COMMUNITY EYE CARE
Year : 1999 | Volume
: 47 | Issue : 2 | Page : 135--141
Rapid assessment of cataract blindness in an urban district of Gujarat
Hans Limburg1, Abhay R Vasavada2, Gautam Muzumdar3, MY Khan4, K Vaidyanathan1, Rupal Trivedi2, Deepa Bhatt2,
1 Danish Assistance to National Programme for Control of Blindness, New Delhi, India
2 Ila Devi Cataract & Intraocular Lens Research Centre, Ahmedabad, India
3 Indian Red Cross Society Eye Bank, Ahmedabad, India
4 Department of Health Services, Government of Gujarat, India
DANPCB, A-1/148, Sj Enclave, New Delhi 110 029
Purpose: To estimate the prevalence of bilateral cataract blindness in persons ≥50 years of age in Ahmedabad district, Gujarat.
Methods: A total of 1,962 persons ≥50 years of age were examined in clusters of 45 people or less. The survey design used a systematic random cluster sampling. The sample size was calculated assuming a prevalence of bilateral cataract blindness (visual acuity <3/60) of at least 3% and design effect of 1.6, to estimate the actual prevalence of cataract blindness with a sampling error of ≤20 at 80% confidence level. Visual acuity was assessed with glasses, where available, and pinhole was used for visual acuity <6/18. Distant direct ophthalmoscopy in semidark condition with undilated pupil was used to assess the lens status.
Results: The age-gender-adjusted prevalence of all blindness was 2.9% in persons ≥50 years of age (6.7% for visual acuity<6/60). The age-gender-adjusted prevalence of bilateral cataract blindness ( visual acuity <3/60) was 1.2% in persons ≥50 years of age. For visual acuity <6/60, the prevalence was 3.1%. The prevalence in females was slightly higher than in males. The prevalence of bilateral and unilateral aphakia and pseudophakia was high. The cataract surgical coverage, an indicator for coverage and service utilization, was 92.9% for persons and 83.1% for eyes. Conclusion: Rapid assessment of cataract blindness in persons ≥50 years of age can be conducted in urban settings with existing resources and at affodable costs, to provide district level data for assessment and monitoring of cataract intervention programs.
|How to cite this article:|
Limburg H, Vasavada AR, Muzumdar G, Khan M Y, Vaidyanathan K, Trivedi R, Bhatt D. Rapid assessment of cataract blindness in an urban district of Gujarat.Indian J Ophthalmol 1999;47:135-141
|How to cite this URL:|
Limburg H, Vasavada AR, Muzumdar G, Khan M Y, Vaidyanathan K, Trivedi R, Bhatt D. Rapid assessment of cataract blindness in an urban district of Gujarat. Indian J Ophthalmol [serial online] 1999 [cited 2022 Dec 5 ];47:135-141
Available from: https://www.ijo.in/text.asp?1999/47/2/135/22789
Age-related cataract remains the single major cause of blindness in most developing countries, including India. Declining birth rates and a rapidly increasing life expectancy have led to an increase of the population above 50 years of age. This, combined with the limited capacity to cover the increased demand for cataract surgical services, cause an increase in blindness from age-related cataract.
In India, the district has become the nucleus for implementation of health programs, including the blindness-control program. District Blindness Control Societies now prepare annual action plans to intensify eye-care services in their districts. Most of the emphasis is on the reduction of cataract blindness. To estimate the need for services and to plan for adequate manpower and logistics, extrapolations are generally made from state-level data. However, prevalence and causes of blindness are likely to vary considerably among districts due to differences in socio-economic situation, in age and gender composition of the population ≥50 years of age, in available resources for surgical services, and in possible risk factors. Therefore, the need to obtain district-level data on the requirements for cataract services and the extent to which the cataract problem is covered became obvious. If repeated every 4-5 years, such rapid assessments could indicate trends in prevalence and coverage and thereby facilitate effective monitoring and planning of cataract-intervention programs at the district level. Existing eye-care resources in most districts of India are adequate to conduct a limited survey using a simple, standardised methodology.
