Screening for Impaired Visual Acuity in Older Adults: Updated Evidence Report and Systematic Review for the US Preventive Services Task Force

Importance  A 2016 review for the US Preventive Services Task Force (USPSTF) found that effective treatments are available for refractive errors, cataracts, and wet (advanced neovascular) or dry (atrophic) age-related macular degeneration (AMD), but there were no differences between visual screening vs no screening on visual acuity or other outcomes.

Objective  To update the 2016 review on screening for impaired visual acuity in older adults, to inform the USPSTF.

Data Sources  Ovid MEDLINE, the Cochrane Central Register of Controlled Trials, and the Cochrane Database of Systematic Reviews (to February 2021); surveillance through January 21, 2022.

Study Selection  Randomized clinical trials and controlled observational studies on screening, vascular endothelial growth factor (VEGF) inhibitors (wet AMD), and antioxidant vitamins and minerals (dry AMD); studies on screening diagnostic accuracy.

Data Extraction and Synthesis  One investigator abstracted data and a second checked accuracy. Two investigators independently assessed study quality.

Results  Twenty-five studies (N = 33 586) were included (13 trials, 11 diagnostic accuracy studies, and 1 systematic review [19 trials]). Four trials (n = 4819) found no significant differences between screening vs no screening in visual acuity or other outcomes. Visual acuity tests (3 studies; n = 6493) and screening question (3 studies; n = 5203) were associated with suboptimal diagnostic accuracy. For wet AMD, 4 trials (n = 2086) found VEGF inhibitors significantly associated with greater likelihood of 15 or more letters visual acuity gain (risk ratio [RR], 2.92 [95% CI, 1.20-7.12]; I² = 76%; absolute risk difference [ARD], 10%) and less than 15 letters visual acuity loss (RR, 1.46 [95% CI, 1.22-1.75]; I² = 80%; ARD, 27%) vs sham treatment, with no increased risk of serious harms. For dry AMD, a systematic review (19 trials) found antioxidant multivitamins significantly associated with decreased risk of progression to late AMD (3 trials, n = 2445; odds ratio [OR], 0.72 [95% CI, 0.58-0.90]) and 3 lines or more visual acuity loss (1 trial, n = 1791; OR, 0.77 [95% CI, 0.62-0.96]) vs placebo. Zinc was significantly associated with increased risk of genitourinary events and beta carotene with increased risk of lung cancer in former smokers; other serious harms were infrequent.

Conclusions and Relevance  This review found that effective treatments are available for common causes of impaired visual acuity in older adults. However, direct evidence found no significant association between vision screening vs no screening in primary care and improved visual outcomes.

Impaired visual acuity is common in older adults. In 2017, an estimated 53 million US adults older than 65 years were at high risk for serious vision loss, which can result in disability, loss of productivity, and reduced quality of life.¹ Rates of severe vision loss are predicted to double or triple as the number of older adults increases.^1-3

In 2016, the US Preventive Services Task Force (USPSTF) concluded that the current evidence was insufficient to assess the balance of benefits and harms of screening for impaired visual acuity in older (≥65 years) adults (I statement).⁴ Although a 2016 USPSTF review found that screening can identify persons with impaired visual acuity and that effective treatments are available for common causes of impaired visual acuity such as refractive error, cataracts, and wet (advanced neovascular [caused by leakage of abnormal blood vessels under the macula]) or dry (atrophic [caused by thinning of the macula]) age-related macular degeneration (AMD), direct evidence found no differences between vision screening in older adults in primary care settings vs no screening in visual acuity or other clinical outcomes.^5,6 This report was conducted to update the 2016 review on screening for impaired visual acuity in older adults, to inform the USPSTF for an updated recommendation.

Detailed methods and additional study details are available in the full evidence report.⁷ Figure 1 shows the analytic framework and key questions (KQs) that guided the review.

