Introduction

Cataract is a leading cause of treatable vision loss globally [1] and in Australia [2]. Not only do individuals with visually significant cataract experience a deterioration of visual acuity, they often experience a loss of contrast sensitivity, difficulties with glare, and altered colour recognition [3, 4]. With an ageing Australian population, a significant increase in the burden of cataract is projected [5]. Thus, to assist in informing the appropriate planning for eye health care delivery, it is important to quantify Australia’s burden of visually significant cataract. Furthermore, an improved understanding of the prevalence of undiagnosed visually significant cataract may help inform more targeted public health strategies to increase eye health awareness.

To date, most previous population-based studies report the prevalence and risk factors of any cataract [6,7,8,9], with only few documenting the burden of visually significant cataract [10, 11]. Internationally, the reported prevalence of visually significant cataract ranges from 1.9% in the Los Angeles Latino Eye Study (LALES) [11] to 10.6% in the Singapore Malay Eye Study [10]. In Australia, two population-based cohort studies conducted in the early 1990s comprehensively reported the prevalence, incidence, and risk factors of any cataract amongst the non-Indigenous population [12, 13]. Since their completion, there have been substantial demographic changes and neither study reported epidemiological data on the prevalence of visually significant cataract. Amongst the Indigenous population, the National Indigenous Eye Health Survey (NIEHS, 2008) reported the national prevalence of bilateral visually significant cataract to be 2.5% [14], while the Central Australian Ocular Health Study (CAOHS, 2010) conducted within remote central Australia reported the prevalence of any (i.e., bilateral and unilateral) visually significant cataract to be in excess of 17% [15]. Given that prevalence of cataract have been reported to be higher in Indigenous communities and the accessibility to cataract surgery services differs considerably across remoteness areas of Australia, careful monitoring of the prevalence of visually significant cataract is required.

The purpose of this study was to estimate the prevalence of visually significant cataract in non-Indigenous and Indigenous Australian adults, stratified by geographic remoteness, as part of the National Eye Health Survey (NEHS).

Methods

Study population

The NEHS (March 2015 to April 2016) is a nationwide population-based survey that investigated the prevalence and causes of vision loss (<6/12) among non-Indigenous Australians aged 50 years and older and Indigenous Australians aged 40 years and older. The younger age criterion was chosen for Indigenous participants due to the earlier onset and more rapid progression of common eye diseases in that population [16]. The sampling and testing methodology has been described in detail elsewhere [17, 18]. In brief, the sample size calculation assumed a similar prevalence of vision loss to historical population-based data for non-Indigenous and Indigenous Australians [19, 20]. Assuming a margin of error of 1.1% for non-Indigenous Australians and 2% for Indigenous Australians, a design effect of 1.5 that adjusted for interclass correlations and a 20% nonresponse rate, the required sample size for the non-Indigenous and Indigenous populations was approximately 3000 and 1500, respectively. Multistage random-cluster sampling was used to select 30 population clusters of 150 participants (100 non-Indigenous and 50 Indigenous per sampling site). Selection of sites utilised Census 2011 data collected by the Australian Bureau of Statistics (ABS) that categorised Statistical Areas into five Remoteness Areas (RAs); Major City, Inner Regional, Outer Regional, Remote, and Very Remote. Participants were recruited door to door and the overall population positive response rate was 71.5% (4836/6760). Ethics approval was obtained from the Royal Victorian Eye and Ear Hospital (RVEEH) Human Research Ethics Committee (HREC-14/1199H) and additional ethical approvals were obtained at the State level to conduct research within Indigenous communities. Study procedures adhered to the tenets of the Declaration of Helsinki as revised in 2013 and participants provided written informed consent to participate.

Procedures

Sociodemographic data, utilisation of eye health services, as well as stroke, diabetes, and ocular history data were collected via an interviewer-administered questionnaire. Participants self-reported whether they had previously undergone cataract surgery in the right and/or left eye. Presenting distance visual acuity (PVA) was measured in each eye using a logMAR chart (Brien Holden Vision Institute, Sydney, Australia) in well-lit room conditions. Pinhole testing was performed on participants with vision loss (PVA < 6/12) in one or both eyes, followed by automated refraction (Nidek ARK-30, Nidek Co., LTD, Japan) if pinhole visual acuity improved to ≥6/12 in either eye. Best-corrected visual acuity (BCVA) was defined as the best of the pinhole or automated refraction measurements for each eye. Examination of the anterior segment was performed using a hand-held slit-lamp (Keeler Ophthalmic Instruments, UK). Participants with PVA < 6/12 in one or both eyes had anterior segment photographs taken using a Digital Retinography System (DRS) camera (CenterVue SpA, Italy). Two-field, 45° colour fundus photographs were taken of each retina, centred on the macula and optic disc, respectively, with the DRS camera in a darkened room to allow for physiological mydriasis. Trained retinal graders from the Centre for Eye Research Australia masked to the identity and clinical characteristics of study participants graded fundus photographs for retinal disease.

