Table 1:

Outcomes of screening for impaired vision in community-dwelling adults aged 65 years and older in primary care settings

Outcome and follow-up periodNo. of studiesCases/screened patientsCases/control patientsRelative risk (95% CI)Absolute difference per 1000 (95% CI)Quality of evidence
Falls (self-report falls as a proxy for fractures)
Follow-up: range 12 mo to 18 mo
2 RCTs
Day et al. (28)
 Any fall*691/547757/5430.88 (0.79 to 0.98)163 fewer (28 to 292 fewer)⊕○○○
Very low
 Falls requiring medical treatment*49/54775/5430.65 (0.46 to 0.91)48 fewer (12 to 75 fewer)
Newbury et al. (29)
 Any fall4/453/441.30 (0.31 to 5.49)20 more (48 fewer to 305 more)
Vision-related function and quality of life (assessed with NEI-VFQ-25; scale from 0 to 100 (22))
Follow-up: median 3.9 yr
1 RCTn = 829n = 978MD 0.4 units higher (1.25 lower to 2.05 higher)⊕⊕○○
Change in visual acuity with objective screening (mean change in high-contrast distance visual acuity) (19), (26) (28)
Follow-up: median 12 mo
4 RCTsn = 764n = 579MD −0.01 logMAR better (−0.05 better to 0.03 worse)⊕⊕⊕○
Impaired visual acuity with objective screening (< 20/40 distance visual acuity: bilateral) (22), (26)
Follow-up: range 12 to 47 mo
2 RCTs290/913394/10540.82 (0.66 to 1.02)67 fewer (from 7 more to 127 fewer)⊕⊕○○
Self-reported vision problems (primarily questionnaire-based impairment tests) (20), (21), (23) (25), (29) (33)
Follow-up: median 20 mo
10 RCTs1042/3767 (27.7%)1296/4916 (26.4%)0.97 (0.90 to 1.05)8 fewer (from 13 more to 26 fewer)⊕⊕⊕○
  • Note: CI = confidence interval, MD = mean difference, NEI-VFQ-25 = National Eye Institute Visual Functioning Questionnaire 25, RCT = randomized controlled trial.

  • * The authors of the study (28) of an intervention to prevent falls in older adults, which included a vision component, showed that only 26 of the 547 participants who were assigned the vision component actually received treatment; therefore, it is unlikely that the vision treatment had the effect on falls quoted in this table. In particular, the authors of this study ascribe differences in the rate of falls to the exercise component of the intervention. In addition, visual acuity improved marginally among the control group and not at all among the intervention group. No other differences were seen in vision measures, which makes the conclusion that vision screening had an impact on falls unlikely.

  • Very serious concerns about an unclear risk of bias, owing to inconsistency from reliance on one trial (28); about indirectness resulting from surrogate outcomes — 75% of participants received an intervention that could have confounded risk; and about imprecision, as the optimal information size was not met.

  • Serious concerns about high risk of bias for not blinding personnel, patients and outcome assessors; about high and differential attrition [42% v. 32% of those alive]; and about inconsistency arising from unknown effects from other studies.

  • § Moderate concerns about four RCTs (19), (26) (28) with unclear risk of bias that used multiple objective screening tools and indirectness, as two of the RCTs (19), (26) included many patients who were receiving home care, and one (19) provided an additional intervention that may have influenced results.

  • Serious concerns about inconsistency in one trial (26) and about imprecision, as the optimal information size of about 200 total events with a control event rate of 0.28 was not met.

  • ** Moderate concerns about risk of bias for not blinding personnel or patients in any study, about attrition in one study, (23) and about allocation concealment. Although there was some inconsistency (2 CIs did not overlap; I2 = 29%), there were no serious concerns regarding this domain.