Prevalence of visual impairment among older Chinese population: a systematic review and meta-analysis

Authors: Zou M, Guo D, Chen A, Young CA, Li Y, Zheng D, Jin G.

Geographical coverage: China

Sector: Burden of disease

Sub-sector: Prevalence

Equity focus: Individuals aged 50+

Study population: General population aged 50+ in China

Review type: Other review

Quantitative synthesis method: Meta-analysis

Qualitative synthesis method: Not applicable

Background: China faces a significant challenge in managing age-related eye diseases which can ultimately cause visual impairment, due to its population increasing in age and size. While previous meta-analyses have examined the prevalence of visual impairment in China, there are many recent studies in this area, often published in Chinese (and so not accessible to the wider research community).

Objectives: To evaluate the prevalence of visual impairment among the elderly Chinese population.

Main findings

Overall, findings of this review show that visual impairment causes a great health burden among Chinese populations, particularly affecting females, those dwelling in rural areas, older people and people with a lower educational level.

A total of 72 studies (involving 90 datasets) with 465,039 individuals out of 14,676 initial records were identified according to the inclusion criteria. Of these 72 studies, 20 were written in English and the rest in Chinese. The target population of all selected studies was clearly defined, and all the samples were representative of the general population. 38 studies used BCVA, 9 used PVA and the rest used both in the diagnoses of blindness and MSVI. The data of several subgroup analyses was not available in some included studies.

The average study quality score was 6.2. Seven studies were given the highest score, 23 were classified as high quality, and the rest as moderate quality. The most common problem of the included studies in the meta-analysis was unclear description of non-responders. After removing each study sequentially for sensitivity analysis, the pooled prevalence of remaining studies did not change significantly compared to the initial results. There was no evidence of publication bias.

The pooled prevalence of moderate and severe visual impairment (MSVI) is 10.9% and blindness is 2.2%, while prevalence of MSVI and blindness by best corrected visual acuity (BCVA) was 5.4% and 2.2%, respectively. Females, rural residents, older age and lower educational level are risk factors for MSVI and blindness.

Authors note the following limitations: missing information in subgroup analysis; potential bias may exist in pooled prevalence of blindness and MSVI by BCVA, which may have some impact on the results; using prevalence of visual impairment (PVA) in China is not common. The authors did not compare the accuracy and practicability of PVA and BCVA but will consider this for the future.

Policy relevant findings were not directly reported by the authors, however, they conclude that “the ageing population is linked with more individuals at risk of blindness causing ocular diseases, such as age-related macular degeneration and diabetic retinopathy, where treatment in late-stage progression is often of limited efficacy. Multiple strategies should be utilised to mitigate these problems, such as better screening strategies and improved education around preventing MSVI and blindness to the general public”.

Methodology:

Studies were included if they met the following criteria: 1) population-based study; 2) utilised recognised definitions and standardised grading method to diagnose and classify VI; 3) accessible full-text in Chinese or English; 4) explicit survey year; and 5) age-specific prevalence data.

Two investigators searched for literature independently in both English (EMBASE, PubMed and Web of Science) and Chinese (SinoMed, VIP and Chinese National Knowledge Infrastructure) databases from 1 January 1999 until 16 June 2020.

The prevalence by presenting visual acuity (PVA) and best corrected visual acuity (BCVA) were pooled separately. Titles and abstracts of all initial searched results were screened independently by two investigators.

Two investigators conducted the data extraction independently and any disagreements were resolved by a discussion with a third investigator. The quality of all selected articles was evaluated by two investigators with a commonly used 8-item assessment tool.

A meta-analysis was conducted to estimate the pooled prevalence of Moderate and Severe Visual Impairment (MSVI) and blindness, using both Presenting Visual Acuity (PVA) and Best Corrected Visual Acuity (BCVA). The Comprehensive Meta-Analysis software was used for this purpose. Prevalence was calculated using random-effects models in case of high heterogeneity, otherwise fixed-effects models were used. Heterogeneity was assessed using the I2 statistic, with I2 >50% indicating high heterogeneity. Age-specific prevalence was also calculated for various age groups. Subgroup analyses were performed to explore potential sources of heterogeneity, including gender, district, geographical location, education level, and survey year. Continuous variables were dichotomised using the median splitting method. Publication bias was assessed using funnel plots, Begg’s tests, and Egger’s tests.

Applicability/external validity: The authors did not evaluate the applicability of their findings to nations outside of China.

Geographic focus: China.

Summary of quality assessment:

The methods used to identify, include, and critically evaluate studies had several shortcomings. Only English and Chinese studies were considered, with no indication of expert consultation, review of relevant article references, or inclusion of unpublished literature. The literature search was limited to works published since 1999, without any provided rationale. Despite the generally robust data analysis methods, no separate analyses were conducted for studies with varying bias risks, although the authors did assess the quality of the included studies. Due to these factors, our confidence in the review’s findings is low.