Ovid MEDLINE, the Cochrane Central Register of Controlled Trials, and the Cochrane Database of Systematic Reviews were searched from January 2015 to February 9, 2021 (eMethods 1 in the Supplement). Searches were supplemented by reference list review of relevant studies; studies from the prior USPSTF review^5,6 that met inclusion criteria were carried forward. Ongoing surveillance was conducted to identify major studies published since February 2021 that may affect the conclusions or understanding of the evidence and the related USPSTF recommendation. The last surveillance was conducted on January 21, 2022, and identified no studies affecting review conclusions.

Two investigators independently reviewed titles, abstracts, and full-text articles using predefined eligibility criteria (eMethods 2 in the Supplement). The population was older adults (65 years or older). Screening was performed with vision tests or questionnaires in primary care settings or were feasible for primary care (did not require eye specialty training or equipment) and compared against no screening. Treatment focused on benefits and harms of wet AMD (intravitreal vascular endothelial growth factor [VEGF] inhibitors) and dry AMD (vitamins and antioxidants). The USPSTF previously determined that treatments for refractive errors and cataracts are effective, and this was not rereviewed.^6,9 Treatment was compared against placebo or sham; in addition, newer VEGF inhibitors (aflibercept and brolucizumab-dbll) were compared against older VEGF inhibitors because of the lack of placebo-controlled trials. Outcomes were visual acuity, vision-related quality of life; functional capacity; and harms (including falls and fractures and other treatment-related harms). An updated version¹⁰ of a systematic review¹¹ on treatment for dry AMD used in the prior USPSTF review was included. Otherwise this report used primary studies. Inclusion was restricted to English-language articles, and studies published only as abstracts were excluded.

One investigator abstracted details about the study design, patient population, setting, interventions, analysis, follow-up, and results from each study. A second investigator reviewed abstracted data for accuracy. Two independent investigators assessed the quality of each study as good, fair, or poor using predefined criteria (eMethods 3 in the Supplement) developed by the USPSTF.⁸ Discrepancies were resolved by consensus. In accordance with the USPSTF Procedure Manual,⁸ studies rated poor quality because of critical methodological limitations were excluded.

For all KQs, the overall strength of evidence was rated “high,” “moderate,” “low,” or “insufficient” based on study limitations, consistency, precision, reporting bias, and applicability, using the approach described in the USPSTF Procedure Manual.⁸ No new evidence suitable for meta-analysis was identified for this review, owing to small numbers of studies and heterogeneity in populations, interventions, and outcomes. However, a random-effects meta-analysis conducted for the prior USPSTF review⁶ on the effects of VEGF inhibitors remained relevant and was carried forward in this review.

Across all KQs, 25 studies (reported in 51 publications, total N = 33 586 participants) were included (13 randomized clinical trials (RCTs), 11 diagnostic accuracy studies, and 1 systematic review) (Figure 2).^12-62 Sixteen studies^{12,14,21,23,24,27,34,35,39,42,44,46,47,52,53,58,61} were carried forward from the 2016 USPSTF review,^5,6 8 studies^{17,19,20,29,33,41,43,57,61,62} were new, and an updated Cochrane systematic review¹⁰ included 19 studies (the previous Cochrane review¹¹ included 13).^{15,16,18,22,25,26,28,30-32,36,37,40,49,54-56,59,60}

Key Question 1. What are the effects of vision screening in asymptomatic older adults vs no screening on visual acuity, morbidity or mortality, general or vision-related quality of life, functional status, or cognition?

Four fair-quality RCTs^{19,23,34,46,47,61} (in 6 publications; n = 4819) compared vision screening in primary care–applicable settings vs no screening, usual care, or delayed screening (eTable 1 in the Supplement; all were included in the 2016 USPSTF review except for 1 small (n = 188) trial.¹⁹ The duration of follow-up ranged from 6 months to 5 years. Screening methods varied: a brief screening questionnaire plus the Glasgow visual acuity chart followed by pinhole testing for persons with visual acuity worse than 6/18 (20/60)²³; assessment of difficulty in recognizing a face, reading normal letters in a newspaper, or both, along with Snellen visual acuity eye chart^46,47; a screening question and clinical summary followed by the Snellen eye chart³⁴; and an Early Treatment Diabetic Retinopathy Study (ETDRS) visual acuity chart, measurement of binocular near vision and visual field testing, along with screening questions.¹⁹ Three of the trials were conducted in community or general practice settings, and screening was conducted by general practitioners, office staff, or trained nurses. The additional trial¹⁹ was conducted in a geriatric day hospital, although screening could be done via home visit if needed. Screening was conducted by study investigators (geriatric medicine or eye specialist) or an orthoptist, but the study was considered primary care–applicable because the screening methods consisted of visual acuity testing, binocular near vision, and visual field confrontation testing. Methodological limitations included unclear allocation concealment and blinding methods and high loss to follow-up (eTable 2 in the Supplement).

