Risk of Admission to a Nursing Home Among Older People With Visual Impairment in Great Britain

Objective  To assess the risk of nursing home admission in a representative sample of older, visually impaired, community-dwelling people 75 years and older living in Great Britain.

Design  Prospective study.

Setting  General practices (n = 53).

Participants  Participants in the Medical Research Council trial of assessment and management of older people in the community (N = 14 037).

Main Outcome Measure  Nursing home admission (nursing or residential care).

Results  After a mean of 4 years' follow-up, 14.1% of visually impaired participants were living in nursing homes (95% confidence interval [CI], 11.9%-16.3%; age- and sex-adjusted risk ratio, 1.47 [95% CI, 1.25-1.73]). Adjusting for a wide range of confounding factors eliminated this significant association (risk ratio, 1.08 [95% CI, 0.91-1.28]).

Conclusions  The association between visual impairment and risk of nursing home admission was eliminated after controlling for a wide range of other confounding factors and comorbidities. This underlines the importance of taking a wider view of an older person's health rather than focusing on a single impairment or disability.

Older people who live in nursing homes (nursing or residential homes) have a higher prevalence of visual impairment compared with people of a similar age living in the community.^1-6 There are 3 possible reasons for this. First, visually impaired people living in the community may be more likely to be admitted to nursing homes than sighted people of the same age; second, access to eye health care services may be more problematic for people admitted to a nursing home; third, people in nursing homes have many other comorbidities, and eye health care may be overlooked or perceived as an unnecessary intervention.

There are limited prospective data on this topic. In the Blue Mountains Eye Study,⁷ visually impaired people had almost twice the risk of nursing home admission after 6 years' follow-up (adjusted for age, sex, home ownership, self-rated health status, walking disability, smoking status, and alcohol consumption). In the Beaver Dam Eye Study,⁸ a 3-fold increased odds of nursing home admission was seen over 5 years (adjusted for age, sex, self-rated health status, and arthritis history). There is limited information about which other factors might be involved. Likely candidates are variables that reflect the degree of social support and those that reflect the level of comorbidities or problems experienced by the visually impaired person.

The aim of this study was to assess the risk of nursing home admission in visually impaired people 75 years and older who were taking part in the Medical Research Council (MRC) trial of the assessment and management of older people in the community.

The MRC trial of the assessment and management of older people in the community was a large cluster randomized trial conducted in 106 general practices from the MRC General Practice Research Framework.⁹ The practices in the study were selected to be representative of the mortality (standardized mortality ratio) and Jarman scores¹⁰ of general practices in Great Britain (England, Wales, and Scotland). The aim of the trial was to evaluate the benefit of different methods of assessment and management of older people in the context of the 1990 contract of service, which required general practitioners in the United Kingdom to offer an annual health checkup to patients 75 years and older.¹¹ The study compared 2 different types of multidimensional assessments (targeted vs universal) and 2 different management models (primary care team vs multidisciplinary geriatric evaluation team). Randomization was at the practice level and stratified by standardized mortality ratio and Jarman score. All patients 75 years or older on the general practitioner list were invited to participate in the trial, unless they were in a long-stay hospital or nursing home or were terminally ill.

Patients of the 53 general practices allocated to the “universal” arm of the trial were given a visual acuity test as part of a detailed health assessment by the practice nurse and are included in this study. Visual acuity was measured at 3 m with a Glasgow Acuity Chart, which measures the minimal angle of resolution on a logarithmic scale.¹² Vision was measured both as presenting vision (using glasses or contact lenses) and in each eye. People with binocular baseline visual acuity better than 20/30 were defined as having “good vision,” and those who came to us with presenting vision of less than 20/30 to 20/60 as having “reduced vision.” Visual impairment was defined as presenting binocular visual acuity of less than 20/60. The prevalence of visual impairment was similar in practices randomized to the 2 different management models: 12.3% in the practices randomized to a primary care team and 12.5% in practices randomized to a multidisciplinary geriatric evaluation team.

In 49 practices, the cause of visual impairment was assessed by medical record review.¹³ Information on nursing home admission was collected by trial nurses from the general practice medical records and by linking information on residence at death (from death certificates) with a national database of nursing homes in the United Kingdom.¹⁴ The baseline assessments were conducted from May 15, 1995, to April 27, 1999; information on nursing home admission was obtained until September 30, 2000.

