Abstract
Rationale/Objective:
To develop a screening and referral algorithm tool to help identify which older institutionalized individuals with dementia need an eye examination.
Methods:
The visual acuity (VA) screening test was developed on an iPad retina display. Three optotypes were used (letters, numbers, and tumbling E’s) to determine whether one works best with dementia. The screening VA results and algorithm decision were validated against those obtained by an optometrist performing a complete eye examination.
Results:
Of the 150 participants, 14.7% did not respond to any optotype, while 85.3% responded to letters, 84.0% to numbers, and 66.0% to tumbling E’s. The VA achieved was superior for letters. The concordance for the screening versus eye examination was >80% for VA and 90% for the algorithm.
Conclusion:
The results indicate that the tool was successful at identifying older individuals with dementia needing an eye examination.
Keywords: dementia, eye examination, long-term care facilities, screening tool, visual acuity
Introduction
In Canada, 14.8% of the population was ≥65 years old in 2011, and it is projected to be around 24% in 2036. 1,2 Similar statistics in the United States indicate that 13.7% of the population was ≥65 years old in 2012, a proportion projected to increase to 20.3% by 2030. 3 This aging of the population is characterized by an increased prevalence of neurodegenerative diseases such as dementia that affects up to a third of those ≥85 years old. 4 Many older individuals affected by dementia will be institutionalized due to decreased autonomy and increased need for personal and health care. 5
The prevalence of ocular disease, visual deficit, and blindness increases with age 6 and is higher in individuals residing in long-term care facilities (LTCFs)/nursing homes (NHs). 7 Yet, vision is a very important sensory modality that should be optimized in older residents, particularly those affected by dementia, as they now represent >60% of people admitted to LTCF. 5 Vision loss can aggravate problems associated with dementia 8 and is an independent factor contributing to the appearance and maintenance of disruptive behavior in LTCF residents. 9 Studies have even indicated that visual deficit is a risk factor for cognitive decline. 10,11 It would therefore be useful to identify, among residents with dementia, those who present an oculovisual problem. A potential first step would be to screen visual acuity (VA) in residents. This screening, together with relevant information on systemic and ocular disease contained in the clinical chart, could help identify residents who would necessitate an eye examination. Our objective was to develop a screening and referral algorithm tool that could be used by noneye care professionals, in older residents affected by dementia, in trying to identify those needing an eye examination.
Methods
Participants were recruited at the Institut universitaire de gériatrie de Montréal (IUGM). They had to be ≥65 years old, to reside in long-term care units, to have a clinical diagnosis of dementia in their clinical chart, and to be able to understand French or English. Residents were excluded if they were terminally ill, not able to respond to any stimulation, and if no guardian could be located if necessary to provide consent. The study protocol was approved by the Research Ethics Committee of the IUGM, and the study was conducted according to the Declaration of Helsinki.
Recruitment
The research assistant (RA) reviewed the clinical chart of all LTCF residents to verify eligibility criteria. The RA then verified whether the person was competent (ie, able to decide for himself or herself) or not as well as the contact person/guardian to be contacted, if required. A first contact was made with the resident and/or contact person to explain in detail the information and consent form. A signed informed consent was obtained from each participant and/or their contact person/guardian.
Development of the Screening Test
The screening test was developed in the form of a dedicated software used on the retina display iPad. It was modeled after an existing tool using tumbling E’s for screening VA in older institutionalized individuals in a project dealing with fall prevention, 12 which was itself adapted from an earlier one developed by the World Health Organization. 13 The new tool also took into account research results, indicating that letters and numbers worked better than tumbling E’s for measuring VA in older institutionalized individuals with dementia. 14 Three types of high-resolution, high-contrast, black optotypes on a white luminous background were used: letters, numbers, and tumbling E’s. Letters were O, H, V, T; numbers were 5, 6, 8, 9; and tumbling E’s were presented in the up, down, right, and left positions. Three levels of VA were used for each of the 3 m and 40 cm test distances. Visual acuity levels were the equivalent of 20/200, 20/60, 20/40 at 3 m and the equivalent of 20/320, 20/120, 20/50 at 40 cm. Four optotypes were presented for each level of VA. The optotypes were displayed one at a time in the center of the iPad screen from the largest to the smallest one, in random sequence for each target size. The software also provided the time taken to complete each test.
