflex; 3, red reflex totally obscured by lens opacity; 4, (pseudo-) aphakia or displaced lens; and 5, red reflex unassessable—for example, because of corneal opacity. Subjects with less than 6/12 presenting visual acuity (with habitual correction if worn) in either eye had their eyes subsequently dilated (after a check for relative afferent pupil defect) and the cup-to-disc ratio recorded and presence or absence of retinal disease noted. Visual field testing was not performed because of logistic considerations. A subject was recorded as having glaucoma if the cup-to-disc ratio was 0.7 or more in either eye in the presence of an IOP of 20.6 mm Hg or more. These values are the 97.5th percentiles calculated from a population-based distribution of intraocular pressure and cup-to-disc ratio measured on a randomly selected subsample of 243 subjects during the main survey. ¹⁴  

All persons with less than 6/18 (0.48 logMAR) presenting visual acuity in the better eye were referred to the nearest eye care facility (district or nongovernment hospital). 

The study achieved a high response rate (91%). Younger men (aged 30–59 years) were less likely to attend. ¹⁴ The research was conducted in full accordance with ethical principles, including the provisions of the World Medical Association Declaration of Helsinki. The Bangladesh Medical Research Council provided written approval of the ethical structure of this survey in March 1999. Oral informed consent was obtained from all participants, rather than written consent (48% of the study population was illiterate) after explanation of the procedures to be conducted and before subjects were examined. 

Data Analysis

Refractive error is presented as spherical equivalent, which equals the sphere power plus half the cylinder power. Statistical analysis was performed on data from right eyes only, in keeping with methodology used in similar refractive error studies, ⁹ as the correlation between the spherical equivalent in right and left eyes was very high (Pearson coefficient, 0.783), with no significant differences between the right and left eyes of any subject (t-test, P = 0.78). Although visual acuity was measured using the logMAR scale, we have added the Snellen equivalents in this article for purposes of comparison with similar studies. 

Spectacle coverage was defined as

 

\[\mathrm{Spectacle\ coverage}\ (\%)\ {=}\ \frac{\mathrm{met\ need}}{\mathrm{total\ need}}\ {\times}\ 100\%\]

where total need = met need + unmet need. 

“Met need” describes the number of subjects who wore distance spectacles and had visual acuity less than 6/12 in the better eye without correction, but who achieved 6/12 or better in the better eye with their present distance spectacles. The methodology used in this survey tested visual acuity with the subjects habitual distance correction if worn. A bespectacled subject was not retested without wearing spectacles at the time of presentation, because of logistic considerations. Therefore, for the purposes of this survey, it was assumed that spectacle wearers would have acuity worse than 6/12 in the better eye without spectacles. 

“Unmet need” was defined as the number of subjects who did not wear spectacles and who had acuity less than 6/12 in the better eye without correction, but who could achieve 6/12 or better in the better eye with correction. 

Odds ratios (OR; presented with 95% confidence intervals) were used in the univariate analysis of spectacle use with key variables, such as sex, literacy, education, occupation, and location of residence. Some subjects were excluded from the analysis of suitability of off-the-shelf spectacles. These include subjects who had undergone cataract surgery or for whom no refractive data were available, principally due to advanced medial opacity. Among the remainder, those who had an uncorrected visual acuity of 6/12 or better in the better eye and those whose best corrected visual acuity was less than 6/12 in both eyes were excluded, as these subjects probably did not need distance refractive correction. The remaining subjects would benefit from spectacles for distance. These included those with astigmatism of more than 1.25 D and anisometropia of moire than 2.0 D, who were judged to be unsuitable for off-the-shelf spectacles. Age and sex standardization of data against a reference population involved national census figures (1998) from the Bangladesh Bureau of Statistics. ¹⁷ This was a direct method of standardization ²⁰ that required the age- and sex-specific rates for the study population and also for the national population of Bangladesh (the standard population). Data management and analysis were performed on computer (EpiInfo, ver. 6.04b, Microsoft Excel; Microsoft, Redmond, WA; and SPSS, ver. 10.0; SPSS Science, Chicago, IL). 

