Visual impairment and socioeconomic factors

P M Livingston, C A McCarty, H R Taylor

Abstract

Background—Information about socio-

economic factors associated with visual

impairment can assist in the design of

intervention programmes. Such infor-

mation was collected by the Melbourne

Visual Impairment Project (Melbourne

VIP).

Methods—The Melbourne VIP was a

population based study of non-

institutionalised permanent residents in

nine suburbs of the Melbourne metropoli-

tan area aged 40 years of age and older. A

standardised eye examination was pro-

vided to eligible residents which included

a structured interview. Variables of inter-

est for this analysis were age, sex, country

of birth, language spoken at home, educa-

tion level, use of private health insurance,

employment status, and living arrange-

ments. Visual impairment was deﬁned as

a best corrected visual acuity <6/18 and/or

visual ﬁeld constriction to within 20° of

ﬁxation

Results—A total of 3271 (83%) residents

participated. Participants ranged in age

from 40 to 98 years; 54% were female.

Forty four (1.34%) were classiﬁed as visu-

ally impaired due to visual acuity and/or

visual ﬁeld loss. To evaluate the independ-

ent association of the signiﬁcant socio-

demographic variables with visual

impairment, a regression model was con-

structed that included age, retirement

status, use of private health insurance,

and household arrangement. The results

showed that age was the signiﬁcant pre-

dictor of visual impairment (OR: 3.19; CI:

2.29–4.43), with the mean age of people

with visual impairment signiﬁcantly older

(75.0 years) compared with people without

visual impairment (58.2 years) (t

test=9.71; p=0.0001). Of the 44 visually

impaired people, 39 (87%) were aged 60

years of age and older.

Conclusion—The results indicate that age

is the most signiﬁcant factor associated

with visual impairment. Of some im-

portance was the ﬁnding that people with

visual impairment were less likely to have

private health insurance. With the aging

of the population, the number of people

aVected by visual impairment will in-

crease signiﬁcantly. Intervention pro-

grammes need to be established before the

onset of middle age to oVset the escalation

of visual impairment in the older popula-

tion.

(Br J Ophthalmol 1997;81:574–577)

Despite the fact that it imposes a signiﬁcant

socioeconomic burden on society, little infor-

mation is available on socioeconomic factors

that may inﬂuence visual impairment. Few

population based studies have investigated the

role socioeconomic factors play in the develop-

ment of visual impairment.

1–5

Socioeconomic

characteristics may highlight the

socioeconomic burden of eye disease in the

community and barriers to utilisation of eye

services in subpopulation groups. This infor-

mation is important to quantify the need for

treatment and rehabilitative services, design

innovative public health education and screen-

ing programmes that are targeted towards cer-

tain groups within the population, and identify

priority areas of research.

The aim of this study was to explore which

socioeconomic factors may be associated with

visual impairment using data from the

Melbourne Visual Impairment Project

(Melbourne VIP).

Method

The Melbourne VIP was a population based

survey of non-institutionalised permanent resi-

dents of the Melbourne metropolitan area aged

40 years or older. The detailed methodology

has been reported elsewhere.

In brief, nine

pairs of adjacent census collector districts were

randomly selected from the Melbourne Statis-

tical Division. A door to door, household cen-

sus was taken to identify all eligible people—

that is, those aged 40 years or older in the

calendar year of examination and who had

lived at that address for 6 months or more.

Eligible individuals were invited to attend a

local examination centre for a standardised

ophthalmic examination and interview.

EYE EXAMINATION

Presenting visual acuity was measured with the

participant’s current distance correction, or if

distant correction was not used, the unaided

distant visual acuity was measured. If the

presenting acuity was <6/6, refraction was

undertaken. Best corrected visual acuity was

determined after objective autorefraction and

subjective reﬁnement.

The visual ﬁeld examination was performed

with a Humphrey ﬁeld analyser (HFA) (Hum-

phrey Instruments Inc, San Leandro, CA,

USA) on all capable participants. Participants

not capable of performing the HFA test and

those examined at their homes were tested with

a Bjerrum screen. For those who were unable

to perform a Bjerrum screen, visual ﬁelds were

assessed by confrontation.