In this article results are presented of one such rapid assessment, conducted in the largely urban district of Ahmedabad in Gujarat. The rapid assessment methodology used in Ahmedabad is a modification of the methodology designed for and tried out in 19 out of 20 districts of Karnataka State. Details of this methodology have been published elsewhere. By limiting the survey to cataract blindness in the population ≥50 years of age, the sample size can remain low, the time and cost of the survey are limited and such surveys can be repeated more frequently. The methodology is kept simple so that it can be carried out by locally available ophthalmologists or well-trained and experienced ophthalmic assistants.
Materials and Methods
Ahmedabad district consists mainly of the city of Ahmedabad and some surrounding villages and small towns. In 1991, the census recorded at a total population of 4,801,812; 74.7% of the population is urban, 25.3% rural. Using the Sample Registration System data, the total population in 1997 was estimated at 5.265 million, with 616,500 persons ≥50 years of age-321,500 males and 295,000 females.
The prevalence of bilateral cataract blindness (visual acuity <3/60) in 1986 was reported to be 0.44% for Gujarat State, against 0.56% for India as a whole. The number of cataract operations in 1995 was over 5,000 per million population in Gujarat and 2,665 for India. This implies that the situation in terms of cataract blindness in Gujarat is likely to be better than the all-India average.
To restrict the sample size it was decided to survey the population ≥50 years of age only. To calculate the sample size, an estimate of the expected prevalence of cataract blindness in the population ≥50 years of age in Gujarat had to be made. In the Survey of Blindness and Visual Impairment in Gujarat, conducted in 1992, 505,000 persons out of the total population in 1992 of 43.5 million were reported to be bilaterally blind (visual acuity <6/60). That gives a prevalence of blindness of 1.16%. People ≥50 years of age are 12.9% of the total population in the state: 5.61 million persons. Of all blindness, 58.3% was caused by cataract.
The WHO-NPCB survey indicated that 95% of all cataract blindness occurs in the age group ≥50 years. Thus, out of the 505,000 people with bilateral blindness, an estimated 294,000 (58.3%) were blind due to cataract and 280,000 (95%) of these people were ≥50 years of age. Therefore, the prevalence of bilateral cataract blindness (visual acuity <6/60) in persons ≥50 years of age in Gujarat was 5% (280,000/5,610,000) in 1992. No data were given on prevalence of cataract blindness for visual auity <3/60, but for sample size calculations, this was estimated at 3%. Allowing a 20% sampling error, the sample size for simple random sampling was calculated to be 1,325 for 80% confidence level.
People in the same cluster have a greater tendency to share the characteristic under examination. The level of variation in distribution of this characteristic in different clusters can be expressed in the design effect, which can be used to increase the sample size required for cluster sampling as compared with simple, random sampling. In this survey a cluster size of 45 was preferred, enabling each team to cover two clusters in one day. The design effect for cluster size 45 was estimated in earlier surveys at 1.6. Therefore, a total sample size of 2,120 would be required. Hence, a sampling design of 45 clusters of 45 persons, giving a total sample of 2,025 persons ≥50 years of age, was selected in Ahmedabad district.
Assuming a sample-based prevalence of cataract blindness to be 3% in persons ≥50 years of age, the probability that the true prevalence would be between 2.4% and 3.6% would be 80%.
Another important consideration related to the logistics required to organize such a survey. For a field survey, the cost, feasibility, and quality of field work have to be balanced with the desired precision.
Six survey teams were formed, three by the Ila Devi Cataract & Intraocular Lens Research Centre and three by the Directorate of Health Services, Government of Gujarat. Each team consisted of an ophthalmologist, a trained paramedical ophthalmic assistant (PMOA), a field supervisor, and the local health worker. All team members attended a one-day training program in survey methodology, field practices and completion of the formats, at the end of which an assessment was made on the inter-observer variation. Since the clinical examination was limited to the examination of the lens (normal; obvious opacity; aphakia; pseudophakia) and visual acuity, the inter-observer agreement was near perfect, as has been reported by other studies also.
The sampling frame was designed using the 1991 census which lists the population of all localities of Ahmedabad city and the adjoining villages. These data were transferred to a spreadsheet, adding a column with cumulative population of the preceding colonies, villages and towns. By dividing the total population of the entire district by the number of clusters (in this case 45) the sampling interval was obtained. A random number was selected between 1 and the sampling interval, and this number was marked on the cumulative list. The colony or village where this marked individual was living was taken as the site for the first cluster and the starting point for the systematic random sampling. The remaining 44 areas were identified by adding the sampling interval to this random number and selecting the village or town in which the marked individuals were residing. Within each area, 45 persons ≥50 years of age were selected by the cluster method, following the procedure recommended for rapid assessments. If the community in the colony or village was too small to provide the required 45 persons ≥50 years of age, the investigators were instructed to go to the next nearest colony or village to complete the required number. By following this procedure, areas are selected with probability proportional to size. This procedure is self-weighing with regard to the total target population.