Results

The Need for Spectacle Correction in Bangladesh

The need for spectacles was calculated by applying age- and sex-standardization rates of unmet need to the Bangladeshi population ¹⁷ (Table 1A) . Overall, 3.3 million Bangladeshi adults were visually impaired (<6/12 in the better eye on presentation) as a result of uncorrected refractive errors, with the need being greater in women than in men (Pearson χ² = 87,451; P < 0.001). 

Spectacle Coverage

In the sample, 282 subjects were corrected for refractive error with spectacles (met need), while 835 subjects had an unmet need, for a 6/12 cutoff. For a 6/18 cutoff, these numbers were 320 and 471, respectively. Hence spectacle coverage was 25.2% for a 6/12 cutoff and 40.5% for a 6/18 cutoff. Spectacle coverage among women was 20.9% for the 6/12 cutoff and 36.0% for the 6/18 cutoff (Table 1B) . In men, these values were 30.2% and 44.8%, respectively. Spectacle coverage was significantly higher in men than in women for both the 6/12 and the 6/18 cutoffs (Pearson χ² = 12.8, P < 0.001 and χ² = 6.3, P = 0.01, respectively). Spectacle coverage among those living in rural areas was 15.9% for the 6/12 cutoff and 27.8% with the 6/18 cutoff. In urban residents, these values were significantly higher, at 58.8% and 75.9%, respectively (Pearson χ² = 185.8, P < 0.001 and χ² = 147.7, P < 0.001, respectively). 

Effect of Refractive Error and Coexistent Ocular Disease on Change of Vision Afforded by Refractive Correction

The amount of visual improvement afforded by refractive correction is likely to be affected by other ocular disease, such as cataract or posterior segment disease. Table 2 presents the change of visual acuity with refractive correction of individuals. The more common ocular diseases in the better eye (at presentation) are illustrated, and changes in visual acuity are shown as change in visual acuity category (categorized as 6/12 or better, <6/12 to ≥6/18, <6/18 to ≥6/60, and worse than 6/60). Only subjects who had less than 6/12 visual acuity in either eye were retested for visual acuity wearing the refractive correction obtained from the automated refraction result. In subjects with only refractive error and no other ocular disease (n = 636), 49.0% had a visual acuity of less than 6/12 in the better eye. Refractive correction reduced this to 1.7% (for less than 6/18, 14.4% and 0.9%, respectively). In those with additional cataract (n = 1026), 78.9% had a visual acuity of less than 6/12 in the better eye, but with refractive correction, this was lowered to 37.6% (for <6/18, 49.8% and 19.2%, respectively). Of the 35 subjects with only age-related macular disease (and a cataract grade of <2A), 80% had visual acuity less than 6/12 in the better eye, but this was reduced to 31.4% with correction (for <6/18: 37.1% and 14.3%, respectively). Those with cataract (grade 2A or more) and age-related macular disease (n = 148), 85.1% had visual acuity less than 6/12 in the better eye, but this was reduced to 46.6% with correction (for <6/18: 49.3% and 22.3%, respectively). Of 357 (54%) subjects with only refractive error and minimal cataract (less than grade 2A) that did not enter a better visual acuity category with refractive correction, only 3 (0.8%) had a refractive error of less than −3 D or more than +3 D, whereas 22 (7.3%) of the 303 subjects whose vision improved by at least one visual acuity category had a refractive error in this range. 

Other less common ocular diseases were not analyzed because of the small number of subjects who had the different diseases. For example, only five had glaucoma as the principal cause of visual acuity of less than 6/12, defined as a cup-to-disc ratio of 0.7 or more and intraocular pressure in the 97.5 percentile or higher. 

