Community screening for visual impairment in older people
References
References to studies included in this review
References to studies excluded from this review
Additional references
References to other published versions of this review
Characteristics of studies
Characteristics of included studies [ordered by study ID]
Jump to:
| Methods | Cluster randomised trial of 12 general practitioners | |
| Participants | Geographic region: Netherlands | |
| Interventions | Each practice was randomised to either: (1) Intervention group: multi‐component screening package including 4 disorders (vision, hearing, urinary incontinence and mobility) using self‐reporting OECD questionnaire as well as diagnostic tests during the first year. When the GP and the patient agreed on the intervention, usual care was provided, n = 483 (2) Control group: no screening (standard care) during the first year, then underwent same multi‐component screening package as the intervention group at year 2, n = 545 | |
| Outcomes | Presence of a visual disorder at 12 months in intervention group versus control group. A visual disorder was defined as "having difficulty recognising a face at 4 m and/ or reading normal letters in a newspaper and/ or impaired vision in both eyes (Snellen chart <0.3 or not being able to read normal newspaper letters at 25cm distance)." | |
| Notes | Funding source: Robert Wood Johnson Clinician Scholars Programme and the National Institute on Aging Geriatric Academic Programme Declaration of interest: not recorded Date study conducted: not recorded Trial registration number: not recorded | |
| Risk of bias | ||
| Bias | Authors' judgement | Support for judgement |
| Random sequence generation (selection bias) | Unclear risk | Method of randomisation not reported. “Randomised in 6 strata pairs of GPs, matched by town/ countryside group/ solo practice, age, sex and number of years practicing as a GP” |
| Allocation concealment (selection bias) | Low risk | Cluster randomised trial |
| Blinding of outcome assessment (detection bias) | Unclear risk | Masking of assessors not performed ‐ knowledge of intervention could have influenced outcome, but judged unlikely to be a material bias |
| Incomplete outcome data (attrition bias) | Low risk | Some imbalance in follow‐up but groups reasonably balanced in terms of reason for loss to follow‐up and thought to be too minimal to significantly affect results. Intervention: 483/732 (66%) Control: 545/738 (74%) |
| Selective reporting (reporting bias) | Low risk | Primary outcomes reported. |
| Other bias | Unclear risk | Uncertain risk of recruitment bias ‐ the exclusions were similar between (6% versus 8%) groups, which suggests that recruitment bias is unlikely to be a problem, but there were fewer people recruited per surgery than 160, according to Table 1, which remains unexplained. |
| Methods | Randomised: random number generator, centrally | |
| Participants | Geographic region: United Kingdom | |
| Interventions | (1) multi‐component home nurse assessment (including social functioning, current medical problems and additional question about vision). Those reporting visual problems given advice and referred to an optometrist (n = 151) | |
| Outcomes | Proportion who 'always' or 'quite often' had difficulty reading ordinary newsprint (with glasses if worn) | |
| Notes | Funding source: Newcastle Health Authority Declaration of interest: not recorded Date study conducted: not recorded Trial registration number: not recorded | |
| Risk of bias | ||
| Bias | Authors' judgement | Support for judgement |
| Random sequence generation (selection bias) | Low risk | Randomly allocated to groups, by random number generator. |
| Allocation concealment (selection bias) | Low risk | Randomisation conducted centrally. |
| Blinding of outcome assessment (detection bias) | Unclear risk | Masking of assessors not performed ‐ knowledge of intervention could have influenced outcome, but judged unlikely to be a material bias. |
| Incomplete outcome data (attrition bias) | Low risk | Attrition rate fully explained and equal between groups. |
| Selective reporting (reporting bias) | Low risk | Primary outcomes reported. |
| Other bias | Low risk | None apparent. |
| Methods | Cluster randomised controlled trial of 26 internists and family physicians | |
| Participants | Geographic region: Greater Los Angeles 8 to 12 patients from each of the internist's or family physician's practice Age: Over 70 Exclusion criteria: not acutely/ terminally ill, able to answer questions 161 patients within 26 practices | |
| Interventions | Each practice was randomised to either: (1) Visual screening as part of multi‐component package. Baseline and follow‐up questionnaires (MOS SF‐36), with screening assessment. This involved an eight item screening questionnaire/assessment including a visual screening question; 'Do you have any difficulty driving or watching television or reading or doing any of the activities because of your eyesight?'. If a positive screen was identified, then Snellen visual acuity assessment performed by physician and unknown interventions provided. (n = 112) (2) N visual screening ‐ standard care. Baseline and follow‐up questionnaires (MOS SF‐6) provided as in the intervention group (n = 149) | |