British Journal of Ophthalmology 1997;81:574–577574

University of

Melbourne,

Department of

Ophthalmology, Royal

Victorian Eye and Ear

Hospital, Melbourne,

Australia

P M Livingston

C A McCarty

H R Taylor

Correspondence to:

Dr P M Livingston,

University of Melbourne,

Department of

Ophthalmology, 1st Floor,

Royal Victorian Eye and Ear

Hospital, 32 Gisborne Street,

East Melbourne 3002,

Australia.

Accepted for publication

26 March 1997

DEFINITION OF VISUAL IMPAIRMENT

Visual impairment was deﬁned as a best

corrected visual acuity <6/18 and/or visual

ﬁeld constriction to within 20° of ﬁxation in

the better eye.

CASES AND CONTROLS

A screening process was conducted on all par-

ticipants during the eye examination. Partici-

pants identiﬁed during the visual acuity or

visual ﬁeld testing who had a best corrected

visual acuity <6/18 and/or visual ﬁeld constric-

tion to within 20° of ﬁxation in the better eye

were deﬁned as cases. A participant qualiﬁed as

a control if he/she was not visually impaired.

SOCIODEMOGRAPHIC STATUS

Sociodemographic status was classiﬁed from

reported characteristics of the individual.

Group characteristics refer to reported charac-

teristics on median household income from

information derived from the Australian Bu-

reau of Statistics.

Sample areas were also

compared.

For the purpose of this analysis, education

level was divided into four mutually exclusive

groups, 6 or less years (primary); greater than 6

but less than 12 years (secondary); trade

apprenticeship; and tertiary education. Em-

ployment status was classiﬁed as currently

employed (either full or part-time), retired or

performing home duties; private health insur-

ance was classiﬁed as either privately insured

or not privately insured; language spoken at

home was categorised as English speaking or

other and country of birth was collapsed into

Australian born or other. These socioeconomic

factors were initially analysed individually

before being condensed for the ﬁnal model.

Data on household income were not col-

lected because of the concerns of the potential

negative impact the questions might have on

participation. Age adjusted rates of visual

impairment for each of the nine sample areas

were calculated and compared with the median

household income as reported in the 1991

census.

Median household income for each

sample area was determined from the Austral-

ian Bureau of Statistics, Basic Community

Proﬁle.

STATISTICAL METHODS

Statistical methods for univariate analyses

included the ÷

test for comparing groups and t

test for comparing means. Multivariate analy-

ses were performed using logistic regression to

compute the relative odds of visual impairment

for the diVerent risk factors while controlling

for potential confounders; analysis of covari-

ance was conducted to compare means while

controlling for confounders. Data analyses

were performed using

SAS (SAS Institute, Cary,

NC, USA). Factors associated with visual

impairment were considered statistically sig-

niﬁcant when p<0.05.

Results

STUDY POPULATION AND PREVALENCE OF VISUAL

IMPAIRMENT

A total of 4273 houses were identiﬁed in the

nine study clusters. Of these 4033 (94%) pro-

vided information and 2391 (59%) had eligible

residents. In the eligible houses there were

3946 eligible people of whom 3271 (83%)

agreed to participate. Non-English speaking

people were signiﬁcantly less likely to partici-

pate (odds ratio (OR): 0.61; conﬁdence

interval (CI): 0.48–0.77), but there were no

other signiﬁcant diVerences between partici-

pants and non-participants and between par-

ticipants and the Melbourne and Australian

populations.

The participants ranged in age

from 40 to 98 years and 54% were female.

The prevalence of visual impairment is

reported elsewhere.

A total 44 (1.34%) people

were classiﬁed as visually impaired with best

corrected visual acuity <6/18 and/or constric-

tion of the visual ﬁeld to within 20° of ﬁxation

in the better eye.

FACTORS ASSOCIATED WITH VISUAL IMPAIRMENT

Individual characteristics

To evaluate the univariate association of

speciﬁc sociodemographic characteristics with

visual impairment, several regression models

were constructed that included age, sex,

education level, retirement status, country of

birth, main language spoken at home, use of

private health insurance, and living arrange-

ments (Table 1).