Cataract surgical coverage
From the data obtained through this rapid assessment we can also calculate the cataract surgical coverage. This indicator measures what proportion of the total problem of cataract blindness has been covered by services in the form of cataract surgery, irrespective of the outcome. This coverage indicator can be used for cataract blind eyes, as well as persons bilaterally blind from cataract. It gives a good impression about the availability, accessibility and utilisation of cataract surgical services over time.,
Besides measurement of visual acuity and assessment of the lens status, persons in the sample, if blind in one or both eyes due to cataract, were asked why they had not been operated so far. The 14 most common reasons are listed in the questionnaire and only the answer closest to the reply of the patient was marked. This provides a first impression of common barriers to cataract surgery. Lastly, persons who were operated for cataract earlier were asked about the age at the time of surgery, whether it was done in a camp or hospital, and about subsequent provision and quality of spectacles.
For the purpose of this assessment, blindness due to cataract was defined as obvious lenticular opacity combined with a visual acuity <3/60 in the better eye with the available correction, or with pinhole. A person was called blind due to cataract if both eyes met these criteria. In India, blindness is defined as visual acuity <6/60 in the better eye with the available correction. These data for Ahmedabad district are also reported.
The teams started off early in the morning and were able to cover a second, nearby, cluster in the afternoon of the same day. With 6 teams, all 45 clusters could be finished in 5 days, with a few extra days to revisit the non-respondents.
Although the survey was preceeded by publicity, some persons ≥50 years of age could not be examined. In these cases, an anecdotal visual status was obtained through interview of members of the same household or nearest neighbours.
Data entry and standardised analysis was done using a specially designed software programme (Epi-Info, version 6.04).
A total of 2,048 eligible persons were listed in Ahmedabad district. The sample was not subdivided in terms of age, rural-urban, or gender. In total, 1,962 persons ≥50 years of age, 799 males and 1163 females, could be physically examined (95.8%). Of the remaining 86 persons who could not be examined, anecdotal visual status was obtained by interviewing members of the same household or the nearest neighbours. These data are not included in the analysis presented here.
Prevalence of cataract blindness
The sample prevalence for all blindness, bilateral cataract blindness, cataract blind eyes, bilateral aphakia and aphakic eyes are given in [Table:1]. The sample size prevalence of bilateral cataract blindness (visual acuity <3/60) was 1.3%, far less than the 3% we assumed in the calculations of the sample size.
Since the composition of the sample, as far as age and gender of the persons examined is concerned, may not be equal to the composition of the actual population of the district, the sample prevalence rates were adjusted for the actual population of the district [Table:2]. The National Census of 1991 provided age and gender composition of the ≥50 years of age population of Ahmedabad district. Using data from the Sample Registration System, adjustments could be made for the year 1997. In making these projections to estimate the actual number of patients or eyes in the district, it was assumed that the prevalence of (cataract) blindness and aphakia in persons younger than 50 years is negligible. Since an estimated 5% of all cataract blindness occurs in persons below 50 years of age, the data given here are likely to be underestimates.
The study results showed an average age-gender-adjusted prevalence of bilateral cataract blindness in persons ≥50 years of age (visual acuity <3/60) of 1.2%-1.1% for males and 1.3% for females [Table:2]. For cataract blindness defined as visual acuity <6/60, the average age-gender-adjusted prevalence was 3.1% (95% confidence interval [CI] 2.2%-4.3%), with a variation of 2.9% for males and 3.2% for females. Prevalence for females was slightly higher compared to males. The prevalence of bilateral (pseudo) aphakia and (pseudo) aphakic eyes was high. More females had been operated for cataract than males.
In people ≥50 years of age, 46% of all severe visual impairment (visual acuity <6/60) was caused by cataract, for blindness (visual acuity <3/60) it was 41.3%, which is substantially lower than the 80.1% in the WHO-NPCB survey of 1986 and the 58.3% from the 1992 Gujarat survey. Of all cataract operations, 62.7% were done on "first" eyes, and 37.3% on patients who had earlier been operated on for cataract in the other eye.