| Outcomes | Frequency of self‐reported improvement in vision via questionnaire Attrition rate: 31 total across both arms, with 17 = refusal, 4 = moving away, 8 = loss to follow‐up, 2 = death Follow‐up: 6 months | |
| Notes | Cluster randomised Funding source: Robert Wood Johnson Clinician Scholars Programme and the National Institute on Aging Geriatric Academic Program Declaration of interest: not recorded Date study conducted: not recorded Trial registration number: not recorded | |
| Risk of bias | ||
| Bias | Authors' judgement | Support for judgement |
| Random sequence generation (selection bias) | Low risk | Random number table used. |
| Allocation concealment (selection bias) | Low risk | Allocation concealment is unlikely to be an issue in cluster‐randomised trials. |
| Blinding of outcome assessment (detection bias) | Low risk | Questionnaire used to assess outcome. |
| Incomplete outcome data (attrition bias) | Low risk | Equal attrition rate (12%) between groups. Explained, but reasons not provided per group. Good follow‐up rate considering the age group. |
| Selective reporting (reporting bias) | Low risk | Primary outcomes reported. |
| Other bias | Unclear risk | Practice was aware of allocation at the time of recruiting participants so there was potential for recruitment bias. On the whole the groups were well balanced apart from membership of a health maintenance organisation which was more common in the intervention group (64% versus 33%). |
| Methods | Centralised cluster computer generated randomisation of general practices | |
| Participants | Geographic region: United Kingdom | |
| Interventions | Participants were randomised to one of two screening strategies | |
| Outcomes | Visual acuity less than 6/18 in either eye and mean composite score of the NEI VFQ‐25 comparing universal with targeted screening | |
| Notes | Cluster randomised Funding source: MRC and Department of Health Declaration of interest: not recorded Date study conducted: not recorded Trial registration number: not recorded | |
| Risk of bias | ||
| Bias | Authors' judgement | Support for judgement |
| Random sequence generation (selection bias) | Low risk | Computer generated randomisation list used. |
| Allocation concealment (selection bias) | Low risk | Randomisation performed centrally. |
| Blinding of outcome assessment (detection bias) | Unclear risk | Masking of assessors not performed ‐ knowledge of intervention could have influenced outcome, but judged unlikely to be a material bias |
| Incomplete outcome data (attrition bias) | Unclear risk | Attrition rate fully explained, but slightly different rates between groups (67.8% versus 57.9%) which could have affected outcome. |
| Selective reporting (reporting bias) | Low risk | Primary outcomes reported. |
| Other bias | Low risk | None apparent. |
| Methods | Randomised using random number tables, to blocks of size 4. Stratified by sex, falls history and recruitment source. Investigator allocating participants by allocation sequence had no contact with study subjects. | |
| Participants | Geographic region: community in Sydney, Australia Contact details obtained from range of care service providers and local advertisements placed. Age: 70 years or older Exclusion criteria: cataract surgery or new spectacle prescription in last 3 months N = 616 | |
| Interventions | (1) Vision tests and eye examinations by an optometrist, with appropriate treatment (new spectacles, referral to ophthalmologist, referral to occupational therapist). n = 309 (2) Standard care, no visual screening, n = 307 Follow‐up = 12 months | |
| Outcomes | Baseline assessment: socio‐economic details, medical history, history of vision and eye problems, falls history, use of psychotropic medications, activities of daily living (ADLs), 25‐item National Eye Institute Visual Function Questionnaire (VFQ‐25), and the Mini‐Mental State Examination (MMSe), visual acuity with near and distance logMAR Attrition rate = 35 (intervention group) versus 49 (control group), reasons provided. Follow‐up: follow‐up questionnaire including VFQ‐25, binocular visual acuity, whether or not had seen an eye‐care practitioner in the past year (if so, when). Data adjusted for baseline VFQ‐25 scores. | |
| Notes | Funding source: National Health and Medical Research Council of Australia Declaration of interest: none Date study conducted: Not recorded Trial registration number: Not recorded | |
| Risk of bias | ||
| Bias | Authors' judgement | Support for judgement |
| Random sequence generation (selection bias) | Low risk | Random number tables used. |
| Allocation concealment (selection bias) | Low risk | Random numbers used by investigator who had no contact with participants. |
| Blinding of outcome assessment (detection bias) | Low risk | Attempts made to mask assessor ‐ research assistants were "unaware of group allocation" |
| Incomplete outcome data (attrition bias) | Low risk | Attrition rate fully explained. |
| Selective reporting (reporting bias) | Low risk | Primary outcomes reported. |
| Other bias | Low risk | None. |
| Methods | Randomisation performed using computer generated random numbers with block design. Participants randomised into 4 groups; vision and hearing screening, vision screening only, hearing screening only, no screening (standard care) No masking was performed. | |
