Article Text

Download PDFPDF

Rapid assessment of avoidable blindness in three counties, Jiangxi Province, China
  1. Baixiang Xiao1,
  2. Hannah Kuper1,
  3. Chunhong Guan2,
  4. Kirsten Bailey3,
  5. Hans Limburg1
  1. 1London School of Hygiene & Tropical Medicine, London, UK
  2. 2The Fred Hollows Foundation, Nanchang, PR China
  3. 3The Fred Hollows Foundation, Sydney, Australia
  1. Correspondence to Dr Xiao Baixiang, London School of Hygiene & Tropical Medicine, Keppel Street, London WC1E 7HT, UK; xiaobaixiang2006{at}


Background A survey was undertaken in 2007 to assess the prevalence and causes of blindness and visual impairment in people aged ≥50 years in three different counties in Jiangxi, South East China (Gao'an, Xin'gan and Wan'zai). The counties were purposively selected to assess the impact of established non-governmental organisation activities in two counties (Gao'an and Xin'gan) compared with a third county (Wan'zai) without such a programme.

Methods Clusters of 50 people aged ≥50 years were sampled with a probability proportional to the size of the population. Because of differences in expected prevalence and resources available for conducting the surveys, the total sample size varied from 4699 in Gao'an (94.0% response rate) to 3834 in Xin'gan (95.9%) and 2861 (95.4%) in Wan'zai. Households within clusters were selected through random walk sampling. Visual acuity (VA) was measured with a tumbling ‘E’ chart. Ophthalmologists examined people with VA<6/18 in either eye.

Results The prevalence of blindness (VA<3/60 in the better eye with available correction) was similar in Gao'an (1.5%, 95% CI 1.1% to 1.8%), Xin'gan (1.8%, 1.4% to 2.2%) and Wan'zai (1.6%, 1.2% to 2.1%), and the prevalence of visual impairment (VA<6/18 and ≥6/60) was approximately fourfold higher. Cataract was the leading cause of blindness in each of the three counties, while uncorrected refractive error was the dominant cause of visual impairment. The majority of blindness was avoidable in Gao'an (84.3%), Xin'gan (71.0%) and Wan'zai (71.7%).

Conclusions The prevalence of blindness in the three counties in Jiangxi, China was lower than expected, yet most of the blindness and visual impairment was avoidable, indicating that the prevalence could be reduced further through adequate programme planning and implementation.

  • RAAB
  • blindness
  • survey
  • China
  • vision
  • public health
  • epidemiology

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.


It was estimated that in 2002, there were 37 million blind people and 124 million people with severe visual impairment or visual impairment globally.1 With population growth and increasing life expectancy, the magnitude of blindness is expected to increase further. Consequently, the WHO and the International Agency for the Prevention of Blindness (IAPB) launched a global initiative in 1999 known as ‘VISION 2020—the Right to Sight’ with the goal of eliminating avoidable blindness by the year 2020. China, as the country with the world's largest population (1.3 billion, 2006), contributes substantially to global blindness.1 It was estimated by WHO that the prevalence of blindness in China was 2.3% in people over the age of 50 years.1

Jiangxi Province is located in southeast China and has a population of 43 million. The province is administratively divided into 12 prefectures/cities and into 87 counties. Over 95% of the hospitals are government-owned in Jiangxi, excluding the village clinics in rural areas. In 1998, the Fred Hollows Foundation (FHF) from Australia set up a cataract surgical training programme in the province, to support high-quality, low-cost cataract surgical services at county hospitals through cataract surgical training, equipment supply, community workers training and team building. By August 2007, when the current survey was conducted, 73 surgeons from 57 county/prefecture hospitals were trained and equipped under the programme. The total number of cataract operations conducted in the province each year is about 15 000 to give a cataract surgical rate (CSR) of 350 operations per million people per year. Forty per cent of these surgeries are performed by surgeons trained under FHF programme since 2002 (data obtained from county hospitals).

