Outcomes of the Veterans Affairs Low Vision Intervention Trial II (LOVIT II): A Randomized Clinical Trial

Key Points

Question  Are low-vision devices with low-vision rehabilitation (including therapy and homework to teach device use, eccentric viewing, and environmental modification) more effective than basic low-vision services (low-vision devices dispensed without therapy) for veterans with macular diseases and visual acuity of 20/50 to 20/200?

Findings  In a multicenter randomized clinical trial, both treatments were found to be effective, but low-vision rehabilitation was more effective than basic low-vision services only for patients with visual acuity worse than 20/63 to 20/200.

Meaning  Basic low-vision services are sufficient for most patients with low vision who have mild visual impairment.

Importance  Randomized clinical trials are needed to compare effectiveness and cost-effectiveness of different low-vision (LV) programs.

Objective  To determine the value of adding LV rehabilitation with a therapist compared with LV services without intervention.

Design, Setting, and Participants  A randomized clinical trial was conducted from September 27, 2010, to July 31, 2014, of 323 veterans with macular diseases and best-corrected distance visual acuity (BCDVAbetter-eye) of 20/50 to 20/200. Masked interviewers administered questionnaires by telephone before and after LV treatment. Using an intention-to-treat design, participants were randomized to receive LV devices with no therapy or LV devices with a rehabilitation therapist providing instruction and homework on the use of LV devices, eccentric viewing, and environmental modification. Visual ability was measured in dimensionless log odds units (logits) (0.14-logit change in visual ability corresponds to ability change expected from a 1-line change in visual acuity).

Interventions  Low-vision devices without therapy and LV devices with therapy.

Main Outcomes and Measures  Comparison of changes (baseline to 4 months) in overall visual ability and in 4 functional domains (reading, visual information, visual motor, and mobility) estimated from responses to the Veterans Affairs Low Vision Visual Functioning Questionnaire (higher scores indicates more ability or less difficulty in performing activities), and comparison of MNREAD changes (baseline to end of treatment) in maximum reading speed, critical print size, and reading acuity (higher number indicates lower visual acuity).

Results  Of the 323 participants, 314 were male (97.2%); mean (SD) age, 80 (10.5) years. Basic LV was effective in improving visual ability. However, the LV rehabilitation group improved more in all visual function domains except mobility. Differences were 0.34-logit reading (95% CI, 0.0005 to 0.69; P = .05), 0.27-logit visual information (95% CI, 0.01 to 0.53; P = .04), 0.37-logit visual motor (95% CI, 0.08 to 0.66; P = .01), and 0.27-logit overall (95% CI, 0.06 to 0.49; P = .01). For MNREAD measures, there was more improvement in reading acuity (difference, −0.11 logMAR, 95% CI, −0.15 to −0.07; P < .001) and maximum reading speed (mean increase of 21.0 words/min; 95% CI, 6.4 to 35.5; P = .005), but not critical print size for the LV rehabilitation group (−0.06 logMAR; 95% CI, −0.12 to 0.002; P = .06). In stratified analyses, the LV rehabilitation group with BCDVAbetter-eye worse than 20/63 to 20/200 improved more in visual ability (reading, visual motor, and overall). Differences were 0.56-logit reading ability (95% CI, 0.08-1.04; P = .02), 0.40-logit visual motor (95% CI, 0.03-0.78; P = .04), 0.34-logit overall (95% CI, 0.06-0.62; P = .02). There was no significant difference between treatment groups for those with BCDVAbetter-eye of 20/50 to 20/63.

Conclusions and Relevance  Both basic LV alone and combined with LV rehabilitation were effective, but the added LV rehabilitation increased the effect only for patients with BCDVAbetter-eye worse than 20/63 to 20/200. Basic LV services may be sufficient for most LV patients with mild visual impairment.

Trial Registration  clinicaltrials.gov Identifier: NCT00958360

Quiz Ref IDLow vision (LV) is defined as any chronic, uncorrectable visual impairment that affects daily life.¹ Low vision interferes with performance of activities such as reading, mobility, recognizing faces, and interacting with family and friends.^2-7 Persons with LV can also experience loss of self-esteem⁸ and personal independence,⁹ as well as decreased quality of life¹⁰ accompanied by a decline in general health¹¹ and an increased risk of depression,^12-14 injury,^15-17 and mortality.¹⁸

Low vision programs include a variety of devices and therapies to improve patients’ performance of tasks limited by visual impairment.² Despite the wide variation in the range and intensity of services provided, different LV programs may demonstrate successful outcomes based on the specific services they provide.² In a systematic literature review, Binns et al³ concluded that there is good evidence that LV devices improve reading ability and are valued by patients, that rehabilitation programs provided by the Veterans Affairs (VA) have a large positive effect, and that other rehabilitation programs have a medium to large effect in improving functional ability.

