Changes in blindness prevalence over 16 years in Malawi: reduced prevalence but increased numbers of blind

Abstract

Background/aims: In the coming two decades significant increases in the burden of blindness are anticipated unless concerted efforts are made to improve eye care in developing countries. Evidence of changing prevalence rates or numbers of blind people are few. The change in blindness prevalence and the number of blind people in an adult population of Malawi was measured over a 16 year period.

Methods: In 1999 a population based survey of blindness in adults (age 50+) was conducted in Chikwawa district of Malawi. Visual acuity and cause of vision loss were recorded for each eye independently. Blindness was defined as presenting better eye vision of <6/60. Findings from a 1983 survey of blindness in the same district (using similar methods) were re-analysed to be comparable with the survey conducted in 1999.

Results: Among 1630 enumerated adults 89% were examined. The age adjusted prevalence of blindness in the adult population was 5.4% and more common in women than men. In each age group the prevalence of blindness was lower in 1999 than in 1983; the overall reduction in blindness was 31%. During this period the 50+ population in Malawi increased almost twofold. Extrapolating the Chikwawa district data to the Malawi population reveals that the number of blind people has increased by 24%; the increase is primarily because of the large increase in the size of the most elderly group, aged 70 and above.

Conclusion: The majority of blind people in Chikwawa (1983 and 1999) are in the age group 70 and over. This group has had the largest proportional increase in population size in this time. Services in this population have improved in the intervening 16 years and yet there was still an increase in the number of blind people. There was little change in excess blindness in women, suggesting that the same barriers that prevented utilisation of services in 1983 probably persist in 1999. Efforts to reach the most elderly and to reach women are needed to lead to a reduction in blind people in settings such as rural Malawi.

The World Health Organization estimates that blindness will increase from the current 45 million to 75 million by the year 2020. These estimates are based on projected growth in global population figures, on projected increases in life expectancy, and an assumption that there will be no further improvement in eye care delivery.¹ In response to these projections district based Vision 2020 plans are aimed at identifying targets for interventions (primarily for cataract and trichiasis surgery) to lead to reductions in blindness. Most districts in countries in Africa have some eye care activities already under way. Few have had the opportunity to evaluate the current magnitude of their eye care needs. Fewer still have been able to assess the change, over time, of the burden of blindness. A notable exception has been the well documented Gambian National Eye Care Programme. Two surveys in 1986 and 1996 demonstrated a relative reduction of blindness of 40% in the 10 year period.²

During the autumn of 1983 the Malawi government, in conjunction with the International Eye Foundation, Johns Hopkins University, and other partners, conducted a population based study of blindness, trachoma, and vitamin A deficiency in the Lower Shire River Valley (Chikwawa and Nsanje districts) in the southern region of Malawi. Findings from this survey have been published^3–5 and were used to develop trachoma and vitamin A deficiency control programmes in the district and elsewhere in Malawi.

We recently conducted a survey of blindness and low vision and trachoma⁶ in Chikwawa District of Malawi to help determine the eye care needs of the population, set targets for reduction of blindness, and to assess the burden of trachoma. We also sought to determine the change in blindness prevalence and burden of blindness in this population compared to the 1983 survey. The objectives of the 1999 Chikwawa survey were also to determine the prevalence of cataract (with vision <6/60 and <3/60) in adults aged 50 and older as well as various measures of cataract surgical coverage and outcome of cataract surgery (not presented here).

METHODS

Study site

There are two tertiary eye care centres in Malawi, one at the Lilongwe Central Hospital and the other at the Queen Elizabeth Central Hospital (Blantyre). QECH serves the southern region of Malawi (population 5.5 million) where there are 11 districts, each of which had an ophthalmic assistant (OA). QECH had one ophthalmologist and one cataract surgeon.