No blindness survey has been conducted in the province, although the National Disabled People Survey arranged by central government in 2006 indicated that the all-age prevalence of visual impairment (including blindness) was 0.9%.2 The 1987 National Epidemiological Survey of Blindness reported a prevalence of blindness in China of 0.4% across people of all ages, mainly caused by cataract (41.1%), corneal diseases (15.4%), trachoma (10.9%) and glaucoma (8.8%).3 Surveys have been conducted in specific areas of mainland China, and these report a prevalence of blindness ranging between 0.1% and 0.7% in all ages.4 Cataract was the leading cause of blindness in most of the surveys.

The Rapid Assessment of Avoidable Blindness (RAAB) has been used for assessment of prevalence and causes of avoidable blindness throughout the world5 6 and has been undertaken successfully in Kunming City, Yunnan Province in southwest China.7 The aim of this study was to assess the prevalence and causes of blindness in people aged ≥50 years in three different counties in Jiangxi (Gao'an, Xin'gan and Wan'zai) using an RAAB and to provide baseline data for future eye care planning in the province. These three counties had different levels of service availability and were not geographically contiguous. Gao'an was one of the most successful counties in the FHF programme with a higher cataract surgery output, an effective patient referral network and a community screening programme that had been in place for 6 years. In Xin'gan, the FHF programme started more recently, with one surgeon trained by FHF in 2006, and a community screening programme was not in place by the time of the RAAB. In the third county, Wan'zai, there were no NGO programmes, only limited ophthalmic services were available, and this county was selected to allow a comparison of the prevalence and causes of blindness with counties where FHF was active. These surveys therefore allow an assessment of the impact of NGO support.


Census and samples

The first stage of sampling, that of selection of clusters, was through probability proportionate to size sampling of enumeration areas from the sampling frame. Updated census data from 2000 formed the sampling frame. The sample size was calculated using an estimated prevalence of blindness in people aged ≥50 of 2.3%, the worst acceptable estimate of 1.7%, with a design effect of 1.5, 95% confidence, an expected response rate of 90% and a minimum population of people aged ≥50 years of 46 341 to give a required sample size of 3800 (76 clusters of 50 people). It was feasible to collect larger samples, and consequently 100 clusters of 50 people over 50 years old were selected through probability proportion to size in Gao'an and 80 clusters in Xin'gan. In Wan'zai, the prevalence was assumed to be higher (3.0% with a worst acceptable result of 2.2%), and so 2805 people were required and 60 clusters selected.

The second stage of sampling, that of selection of individuals within clusters, was through the random walk method. Each survey team selected one cluster of 50 adults aged ≥50 years per day. The examination team arrived in the cluster and started at a place selected by the guide. They randomly chose one of the houses as the starting-point by spinning a bottle. The neighbouring houses were then included sequentially, moving in the same direction as the bottle had indicated, until the end of the road. Then, the bottle was spun again to determine whether to make a left or right turn, and the next nearest house was selected and sampling continued. Sampling ceased when 50 subjects were selected. In smaller villages with fewer than 50 subjects aged ≥50, the nearest village would then be selected and sampling continued until 50 subjects were recruited. Eligible survey subjects were residents aged ≥50 years living in the selected places over 6 months. If an eligible household member was absent, at least two return visits were made during that day. Information about visual status was ascertained from relatives or neighbours for non-responders, that is people who were not available after repeated visits, refused to or were unable to cooperate. If there was over 10% non-response in a cluster, the team members returned to the village on another day to revisit the houses of missing subjects. The examination teams were accompanied in the cluster by the village doctor, women's representative or head of the village.