Factors Associated with Lack of Refractive Correction

Of all the subjects examined, 340 (3.04%) wore distance spectacles on presentation: 192 men (3.62% of all men examined) and 148 women (2.59% of all women examined). 

Overall, men were significantly more likely to wear spectacles than women (OR [95% CI], 1.36 [1.09–1.70]; P = 0.005). Men aged 60 to 69 years were more likely to wear spectacles than women of the same age (OR, 1.90 [1.13–3.22]; P = 0.01). In the other age groups, the gender difference was not significant. Older subjects were more likely to wear spectacles than younger subjects. This was significant (P < 0.001) in all older age groups compared with those aged 30 to 39 years (40–49 years: OR, 2.77 [1.97–3.90]; 50–59 years: OR, 4.16 [2.92–5.95]; 60–69 years: OR, 5.82 [4.05–8.35]; 70+ years: OR, 4.28 [2.65–6.89]). Hyperopes (>+0.5 D) were more likely to wear spectacles than myopes (less than −0.5 D; OR, 1.81 [1.39–2.37]; P < 0.001). Younger adults (aged 30–49 years) were significantly more likely to wear spectacles than older adults (aged ≥50 years) in the following categories of spherical equivalent: less than −5 D (OR, 3.09 [0.86–10.55]; P = 0.04); –3 to –5 D (OR, 3.72 [1.09–12.1]; P = 0.03); and –0.5 to +0.5 D (OR, 0.40 [0.24–0.67]; P = 0.0001). There were no other significant age differences in the remaining range of refractive error. Compared with women, men were significantly more likely to wear spectacles in the –1 to –3 D (OR, 2.69 [1.08–6.96]; P = 0.02) and –3 to –5 D (OR, 2.97 [1.01–9.36], P = 0.03) myopic groups, and the more than +0.5 to +1 D (OR, 1.95 [0.97, 3.93]; P = 0.04) and more than 1 to 3 D (OR, 2.89 [1.90–4.40]; P < 0.001) hypermetropic groups. No gender difference existed in the other categories of refractive error. When we fit a logistic regression model for age and spherical equivalent, the wearing of spectacles was statistically more common among older individuals (P < 0.001) and hyperopes (P < 0.002), and this was also the case when adjusting for the interaction between these two variables. 

Table 3 presents other factors associated with the wearing of distance spectacles, by comparing those subjects who achieved 6/12 or more in the better eye with their distance correction with those who needed spectacles (habitually uncorrected but would have achieved 6/12 or better in the better eye with correction). Illiterate subjects were significantly more likely to have uncorrected refraction than literate individuals (OR, 22.49 [14.5–34.9]; P < 0.001), as were subjects living in rural areas compared with urban inhabitants (OR, 7.56 [5.5–10.5]; P < 0.001). Those who had not attended school were much more likely to have uncorrected refraction than those who had ever gone to school (OR, 17.74 [11.7–26.9]; P < 0.001). In those who had attended school, those who had not progressed to secondary school education or college and university education were more likely to have uncorrected refraction than those who had (secondary school: OR, 7.51 [3.7–15.5]; P < 0.001; college or university: OR, 5.22 [3.78–7.22]; P < 0.001). Unemployed subjects were much more likely to have uncorrected refraction (OR, 7.21 [4.9–10.5]; P < 0.001) when compared with professionals employed in nonmanual labor. Likewise, correction among manual workers (OR, 11.83 [7.3–19.2]; P < 0.001) was less common than in professional (nonmanual) workers. Manual workers were more likely to have uncorrected refraction than those who were unemployed (OR, 1.64 [1.08–2.49]; P = 0.013). 

Accuracy of Current Habitual Refraction in Those Who Wore Spectacles

This subanalysis was limited to those subjects who were wearing distance spectacles on presentation, and who were retested with the objective refractive correction provided by the autorefractor. This analysis was therefore limited to those bespectacled subjects with visual acuity less than 6/12 (0.3 logMAR) in one or both eyes on presentation (92 subjects). Figure 1 compares the visual acuity of the right eye with habitual distance spectacles with the visual acuity achieved wearing the result of the automated refraction in trial frames. This graph shows that the majority of eyes (81%) achieved a better result wearing the autorefractive correction in trial frames than when they were using their habitual distance spectacles. 