Screening VA for Residents
The screening session for each participant took place in a room dedicated to that purpose. The test was performed with the habitual ophthalmic correction of each participant, as used on a daily basis. The test was performed monocularly, starting at 3 m, testing the right then the left eye, and then testing the 40 cm distance in the same fashion. The nontested eye was covered with a black eye patch.
The 3 m distance from the ophthalmic correction plane was verified with a laser meter and the 40 cm distance with a measuring tape. When bifocal type lenses were worn, the experimenter ensured that the participant was looking through the appropriate portion of the lenses for each distance. The optotype to be used, eye to be evaluated, and viewing distance were chosen from the software menu. The start button was pressed and the first symbol for the lower acuity level was displayed in the center of the screen. The symbol was changed by sliding a finger on the screen. The experimenter could go forward or backward, and the participant was always given a chance to repeat a symbol that was missed in the first place. Each correct or incorrect response given by the participant was recorded as the test progressed. The test continued until the 12 optotypes in a series were presented, which was indicated by a sound on the iPad. The menu button to test the second eye could then be activated. Once the VA for letters was complete, the experimenter proceeded with the numbers or tumbling E’s in random order and then tested the 40 cm distance.
Care was taken to do the test sessions at a time of the day when the person usually was at their best according to the person himself or herself, family, or nursing staff. All testing was performed in 1 session if possible, but more often than not, 2 or 3 sessions were required, because the participant was too tired, too sleepy, or too frustrated to continue.
Eye Examination for Residents
The eye examination was performed by the research optometrist in a fully equipped examination room at the time of the day that best suited each participant. The eye examination was usually performed on 2 different days, 1 for the basic eye examination and 1 for the ocular health assessment under pupillary dilation. Baseline VA for distance and near vision was performed through the habitual ophthalmic correction of participants, that is, the same one that was also used during the screening test. The optometrist was not aware of the results of the screening test and was not limited to any single VA chart to conduct the examination, but rather chose the one that worked for each person, as is done in any regular eye examination.
Referral Algorithm
A referral algorithm was built to determine whether a participant should have been referred after testing, based on the VA level obtained (screening for RA; eye examination for optometrist) and on the information regarding systemic and ocular disease contained in the resident’s chart (Figure 1). The presence of diabetes, cataract, glaucoma, or age-related macular degeneration (ARMD) in the chart was considered a reason for referral for an eye examination. The referral algorithm was used by the RA and the optometrist, independently, without knowing the results obtained by each other.
Figure 1.
Algorithm used to decide whether a participant should be referred for a complete eye examination. D indicates diabetes; far, far distance; near, near distance; OP, ocular pathology; VA, visual acuity.
Statistical Analyses
Analyses were performed to verify whether there were any differences in the (1) response rate to each set of optotypes used in the screening test; (2) level of VA obtained with, and time taken to respond to, each of the 3 sets of optotypes; (3) VA results obtained during screening versus eye examination; (4) VA obtained monocularly versus binocularly during the eye examination; and (5) classification of participants meeting the criterion level of VA for referral between the RA and the optometrist. Finally, the results obtained by the RA on the referral algorithm were compared to those obtained by the optometrist to evaluate the validity of the screening tool. Paired t tests and McNemar tests were conducted to compare dependent samples. Chi-square test was used to examine variables with independent samples. Differences with a P value <.05 were considered significant.
Results
In total, 150 residents (110 women) aged 67.9 to 102.3 years (mean = 86.7 ± 7.2) with a clinical diagnosis of dementia were recruited. There were 45 competent (29 women) participants (mean age = 85.7 ± 6.5 years old) and 105 noncompetent (81 women) participants (mean age = 87.2 ± 7.5 years old).
Of the 150 participants, 128 (85.3%) were able to respond to letters, 126 (84.0%) to numbers, and 99 (66.0%) to tumbling E’s, while 22 (14.7%) did not respond to any optotype (Table 1). These percentages did not differ for letters and numbers (P > .05), but it was inferior for tumbling E’s (P = .003). Table 1 also presents results for competent and noncompetent participants. The data further showed that the averaged VA for letters was always superior (P = .000) to that measured for numbers and tumbling E’s, which did not differ from each other (P > .05; Table 2). The time taken to respond to letters and numbers did not differ (P > .05), while it was longer for tumbling E’s (P = .000; Table 3).