Diagnosis of visually significant cataract

Visually significant cataract was defined as any eye with BCVA of <6/12 and cataract as the primary cause of vision loss in that eye. This definition is in line with previous Australian [14, 15] and international reports [10, 11] and was adopted as 6/12 is a legal requirement for driving in Australia and therefore these cases should be prioritised for targeted resource allocation. To ensure an accurate diagnosis of visually significant cataract, a two-step protocol was utilised. First, two optometrists independently assessed anterior segment photographs and fundus photographs to categorise participants into one of three groups: (1) no cataract, (2) probable cataract, or (3) definite cataract. High inter-rater reliability (85%) and intra-rater reliability (94 and 96%) were achieved. Any discrepancies were adjudicated by an ophthalmologist. In cases where photographs were unavailable, a cataract grade was assigned on the basis of an anterior segment examination by a trained clinician using a hand-held slit-lamp (Keeler Ophthalmic Instruments, UK). Second, a single ophthalmologist reviewed relevant case notes (retinal and anterior photographs, grading data and clinical examination results) of all participants with best-corrected VA of <6/12 in one or both eyes, and assigned the main cause of unilateral or bilateral vision loss. Participants with visual impairment for whom the cause of vision loss was not determinable and were not pseudophakic were excluded from the analyses.

Statistical analysis

Differences between included and excluded participants were compared using a chi-squared test or Wilcoxon test for categorical and continuous variables, respectively. The 95% confidence interval (CI) for crude prevalence was estimated using robust standard errors to account for clustering within study site. Population proportions were estimated using post-stratification for age in addition to the survey weights which were derived separately for Indigenous and non-Indigenous participants according to study site. Population values were derived using data from the 2016 Australian Census (source: Australian Bureau of Statistics TableBuilder www.abs.gov.au, accessed 19 April 2018).

Age- and sex-adjusted logistic regression was performed separately for Indigenous and non-Indigenous participants to investigate the following risk factors for visually significant cataract: educational attainment, language, place of birth (for non-Indigenous participants only), remoteness area, self-reported diabetes, and time since last eye exam. Robust standard errors were used to account for clustering within study site. All analyses were performed using Stata/SE version 15.1 (StataCorp LP, College Station, TX, USA).

Results

To explore potential response bias, we compared sociodemographic characteristics (age, gender, education status, and country of birth) of NEHS participants to the target population using Australian Census 2016 data. The Indigenous sample in the NEHS had a lower proportion of males, people in the youngest age group (40–49 years), and an overall lower level of educational attainment when compared to those in the target Australian population (p < 0.001 for each). Non-Indigenous participants in the NEHS were less likely to be born outside of Oceania or Europe, had an overall lower educational attainment, and there was a lower proportion of people in the youngest age group (50–59 years) when compared to those in the target population (p < 0.001 for each).

A total of 4836 individuals were recruited and examined in the NEHS, including 3098 non-Indigenous and 1738 Indigenous Australians. Of these, 1211 (25%) participants had vision loss in at least one eye. The cause of vision loss could not be ascertained in 39 phakic participants (3.2%), leaving a total of 4797 (99.2%) participants eligible for inclusion. This included 3079 non-Indigenous (53.7% female, median age (interquartile range (IQR)) 66.2 (59.6–73.5) years) and 1718 Indigenous (59.0% female, median age (IQR) 54.2 (47.5–61.8) years) Australians. There were no differences in the distribution of gender (p = 0.39) and remoteness (p = 0.41) amongst participants who were and were not excluded. However, on average, excluded participants were older (p = 0.04) and more likely to be Indigenous (p = 0.045).

Weighted prevalence of visually significant cataract by Indigenous status

In the non-Indigenous population aged 50 years and over, the weighted prevalence of visually significant cataract was 2.7% (95% CI: 2.0, 3.5) (Table 1). Among Indigenous Australians aged 40 years and over, the weighted prevalence of visually significant cataract was 4.3% (95% CI: 3.1, 5.9). Bilateral visually significant cataract was present in 29 non-Indigenous participants and 43 Indigenous participants. The weighted prevalence of visually significant cataract increased with age in non-Indigenous participants, with the following age-specific prevalence: 0.9% in those <60 years, 2.5% in those aged 60–69 years, 4.8% in those aged 70–79 years, and 5.0% in those aged 80–99 years. In the Indigenous population, the prevalence of visually significant cataract for the age groups of <60 years, 60–69 years, 70–79 years, and 80–99 years were 1.9%, 7.0%, 12.8%, and 34.1%, respectively. Visually significant cataract was found in 3.2% of males and 2.2% of females in the non-Indigenous population. Among Indigenous Australians aged 40 years and over, the prevalence of visually significant cataract was 3.7% for males and 4.6% for females.