The study site, Chikwawa District, had a population of 260 000 in 1983 and 375 000 in 1999. Most of the population lives in the rural area; trading centres are found in Ngabu, Nchalo, and Chikwawa. Between 1983 and 1999 eye care services in Chikwawa District were provided by one OA based at the district hospital (30 minutes from Blantyre) occasionally supplemented by a second OA, and one visiting cataract surgeon (from Blantyre). There was no district eye staff in Chikwawa before 1983. There has been no ophthalmologist resident in Chikwawa; for most of the period between 1983 and 1999 the (non-physician) cataract surgeon performed intracapsular cataract surgery with provision of aphakic spectacles. Since the late 1990s a visiting ophthalmologist (from Blantyre) has been providing cataract surgery with IOL implantation.

A previous assessment of surgical uptake in Chikwawa showed relatively low uptake of services as well as significant barriers to use of services.⁷ At the time of the 1999 survey the output of the eye units in the southern region of Malawi was 1157 cataract operations (88% with an IOL), low when compared to the need and the potential of a population of 5.5 million.

Sample size and sampling

Methods used in this study have been described previously.⁶ Briefly, the sample size calculation for the 1999 survey was based on the assumption that the prevalence of blindness (visual acuity <6/60) in people 50 or older was 10%. Given an α value of 0.05 and β value of 0.2, about 864 people over 50 years old were needed for the survey. In order to adjust for clustering (design effect of 1.5) for clusters of 60 residents and a response rate of 90% the total sample size needed was estimated to be 1500.

The demographic data of the 1987 census (updated by the Chikwawa District Commission 1997) were employed as the frame for sampling. A list of villages and their population was created. Sampling clusters were made in order to yield about 60 people aged 50 years or older. The sampling clusters were created by grouping villages with less than 550 population and subdividing villages with more than 1500 population into segments. The actual geographic boundaries of these segments, based on the local layout, were defined only for the chosen segments. Twenty five clusters were randomly selected using a simple random sampling of the clusters.

Data collection and quality control

Data were collected at the cluster level, household level, and individual level. At the household level, enumerators completed the household form by listing all the details (age, sex, marital status, occupation) for those 50 years of age and over; these individuals were requested to come for examination at a central site in the village. Trained OAs using a standard protocol and eye examination record conducted examinations. Examination included testing of presenting visual acuity using a Snellen tumbling E chart at a distance of 6 metres. Pinhole vision was taken for those individuals with a vision of <6/18 (either eye). Assessment of cause of vision loss (<6/18 after pinhole correction) was undertaken using a standard protocol with a direct ophthalmoscope after pupil dilatation. One of us (NM) supervised OA examinations to assess causes of vision loss in cases that could not be determined by the OA. A pilot study to assess the quality of OA examination and vision taking was done in two non-study villages, one by each of the clinical teams. The survey was conducted from July to October 1999.

Data from the 1983 survey, archived at Johns Hopkins University, were abstracted and findings reanalysed by age and sex (limited to Chikwawa only) according to a definition of blindness of <6/60 (presenting better eye).

Analysis

For the purpose of this study, blindness was defined as presenting vision in the better eye of <6/60. This cut off was chosen because previous work in Malawi⁸ had shown significant reductions in visual function at <6/60. The principal causes of blindness were grouped into causes related to lesions of the anterior segment, related to the lens, the posterior segment and related to the whole globe. The causes of blindness related to lens included cataract and uncorrected aphakia. The causes of anterior segment blindness were trachomatous and non-trachomatous corneal opacities. The posterior segment causes of blindness include glaucoma, optic atrophy, macular degeneration, retinal detachment, congenital anomalies, vascular retinopathy, refractive error, and uveitis. The causes of blindness related to the whole globe included phthisis, disorganised, and absent eyeball.

Ethical approval

The study was approved by the Malawi committee of medical research. After explaining the purpose of the survey, village leaders were requested to consent to survey in their villages. Villagers were also requested to participate and verbal consent was obtained.