Another multivariate regression model was

constructed which included the factors signiﬁ-

cant in the univariate analysis. The results

showed that age was the only signiﬁcant

predictor of visual impairment (OR: 3.19; CI:

2.29–4.43). However, not having private health

insurance approached statistical signiﬁcance

(OR: 1.65; CI: 0.86–3.20) (Table 2).

People with visual impairment were signiﬁ-

cantly older (75.0 years) compared with people

without visual impairment (58.2 years)

(t test = 9.71; p = 0.0001). Further analyses

Table 1 Univariate logistic regression assessing the odds of

visual impairment associated with socioeconomic factors

Socioeconomic factors:

individual characteristics Odds ratio Conﬁdence intervals

Born elsewhere 1.23 0.70–2.40

Other language spoken at

home 1.28 0.61–2.69

Sex (females) 1.67 0.89–3.12

Education (secondary) 1.84 0.95–3.54

No private health insurance 2.43 1.29–4.56

Age (10 years) 3.27 2.48–4.32

Living arrangement (alone) 3.28 1.75–6.25

Education (trade) 5.26 0.69–40.00

Education (tertiary) 5.26 0.70–11.40

Retired 10.31 3.17–33.33

Table 2 Multivariate logistic regression assessing the odds

of visual impairment associated with socioeconomic factors

(individual characteristics)

Socioeconomic factors:

individual characteristics Odds ratio Conﬁdence intervals

Living arrangement (alone) 1.04 0.51–2.12

Retired 1.51 0.41–5.58

No private health insurance 1.65 0.86–3.20

Age (10 years) 3.19 2.29–4.43

Visual impairment and socioeconomic factors 575

indicated that of the 44 people visually

impaired, 39 (87%) were aged 60 years of age

and older. People aged 60 years of age and

older were 10 times more likely to have visual

impairment compared with those aged under

60 years of age.

Thirty ﬁve per cent of people with visual

impairment had private health insurance com-

pared with 57% of people without private

health insurance.

Geographic location

Visual impairment varied by geographic loca-

tion (Table 3). However, this diVerence was

not signiﬁcant (÷

=10.04; p=0.26). Owing to

the relatively small sample size in each test site,

there was insuYcient power to determine the

local predictors of visual impairment by

sample area. It was impossible to assess any

clustering eVects because too few cases resided

in any sample area; however, all clusters

showed an increase in visual impairment with

increasing age.

Median household income

The distribution of the median household

income is shown in Figure 1. Overall, people

with visual impairment were more likely to live

in an area where the median household income

was less than A$35 000 per year (Windsor,

Albert Park, Werribee, and East Bentleigh)

compared with people with median household

incomes of $35 000 or more (Frankston,

Edithvale, Glenroy, Wantirna, and Doncaster)

(÷

=30.90; p=0.001). People who lived in the

sample area Windsor which had the lowest

median household income ($22 000) had

twice the rate of visual impairment compared

with overall sample rate of 1.34%.

In the univariate analysis looking at visual

impairment by median household income, the

analysis showed that people were 68% less

likely to have visual impairment if they lived in

an area with a median household income of

$35 000 or more. However, this diVerence was

not signiﬁcant (Table 4).

Discussion

Although the prevalence of visual impairment

was low, the similarities between the Mel-

bourne VIP with the Melbourne and Austral-

ian communities indicate that this level of

impairment is representative of the Australian

community. The Melbourne VIP age standard-

ised rate of blindness (US deﬁnition <6/60)

was 0.34%, which was not signiﬁcantly diVer-

ent from other (age standardised to the

Melbourne VIP) rates reported in white

cohorts from three studies conducted in urban

and rural areas in the United States—the Bea-

ver Dam Eye Study (0.42%),

the Baltimore

Eye Survey (0.79% in whites),

and the Appa-

lachian community (0.61%).

This rate was

also similar to the rate from the Blue

Mountains Eye Study in New South Wales

(0.39%).

Previous studies have analysed the role of

individual characteristics to oVer insight into

the association between socioeconomic status

and visual impairment.

1–5

After adjustment for

potential confounders, age was the most

signiﬁcant factor associated with visual impair-

ment.