Table 1.
Percentage of Participants Responding to the Various Optotypes.
| All | Competent | Noncompetent | |
|---|---|---|---|
| Letters, n (%) | 128 (85.3) | 45 (100.0) | 83 (79.0) |
| Numbers, n (%) | 126 (84.0) | 45 (100.0) | 81 (77.1) |
| E’s, n (%) | 99 (66.0) | 40 (88.9) | 59 (56.2) |
| Nothing, n (%) | 22 (14.7) | 0 (0.0) | 22 (21.0) |
Table 2.
Averaged Visual Acuity Obtained With the Optotypes Used in the Screening.
| 3 m | 40 cm | |||
|---|---|---|---|---|
| OD | OS | OD | OS | |
| Letters | 20/56 | 20/62 | 20/75 | 20/80 |
| Numbers | 20/65 | 20/71 | 20/87 | 20/91 |
| Tumbling E’s | 20/64 | 20/69 | 20/84 | 20/90 |
Abbreviations: OD, right eye; OS, left eye.
Table 3.
Time (Seconds) Taken to Respond to the Various Optotypes Used in the Screening.a
| 3 m | 40 cm | |||
|---|---|---|---|---|
| OD | OS | OD | OS | |
| Letters | 106.2 ± 105.2 | 88.7 ± 85.1 | 72.6 ± 59.8 | 71.4 ± 65.5 |
| Numbers | 95.2 ± 80.8 | 89.3 ± 73.6 | 68.5 ± 50.6 | 76.7 ± 55.8 |
| Tumbling E’s | 129.4 ± 92.1 | 116.6 ± 90.7 | 110.0 ± 70.9 | 108.3 ± 78.0 |
Abbreviations: OD, right eye; OS, left eye.
aValues are average ± standard deviation (SD).
The VA results obtained for letters were compared between the screening versus eye examination (Table 4). The VA at far distance did not differ (P > .05), while it was superior during the eye examination at near distance (P = .000). The VA obtained during the eye examination was superior in the binocular versus monocular condition (P = .000) for both distances.
Table 4.
Averaged Visual Acuity Obtained During the Screening and the Eye Examination.
| Far Distance | Near Distance | |||||
|---|---|---|---|---|---|---|
| OD | OS | OU | OD | OS | OU | |
| Screening | 20/56 | 20/62 | n/a | 20/75 | 20/80 | NA |
| Eye examination | 20/56 | 20/60 | 20/47 | 20/58 | 20/59 | 20/50 |
Abbreviations: OD, right eye; OS, left eye; OU, both eyes.
The RA and optometrist classified the participants as having or not a decrease in monocular VA for far and near distances, based on the VA criteria for referral (Figure 1). The concordance was 83.6% (far) and 81.2% (near) between the results of the screening versus eye examination (P = .000). The results obtained by the optometrist indicated that 33 (at far) and 23 (at near) participants who did not meet the VA criteria (14 having a deficit at both distances) would have been missed if monocular VA had not been measured (P = .001).
There was a 90.0% (n = 135 of 150) concordance in the algorithm results between the screening versus eye examination (P = .007). These concordant data indicated that 85.3% of residents (n = 116 of 135) would have needed a complete eye examination, based on the criteria indicated in Figure 1. This represents 77.3% (n = 116 of 150) of the population under study. The algorithm obtained by the optometrist further showed a difference when binocular versus monocular VA was considered (P = .000), 14 participants needing an eye examination not being detected under binocular condition.
Discussion
Our data indicate that using letters and numbers elicit a higher percentage of response than tumbling E’s. They also show that the time taken to measure VA was similar for letters and numbers, while it was longer for tumbling E’s. The data further indicate that letters provide a higher level of VA compared to numbers or tumbling E’s. These combined data therefore indicate that letters should be privileged when screening VA in older individuals with dementia. Further, there was a fairly good concordance between the screening and eye examination results, indicating that the tool could be used by noneye care professionals for detecting older individuals with dementia needing an eye examination.