Meeting the Need for Refractive Correction in Bangladesh

Of the 11,189 subjects, who had not undergone cataract surgery and for whom refractive error data were available, 10,047 (89.8%) would not have benefited from a distance correction, on account of good uncorrected (6/12 or better) vision in the better eye (n = 9,475) or poor corrected visual acuity (<6/12 in both eyes, n = 572). Of the remaining 1142 subjects, 827 (72.4%) had astigmatism of 1.25 D or less or anisometropia of 2.00 D or less, and were judged suitable for off-the-shelf spectacles (Table 4A) . Those adults with vision suitable for off-the-shelf spectacles included 340 (41.1%) hyperopes (more than +1 D) and 263 (31.8%) myopes (less than −1 D; Table 4B ). The remaining 224 suitable subjects had low spherical or astigmatic refractive error. Table 4B presents the number of myopic (less than −1 D) and hyperopic (more than +1 D) subjects whose refraction was correctable with off-the-shelf spectacles if the dioptric range −3 to +3 D (in 0.5-D steps) were available. For all adults (30 years of age and older) this range would provide suitable stock spectacles for 316 (92.9%) subjects with hyperopia of more than +1 D and 195 (74.1%) subjects with myopia of less than −1 D. 

Discussion

The Bangladesh National survey provides the only nationally representative data on refractive error correction in a South Asian population. The large sample size and high response rate lend considerable statistical weight to the overall study. 

There are a large number of people who need spectacle correction in Bangladesh based on the magnitude of uncorrected refractive errors detected by the survey. In our study, unmet need was higher in women than in men and those aged 50 years or more. When the visual acuity cutoff was lowered from 6/18 to 6/12, the need increased, as would be expected. The cutoff of 6/12 was chosen, because it is probably more representative of the visual needs of modern life, such as driving. This study did not test a bespectacled subject without spectacles on presentation. Therefore, for the purposes of this survey and the definition of met need within spectacle coverage, it was assumed that spectacle-wearers would have acuity worse than 6/12 in the better eye without spectacles. Although this may introduce some error, the relatively small number of spectacle wearers would lead one to assume this error to be small relative to the need for spectacles in this population. 

The analysis of refractive error and coexistent ocular disease showed that refractive correction reduced the percentage of subjects with a visual acuity of less than 6/12 with refractive error as the principal disease from 49% to 2%, whereas in those with additional cataract and/or age-related macular disease, the percentage was approximately halved. This demonstrates the importance of refractive correction, not only in those with refractive error alone, but also in those with other disease such as cataract, which is more prevalent in older age groups. Indeed, refractive care can also provide an excellent opportunity for access to the population for the detection of other ocular diseases ^{21 22} The severity of ocular disease is also undoubtedly important. There was a higher proportion of subjects with more severe refractive error among those whose refraction improved by at least one visual acuity category than among those whose acuity remained unchanged. 

Table 3 presents factors associated with the wearing of spectacles in the population and compares those who need spectacles with those who have spectacles that meet a need (6/12 or better in this case). This approach is unlike that used in other studies ²³ in which those wearing spectacles and those not wearing spectacles were compared within the entire study sample. The latter approach is less useful, because it includes subjects who may not actually need spectacles and hence should be excluded from the analysis, just as it includes wearers of spectacles who do not achieve 6/12 visual acuity despite wearing glasses. The analysis of factors associated with wearing of spectacles demonstrated that inequalities exist within the population. 