None of the trials, including the trials added for this update, found beneficial effects of screening on visual acuity, likelihood of vision disorders, or vision-related functional impairment or quality of life (Table 1). In the largest (n = 3249) trial, universal vision screening identified about 10 times as many patients with impaired visual acuity and correctable impairment compared with targeted screening, but there was no significant difference in the likelihood of visual acuity worse than 20/60 at 3- to 5-year follow-up (relative risk [RR], 1.07 [95% CI, 0.84-1.36]).²³ Another large (n = 1121) trial found no significant difference between immediate vs delayed screening in the likelihood of visual disorders at 2 years (51% [95% CI, 45%-58%] vs 47% [95% CI, 42%-52%]; P = .68).^46,47 Potential reasons for lack of screening benefit may include attrition (24% to nearly 60% in the larger trials at 2 to 5 years),^23,46,47 similar frequency of vision disorder detection and treatment in the screening and control groups,³⁴ use of a suboptimal method (a question) for initial screening,³⁴ low uptake of recommended follow-up or interventions,^23,47 or high rates of antecedent eye professional care.¹⁹

Key Question 2. What are the harms of vision screening in asymptomatic older adults vs no screening?

No screening study reported harms.

Key Question 3. What is the diagnostic accuracy of screening for impaired visual acuity due to uncorrected refractive error, cataracts, or AMD?

Eight fair-quality studies (n = 7398) examined the accuracy of screening tests for impaired visual acuity due to visual conditions such as cataracts, refractive error, and AMD in older adults (eTables 3-4 in the Supplement). Seven (reported in 6 publications)^{12,14,21,35,39,58} were in the prior USPSTF review⁶ and 1 study (n = 104)⁵⁷ was added. Screening was conducted using an eye chart (Snellen or logarithm of the minimum angle of resolution [logMAR], 3 studies),^12,14,58 a computerized tool based on 4 tests of vision function (2 studies),³⁹ the Minimum Data Set Vision Patterns section score (1 study),²¹ geriatrician examination (1 study),³⁵ the Amsler grid (a grid of horizontal and vertical lines used for central visual field monitoring) (1 study),⁵⁸ or a mobile application.⁵⁷ Methodological limitations included failure to apply the reference standard in all patients, interpretation of the reference standard not independent from screening test results, and thresholds for a positive screening test result not prespecified (eTable 5 in the Supplement).

Three studies (n = 6493) evaluated screening visual acuity tests compared with a complete ophthalmologist examination. Based on a visual acuity threshold on screening of less than 20/30 or less than 20/40, sensitivity ranged from 0.27 to 0.75 and specificity from 0.51 to 0.87. One study each found low accuracy of a computer-based screening tool or the Minimum Data Set MDS Vision Patterns section score.^21,39 One study (n = 50) in the prior USPSTF review found a geriatrician examination had sensitivity of 1.0 (95% CI, 0.69-1.0) for cataract and 0.80 (95% CI, 0.28-0.99) for AMD compared with ophthalmologist examnation, with no false-positive results, but estimates were imprecise.³⁵ One new study found visual acuity screening using a mobile application associated with sensitivity of 0.98 (95% CI, 0.91-1.00) and specificity of 0.94 (95% CI, 0.82-0.99) for identifying visual acuity 20/40 or less compared with a visual acuity chart.⁵⁷

Key Question 4. What is the accuracy of instruments for identifying patients at higher risk of impaired visual acuity due to uncorrected refractive error, cataracts, or AMD?