Ophthalmic examination

Presenting visual acuity (VA) was measured using a tumbling E-chart with size 6/18 on one side and 6/60 on the other side at 6 m or 3 m, so that attainment of VA of 6/18, 6/60 and 3/60 could be assessed. The chart was shown to the participant five times, varying the direction of the E, and the participant had to read the chart correctly at least four times to be categorised as attaining that level of VA. VA was tested outside the house in full daylight by a trained nurse with the help of the coordinator. If VA was below 6/18 in either eye, a pinhole was then used as a proxy for best-corrected VA (BCVA). All the subjects were examined by the ophthalmologist in the house using a torch, direct ophthalmoscope and/or slit lamp. The pupil was dilated, if needed, for diagnosis. People with a VA less than 6/18 in either eye had the principal cause of the visual impairment/blindness diagnosed using the WHO convention whereby the major cause is assigned to the disorder that is easiest to treat.9 Blindness was defined as VA<3/60 in the better eye with available spectacle correction, severe visual impairment (SVI) as VA<6/60 to 3/60 and visual impairment (VI) as VA<6/18 to 6/60. People with a visual impairment from cataract were asked why they had not gone for surgery, and people with aphakia or pseudophakia were questioned about the details of their cataract surgery.

The standardised protocol for RAAB was used.5 6 All the survey teams for the three counties were trained together in Gao'an for 1 week by an experienced trainer. Each county had four teams, each including one coordinator, one ophthalmologist and one nurse. The interobservation agreement between teams was tested separately in the three counties. The kappa coefficients were all over 0.6 for all records, indicating good to very good agreement.

RAAB software version 4.02 software was used for data entry and standardised data analysis. Data entry was undertaken on the same day after the data collection, and inconsistencies were checked and corrected where needed. All data were then cleaned by the survey coordinator before the reports were generated.

Ethical approval

The Jiangxi Provincial Health Bureau gave approval for these surveys, which were planned as part of FHF programme activities for 2007. All subjects in the selected clusters were informed before the arrival of the survey team. All the subjects gave verbal consent before examination. Cataract patients identified during the survey who needed operations were referred to the county hospital for cataract surgery at reduced cost. All other subjects with eye conditions needing treatment were referred to the county hospital.


The response rate for RAAB in the three counties was high: 94.0% in Gao'an, 95.9% in Xin'gan and 95.4% in Wan'zai. Among those not examined, most were not available (5.7% in Gao'an, 3.9% in Xin'gan and 4.2% in Wan'zai), and few refused (0.2%, 0.2%, 0.2%) or were incapable of being examined (0.1%, 0%, 0.3%). In terms of age and gender distribution, the examined population represented the population aged ≥50 years in the three counties (table 1).

Table 1

Age and gender composition of survey area and sample population

The prevalence of blindness with available correction was low in Gao'an (1.5%; 95% CI 1.1% to 1.8%), Xin'gan (1.8%, 95% CI 1.4% to 2.2%) and Wan'zai (1.6%, 95% CI 1.2% to 2.1%) (table 2). The prevalence of severe visual impairment (SVI) was similar to blindness in all three counties, while visual impairment (VI) was more common. In Gao'an and Xin'gan, the prevalence of blindness and SVI in women was approximately twice as high as that in men, but in Wan'zai there was no apparent difference. The prevalence of blindness, SVI and VI increased with age in all three counties (figure 1).

Table 2

Sample results for the avoidable blindness survey (with available correction)

Figure 1

Prevalence of blindness, severe visual impairment (SVI) and visual impairment (VI) by age group in Gao'an, Xin'gan and Wan'zai.

Cataract was the leading cause of blindness and SVI in all three counties, followed by posterior segment disease (table 3). Uncorrected refractive error was the main cause of visual impairment in each of the three counties, closely followed by cataract. Overall, 71–92% of all blindness, SVI and VI was curable or could have been prevented (‘avoidable’).