One of the principal findings is the low level of spectacle coverage. Using the definitions presented in this report, a coverage of between 21% to 45% was found in men and women, and significantly lower coverage was found in rural compared with urban areas. The concept of spectacle coverage has not been defined previously. We suggest that this definition may be used in similar studies. The lack of a definition leads to difficulties in comparing the results in Bangladesh with those in other regions. A study in Andhra Pradesh ²³ reported that 2305 (41.2%) of 5587 adults (aged 30 years or more) wore spectacles, but the study did not consider this number in relation to the number who would have benefited from spectacles. 

The finding that older subjects were more likely to wear spectacles than younger subjects concurs with that of studies in India ²³ and in whites, ²⁴ the latter reporting a risk of undercorrected refractive error (defined as an improvement of five letters or more in best corrected visual acuity) increasing by 1.8 times for every decade of life starting at 40 years of age. Spectacles were less frequently worn by women than men across most of the range of refractive error. It is interesting that this differs from an Indian study (subjects aged 15 years or more) in which it was reported that women were significantly more likely to wear spectacles than men. ²³ The differences may be explained by cultural differences between these two populations and also perhaps by life expectancy, because women have a higher life expectancy than men in India, whereas the opposite is the case in Bangladesh. The sex that has longer life expectancy may have a greater need for spectacles. 

The current analysis also showed that most subjects wearing spectacles were not wearing the best possible correction. It is acknowledged that the accommodation that occurs with the use of the autorefractor may have given a more myopic result ^{5 8} (particularly in younger adults) than would have been obtained with a subjective refraction; hence correction with the results of the automated refraction alone may have resulted in poorer distance visual acuity in some subjects. There may be various reasons why subjects were wearing the incorrect prescription. Spectacles may have been prescribed some time before the survey and the refractive error may have changed. The person may be prevented from obtaining a new prescription on account of cost or availability of services or may think that the current spectacles are giving a reasonable level of vision for his or her daily life. The refractive prescription may have been inaccurate in the first instance or the spectacles may have become damaged since they were prescribed. Several of these reasons were explored by the Andhra Pradesh study, ²³ which reported that a significant proportion of subjects had discontinued wearing spectacles, with explanations that included loss of spectacles, discomfort, and a perception that they were unnecessary. 

Hyperopes (more than +0.5 D) were more likely to wear spectacles than myopes (less than −0.5 D). Hyperopia may be more debilitating in a population that is predominantly rural and either unemployed or involved in manual work. Higher magnitudes of myopia were more likely to be corrected in younger adults than in older age groups, perhaps because high myopia is related to the onset of cataract in older age; hence, a refractive correction would result in less visual improvement. Alternatively, other barriers to the uptake of refractive correction in old age may be involved, such as lack of accompaniment to a treatment center. As would be expected, those who were literate, had attended school, who were of higher educational status, and who lived in urban areas were more likely to wear refractive correction. These observations were similar to those in the study in Andhra Pradesh, ²³ where, in addition, those of extreme lower socioeconomic status were less likely to be wearing spectacles. 

To address how the need for refractive correction can be met, one must understand the current situation in Bangladesh with regard to refractive services. WHO recently ²⁵ recommended a target of 1 ophthalmologist per 500,000 persons, 1 ophthalmic nurse per 400,000, and 1 refractionist per 250,000 persons in the year 2000. Refractive services in Bangladesh are almost exclusively provided by 400 to 500 ophthalmologists (apart from one regional center in Chittagong), of whom 100 to 150 are surgically trained ²⁶ and of whom only approximately 5% have had formal training in refraction. Approximately 30 ophthalmic technicians can measure refraction. Within cities, refraction takes place in hospitals and private ophthalmic clinics, whereas the optician outlets sell spectacles only. Outside the cities, more refractive services are offered by clinics sponsored by nongovernmental organizations than government district hospitals, and some level of refraction takes place at opticians’ shops, where over-the-counter presbyopic spectacle corrections are available. There are practically no career advancement opportunities for midlevel personnel trained in refraction. The cost of a pair of spectacles ranges from approximately US$2.50 (150 Bangladeshi taka) to US$100.00. Average per capita income is US$280 to US$300. It is possible that spectacles could be provided at low cost or without cost. This prospect is the subject of a nongovernmental organization pilot initiative planned in the Sirajganj District. At present, no facilities exist for recycling of spectacle lenses. After cataract surgery, no or limited refractive correction is provided free of charge in the government sector; however, in eye camps (operations performed in locations distant from a hospital or clinic) an aphakic correction is usually provided. Spectacle frames are both made in-country and imported. However, spectacle lenses are almost exclusively imported. There is negligible coverage in terms of contact lenses or laser refractive corrective surgery. 