Two studies^42,46,47 (n = 1121 and n = 3997) included in the prior USPSTF review and 1 new study³³ (n = 85), all fair quality, found that screening questions were not accurate for identifying older persons with impaired visual acuity compared with an eye chart; all studies reported low sensitivity, low specificity, or both (eTables 6-8 in the Supplement). Sensitivities ranged from 0.17 to 0.81 and specificities from 0.19 to 0.84. Questions included asking about trouble recognizing faces, reading the newspaper, or seeing.

Key Question 5. What are the effects of treatment for wet or dry AMD vs placebo or no treatment on visual acuity, morbidity, mortality, general or vision-related quality of life, functional status, or cognition?

Four good-quality RCTs (n = 2086; reported in 5 publications), all included in the prior USPSTF review, evaluated intravitreal injection with VEGF inhibitors vs sham injection.^{24,27,44,52,53} At 1 year, intravitreal VEGF inhibitors were significantly associated with greater likelihood vs sham of 15 letters or more of visual acuity gain (RR, 2.92 95% CI, 1.20-7.12], I² = 76%; absolute risk difference [ARD], 10%); less than 15 letters of visual acuity loss (RR, 1.46 [95% CI, 1.22-1.75]; I² = 80%; ARD, 27%); and having vision 20/200 or better (RR, 1.47 [95% CI, 1.30-1.66]; I² = 42%; ARD, 24%) (eFigures 1-3 and eTables 9-10 in the Supplement).^24,27,44 In 1 trial,⁵² VEGF inhibitors were significantly associated with better vision-related function and quality-of-life measures vs sham injection at 1 and 2 years. Differences on the National Eye Institute–Vision Function Questionnaire 25 (NEI-VFQ) composite and subscales were about 8 points on a 0 to 100 scale, or above the proposed threshold for a clinically important difference (4 to 6 points).⁶³

The large (n = 3640), good-quality Age-Related Eye Disease Study⁵⁹ (AREDS), included in prior USPSTF reviews,^5,6 remains the key trial on treatment for dry AMD (eTables 11-12 in the Supplement). At 6.3 years, it found an antioxidant plus zinc combination significantly associated with decreased risk of progression to advanced AMD vs placebo (odds ratio [OR], 0.72 [99% CI, 0.52-0.98]).⁵⁹ In patients with more advanced (category 3 or 4) AMD, antioxidants plus zinc were significantly associated with decreased risk of visual acuity loss of 15 lines or more on the ETDRS (OR, 0.73 [99% CI, 0.54-0.99]). Ten-year results⁶⁴ were consistent with 6.3-year results.

An updated (2017) Cochrane systematic review¹⁰ included 19 trials^{15,16,18,22,25,26,28,30-32,36,37,40,49,54-56,59,60} (n = 11 162; 13 trials in the prior [2012] version¹¹) of antioxidant multivitamins, zinc, lutein and zeaxanthin, or vitamin E for dry AMD; results were heavily influenced by AREDS (eTables 13-14 in the Supplement). Besides AREDS, the systematic review included the large (n = 4203) AREDS2 trial,^51,60 which evaluated the AREDS formulation or a variation of it (elimination of beta carotene, lowering of zinc dose, or both), and the Vitamin E, Cataract, and Age-related Maculopathy (VECAT) study (n = 1193).¹⁸ In the other trials, sample sizes ranged from 14 to 433. The review found antioxidant multivitamins significantly associated with decreased risk of progression to late AMD (3 trials, n = 2445; OR, 0.72 [95% CI, 0.58-0.90]; 73% of patients from AREDS) and 3 lines or more visual acuity loss (1 trial [AREDS], n = 1791; OR, 0.77 [95% CI, 0.62-0.96]) vs placebo. Zinc was significantly associated with decreased risk of progression to late AMD vs placebo (3 trials, n = 3790; OR, 0.83 [95% CI, 0.70-0.98]; 96% of patients from AREDS) and decreased risk of 3 lines or more of visual acuity loss that was of borderline statistical significance (2 trials, n = 3791; RR, 0.87 [95% CI, 0.75-1.00]; 96% of patients from AREDS). Lutein and zeaxanthin or vitamin E were associated with little or no effect on risk of AMD progression. Data on effects of multivitamins on vision-related function were limited, with most trials showing no statistically significant differences.^18,25,28,65 AREDS found no differences between antioxidants, zinc, both, or placebo in measures of cognition at 6.9 years.¹³

Two additional fair-quality trials not included in the systematic review^20,29 evaluated an antioxidant combination or α-lipoic acid, but were small (n = 80 and 100) with imprecise estimates, and did not affect the findings of the systematic review (eTables 15-16 in the Supplement).