Deterioration of vision threatens one’s inde-

pendence and lifestyle. Good vision is a key

element in one’s ability to live an active and

independent life. Informing the community of

age related eye disease, regular eye examina-

tions (that is, every 2 years after age 40), and

appropriate treatment in the prevention of

visual impairment needs to be tailored to an

aging population.

Of importance was the ﬁnding that those

who did not have private health insurance were

at increased risk of being visually impaired, but

this ﬁnding was not statistically signiﬁcant. An

additional 75 cases of visual impairment would

have been necessary to demonstrate, with 80%

power, the association between private health

insurance and visual impairment. It is interest-

ing to note that 45% of people in Australia over

55 years of age have private health insurance.

The rate in this study was higher for those

who did not have visual impairment (57%) but

was also lower among those who did have

visual impairment (35%). These percentages

Table 3 Prevalence of visual impairment by sample area

Sample area Prevalence (%)

Windsor 2.7

Albert Park 2.2

Edithvale 1.8

East Bentleigh 1.4

Werribee 1.6

Frankston 1.2

Doncaster 0.9

Wantirna 0.6

Glenroy 0.6

Figure 1 Visual impairment by median household

income.

22 000

Median household income (S)

27 500

32 500

35 000

2.5

1.5

41 250

46 250

1.34%

0.5

Table 4 Univariate logistic regression assessing the odds of

visual impairment associated with median household

income

Median household income

(A$) Odds ratio Conﬁdence intervals

<35 000 1.00 —

>35 000 0.68 0.37–1.25

576 Livingston, McCarty, Taylor

are much higher in the urban communities

compared with 42% of people in rural

communities who have private health insur-

ance.

Use of private health insurance is also

reﬂected by the economic circumstances of the

individual. Only 32% of people in households

with median household incomes less than

$32 500 in Australia have private health insur-

ance compared with 67% of those with median

household incomes over $32 500.

Little infor-

mation is available on these economic consid-

erations in Australia. These costs may reﬂect

the issues of access and lack of economic

resources to participate in preventive eye

health care. Further research is required to

ascertain the possible barriers to access of eye

services among people who have the potential

to develop visual impairment.

In previous studies education level, employ-

ment status, and living arrangement appeared

to inﬂuence visual impairment.

135

These

socioeconomic factors were not demonstrated

to be signiﬁcantly associated with visual

impairment in this analysis; however, this may

be attributable to the small number of cases.

Although an additional 115 cases would be

necessary to demonstrate the signiﬁcant im-

portance of retirement status, the point esti-

mates of the odds ratios indicate that being

retired may have been associated with visual

impairment.

It should be noted that although living alone

and not having private health insurance were

not signiﬁcant in the multivariate analysis, this

may be attributable to the variables being

covariate—that is, people who are uninsured

are alone and elderly. The Melbourne VIP

found that 65% of people who lived alone and

aged 60+ did not have private health insurance

compared with 45% of people, aged 60 years

and older, who lived with others.

While it is recognised that there are some

limitations in comparing group or sample area

characteristics, such as the ecological bias, the

information can be used to conﬁrm trends

from the individual characteristics. In the

present study there was no signiﬁcant diVer-

ence between people with a median household

income of less than $35 000 compared with

households with a median household income

over $35 000. However, this may be attribut-

able to the small sample size. An additional

140 cases would have been necessary to

demonstrate an eVect of median household

income with visual impairment. Australian

evidence

13 14

suggests that the cost of services

can be a great deterrent in the use of services

by lower socioeconomic groups, although

American researchers suggested that this situa-

tion may be attributable to the reduced earning

potential of people with visual impairment.

Further research is necessary to address this

situation.

The implications of this research for health

services planning and delivery are noteworthy.

The Australian population is aging

and as

individuals grow older they are more likely to

have impaired vision. Assuming age speciﬁc

rates of visual impairment continue, popula-

tion growth projections

indicate, without

intervention, the number of people with visual

impairment will double over the next 25 years.

A planned, systematic, educational interven-

tion programme needs to be designed and

implemented to reduce the prevalence and

incidence of visual impairment in the commu-

nity. The aim should be to shift the emphasis

away from the end stage of disease towards

increased knowledge of the dangers associated

with age related eye conditions. The aim must

be to improve community awareness on the

consequences of undiagnosed and untreated

age related eye disease and how to access

primary and secondary eye health care. This

would also provide a more eVective use of the

eye healthcare system.