Table 1 Crude and estimated population prevalence of visually significant cataract for Indigenous Australians 40 years of age and above and non-Indigenous Australians 50 years of age and above, National Eye Health Survey (n = 4797, 2015–2016)
Full size table

Associations of visually significant cataract

After adjusting for age and gender, the odds of visually significant cataract were almost three times higher among Indigenous participants compared to non-Indigenous participants (adjusted odds ratio (OR) 2.95, 95% CI: 2.03, 4.29). In the non-Indigenous population, adjusted logistic regression analysis revealed that increasing age, male gender, and fewer years of education were associated with visually significant cataract. Among Indigenous participants, increasing age, speaking a language other than English, and having never had an eye examination were associated with higher odds of visually significant cataracts (Table 2). All significant associations of visually significant cataract remained in the fully adjusted model (Supplementary Table 1).

Table 2 Estimated association between participant characteristics and visually significant cataract in the National Eye Health Survey, 2015–2016 (n = 4797)
Full size table

Undiagnosed visually significant cataract

In the non-Indigenous population, 54.8% (57/104) of all visually significant cataract cases self-reported a history of being previously diagnosed with cataract. Of all undiagnosed cases in the non-Indigenous population, only 28.3% (13/46) had accessed an optometry or ophthalmology service in the past 12 months (self-report). Among the Indigenous population, 38.9% (28/72) of cases with visually significant cataract self-reported a history or being previously diagnosed with cataract. Of all undiagnosed cases, 34.4% (11/32) had accessed an optometry or ophthalmology service in the past 12 months (self-report). The presence of undiagnosed visually significant cataract was associated with older age and longer time since last eye exam in the non-Indigenous population, and older age and speaking a language other than English in the Indigenous population (Table 3).

Table 3 Fully adjusted model examining associations between participant characteristics and undiagnosed visually significant cataract in the National Eye Health Survey, 2015–2016 (n = 4797)
Full size table

Discussion

This paper presents the prevalence of visually significant cataract in a population-based, national sample of non-Indigenous and Indigenous Australian adults. After adjusting for age, the odds of visually significant cataract were almost 3 times higher in Indigenous Australians compared to non-Indigenous Australians. The prevalence of visually significant cataract rose sharply with age, with Indigenous and non-Indigenous persons aged 70–79 years being approximately 4 times more likely to have visually significant cataract than persons aged under 50–59 years of age.

The prevalence of visually significant cataract in non-Indigenous Australians in the NEHS (2.7%) is similar to that reported in the LALES (1.9%) [11] and is low by comparison to other population-based reports from developed nations [1, 10]. The relatively low prevalence observed in non-Indigenous Australians is perhaps not surprising given cataract surgery is the most commonly performed elective surgical procedure in Australia and is subsidised by government funding [21]. Furthermore, this finding is in line with our previous work that reports accessibility and utilisation rates of cataract surgery services [22] and eye health care services in general [23] are high amongst the non-Indigenous Australian population. Nonetheless, our finding of a significantly higher prevalence of visually significant cataract amongst non-Indigenous males in the NEHS is noteworthy. This observed difference appears to be partly attributable to a less frequent utilisation of eye care services (previous examination > 2 years ago or never; males = 20.8% vs. females = 12.0%) and a higher prevalence of self-reported diabetes (males = 16.8% vs. 11.3%) amongst non-Indigenous men in the NEHS, placing them at a higher risk of vision loss from cataract. It may also be speculated that a known higher exposure to modifiable cataract-related risk factors, including ultraviolet-B exposure [24] and smoking [25], amongst Australian males may contribute to this gender difference.

In line with previous international reports [26,27,28], lower educational attainment was associated with greater risk of visually significant cataract amongst non-Indigenous Australians in the NEHS. Educational attainment is an important surrogate of socioeconomic status [29], and therefore this finding may reflect the role of poorer lifestyle choices (e.g., smoking), a lower health literacy, or real or perceived costs of cataract surgery amongst non-Indigenous Australians with lower levels of education. Given Australia has a universal health insurance scheme (i.e. Medicare) and the availability of ophthalmology clinics in public hospitals, our results provide evidence that targeted eye health promotion may be warranted to improve awareness and eye health literacy.