Key Question 6. What are the effects of newer (aflibercept or brolucizumab-dbll) vs older VEGF inhibitors for wet AMD on visual acuity, morbidity, mortality, general or vision-related quality of life, functional status, or cognition?

Three new good-quality trials (n = 2738; reported in 5 publications) compared aflibercept vs the older VEGF inhibitor ranibizumab (eTables 9-10 in the Supplement).^{17,41,43,45,62} The duration of follow-up ranged from 1 year to 4 years. Aflibercept was noninferior to ranibizumab in likelihood of less than 15 ETDRS letters of visual acuity loss or 15 letters or more of visual acuity gain, and 2 trials (n = 2457) found similar improvements in vison-related function. No trial compared brolucizumab-dbll vs an older VEGF inhibitor.

Key Question 7. What are the harms of treatment for early impaired visual acuity due to wet or dry AMD?

There were no significant differences between VEGF inhibitors vs sham treatment in likelihood of withdrawal due to adverse events (eTables 9-10 in the Supplement). Evidence on the effects of VEGF inhibitors on other harms was limited.⁶ Serious ocular harms were infrequent, and incidence of endophthalmitis (2 trials, n = 1924; RR, 5.49 [95% CI, 0.30-99] and RR, 8.33 [95% CI, 0.50-140]), ocular hemorrhage (1 trial, n = 184; RR, 0.52 [95% CI, 0.08-3.61]), and retinal detachment (2 trials, n = 1924; RR, 0.17 [95% CI, 0.01-4.07], and RR, 3.67 [95% CI, 0.20-65]) were similar in VEGF and sham treatment groups.^6,24,27,44 The studies were not sufficiently powered to assess rates of cardiovascular events or other serious adverse events, although no statistically significant differences were reported.^24,27,44,66

Three trials (n = 2738; reported in 2 publications) found that serious ocular adverse events and cardiovascular events were infrequent and occurred in similar proportions of patients randomized to aflibercept or ranibizumab (eTables 9-10 in the Supplement).^43,62

AREDS found zinc use associated with increased risk of hospitalization due to genitourinary causes vs nonuse (7.5% vs 4.9%; RR, 1.47 [95% CI, 1.19-1.80])³⁸ and antioxidant use significantly associated with increased risk of yellow skin vs nonuse (8.3% vs 6.0%; RR, 1.38 [95% CI, 1.09-1.75]).⁵⁹ No active treatment in AREDS (antioxidants, zinc, or both) was associated with increased risk of other serious adverse events, which were uncommon (eTable 17 in the Supplement). In AREDS2, there were no differences between AREDS formulation variations and risk of serious adverse events.⁶⁰ However, in an analysis in which current smokers were excluded, the AREDS formulation with beta carotene was significantly associated with increased risk of lung cancer vs without beta carotene (2.0% vs 0.9%, P = .04). Almost all (91%) of the lung cancers occurred in former smokers.

VECAT (n = 1193), the largest trial other than AREDS and AREDS2, reported no serious adverse events with vitamin E or placebo, and no differences in risk of withdrawal due to adverse events or specific adverse events.¹⁸ Evidence on harms from other trials was limited because of suboptimal reporting and imprecision but did not indicate increased risk of serious adverse events or withdrawal due to adverse events.

