Methodological quality of the review: High confidence

Author: Ye Q, Chen Y, Yan W, Wang W, Zhong J, Tang C, Muller A, Qui B.

Region: Bangladesh, Bhutan, India, Nepal and Pakistan

Sector: Service delivery

Subsector: Access

Equity focus: Gender

Review type: Other review

Quantitative synthesis method: Meta-analysis

Qualitative synthesis method: Not applicable

Background:
In 2010, it was estimated that there were 39 million people blind globally, with 51% being attributed to cataract. In particular, the female population accounts for approximately 60% of blindness and 57% of moderate-to-severe visual impairment. South Asia, which accounts for 24.6% of the world’s population, has undergone fundamental transitions in society and economic development over the past two decades. Although this region has achieved significant progress in blindness prevention, it still accounted for 32.7% of the world’s blind population in 2010. In addition, the expansion rate of the blinded population was four times higher in females than in males (16.1% versus 3.7%) from 1990 to 2010, which poses significant public health concerns.

Objectives:
To determine whether the female gender has been a barrier to accessing cataract surgery services in South Asia in the last two decades.

Main findings:
Sixteen independent studies, with a total of 135,972 subjects, were included in the review, of which, 14 studies were published in peer-reviewed journals, while two were unpublished. One study was national, others were regional. The number of examined people ranged from 1,076 to 39,908. Two studies were conducted in Bangladesh, one in Bhutan, four in India, five in Nepal and four in Pakistan. All 16 studies were graded low risk of bias.

Meta-analysis was used for 15 studies. Based on the meta-analysis of CSC by sex, on a person basis, visual acuity (<3/60) showed a pooled OR of 1.46 (95% CI:1.23 – 1.75) for males’ use of cataract surgery compared with females. Authors reported significant heterogeneity (P<0.001, I²=67.6%). Authors reported similar results of CSC by sex, on a person basis, at visual acuity <6/18, with a pooled OR of 1.14 (95% CI: 1.05 – 1.24) with no significant heterogeneity among studies (P=0.349; I²=10.4%).

Regarding the CSC on an eye basis at visual acuity <3/60, authors reported statistically significant associations with a pooled OR of 1.40 (95% CI: 1.16–1.70).

Authors note that only five studies were available for meta-anlaysis on CSC on an eye basis at visual acuity <6/18 where the pooled OR was not significant due to substantial heterogeneity among studies.

Authors reported that subgroup analyses by design and location types produced similar results to the primary analyses. In additon, authors found no evidence of publication bias, as indicated by nonsignificant Begg’s tests (P=0.488) and Egger’s tests (P=0.173) in persons at VA <3/60. The values of ARP at a person and eye basis were 6.28% and 7.48%, respectively.

Methodology:
Authors conducted a systematic literature search in PubMed, Embase.com, and ISI Web of science. Google Scholar Engine, Rapid Assessment of Avoidable Blindness (RAAB) repository, and the WHO website were also used to identify eligible studies. The reference lists of the included articles and related reviews were screened to identify additional eligible studies.

Studies were included for analysis if they met the following criteria: (1) they were performed after 2000; (2) the study design was population-based or in the form of RAAB, national registry, or national/subnational survey of all populations; (3) the study was conducted in a general adult population; (4) the study reported the number of cataract surgery performed (person/eyes) and the number requiring cataract surgery (person/eyes) at visual acuities of <3/60 (20/400) or 6/18 (20/60), or provided the data needed to calculate them.

Screening of studies for inclusion, data extraction and quality assessment was conducted by two authors independently. Quality of studies was assessed using a risk and bias tool in prevalence studies. Each item was scored 1 or 0 to represent a low risk of bias or high risk of bias. The total score ranged from 0 to 10, with 0-3 for high risk, 4-7 for moderate risk, and 7 and greater for low risk.

For the meta-analysis, authors used the weighted average (mean) odds ratio. The meta-analysis was performed using raw data to calculated pooled odds ratio based on person (eyes) at a visual acuity of <3/60 and 6/18. A pooled relative risk was calculated and used to estimate the blindness (visual acuity <3/60) population attributable risk percentage (ARP). Statistic heterogeneity between studies was tested using Cochran’s Q and I² statistics. A random effect model was used. Subgroup analysis was performed by study design and location type.

Applicability/external validity:
Authors conducted subgroup analysis by study location and note that findings are applicable to South Asia.

Geographic focus:
Overall findings of this review should be applicable to South Asia, as subgroup analyses location produced similar results to the primary analyses.

Summary of quality assessment:
High confidence was attributed to this review, as authors used rigorous and reliable methods to conduct the review.