Table 3

Causes of bilateral blindness and bilateral visual impairment (VI) with available correction

Cataract surgical coverage (CSC) is an indicator of the proportion of people or eyes with operable cataract that have been operated at the time of the survey, whereby the VA indicates the threshold for ‘operable’ cataract. As preoperative VA was not known, the assumption is made (in turn) that the people/eyes were operated on at VA<3/60, <6/60 or <6/18, and this is compared with the total need at that level of VA (ie, existing plus operated cases). The CSC was higher in Gao'an and Xin'gan than Wan'zai for blindness and SVI, both for persons and for eyes (table 4). There was less difference between the three counties using VI as the cut-off level for surgery. The CSC was consistently higher in men in Gao'an and Xin'gan, while in Wan'zai the CSC was higher in women.

Table 4

Cataract surgical coverage (adjusted for age and gender) by person and by eye in people aged 50+

Individuals who were blind due to cataract were asked why they had not attended for cataract surgery. Lack of awareness that treatment was possible and inability to afford surgery were leading barriers in Gao'an (25% and 22% respectively), Xin'gan (17% and 29%) and Wan'zai (23% and 42%). A quarter (25%) of people who were blind from cataract in Gao'an reported that they had not attended for surgery because they were too old or felt no need. Other key barriers in Xin'gan were reported as contraindication to surgery (21%) or that the participant was waiting for the cataract to mature (13%). In Wan'zai, the remaining key barriers were not knowing how to obtain the surgery or that the participant felt that they were too old for surgery.

The proportion of eyes operated on with poor outcome (VA<6/60) was much higher than the 5% recommended by the WHO. One in four eyes in Gao'an and Wan'zai, and two out of five operated eyes in Xin'gan had a poor outcome with presenting vision (table 5). In all three counties, the poor outcomes were much more likely in eyes without an IOL compared with eyes with an IOL. Only five or six eyes out of ten achieved a good outcome in each of the three counties with presenting vision. With best correction, the visual outcome could improve considerably, mainly in eyes with an IOL implanted, indicating that a lack of spectacle correction was an important cause of poor outcome.

Table 5

Cataract surgical outcome in sample with available correction

Using the population of the three counties to extrapolate the age- and gender-adjusted prevalence of blindness and visual impairment (table 6) to estimate the magnitude of blindness and visual impairment indicated that in Gao'an, there were approximately 2950 people aged ≥50 years who were bilaterally blind, 2600 with severe visual impairment and 12 230 people with visual impairment. In Xin'gan, there were an estimated 660 blind people aged ≥50 years, 490 with severe visual impairment and 2220 with visual impairment. In Wan'zai, the number of blind people aged ≥50 years was estimated to be 1610, and there were a further 1290 with severe visual impairment and 5310 with visual impairment. Assuming that people will be eligible for cataract surgery if they have VA<6/60 in the better eye (ie, are blind or SVI) and using the proportion of cases caused by cataract at different levels of VA from table 3, then there are 3050 people needing surgery in Gao'an, 500 in Xin'gan and 1550 in Wan'zai. With a target of operating 20% of the backlog per year, and a population of 0.79 million in Gao'an, 0.29 million in Xin'gan and 0.49 million in Wan'zai, the target CSRs for the counties are 771 in Gao'an, 350 in Xin'gan and 640 in Wan'zai.

Table 6

Age- and gender-adjusted prevalence and estimated magnitude of blindness and visual impairment (VI) (available correction)


The prevalence of blindness, severe visual impairment and visual impairment in three counties in Jiangxi Province were all much lower than WHO's estimation in 2002.1 Most of the cases of blindness, SVI and VI remained avoidable, thus showing that the magnitude of disease could be reduced further still. Gao'an and Xin'gan had relatively well-established programmes compared with Wan'zai; the current CSRs in Gao'an and Xin'gan (800 and 400 per million people per year) were higher than in Wan'zai (200) where the CSR was low (data obtained from county hospitals). This could be mainly attributable to the success of the ‘China–Australia Cataract Surgical Training Cooperation Program’ between the Provincial Health Bureau and the FHF as well as the service improvement in the whole province. Surprisingly, this did not translate into a lower prevalence of blindness in Wan'zai. Our estimates show that the CSR was adequate in Gao'an and Xin'gan but needs to be raised substantially in Wan'zai, supporting the impact of the NGO project. The quality of cataract surgery is of some concern, as long-term outcomes were not satisfactory in any of the counties, and this deserves attention, and one strategy to improve outcome is through the provision of spectacles after surgery or biometry (not routinely available). Equality of access of treatment needs to be considered in Gao'an and Xin'gan, where the prevalence of blindness and SVI in women was approximately twice as high as that in men, and the CSC was lower in women. The information on barriers showed that increasing awareness and considering reducing the cost of the surgery may increase uptake, and this could be considered during the planning process.