Education programmes need to be estab-

lished before the onset of middle age to oVset

the escalation of visual impairment in the older

population. This is an important goal in the

promotion of preventative ophthalmic care in

an aging population.

The authors wish to acknowledge the contributions of the

following people: Ms Sharon Bayley, Ms Marie Bissinella, Dr

Charles Guest, Ms Cara Jin, Ms Sharon Lee, Ms Claire Mc-

Kean, Dr Yury Stanislavsky, Mrs Catherine Walker, and Mr

Matthew Wensor. The Melbourne VIP is supported in part by

the Victorian Health Promotion Foundation, the Ansell

Ophthalmology Foundation, and the National Health and

Medical Research Council, including the Sir John Eccles Award

to Professor Hugh Taylor. We also acknowledge the support of

Carl Zeiss in their donation of Humphrey equipment for use by

the project.

1 Klein R, Klein BEK, Jensen SC, Moss SE, Cruikshanks KJ.

The relation of socioeconomic factors to age-related

cataract, maculopathy and impaired vision. Ophthalmology

1994;101:1969–79.

2 Salive ME, Guralnik J, Christen E W, Glynn RJ, Colsher P,

Ostfeld AM. Functional blindness and visual impairment

in older adults from three communities. Ophthalmology

1992;99:1840–7.

3 Klein R, Klein BEK, Linton KLP, DeMets DL. The Beaver

Dam Eye study: visual acuity. Ophthalmology 1991;98:

1310–5.

4 Tielsch JM, Sommer A, Katz J, Quigley H, Ezrine S and the

Baltimore Eye Survey Research Group. Socioeconomic

status and visual impairment among urban Americans.

Arch Ophthalmol 1991;109:637–41.

5 Dana MR, Tielsch JM, Enger C, Joyce E, Santoli JM, Taylor

HR. Visual impairment in a rural Appalachian community.

Prevalence and causes. JAMA 1990;264:2400–5.

6 Livingston PM, Carson CA, Stanislavsky YL, Lee SE, Guest

CS, Taylor HR. Methods for a population-based study: the

Melbourne Visual Impairment Project. Ophthalmic Epide-

miol 1994;1:139–48.

7 Livingston PM, Guest CS, Bateman A, Woodcock N, Taylor

HR. Cost eVectiveness of recruitment methods in a

population-based epidemiological study. Aust J Pub Health

1994;18:314–8.

8 Taylor HR, Livingston PM, Stanislavsky YL, McCarty CA.

Visual impairment in Australia: distance and near visual

acuity and visual ﬁeld ﬁndings of the Melbourne Visual

Impairment Project. Am J Ophthalmol 1997;123:328-37.

9 Australian Bureau of Statistics. Basic Community Proﬁle.

Unpublished data; 1995.

10 Livingston PM, Lee SE, McCarty CA, Taylor HR. A com-

parison of participants with non-participants in a

population-based epidemiologic study: the Melbourne

Visual Impairment Project. Ophthalmic Epidemiol 1997 (in

press).

11 Tielsch JM, Sommer A, Witt K, Katz J, Royall RM,

Baltimore Eye Survey Research Group. Blindness and

visual impairment in an American urban population. The

Baltimore Eye Survey. Arch Ophthalmol 1990;108:286-90.

12 Attebo K, Mitchell P, Smith W. Visual acuity and the causes

of visual loss in Australia. The Blue Mountains Eye Study.

Ophthalmology 1996;103:357-64.

13 Donnovan J, d’Espaignet E, Merton C, van Ommeren M,

eds. Immigrants in Australia: a health proﬁle. Australian

Institute of Health and Welfare. Ethnic Health Series, No

1. Canberra: AGPS, 1992.

14 National Health Strategy. Inequalities in health. Issues Paper

No 4. Melbourne: National Health Strategy; 1992.

15 Australian Bureau of Statistics. Projections of the populations

of Australia, states and territories 1993 to 2041. Canberra:

AGPS, 1994.

Visual impairment and socioeconomic factors 577