RESULTS

Among the 1630 residents 50+ years of age who were enumerated, 1384 were examined during the survey, yielding a participation rate of 88.8%. Participation was slightly higher for women (90.8%) than for men (86.3%). There was no regional variation in rate of participation.

The age adjusted prevalence of bilateral blindness (<6/60) in the survey population was 53.9/1000 among those 50 years of age or older (Table 1). Overall, blindness was more common in women than men; the female: male ratio for bilateral blindness was 1.94:1. Only 12.3% of bilateral blind were between 50–59 years of age. 21.5% were in the 60–69 age group, and more than half of the blind were in the 70 years and above age group.

Table 1.

Prevalence of bilateral blindness (visual acuity <6/60) by age and sex (1999)

	Bilateral visual acuity <6/60	Bilateral visual acuity <3/60	Unilateral visual acuity <6/60
Age group	No (%)	No (%)	No (%)
Males
50–59	1 (0.4)	1 (0.4)	16 (6.4)
60–69	6 (4.1)	4 (2.9)	11 (8.1)
70+	15 (11.9)	10 (7.9)	20 (15.9)
Age adjusted	4.8%	3.3%	10.2%
Females
50–59	7 (2.1)	5 (1.5)	15 (4.4)
60–69	8 (4.0)	5 (2.5)	14 (7.0)
70+	28 (16.4)	14 (8.2)	21 (12.3)
Age adjusted	6.1%	4.0%	8.3%

The prevalence of unilateral blindness in Chikwawa was 79.2/1000 population. Unilateral blindness was more common in men than women (Table 1).

Sixty five bilaterally blind (visual acuity <6/60) people were identified in the survey. The lesions related to the lens were the most common, followed by others and corneal opacities (Table 2). Overall, more than 79% of the identified cases were either curable or preventable by appropriate interventions. Among the 97 unilaterally blind people identified in the survey the principal causes of their blindness were lens related followed by others and corneal opacities.

Table 2.

Causes of bilateral blindness (visual acuity <6/60) by sex

	Men	Women
Cause	No	No	Total No (%)
Unoperated cataract	16	24	40 (61.5)
Uncorrected aphakia	1	2	3 (4.6)
Corneal opacity	0	6	6 (9.2)
Globe	1	2	3 (4.6)
Other	4	9	13 (20.0)
Total	22	43	65

Comparison of 1999 and 1983 blindness data

Comparing the 1983 blindness (using the same definition of presenting vision <6/60 in the better eye) with the 1999 data revealed reductions in the prevalence of blindness in all adult age groups (Table 3). Overall, there was a 31% reduction in blindness in the 16 years between 1983 and 1999. The odds of blindness for women compared to men changed little in this period; women had an odds of blindness of 1.32 (95% CI 0.52 to 3.40) compared to men in 1983 and of 1.43 (95% CI 0.82 to 2.51) compared to men in 1999.

Table 3.

Comparison of 1983 and 1999 age specific prevalence of blindness (<6/60)

	Prevalence and 95% CI
Age group	1983	1999
50–59	2.13 (0.77 to 3.69)	1.36 (0.87 to 1.84)
60–69	10.34 (7.53 to 13.08)	4.19 (3.13 to 5.21)
70+	16.67 (11.35 to 21.97)	14.68 (12.91 to 16.37)
Age adjusted	7.72 (6.34 to 9.03)	5.33 (4.80 to 5.83)

Between 1983 and 1999 the total Malawi population of the 50+ age group increased from 598 051 to 961 614 people. Applying the Chikwawa blindness data to the census data reveals that although blindness rates have decreased, the overall number of blind residents (age 50+) had increased from 41 316 to 55 337, a 24.1% increase (Table 4). The increase is because of the large increase in the eldest age group.

Table 4.