The trial and additional data collection on the cause of visual loss were approved by the relevant local research ethics committees.

All analyses were performed using Stata statistical software, version 9.0 (Stata Corp, College Station, Texas), and took into account the cluster design of the study using “svy” commands. People with a Mini-Mental State Examination (MMSE) score of less than 12 were excluded from these analyses because measurement of visual acuity in this group is likely to be inaccurate. The risk ratio of nursing home admission associated with visual impairment was estimated using Poisson regression. In all analyses, the reference group consisted of participants with good vision. The following potential confounding factors and effect modifiers were considered: age, sex, marital status (single, married, or widowed), living alone, housing tenure (home owner, not a home owner, or sheltered accommodation), financial difficulties (difficulties making ends meet and/or managing finances), looked after someone with a serious illness in the last year, death of a loved one in the last year, social support (no relative or friend to call on and/or no help at night), alcohol consumption (never, former drinker, currently below median, or currently above median), smoking status (never, former smoker, or current smoker), body mass index (quintiles; calculated as weight in kilograms divided by height in meters squared), depression (score of ≥6 on the Geriatric Depression Scale), having diabetes mellitus, hearing impairment (failed whispered voice test), reported major illness (heart attack, stroke, Parkinson disease, or cancer), self-reported health status (excellent or very good, good, or fair or poor), activities of daily living (ADL) score (unable to complete 0-1, 2-4, or 5-8 ADL), falls in the last 6 months (none, 1, or ≥2), self-reported activity level (very or fairly active vs not very or not at all active), and cognitive impairment (MMSE score of 12-17, 18-23, or 24-30).

In addition to age and sex, only those variables statistically associated with both nursing home admission and visual impairment in this data set (after controlling for age and sex) were considered as confounders. These were marital status, housing tenure, financial difficulties, alcohol consumption, body mass index, depression, diabetes mellitus, reported major illness, self-reported health status, number of falls, ADL score, self-reported activity level, and MMSE score. Because vision impairment may lead to problems with self-reported health status, activity level, ADL score, number of falls, depression, and body mass index, and because these could be considered to be on the “causal pathway,” analyses were carried out with and without these variables.

Interaction terms for all the variables considered as effect modifiers were entered into the Poisson regression model. The significance of these terms was assessed using the adjusted Wald test.

Response rates and characteristics of nonresponders have been published elsewhere.¹⁵ Of 21 762 eligible people, 15 336 (70.5%) in practices randomized to the universal arm of the MRC trial had a detailed assessment. People taking part were slightly younger and more likely to be men compared with those who did not take part. Of 15 336 people with a detailed assessment, 14 037 (91.5%) were included in the current analyses. Reasons for exclusion were already living in a nursing home at baseline (333 [25.6%]), MMSE score of less than 12 (379 [29.2%]), no visual acuity data (344 [26.5%]), and data on baseline/follow-up were not available (243 [18.7%]). Table 1 shows the characteristics of the study population.

Table 2 shows the risk of nursing home admission among visually impaired participants. After a mean of 4 years' follow-up, 14.1% of visually impaired participants were living in nursing homes (95% confidence interval [CI], 11.9%-16.3%) compared with 5.4% of participants with good vision (4.6%-6.3%).

Table 3 shows the effect of controlling for confounding factors. The risk ratio adjusted for age and sex was 1.47 (95% CI, 1.25-1.73). Adjusting for a wide range of social and comorbidity confounders eliminated this increased risk of nursing home admission among visually impaired participants. The risk ratio adjusting for all confounders was 0.92 (95% CI, 0.73-1.16). We repeated the analyses, excluding variables potentially on the causal pathway. The risk ratio for nursing home admission was 1.08 (95% CI, 0.91-1.28). Analyses were repeated for 11 551 participants who had complete data for all confounders. The risk ratio adjusting for age and sex was 1.37 (95% CI, 1.11-1.69), and the risk ratio adjusting for confounders excluding those on the causal pathway was 1.20 (0.97-1.49).

Three variables were found to interact with visual impairment in determining the risk of nursing home admission: sex, ADL score, and MMSE score. Table 4 shows the results of stratified analyses with these variables. The risk of nursing home admission with visual impairment was greater in men, people with the least difficulties with ADL, and people with no evidence of cognitive impairment.