In the sample, 776 eyes had undergone cataract surgery: 587 eyes were aphakic and 189 eyes were pseudophakic. The visual acuity of these eyes postoperatively, with the available correction, is given in [Table:3].
The mean postoperative period was 5.5 years with a range of 0-30 years. Since IOL implants are a recent development, the average follow-up period for IOLs is likely to be shorter than for non-IOLs. Of those with IOLs, 66% could see 6/18; of those without IOLs, 43% could see 6/18. Of those with IOLs, 11.6% could not see 6/60, among the non-IOLs this figure rose to 28%. Reasons for poor visual outcome were not assessed in this study.
In total, 16 eyes with pseudophakia also had posterior capsule opacification-8 (9%) male eyes and 8 (7.9%) female eyes. Six eyes had visual acuity >6/18; eight eyes could see 6/60, but did not improve with pinhole, and 2 eyes had a visual acuity <3/60 with no improvement on pinhole. There was no relationship between visual acuity and the duration of the interval between surgery and lens assessment in this survey.
For calculation of cataract surgical coverage, aphakia and pseudophakia were considered identical [Table:4]. It is not possible in population-based surveys to determine retrospectively what the visual acuity was at the time of surgery. The cataract surgical coverage for visual acuity <3/60 was 92.9% [Table:4].
If we assume that only eyes with visual acuity <3/60 were operated upon in Ahmedabad district, then 92.9% of those who could have potentially been blind due to cataract, were operated in one or both eyes. There was no difference between males and females. The cataract surgical coverage for eyes was on average 83.1% for Ahmedabad with minimal difference for gender. If we assume that only eyes with visual acuity <6/60 were operated upon, then the cataract surgical coverage was 84.4% for persons and 71.5% for eyes.
Barriers to cataract surgery
Those persons who were blind in one or both eyes from cataract, were asked why they had not come forward to have their eye(s) operated so far. In the entire sample, there were only 26 persons bilaterally blind from cataract (visual acuity <3/60). 'Medical contraindications' was the barrier for 7 patients, 'waiting to mature' for 6, and a variety of other reasons for the remaining 13 patients. Of the 133 patients with unilateral cataract blindness, 'cataract not yet mature' was mentioned by 31, 'need not felt' by 30, 'contraindications' by 20, 'fear of operation' by 18 and 'one eye operated' by 13. The remaining 21 mentioned one of various other reasons.
For people with severe visual impairment from cataract (visual acuity <6/60), the reasons for not being operated on are given in [Table:5]. In males, 'waiting to mature' was the most common barrier. This may either mean that these persons had requested surgery, but were advised to come at a later stage, or that these patients themselves felt that they had to be totally blind before qualifying for surgery. 'No felt need' was mentioned as frequently by males as by females. 'Fear' was prominent in females. 'Not knowing where to go' was not mentioned, indicating a high level of public awareness.
A total of 558 patients, operated for cataract in one or both eyes, were also questioned on different aspects of the services provided to them. In Ahmedabad district, services are more or less equally shared by the Government sector, non-governmental organisations (NGOs), and the private sector, and are mainly institution based.
Of the total 776 operations performed, 27% were in Government hospitals, 37% through NGOs, 31% by the private sector and 6% in eye camps. Slightly more females were operated on in eye camps, but otherwise the distribution for males and females was equal.
All patients interviewed received spectacles after surgery. Spectacles were provided free of cost to 48%, against subsidised rates to 22%, and 30% of the patients paid the full cost. Spectacles and IOLs were considered equal in this case.
Patients who paid for their spectacles used them most; 53%. Since these are data from a population-based survey, covering people operated recently, but also as much as 30 years ago, some of these people may have replaced their first pair of spectacles. Of patients who got free spectacles, 24% were still using them. A total of 21% of patients had no spectacles at the time of interview. About half of these were operated on with IOL implants. The quality of spectacles was appreciated by 80% of the patients, although more women complained about poor quality spectacles.
The precision of the estimate of cataract blindness was lower than anticipated. We assumed a prevalence of 3% for bilateral cataract blindness with visual acuity <3/60 in people ≥50 years of age. We found a prevalence of 1.3% in the sample. The 95% CI was 0.9-1.9%, a variation of 31%. The design effect was lower than assumed, 1.0, indicating equal distribution of cataract blindness in all clusters.