| Participants | Geographic region: Sydney, Australia | |
| Interventions | (1) Visual acuity screening (including groups with visual screening only and visual and hearing screening combined). VA testing (logMAR), binocular near testing, visual field testing (confrontation), n = 96 Under‐corrected refractive error (pinhole VA improved at least 10 letters in those presenting with VA < 6/6), bilateral visual impairment (better eye VA < 6/12) or self‐reported visual problems were recommended to have further assessment by eye‐care professionals (2) No visual acuity screening (including groups with hearing screening only and no screening), n = 92 Routine aged care assessment and interview using a standardised questionnaire All participants had MMSE, sociodemographic information, self‐rated health, past medical histories, use of community support services and details of informal help | |
| Outcomes | Improvement in visual acuity in one or both eyes at follow‐up (1 year) Number of individuals that were followed up (n = 121) not provided per group, but authors report "the mean VA in participants who had VA assessed (intervention) at baseline (39 letters) was non‐significantly better than those who did not have their VA assessed (35 letters, p = 0.25)." Attrition: 85/206 = 40% dropped out after randomisation. Rates between groups not provided. | |
| Notes | Funding sources: University of Sydney SESQUI Ophthalmic Research Institute of Australia (ORIA grant 2004), University of Sydney Postgraduate Award, Westmead Millennium Foundation Research Scholarship Stipend Enhancement Grant Declaration of interest: not recorded Date study conducted: not recorded Trial registration number: not recorded | |
| Risk of bias | ||
| Bias | Authors' judgement | Support for judgement |
| Random sequence generation (selection bias) | Low risk | Contact from author: randomisation performed using computer generated random numbers with block design. |
| Allocation concealment (selection bias) | Unclear risk | No information provided. |
| Blinding of outcome assessment (detection bias) | Unclear risk | Masking of assessors not performed ‐ knowledge of intervention could have influenced outcome, but judged unlikely to be a material bias. |
| Incomplete outcome data (attrition bias) | Unclear risk | 121/206 = 59% randomised participants seen at follow‐up. Follow‐up rate by intervention group not reported. |
| Selective reporting (reporting bias) | Low risk | Primary outcomes reported. |
| Other bias | Low risk | None. |
| Methods | Randomised: random numbers generator, centrally | |
| Participants | Geographic region: the Netherlands | |
| Interventions | (1) Four visits per year for 3 years by trained nurses. One question about vision: 'How do you assess your vision at present?' Possible answers: excellent, good, fair, not so good or bad. Those answering ‘fair’, 'not so good’ or ‘bad’ to the screening question advised to contact an optometrist (n = 292) | |
| Outcomes | Proportion answering ‘fair’, ‘not so good’ or ‘bad’ to the screening question at the end of the study | |
| Notes | Funding source: Netherlands Ministry of Welfare, Health and Cultural Affairs, the Foundation for Research and Development of Social Care (STOOM) and Het Praeventiefonds Declaration of interest: not recorded Date study conducted: not recorded Trial registration number: not recorded | |
| Risk of bias | ||
| Bias | Authors' judgement | Support for judgement |
| Random sequence generation (selection bias) | Low risk | Computer generated random numbers. |
| Allocation concealment (selection bias) | Low risk | Randomisation conducted centrally. |
| Blinding of outcome assessment (detection bias) | Low risk | Attempts were made to mask assessor: "interviews were conducted by trained interviewers, who were unaware of whether a participant had been regularly visited by a nurse or not". |
| Incomplete outcome data (attrition bias) | Unclear risk | Analysed data per protocol and ITT but no comment on differences between these results and unclear which are reported. |
| Selective reporting (reporting bias) | Low risk | Primary outcomes reported. |
| Other bias | Low risk | None. |
| Methods | Randomised by household: random number tables, centrally. Household randomisation undertaken because it was felt it would be difficult for the health visitor to intervene on behalf of one member of a household and not for another | |
| Participants | Geographic region: United Kingdom | |
| Interventions | (1) Annual assessment at home by a health visitor. Two questions about glasses and difficulty seeing. Those reporting difficulties seeing were referred to an optometrist or to their general practitioner and were offered advice from the health visitor (n = 577) | |
| Outcomes | Proportion with a positive response to the question at interview: ‘Do you have any difficulty seeing (even when wearing your glasses)' | |
| Notes | Funding source: Welsh Office and Department of Health and Social Security Declaration of interest: not recorded Date study conducted: not recorded Trial registration number: not recorded | |
| Risk of bias | ||