These indices demonstrate that insufficient numbers of personnel trained in refraction are available, with a need for an expansion in the number of midlevel ophthalmic personnel. Distribution of resources may also be limited. Most of the optical services are located in urban areas, which was reflected in the finding in this survey that urban individuals were much more likely to wear spectacles. Integration of primary eye care into primary health care would improve the situation in rural areas. Other factors such as wealth and education are also undoubtedly important. To meet this need for refractive correction, these barriers must be identified and overcome. 

Two studies have examined the suitability of off-the-shelf spectacles in whites. ^{22 27} A study of an Australian adult population (age, 40–60 years) ²⁷ reported that 26.8% of subjects had low astigmatic or nonastigmatic refractive error and were suitable candidates for off-the-shelf spectacles. The Australian study used definitions different from those in the Bangladesh study; however, if the same definitions are used, this percentage is remarkably similar when considering the same age range in Bangladesh (29.1%). In defining a suitable group, we chose to exclude those who did not need spectacles on account of good uncorrected vision in the better eye, unlike the Australian study, which also excluded best corrected vision of less than 6/12 in either eye. These definitions were thought to be more pertinent to the Bangladeshi population, where correcting refraction in an individual with an uncorrected visual acuity of 6/12 or better in the better eye is less of a priority in terms of need than correction in someone with an uncorrected visual acuity of less than 6/12 in both eyes. We also increased the cutoff for suitability of anisometropes for off-the-shelf spectacles to 2.0 D or less (rather than <0.5 D ²⁷ ), because it is known that patients can tolerate up to 2.0 D of anisometropia. ²⁸ According to this definition, refraction in 827 (72.4%) of the 1142 adults who needed a distance correction could be corrected by using off-the-shelf spectacles, with a stock of spectacles ranging from −3 to +3 D providing for most of the need among myopes and hyperopes outside a range of −1 to +1 D. 

In this study, we did not investigate the need for presbyopic correction among adults, which would add considerably to the number of persons benefiting from off-the-shelf spectacles. In consideration of the low number of refraction-trained ophthalmologists and low level optometric services in Bangladesh, off-the-shelf spectacles could be provided by paramedical staff with basic training in subjective refraction and dispensing. These spectacles could either be manufactured locally or imported either at cost or by donation. It would also be important that follow-up care that involves repair of spectacles and repeat dispensing be available. 

We addressed the met and unmet needs for refractive services at a nationally representative level. It is intended that these and other data ¹⁵ that have described the prevalence and magnitude of refractive error in this population be used to plan eye care health policy in Bangladesh and the surrounding region. 

 Supported by Sight Savers International.

 Submitted for publication February 6, 2003; revised June 26 and August 17, 2003; accepted August 28, 2003.

 Disclosure: R.R.A. Bourne, None; B.P. Dineen, None; D.M. Noorul Huq, None; S.M. Ali, None; G.J. Johnson, None

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Table 1A.

 

Estimated (or Extrapolated) Magnitude of the Unmet Need for Refractive Correction in Bangladesh by Age and Presenting Visual Acuity Category

Visual Acuity Cutoff (Better Eye)	30–49 Years			50 Years and Older			Total (% of Total Population) Both Sexes
		Men	Women	Men	Women
<6/12	239,000	438,000	1,230,000	1,427,000	3,334,000 (7.9)
<6/18	76,000	138,000	811,000	870,000	1,895,000 (4.5)

 Data are rounded to the nearest 1000, after age- and sex-standardization of the study sample results to the population of Bangladesh.