Table 5 shows the risk of nursing home admission by cause of visual impairment. People with refractive error and visual impairment of unknown cause were at greater risk, but after adjustment for confounders there was no significant association by cause of vision impairment.

We found little evidence for an increased risk of nursing home admission among visually impaired older people in the MRC trial during 2 to 5 years' follow-up. The current study differs from previous studies in that we had access to information on a wide range of other factors associated with nursing home admission, such as sociodemographic factors, lifestyle factors, social support, and comorbidities. Although we found an age- and sex-adjusted risk of 50%, controlling for these factors effectively removed the apparent increased risk of nursing home admission among visually impaired participants. This suggests that as people get older and many difficulties are experienced, visual impairment does not add significant extra disadvantage leading to nursing home admission. In one previous study,⁶ nursing home residents (or their families and guardians) were asked whether vision problems influenced admission to the nursing home. Only 10% of people responded yes, despite a high prevalence of visual impairment among nursing home residents included in the study (57% had visual acuity of <20/40).

Our study had some limitations. We did not have data on time to institutional admission, which means that we could not take into account that people who died therefore had less time at risk of admission. However, there was only a weak association between visual impairment and death in this study,¹⁶ and it is unlikely that this explains the lack of association between visual impairment and nursing home admission. Our measure of visual impairment was based on distance visual acuity. It is possible that some people with visual field loss or low contrast sensitivity had good distance visual acuity. We aimed to minimize this misclassification by counting people registered as blind or partially sighted as visually impaired, even if they had good visual acuity.

In the Blue Mountains Eye Study in Australia, there was a relative risk of nursing home admission of 1.8 (95% CI, 1.1-2.9) among people with visual impairment after 6 years' follow-up.⁷ This was adjusted for age, sex, home ownership, self-rated health status, walking disability, smoking status, and alcohol consumption. Follow-up of participants in the Beaver Dam Eye Study showed a higher risk among people with visual impairment, with an odds ratio of 3.20 (95% CI, 1.85-5.56) after 5 years' follow-up.⁸ This was controlled for age, sex, self-rated health status, and arthritis history.

One unexpected result was that the risk of nursing home admission associated with visual impairment was higher among people with no evidence of cognitive impairment and in the highest quintile for ADL score. We expected that people with cognitive impairment would have more difficulties coping with visual impairment and therefore be more likely to need nursing home admission. One possible interpretation is that when multiple problems exist, visual impairment alone does not significantly influence nursing home admission.

Analyses of the causes of visual impairment and nursing home admission suggest that people with refractive error were at increased risk of nursing home admission. However, this finding was not significant and may have occurred owing to chance. Similarly, the Blue Mountains Eye Study found a slightly higher risk of nursing home admission among people with visual impairment because of refractive error. Similarly, this finding was not statistically significant.

Other studies have found that nursing home residents face considerable barriers to accessing eye care.¹⁷ In addition to transportation and patient advocacy, there is often a reluctance to consent to intervention, particularly surgical intervention. Similarly, studies of screening for visual impairment have found little effect on the prevalence of visual impairment in this age group.¹⁸ This lends support to the hypothesis that the high prevalence of visual impairment in nursing homes may be partly attributable to barriers to accessing eye care.

Cross-sectional studies of people in nursing homes suggest a much increased prevalence of visual impairment^1-6 (Table 6). Our results would suggest that, in the United Kingdom at least, although visually impaired people are at higher risk of nursing home admission, this may well be because they are older and thus more frail, with an increased number of comorbidities. It is unclear whether intervention to reduce visual impairment alone among older, community-dwelling adults would either reduce the risk of nursing home admission or reduce the prevalence of visual impairment in nursing homes.

Correspondence: Jennifer R. Evans, PhD, International Centre for Eye Health, London School of Hygiene and Tropical Medicine, Keppel Street, London WC1E 7HT, England (jennifer.evans@lshtm.ac.uk).

Submitted for Publication: July 13, 2007; final revision received April 3, 2008; accepted April 9, 2008.

Author Contributions: Drs Evans and Fletcher had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

Financial Disclosure: None reported.

Funding/Support: The MRC trial of the assessment and management of older people in the community was funded by the United Kingdom Medical Research Council, the Department of Health, and the Scottish Office. Collection of data on causes of visual impairment and additional analyses were funded by the Thomas Pocklington Trust.