Previous research has inferred a disproportionately higher prevalence of visually significant cataract [14, 15, 30], and longer wait times for cataract surgery, amongst Indigenous Australians [31, 32]. As such, over the past decade, several targeted interventions have been implemented to improve the accessibility of cataract surgery services and reduce the burden of visually significant cataract in this population [33]. To allow for more robust comparisons with the CAOHS, the weighted prevalence of cataract in the Indigenous population was recalculated to include those residing in remote and very remote areas only. After adjustments, the prevalence was notably lower than the CAOHS (CAOHS = 17% vs. NEHS = 5.3%). We speculate that the observed differences are, at least in part, explained by contrasting regions of Australia that were sampled. That is, the CAOHS recruited participants from inland regions only, where greater barriers (e.g., increased travel distance) to accessing cataract surgery services exist [34]. We report a prevalence of bilateral visually significant cataract that is comparable with that of the NIEHS (2.5%) [14]. However, as cataract is strongly age-related and Indigenous participants in the NEHS were, on average, older (mean age; NEHS = 58 years vs. NIEHS = 50 years) than NIEHS participants, this finding may in fact reflect a modest improvement in the prevalence of visually significant cataract in the ageing Indigenous population [35]. Nonetheless, our finding that the age-adjusted weighted prevalence of visually significant cataract was almost 3 times higher (OR = 2.95) in Indigenous Australians than non-Indigenous Australians highlights that continued efforts are required to build sustainable cataract surgery services within Indigenous communities.

Very few population-based studies have investigated the prevalence of undiagnosed visually significant cataract [10, 11]. In the present study, approximately 60% of Indigenous Australians and 45% of non-Indigenous Australians with visually significant cataract did not have a known history of cataract. The frequency of undiagnosed visually significant cataract in Indigenous Australians is similar to that reported in the Singapore Malay Eye Study (70%) and the LALES (60%) [10, 11]. It is plausible that the high undiagnosed rates in the present study are attributable to the fact that most cases of age-related cataract are slow progressing, and therefore gradual changes in visual acuity go unnoticed. In addition, approximately 70% of non-indigenous and Indigenous Australians with undiagnosed visually significant cataract had not accessed an optometry or ophthalmology service in the past 12 months. These results emphasise that improved awareness of the importance of regular eye examinations is warranted, particularly given vision loss from cataract is readily treatable.

Unlike findings from the NIEHS (2008) [14], our results suggest that Indigenous Australians who do not speak English as their primary language at home are at a notably higher risk of visually significant cataract (including undiagnosed). Given an estimated 11% of Indigenous Australians speak an Indigenous language at home [36], this finding reinforces the importance of understanding linguistic and cultural backgrounds when planning eye health services, and providing access to Indigenous language interpreter services for clinical discussions.

The strengths of this study include its sampling design, stratification by Indigenous status, and the large sample size. Some limitations must also be considered. Firstly, we did not utilise a traditional slit-lamp grading method (e.g., Wisconsin cataract grading system, Lens Opacities Classification System, The Age-Related Eye Disease Study System) to diagnose cataract, rather we adopted a less common method using a nonmydriatic fundus camera [37]. Nonetheless, a high sensitivity and specificity for detecting visually significant cataract has been demonstrated using this method [37]. Second, prior diagnosis of cataract was obtained via self-report. Therefore, the influence of recall and/or misclassification bias on our analysis of ‘undiagnosed’ cataract cannot be discounted. Third, a failure to capture several known and potential cataract-related risk factors (ultraviolet-B exposure, smoking, cortico-steroid use, systemic diseases, diet) limited our ability to conduct a more meaningful risk factor analysis. Lastly, sufficient information from all the non-responders was not available to quantify response bias. However, a comparison of key sociodemographic characteristics to the target Australian population revealed that, on average, participants in the NEHS were older and had a lower educational attainment. Given both of these factors were associated with a greater risk of visually significant cataract in the present study, this may have resulted in an overestimation of the true prevalence of cataract. Strategies to address potential response bias included post-survey adjustments of prevalence estimates to the age structure of the nation for both non-Indigenous and Indigenous populations.

In conclusion, this report described the age-, gender-, and Indigenous-specific prevalence estimates of visually significant cataract in a national population-based sample of Australian adults. Our data suggest that several high-risk groups, including those aged over 70 years, Indigenous Australians who speak a language other than English at home and non-Indigenous males and those with fewer years of education may benefit from targeted eye health awareness campaigns and improvements in the provision of cataract surgery services. With a significant ageing of the Australian population and an increase in life expectancy, high cataract surgery rates are critical to reduce cataract-related vision loss in Australia.

Summary

What was known before

  • In Australia, knowledge of the epidemiology of visually significant cataract remains scarce due to a paucity of recent population-based data.

What this study adds

  • The National Eye Health Survey provides an up-to-date estimate of the prevalence of visually significant cataract in non-Indigenous and Indigenous Australian adults.