The results showed that the main cause of blindness and severe visual impairment was cataract, which was consistent with other surveys in China.10–12 The high proportion of visual impairment due to uncorrected refractive error has also been documented in these surveys. The high frequency of myopic retinopathy contributing to blindness and visual impairment has not been previously documented and may point to the need for further research on this topic.

The prevalence estimated in these surveys was lower than the estimates from the RAAB in Kunming conducted in 2006 (bilateral blindness prevalence=3.7%, 2.8–4.6)7 or a survey of people ≥50 in Doumen, Guandong Province conducted in 1997 (prevalence of VA<6/60 with pinhole=4.4%, 3.7–5.1).11 However, our estimates were similar to those obtained from surveys in Shunyi, Beijing, China in 1996 (aged ≥50 years, blindness prevalence=2.8%),10 Nantong, Jiangsu Province (≥60 years, 1.4%),12 Harbin (≥50 years, 1.9%)13 and Chongqing (≥50 years, 1.8%).8 The estimates from the current RAABs are therefore consistent with previous studies and have been used in planning by the Jiangxi Provincial Health Bureau.

There were a number of limitations to the study, which are mainly inherent in the RAAB methodology. Ophthalmic examination was undertaken using a direct ophthalmoscope and/or slit lamp to reduce the time and cost of the survey. This method allows assessment of avoidable causes of blindness but is insufficiently detailed to identify most of the causes for posterior segment diseases, and therefore future surveys may benefit from using methodologies that focus on assessment of posterior segment disease, particularly as the prevalence of cataract falls and these conditions make up a larger proportion of total cases. During the study, many people expressed concern about presbyopia, yet this was not assessed in the RAAB. Even if presbyobia does not lead to the blindness, in reality it causes enormous concern and inconvenience for people in this age group, which should be noted by the blindness-prevention programme and normal eye-service delivery. The results showed a lower prevalence of blindness than expected in the three counties based on the estimates from the WHO.1 One implication of this is that the sample sizes obtained in the surveys were too small, since they were estimated using a higher prevalence than that obtained, and this would have resulted in wide CIs.

There were also several strengths to the study and the RAAB design. The response rate was high in each country, and this was achieved through collaborating closely with communities, revisiting clusters where the response rate was less than 90% and conducting the survey during a favourable season when people are close to their homes. The surveys were undertaken relatively quickly and at low cost. A standardised methodology was used in the three counties, which made the comparison between them (and with other RAABs) possible.

In summary, the studies in the three counties in Jiangxi Province, China at the age of over 50 years have documented a low prevalence of blindness and visual impairment. The causes remain largely avoidable, and so the prevalence of these conditions can be almost halved. The results provide useful baseline information for the future studies and planning for intervention programmes in these three counties.


The authors acknowledge XH Tu, F Shan, YG Zhou, H Zhou, XP Zhou, WQ Liao, PS Zhu, XF Luo, Y Zhang and B Zhang for data collection in the field, as well as the nurses, other paramedical staff and drivers from these three counties.



  • Funding The Fred Hollows Foundation funded the study.

  • Patient consent Obtained.

  • Ethics approval Ethics approval was provided by the Provincial Bureau of Health through an annual project agreement.

  • Provenance and peer review Not commissioned; externally peer reviewed.