This report evaluated evidence regarding screening for impaired visual acuity in older adults; the findings are summarized in Table 2. As in the prior review for the USPSTF, direct evidence on screening older adults for impaired visual acuity in primary care settings vs no screening, delayed screening, or usual care found no benefits on vision-related or other outcomes.^{19,23,34,47,61} Potential reasons for lack of benefit in the screening trials may include high attrition, use of suboptimal screening interventions, low uptake of recommended interventions, or high rates of antecedent eye professional care. Recent reviews of vision screening in older adults in broader (eg, community and home-based) settings^67,68 also found no differences between screening vs no screening in vision or vision-related outcomes, even though they included a number of trials that did not meet inclusion criteria for this report because they did not evaluate the vision screening component separately or screening was conducted by an eye specialist and was not primary care feasible.

Conclusions regarding the suboptimal diagnostic accuracy of vision screening tests for identifying conditions associated with impaired visual acuity in primary care settings are also unchanged from the prior review for the USPSTF. No screening question is comparable in accuracy to tests of visual acuity for identifying impaired visual acuity,^42,46,69-71 and visual acuity testing with a chart is inaccurate for identifying visual conditions identified on a comprehensive ophthalmological examination. However, it is not known whether identification of cataracts or AMD prior to the development of impaired visual acuity is associated with improved clinical outcomes compared with identification after the development of mildly impaired visual acuity. Data on other screening tests was limited or indicated suboptimal performance.^21,39,57 There remains insufficient evidence to assess the accuracy or utility of pinhole testing, the Amsler grid, visual acuity tests other than the Snellen or ETDRS, physical examination, or funduscopic examination performed in primary care settings.

As in the prior review for the USPSTF, strong evidence supports the effectiveness of treatments for common causes of impaired visual acuity. The USPSTF previously determined that a very high proportion of patients experience favorable vision-related outcomes and improvement in vision-related quality of life following treatment for impaired visual acuity due to refractive error and cataracts; therefore, this evidence was not rereviewed for this update.⁷² For dry AMD, evidence showing the effectiveness of antioxidant vitamins and minerals for slowing progression of disease or improving visual acuity remains largely based on the large AREDS trials, which included extended (10-year) follow-up.^49,59,73 Based on AREDS2 and other evidence⁷⁴ indicating an association between use of beta carotene and increased risk of lung cancer in smokers, recommendations⁷⁵ for current and former smokers are to avoid the AREDS formula with beta carotene, using lutein and zeaxanthin in its place. For wet AMD, this update focused on VEGF inhibitors, which are first-line treatment in most patients. As in the prior review for the USPSTF, VEGF inhibitors were associated with improvement in visual acuity–related outcomes, with a relatively low incidence of serious harms, although data on effects on vision-related quality of life or function are limited and inconclusive. One area of concern with VEGF inhibitors has been a potential association with increased risk of cardiovascular events.⁷⁶ Although randomized trials of VEGF inhibitors for AMD did not report increased risk of cardiovascular events, they were not designed to evaluate these outcomes and the number of events were small. Although new sham-controlled trials of VEGF inhibitors were not identified, head-to-head trials^43,77 of the recently approved US Food and Drug Administration (FDA)–approved VEGF inhibitor aflibercept vs an older VEGF inhibitor indicated similar effects on visual acuity–related outcomes and no difference in serious harms. No trial of the recently FDA-approved VEGF inhibitor brolucizumab-dbll met inclusion criteria. However, in May 2021, several ongoing brolucizumab-dbll trials were discontinued because of higher rates of intraocular inflammation, including retinal vasculitis and retinal vascular occlusion.⁷⁸