The mean VA obtained for letters during the screening was comparable to the one measured by the optometrist for far distance, while it was lower by about 1 acuity line during the screening versus eye examination at near distance. Furthermore, although there was a good concordance (83.6%/81.2%) in the classification of participants meeting or not the referral VA criteria, there was also some discrepancy. This indicates that for the remaining participants, either the screening or eye examination found no visual deficit, while the other did. A closer look at the discordant results indicate that about half of the time, the data were very close, that is, within 1 line of VA between the screening versus eye examination, and about half of the time they varied substantially by ≥3 lines of VA. This latter result may have been due of a lack of attention or collaboration from the participant, potentially linked to dementia. This suggests that for older individuals with dementia, if the acuity measured is within 1 line of the cutoff for visual deficit, and in the absence of ocular disease, diabetes, or any sign/symptom of an oculovisual problem, a follow-up in a year rather than a complete eye examination might be considered.
The algorithm results used to decide whether a person should be referred for an eye examination showed a high concordance (90.0%) between the RA versus optometrist, but it also showed a discrepancy 10.0% of the time. Again, the discordant data indicated that the results obtained were either within 1 line of VA or varied by ≥3 lines. On the other hand, the concordant data clearly indicated that an eye examination was deemed necessary for most participants (85.3%, representing 77.3% of all participants), while for the others, only a follow-up in a year was necessary. To our knowledge, there are no other studies having used such a tool in an older LTCF/NH population comprised exclusively of residents affected by dementia. A few studies have used a vision screening tool in older individuals, all or some of whom resided in NH, and concluded on the effectiveness of such screening to detect and refer patients needing eye care. 12,15,16 Our data support these findings and further indicate that the tool can also be used successfully in older residents affected by mild to severe dementia. The previous studies, however, used only tumbling E’s to screen VA and they tested it under binocular condition. Our results therefore present novel information, showing that tumbling E’s don’t work as well for people with dementia and that letters should be used. Our results further indicate that the screening should be performed monocularly, since the optometrist’s data showed that participants with deficient monocular VA, as well as some referrals, would have been missed with VA measured under binocular condition only. This is important as changes in refraction may happen in 1 eye without the person noticing it, but more importantly, because ocular disease can progress slowly and insidiously in 1 eye without affecting binocular VA. Decreased VA related to refractive changes or cataract is usually correctable, which is not the case in diseases such as glaucoma or ARMD. Earlier detection of disease may help prevent or at least slow its progression and this requires monocular evaluation. This should certainly not be overlooked from a vision standpoint but also because dual visual and cognitive impairment may lead to negative consequences such as disorientation, loss of independence, risk of isolation, or hallucinations. 17
The present results thus indicate that the screening and algorithm tool could be offered to LTCF/NH, together with adequate training for their noneye care professionals who would use it, to help identify older residents needing an eye examination. However, our data also indicate that the majority of residents would need a full eye examination. This suggests that even if screening VA works in people affected by mild, moderate, or severe dementia, LTCF/NH should work at implementing regular eye care services for their residents affected by dementia.
These study results could easily be extended to the community to screen VA in older vulnerable individuals who, for whatever reason, don’t have easy access to eye care. This tool could easily be used by health or social workers doing home visits or in individuals admitted to acute geriatric care or hospital, for example, following a fall. Knowing that this tool can screen VA at any stage of dementia, it certainly could be used confidently in the community even if the cognitive status of an older person was not known. Help could then be offered to the person testing positive, in order for a full eye examination to be provided. Any oculovisual problem that can be detected and managed while in the community would be of direct benefit to the older person now and later in life if the person ever has to be admitted to a long-term care institution. 18
In conclusion, the present results revealed that the screening and referral algorithm tool developed in this study can be used in older individuals at all levels of dementia. They have further shown, however, that most LTCF/NH residents would rather benefit from a full eye examination.
Acknowledgments
The authors wish to thank the ethics committee and Geneviève Ducharme (site coordinator) for facilitating the implementation of this project within the IUGM as well as all participants and their families for their generous contribution to this project.
Footnotes
The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding: The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This study was funded by a research grant from the Alzheimer Society of Canada.