Table 1B.

 

Spectacle Coverage by Gender and According to Presenting Visual Acuity Category

Visual Acuity Cutoff (Better Eye)	Met Need (n) (%)					Unmet Need (n) (%)					Coverage (%)
		Men	Women	All	Men	Women	All	Men	Women	All
<6/12	158 (56.0)	124 (44.0)	282 (100.0)	365 (43.7)	470 (56.3)	835 (100.0)	30.2	20.9	25.2
<6/18	180 (56.3)	140 (43.8)	320 (100.0)	222 (47.1)	249 (52.9)	471 (100.0)	44.8	36.0	40.5

Table 2.

 

Uncorrected and Corrected Visual Acuity in Subjects with and without Ocular Disease in Addition to Refractive Error

Visual Acuity (Better Seeing Eye)			Ocular Disease in Addition to Refractive Error in the Better-Seeing Eye* at Presentation
				Refractive Error Only n (%) (n = 636)	Cataract ≥ Grade 2A, † n (%) (n = 1026)	Age-Related Macular Disease, ‡ n (%) (n = 35)	Age-Related Macular Disease and Cataract ≥ Grade 2A n (%) (n = 148)
Uncorrected	<6/12 to ≥6/18	220 (34.6)	299 (29.1)	15 (42.9)	53 (35.8)
	<6/18 to ≥6/60	90 (14.2)	466 (45.4)	13 (37.1)	64 (43.2)
	<6/60	1 (0.2)	45 (4.4)	0 (0.0)	9 (6.1)
Corrected	6/12 or better	625 (98.3)	640 (62.4)	24 (68.6)	79 (53.4)
	<6/12 to ≥6/18	5 (0.8)	189 (18.4)	6 (17.1)	36 (24.3)
	<6/18 to ≥6/60	4 (0.6)	185 (18.0)	5 (14.3)	31 (20.9)
	<6/60	2 (0.3)	12 (1.2)	0 (0.0)	2 (1.4)

*  In Persons in whom presenting vision was the same in both eyes (264 subjects), the right eye was chosen.

†  These subjects had no evidence of age-related macular disease. In six of them, the fundus could not be visualized despite dilation.

‡  Includes age-related maculopathy and age-related macular degeneration. No cataract was identified in these subjects.

Table 3.

 

Factors Associated with the Wearing of Spectacles, the Met Need (282 Subjects) and Unmet Need (835 Subjects)

Associated Factor	Wears Distance Spectacles and Achieves ≥6/12 in the Better Eye* (Met Need) n (%)	Need Spectacles to Achieve ≥6/12, † (Unmet Need) n (%)	Odds Ratio (95% CI) of Need for a Spectacle Correction	P
Literacy, †
Easily reads and writes	236 (54.3)	199 (45.7)	1.00
Reads with difficulty	17 (16.8)	84 (83.2)	5.86 (3.3–10.6)	< 0.001
Illiterate	29 (5.0)	550 (95.0)	22.49 (14.5–34.9)	< 0.001
Location
Urban	143 (58.8)	100 (41.2)	1.00
Rural	139 (15.9)	735 (84.1)	7.56 (5.5–10.5)	< 0.001
Level of Education
College/university	103 (89.6)	12 (10.4)	1.00
Secondary school	96 (53.3)	84 (46.7)	7.51 (3.7–15.5)	< 0.001
Primary school only	50 (24.2)	157 (75.8)	26.95 (13.1–56.4)	< 0.001
No schooling	32 (5.2)	579 (94.8)	155.30 (73.9–332.9)	< 0.001
Occupation
Nonmanual labor	116 (63.7)	66 (36.3)	1.00
Manual labor	37 (12.9)	249 (87.1)	11.83 (7.3–19.2)	< 0.001
Inactive/unemployed	126 (19.6)	517 (80.4)	7.21 (4.9–10.5)	< 0.001

*  Among those who wore spectacles, data were missing for one subject on educational level and three subjects on occupation.