This evidence review has several limitations. First, a previously published systematic review¹⁰ on antioxidant multivitamins and minerals for dry AMD was used. The reliability of systematic reviews depends on how well they are designed and conducted. Therefore, the systematic review was required to meet a quality threshold based on predefined criteria,⁷⁹ and data abstraction and quality assessment of included trials was independently verified. Second, evidence on effectiveness of treatment for dry AMD relied heavily on results of a single trial—the large, well-conducted AREDS trial.⁵⁹ Third, non–English–language studies were excluded, which could introduce language bias. However, no relevant non–English-language studies that appeared likely to affect conclusions were identified. Fourth, there were too few randomized trials to perform formal assessments for publication bias with graphical or statistical methods for small sample effects. However, unpublished trials likely to affect findings were not identified. Fifth, there was statistical heterogeneity in some pooled analyses of VEGF inhibitors vs sham. However, inconsistency was in the magnitude of benefit, not direction of effect, which consistently favored VEGF inhibitors. In addition, because of anticipated heterogeneity, a random-effects model was used for pooling. Sixth, trials of screening vs no screening had methodological limitations, including high attrition and use of a suboptimal screening test. In some trials, low uptake of recommended interventions or a high rate of eye specialist care prior to screening could have attenuated potential benefits. In addition, the screening trials were published between 1997 and 2006, potentially reducing applicability to current clinical practice.

This review found that effective treatments are available for common causes of impaired visual acuity in older adults. However, direct evidence found no significant association between vision screening vs no screening in primary care and improved visual outcomes.

Corresponding Author: Roger Chou, MD, Oregon Health & Science University, 3181 SW Sam Jackson Park Rd, Mail Code: BICC, Portland, OR 97239 (chour@ohsu.edu).

Accepted for Publication: April 5, 2022.

Published Online: May 24, 2022. doi:10.1001/jama.2022.6381

Correction: This article was corrected on August 9, 2022, for an incorrect reference (reference 10).

Author Contributions: Dr Chou had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data.

Concept and design: Chou, Jonas.

Acquisition, analysis, or interpretation of data: Chou, Bougatsos, Jungbauer, Grusing, Blazina, Selph, Tehrani.

Drafting of the manuscript: Chou, Bougatsos, Jungbauer, Grusing, Blazina.

Critical revision of the manuscript for important intellectual content: Chou, Blazina, Selph, Jonas, Tehrani.

Statistical analysis: Chou, Blazina.

Obtained funding: Chou, Bougatsos, Jonas.

Administrative, technical, or material support: Bougatsos, Jungbauer, Grusing, Blazina, Jonas, Tehrani.

Supervision: Chou, Bougatsos, Jonas, Tehrani.

Conflict of Interest Disclosures: None reported.

Funding/Support: This research was funded under contract HHSA-290-2015-00011-I, Task Order 75Q80119F32015, from the Agency for Healthcare Research and Quality (AHRQ), US Department of Health and Human Services, under a contract to support the US Preventive Services Task Force (USPSTF).

Role of the Funder/Sponsor: Investigators worked with USPSTF members and AHRQ staff to develop the scope, analytic framework, and key questions for this review. AHRQ had no role in study selection, quality assessment, or synthesis. AHRQ staff provided project oversight, reviewed the report to ensure that the analysis met methodological standards, and distributed the draft for peer review. Otherwise, AHRQ had no role in the conduct of the study; collection, management, analysis, and interpretation of the data; and preparation, review, or approval of the manuscript findings. The opinions expressed in this document are those of the authors and do not reflect the official position of AHRQ or the US Department of Health and Human Services.

Additional Contributions: We thank the following individuals for their contributions to this project: Pacific Northwest Evidence-based Practice Center Librarian, Tracy Dana, MLS; Agency for Healthcare Research and Quality Medical Officer, Justin Mills, MD, MPH; as well as the US Preventive Services Task Force. We also acknowledge past and current USPSTF members who contributed to topic deliberations. The USPSTF members, external reviewers, and federal partner reviewers did not receive financial compensation for their contributions.

Additional Information: A draft version of this evidence report underwent external peer review from 4 content experts (April Maa, MD, Emory University School of Medicine, Emory Eye Center; Atlanta VA Medical Center; Nancy Weintraub, MD, David Geffen School of Medicine at University of California at Los Angeles; Jennifer Evans, PhD, MSc, London School of Hygiene and Tropical Medicine; and 1 nondisclosed reviewer) and federal partners representing the Centers for Disease Control and Prevention. Comments were presented to the USPSTF during its deliberation of the evidence and were considered in preparing the final evidence report.

Editorial Disclaimer: This evidence report is presented as a document in support of the accompanying USPSTF Recommendation Statement. It did not undergo additional peer review after submission to JAMA.