References
- 1. Statistics Canada. Analytical Document. The Canadian Population in 2012: Age and Sex. Ottawa, ON: Statistics Canada. Catalogue no. 98-311-X2011001. [Google Scholar]
- 2. Statistics Canada. Population Projections for Canada, Provinces and Territories. 2009 to 2036. Ottawa, ON: Minister responsible for Statistics Canada © Minister of Industry; 2010. Catalogue no. 91-520-X. [Google Scholar]
- 3. Ortman JM, Velkoff VA, Hogan H. An Aging Nation: The Older Population in the United States. Washington, DC: Population Estimates and Projections, US Census Bureau; 2014. [Google Scholar]
- 4. Canadian Study of Health and Aging: study methods and prevalence of dementia. CMAJ. 1994;150(6):899–913. [PMC free article] [PubMed] [Google Scholar]
- 5. Un milieu de vie de qualité pour les personnes hébergées en CHSLD. Orientations ministérielles. 2003. http://www.msss.gouv.qc.ca/sujets/groupes/personnes_agees.php#milieu. Accessed September 15, 2016.
- 6. Congdon N, O’Colmain B, Klaver CC, et al. The eye diseases prevalence research group. Causes and prevalence of visual impairment among adults in the United States. Arch Ophthalmol. 2004;122(4):477–485. [DOI] [PubMed] [Google Scholar]
- 7. Owsley C, McGwin G, Scilley K, Meek GC, Dyer A, Seker D. The visual status of older persons residing in nursing homes. Arch Ophthalmol. 2007;125(7):925–930. [DOI] [PubMed] [Google Scholar]
- 8. Coons DH, Weaverdick SE. Wesley hall: a residential unit for persons with Alzheimer’s disease and related disorders. Phys Occup Ther Geriatr. 1986;4(3):29–53. [Google Scholar]
- 9. Horowitz A. The relationship between vision impairment and the assessment of disruptive behaviors among nursing home residents. Gerontologist. 1997;37(5):620–628. [DOI] [PubMed] [Google Scholar]
- 10. Lin MY, Gutierrez PR, Stone KL, et al. ; Study of Osteoporotic Fractures Research Group. Vision impairment and combined vision and hearing impairment predict cognitive and functional decline in older women. J Am Geriatr Soc. 2004;52(12):1996–2002. [DOI] [PubMed] [Google Scholar]
- 11. Reyes-Ortiz CA, Kuo YF, DiNuzzo AR, Ray LA, Raji MA, Markides KS. Near vision impairment predicts cognitive decline: data from the Hispanic established populations for epidemiologic studies of the elderly. J Am Geriatr Soc. 2005;53(4):681–686. [DOI] [PubMed] [Google Scholar]
- 12. Bell S, Hawranik PG, McCormac K. Focus on falls prevention: a quality improvement initiative. AER J. 2011;4(3):133–138. [Google Scholar]
- 13. Keeffe JE, Lovie-Kitchin JE, Maclean H, Taylor HR. A simplified screening test for identifying people with low vision in developing countries. Bull World Health Organ. 1996;74(5):525–532. [PMC free article] [PubMed] [Google Scholar]
- 14. Chriqui E, Kergoat MJ, Champoux N, Leclerc BS, Kergoat H. Visual acuity in institutionalized seniors with moderate to severe dementia. J Am Med Dir Assoc. 2013;14(4):275–279. [DOI] [PubMed] [Google Scholar]
- 15. Nottle HR, McCarty CA, Hassell JB, Keeffe JE. Detection of vision impairment in people admitted to aged care assessment centres. Clin Exp Ophthalmol. 2000;28(3):162–164. [DOI] [PubMed] [Google Scholar]
- 16. Camicelli A, Keeffe J, Martin K, Carbone J, Balding C, Taylor H. Vision screening for older people: the barriers and the solutions. Australas J Aging. 2003;22(4):179–185. [Google Scholar]
- 17. Lawrence V, Murray J, Banerjee S. The Experiences and Needs of People with Dementia and Serious Visual Impairment: A Qualitative Study. London, UK: Occasional Paper Thomas Pocklington Trust; 2008;16:1–23. [Google Scholar]
- 18. Armstrong RA, Kergoat H. Oculo-visual changes and clinical considerations affecting older patients with dementia—invited review. Ophthal Physiol Opt. 2015;35(4):352–376. [DOI] [PubMed] [Google Scholar]