†  Among those who needed spectacles, data were missing for two subjects on literacy and three subjects on educational level and occupation.

Figure 1.

 

Visual acuity with habitual distance spectacles and wearing the objective result in trial frames (right eyes of 92 subjects). Solid line: unity. Several of the right eyes had a visual acuity of 6/12 or better on presentation. The left eyes of these subjects had acuity less than 6/12; hence, the subjects were retested for corrected visual acuity.

Table 4A.

 

Suitability of Off-the-Shelf Spectacles for Meeting the Need for Spectacles

	Age									All Ages
		30–39 y	40–49 y	50–59 y	60–69 y	70+ y
Number examined, *	4,829	2,947	1,660	1,167	586	11,189
Would not benefit from a distance correction	4,726	2,722	1,333	843	423	10,047
% of number examined	97.9	92.4	80.3	72.2	72.2	89.8
Uncorrected visual acuity 6/12 or better in better eye	4,702	2,687	1,256	640	190	9,475
% of those who would not benefit	99.5	98.7	94.2	75.9	44.9	94.3
Best corrected visual acuity of <6/12 in both eyes	24	35	77	203	233	572
% of those who would not benefit	0.5	1.3	5.8	24.1	55.1	5.7
Would benefit from a distance correction, †	103	225	327	324	163	1,142
% of number examined	2.1	7.6	19.7	27.8	27.8	10.2
Astigmatism >1.25D, ‡ , §	33	33	53	76	69	264
% of those who would benefit	32.0	14.7	16.2	23.5	42.3	23.1
Anisometropia >2.00 D, §	3	11	25	35	31	105
% of those that would benefit	2.9	4.9	7.6	10.8	19.0	9.2
Suitable for “off the shelf” spectacles	71	186	260	230	80	827
% of those who would benefit	68.9	82.7	79.5	70.9	49.1	72.4

*  Excludes post-cataract operated subjects (123 subjects) and those in whom refractive error could not be recorded (132 subjects).

†  Includes the remaining subjects after exclusion of those who do not need spectacles and those with a best corrected visual acuity of less than 6/12 in the better eye.

‡  of the right eye

§  Some subjects had both anisometropia and astigmatism.

Table 4B.

 

Myopes (less than −1D) and Hyperopes (more than +1D) among the Group Suitable for Off-the-Shelf Spectacles

Number Suitable by Type of Refractive Error	Age (y)									All Ages
		30–39	40–49	50–59	60–69	70+
Hyperopes, *	8	91	134	88	19	340
% of all suitable	11.3	48.9	51.5	38.3	23.8	41.1
(Range of > +1 to +3 D)	6	87	120	84	19	316
Myopes, *	16	25	72	99	51	263
% of all suitable	22.5	13.4	27.7	43.0	63.8	31.8
(in range of < −1 to −3 D)	13	18	47	87	30	195

 The number who would benefit from off-the-shelf spectacles in the range −3 to +3 D in 0.5-D steps is also given.

*  spherical equivalent of the right eye of a subject is presented as the refractive error for that subject.

The authors thank the three teams who performed the survey (all recruited from the National Institute of Ophthalmology, Dhaka), led by Syed Abdul Wadud, Shajahan Ali, and Lakshman Kumar Kar (ophthalmologists), and the team ophthalmic nurses Smrity Arinda, Nilufa Begum and Benju Rani Talukder; and Zakir Hussain Khan, and Kaesur Rahman for providing logistic support; Clare Gilbert and Allen Foster for reading and providing advice on preparation of the manuscript and assistance with the definition of spectacle coverage.