| Bias | Authors' judgement | Support for judgement |
| Random sequence generation (selection bias) | Low risk | Randomly allocated to groups, using random number tables. |
| Allocation concealment (selection bias) | Low risk | Randomisation conducted centrally. |
| Blinding of outcome assessment (detection bias) | Unclear risk | Masking of assessors not performed ‐ knowledge of intervention could have influenced outcome, but judged unlikely to be a material bias. |
| Incomplete outcome data (attrition bias) | Low risk | Attrition rate fully explained. |
| Selective reporting (reporting bias) | Low risk | Primary outcomes reported. |
| Other bias | Low risk | None. |
| Methods | Randomised: random number tables, centrally. Household randomisation undertaken because part of intervention included improvements in the home environment | |
| Participants | Geographic region: United Kingdom | |
| Interventions | (1) Annual assessment at home by a health visitor, specifically aimed at reducing falls and fractures. Two questions about glasses and difficulty seeing, and third question about recent eye tests. Those reporting difficulties seeing were referred to an optometrist or to their general practitioner, and were offered advice from the health visitor (n = 350) | |
| Outcomes | Proportion with a positive response to the interview question ‘Do you have any difficulty seeing (even when wearing your glasses)’ | |
| Notes | Funding source: the Grand Charity and the Welsh Office Declaration of interest: not recorded Date study conducted: not recorded Trial registration number: not recorded | |
| Risk of bias | ||
| Bias | Authors' judgement | Support for judgement |
| Random sequence generation (selection bias) | Low risk | Random number table used. |
| Allocation concealment (selection bias) | Low risk | Use of study numbers without direct contact from participants. |
| Blinding of outcome assessment (detection bias) | Unclear risk | Masking of assessors not performed ‐ knowledge of intervention could have influenced outcome, but judged unlikely to be a material bias. |
| Incomplete outcome data (attrition bias) | Low risk | Attrition rate fully explained. |
| Selective reporting (reporting bias) | Low risk | Primary outcomes reported. |
| Other bias | Low risk | None. |
| Methods | Randomised: random number table, independent of trialists or participants | |
| Participants | Geographic region: United States | |
| Interventions | (1) Invited for a multi‐component nurse assessment (including vision) aimed at reducing disability and falls. Those reporting problems received information about resources in the community designed to assist those with poor vision (n = 635) | |
| Outcomes | Proportions reporting visual problems on a mailed questionnaire | |
| Notes | Funding source: Centres for Disease Control Declaration of interest: not recorded Date study conducted: not recorded Trial registration number: not recorded | |
| Risk of bias | ||
| Bias | Authors' judgement | Support for judgement |
| Random sequence generation (selection bias) | Low risk | Randomly allocated to groups, using random number table. |
| Allocation concealment (selection bias) | Low risk | Allocation independent of trialists. |
| Blinding of outcome assessment (detection bias) | Low risk | Questionnaire used to assess outcome. |
| Incomplete outcome data (attrition bias) | Low risk | Attrition rate very low and fully explained. |
| Selective reporting (reporting bias) | Low risk | Primary outcomes reported. |
| Other bias | Low risk | None. |
General practice is equivalent to family practice
Characteristics of excluded studies [ordered by study ID]
Jump to:
| Study | Reason for exclusion |
| Visual outcomes not measured | |
| Intervention did not include assessment of vision | |
| Visual outcomes not measured | |
| Visual outcomes not measured | |
| Visual outcomes not measured | |
| No control group | |
| Visual outcomes not measured | |
| Not a general population group ‐ participants were recently discharged from emergency department and study was aimed at falls prevention rather than vision improvement | |
| Visual outcomes not measured | |
| No control group | |
| Visual outcomes not measured | |
| Participants aged 64 years and under only | |
| Visual outcomes not measured | |
| Visual outcomes not measured | |
| Visual outcomes not measured | |
| Visual outcomes not measured. | |
| Visual outcomes not measured |
Data and analyses
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 1 Not seeing well (as defined by each trial) Show forest plot | 6 | 4522 | Risk Ratio (M‐H, Fixed, 95% CI) | 1.05 [0.97, 1.14] |
| Analysis 1.1  Comparison 1 Vision screening as part of multi‐component screening package versus no vision screening (standard care), Outcome 1 Not seeing well (as defined by each trial). | ||||
Risk of bias summary: review authors' judgements about each risk of bias item for each included study.
Comparison 1 Vision screening as part of multi‐component screening package versus no vision screening (standard care), Outcome 1 Not seeing well (as defined by each trial).
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 1 Not seeing well (as defined by each trial) Show forest plot | 6 | 4522 | Risk Ratio (M‐H, Fixed, 95% CI) | 1.05 [0.97, 1.14] |