The prevalence of bilateral cataract blindness with visual acuity <6/60 was 3.5% in the sample. The 95% CI was 2.6-4.7%, a variation of 26%. The design effect here was 1.5.
For bilateral aphakia, the sample prevalence was 14.9%, with a 95% CI of 13.4-16.5%, an 11% variation. The design effect here was 2.8.
Rapid assessments do not aim to replace detailed surveys on magnitude and causes of blindness. Detailed surveys are large, complicated exercises that need specialised staff, considerable planning, time and funds. They are usually complex and cannot be conducted by local staff alone. Analysis of data requires expertise and the results are mainly used to direct future policies.
The rapid assessment presented here had a limited objective of assessing the prevalence of cataract blindness in the population ≥50 years of age in a district. The methodology was designed specifically for implementation by local ophthalmologists or ophthalmic assistants after proper training.
The eye examination was limited to measuring visual acuity, with available correction, and, if <6/18, also with pinhole, and assessment of the lens status by distant direct ophthalmoscopy, without dilatation of the pupil. Cataract blindness was defined as an eye with obvious opacification of the lens and a visual acuity <3/60, that did not improve with pinhole. A person was termed blind due to cataract if both eyes met these criteria. The posterior segment was not examined. A similar procedure was followed in the 1981 Nepal Blindness Survey, in which only visual impairment could not be explained by any anterior segment findings, the pupil was dilated and the posterior segment examined by ophthalmoscope.
Some patients could be blind due to an obvious lens opacity and co-existing retinal or optic nerve disease. Accurate estimation of this can be provided only be detailed dilated eye examination. Posterior segment disease or disorder were reported to cause 3.6% of blindness in people in India and 9.9% of all blind eyes in Nepal. The last study also reported that less than 10% of the blind persons had two or more co-existing blinding disorders.
A standard software programme, with data entry and automatic data analysis has been tested extensively with these data and is now available for use elsewhere. This will enable district staff to conduct such rapid assessments in a standardised way in their own districts. However, the precision achieved in these rapid assessments is limited and therefore the data obtained should be regarded as indicative.
Blindness (visual acuity <6/60) in Ahmedabad district in 1997 is estimated at 3.1% (95%CI 2.2-4.3%) in people ≥50 years of age, which seems considerably less than the 5% in 1992. However, in the absence of confidence intervals of the 1992 study, it is not possible to say whether the difference is statistically significant.
Also, the 1992 survey gives average data for Gujarat state, which also includes rural districts where prevalence of blindness would be much higher. Ahmedabad district is largely urbanised, has an abundant availability of ophthalmologists and eye-care facilities and a high level of public awareness.
Cataract blindness in persons ≥50 years of age in Ahmedabad district accounts for less than 50% of all blindness, indicating that the available services are well used by people who need them. This is further expressed by the high prevalence of aphakia. Population-based prevalence surveys look back at a number of years and thereby recent trends towards more IOL surgery are diluted by earlier practices.
IOLs gave a better visual outcome, with 66.1% having visual acuity >6/18, compared to 48.6% for the non-IOLs. Of the 776 operated eyes, 186 (24%) could not see 6/60. Of the 189 eyes with IOLs, 11.6% could not see 6/60; of the 587 aphakic eyes with spectacles, 28% could not see 6/60.
Of the aphakics with spectacle correction, 8% did not have any and 25% had "poor" quality spectacles. However, subsequent pinhole examination could only improve the visual acuity in 1.4% of those with visual acuity <6/18. Further research on the causes of poor visual outcome is needed. Inadequate correction of postoperative refractive errors may be an important cause.
In the 1986 National survey, the prevalence of cataract blindness in females was 30-50% higher as compared with males; in the Karnataka study it was 60-80% higher. In Ahmedabad, the prevalence of cataract blindness was only 10-20% higher in females as compared to males. The number of aphakic eyes in females was higher than in males.
The barriers in Ahmedabad district were also different from those in Karnataka. Lack of information was not mentioned. The main barriers were 'no felt need', 'waiting to mature' and 'fear'.