Change in number of blind residents in Malawi extrapolating from Chikwawa data

Age group	1983 population	Estimated blind	1999 population	Estimated blind
50–59	320 450	6826	414 072	5631
60–69	186 192	19 252	292 404	12 252
70+	91 409	15 238	255 138	37 454
Total	598 051	41 316	961 614	55 337

DISCUSSION

As expected, the majority of blindness in Chikwawa (in both 1983 and 1999) occurred among those in the oldest age group (70+ years). This group had the largest proportional increase in population size in this time period, similar to other developing countries. Thus, the decrease in blindness, which would be enviable in many other settings, failed to result in an overall decrease in the number of blind people.

The proportion of blindness due to cataract in 1983 was 40% while in 1999 it was 61% (66.1% if uncorrected aphakia is included). Corneal opacity related blindness decreased proportionately, from 30% to 9%. These changes may be due to a number of factors. Firstly, although re-analysis of the 1983 data to generate age and sex rates of blindness at the <6/60 cut off was performed, it was not possible to generate cause of blindness data for this vision cut off; accordingly, the 1983 data we are using to determine cause of blindness in 1983 (<3/60) may be slightly different from our 1999 data. Similarly, there may have been some differences in the study protocol; however, the procedures in 1999 were based on the 1983 protocol. Corneal opacities, secondary to infectious causes (vitamin A deficiency and measles in particular) have decreased significantly, probably as a result of Ministry of Health and International Eye Foundation programmes to improve nutritional status and vitamin A status and collaborative programmes with traditional healers.⁹ It should be recognised that routine cataract surgery in Chikwawa District was not practised before 1983; although there is no information on cataract surgical coverage from 1983, it is likely to have been much less than recorded in 1999. Since 1999 there have been considerable changes in eye care delivery in Chikwawa and the southern region of Malawi. In particular, there has been a significant expansion of IOL surgery in the districts; in 2001 there were 1560 cataract surgeries in the southern region (96% with an IOL).

Interpretation of our findings is limited by the 16 year span between the two surveys. Although protocols and definitions used were similar there is a potential for bias in case definition or allocation, which may have led to overestimation or underestimation of blindness in the two surveys.

Services in this population have improved in the past 16 years, and yet we still found an increase of 24% in the number of blind. The implications of our findings to other settings are dependent upon a number of factors. The size and growth patterns of the ageing population vary globally and there is concern in Africa that the HIV/AIDS epidemic will lead to significant reductions in the adult population. Even if HIV/AIDS has a significant impact on demographic patterns in sub-Saharan Africa, however, it is likely that this impact will not be realised for the elderly population for the next 10–20 years. With changing demographic profiles, it will be hazardous to extrapolate blindness prevalence rates in countries like Malawi after the year 2020. Within the next 15 years, however, the socioeconomic consequences from a reduction in the working age population due to HIV/AIDS may have a significant impact upon utilisation of services by the elderly. Any extrapolation of our findings is also dependent upon the availability, accessibility, and acceptance of eye care services by the elderly. Over the 16 year period from 1983 cataract surgical services became much more available and accessible to the population; however, it is not clear how accepting the population (especially the most elderly) has been to these services. Even other settings in Malawi where cataract surgical services have been locally available for many years have found very low uptake of services.¹⁰

We do not have sufficient power in our study to compare sex specific blindness rates; however, our findings are consistent with a meta-analysis of blindness surveys showing a significant excess of blindness among women compared to men.¹¹ Interestingly, there was little change in the excess of blindness among women in the 16 years, suggesting that the same risk factors predisposing blindness in women and the same barriers that prevented utilisation of services to prevent blindness in 1983 may have continued to exist in 1999.

The Chikwawa survey of 1999 shows that, even with reductions in age-specific blindness rates among the 50+ age group, there has been an increase in the number of blind people in the country. Programmes recently put in place to increase the productivity, efficiency, and quality of cataract surgery will probably have significant impact on changes in blindness prevalence over the next 16 years. It will be critical, however, to target services to the most elderly and to women to realise the Vision 2020 goals and further reduce blindness in the population.