It appears from these results that Ahmedabad district is close to controlling cataract blindness. If we assume all operations to be done on eyes with visual acuity <3/60, then 83.1% of the eyes and 92.9% of all persons with bilateral cataract blindness have been operated upon. If we assume all operations to be done on eyes with visual acuity <6/60, then 71.5% of the eyes and 84.4% of all persons have been covered
It will never be possible to operate all persons with cataract blindness. There will always remain a hard-core group of patients who do not wish to undergo an operation or have contraindications for surgery. It seems most patients with cataract blindness in Ahmedabad today belong to that group.
In this rapid assessment, 1,962 persons ≥50 years of age, in 45 clusters of 45 persons each, were examined. After a one-day training session, 6 teams, each consisting of one ophthalmologist, one PMOA and a local health worker, could examine two clusters in one day. The entire fieldwork in Ahmedabad was finished in 5 days. Data entry was done simultaneously. Because the period of field work was so short, the survey teams were very enthusiastic and alert throughout. Had it taken 3 or more weeks, a routine and casual attitude might have crept in, which would have made the findings less acccurate.
This was the first time this rapid assessment methodology was tested in a large city. No major logistic problems were encountered. When people live in apartment blocks, strict adherence to survey protocols, regarding the random selection of clusters, is essential.
The prevalence of bilateral cataract blindness in the sample was much less than assumed in the design. The design effect was slightly lower than expected. A larger sample size is advisable to achieve a higher level of precision in urban areas with a high level of public awareness and adequate eye-care services.
Sample surveys, including rapid assessments, are always a compromise between accuracy and feasibility. In our opinion, the methodology presented here gives reliable and relevant data at low cost, within an acceptable period of time, without interfering too much with normal routine work. In case a higher level of accuracy is required, the sample size can be increased. The design effect can be estimated as 1.4-1.7 for a cluster size of 40-60.
The authors wish to thank Dr. P. Belani, Dr. G. Janaswamy, Dr. L. Shukla, Dr. P. Soni, Dr. A. Zaveri, the ophthalmic assistants and field staff from the Ila Devi Cataract & Intraocular Lens Research Centre, the Indian Red Cross Society Eye Bank, Dholka Branch and the Government of Gujarat, for the collection of field data; and Mr. K. Parik, Ms. P. Brahmbhatt and Ms. S. Shah for data entry.
|1||Foster A, Johnson GJ. Blindness in the developing world. Br J Ophthalmol 1993;77:398-99.|
|2||Government of India. Guidelines for District Blindness Control Societies. New Delhi: Ophthalmology Section, Directorate General of Health Services, Govt. of India, 1995.|
|3||Limburg H, Kumar R, Indrayan A, Sundaram KR. Rapid assessment of prevalence of cataract blindness at district level. Int J Epidemiol 1997;26:1049-54.|
|4||Mohan M. Survey of Blindness-India, Summary and Results. New Delhi: Dr. Rajendra Prasad Centre for Ophthalmic Sciences; All India Institute for Medical Sciences. 1989.|
|5||Statistical Unit, Central Ophthalmic Cell, Ministry of Health and Family Welfare, New Delhi, 1997.|
|6||Government of Gujarat. Blindness and Visual Impairment in Gujarat. Ahmedabad, India: State Ophthalmic Cell, Commissionerate of Health, Medical Services and Medical Education, Government of Gujarat; 1992.|
|7||Venkataswamy G, Lepkowski JM, Tulsiraj R, Brilliant GE, Shanmugham CAK, Vaidyanathan K, et al. Rapid epidemiological assessment of cataract blindness. Int J Epidemiol 1989;18 (Suppl 2):S60-67.|
|8||Bennet S, Woods T, Liyanage WM, Smith DL. A simplified general method for cluster-sample surveys of health in developing countries. Wld Hlth Statist Quart 1991;44:98-106.|
|9||Limburg H, Kumar R, Bachani D, Monitoring and evaluating cataract intervention in India. Br J Ophthalmol 1996;80:951-55.|
|10||Limburg H, Foster A. Cataract surgical coverage - an indicator to measure impact of cataract intervention programmes. J Comm Eye Health 1998;25:3-6.|
|11||Registrar General India. Sample Registration Bulletin. New Delhi: Registrar General India;1996. Vol 30, No. 1.|
|12||The Seva Foundation. The Epidemiology of Blindness in Nepal. Report of the 1981 Nepal Blindness Survey. Chelsea, MI, USA; The Seva Foundation; 1988.|