Veterans Affairs services for blind and visually impaired veterans include comprehensive inpatient rehabilitation programs, as well as multidisciplinary and basic LV programs.¹⁹ Randomized clinical trials are needed to compare the effectiveness and cost-effectiveness of LV programs, guide policy, and identify individuals who benefit most from different services.²⁰ The Veterans Affairs Low Vision Intervention Trial (LOVIT) evaluated an intense outpatient LV rehabilitation program for legally blind veterans with macular diseases (habitual distance visual acuity better-eye worse than 20/100 to 20/400) compared with a waiting-list control group.²¹ LOVIT II complemented LOVIT by comparing the outcomes of 2 types of LV programs for veterans less severely visually impaired from macular diseases (best-corrected distance visual acuity better-eye [BCDVAbetter-eye] 20/50-20/200).²⁰

The inclusion criteria were eligible for VA benefits, diagnosis of any macular disease, and BCDVAbetter-eye of 20/50 to 20/200. Exclusion criteria were no access to telephone, less than fifth grade level achieved on the Dolch English Literacy Test,²² and no vision loss since previous LV rehabilitation, Telephone Interview for Cognitive Status²³ screening score less than 30, unable or unwilling to attend clinic visits, hearing impairment that interferes with telephone questionnaires, visual field better-eye less than 20° in diameter, vitreous hemorrhage affecting line of sight, cataract extraction planned within 4 months, receiving macular disease treatment expected to improve vision,²⁰and participating in another study that does not allow dual enrollment.

LOVIT II was conducted at 9 VA medical facilities. The study rationale and methods have been published.^20,24 The protocol (available in the Supplement) and written informed consent were approved by the VA Central Institutional Review Board. Participants gave written informed consent after the purpose and procedures of the trial were explained, and financial compensation was provided. Study oversight was provided by an independent data and safety monitoring committee and the VA Cooperative Studies Program Coordinating Center.

Patients with macular diseases were screened by medical records review for major inclusion and exclusion criteria. Eligible patients received study information from clinical health care professionals or a letter sent by mail. Site coordinators approved patients for enrollment after screening to determine eligibility using the Early Treatment of Diabetic Retinopathy Study visual acuity chart,²⁵ Dolch English Literacy Test words,²² and the Telephone Interview for Cognitive Status.²³

Low vision devices were prescribed based on standard of care after a LV examination.²⁶ Contrast sensitivity was measured with the Pelli-Robson Contrast Sensitivity Test.²⁷ Central and juxtafixational scotomas (defined as 4 contiguous points not seen) were measured with the Johns Hopkins University and Erickson Visual Field Test.²⁸ The MNREAD test²⁹ was administered at baseline. On the test, a higher number for reading acuity indicates lower visual acuity; maximum reading speed is patient’s reading speed when reading is not limited by print size (units are words/min), and critical print size is the smallest print size the patient can read with their maximum reading speed at 20 cm with +5.00 diopters adjusted for nonstandard viewing distances; a higher number indicates larger critical print size; positive changes from baseline indicate worsening and negative changes indicate improvement. The VA Low Vision Visual Functioning Questionnaire (VA LV VFQ-48),^30-34 Short Form-36,³⁵ and EuroQol-5D³⁶ were administered by telephone before randomization. A higher score on the VA LV VFQ-48 indicates better ability or less difficulty in performing activities, Higher scores on both the Short Form-36 and EuroQol-5D indicate better quality of life.

Quiz Ref IDEligible and consenting patients were assigned randomly to the treatment groups. In basic LV services, the optometrist dispensed LV devices without therapy or assigned homework. In LV rehabilitation, patients received basic LV services plus 1 to 3 therapy sessions including instruction in eccentric viewing, use of LV devices (at near, intermediate, and far distances), environmental modification, integration of LV devices into lifestyle, and assigned homework to practice using LV devices for everyday tasks.

Changes in visual ability and quality of life were assessed by telephone 4 months from baseline with the VA LV VFQ-48, Short Form-36, and EuroQol-5D. Changes in MNREAD measures were assessed after treatment.

The coordinating center created a computer-generated permutated block randomization with random block sizes. Study site coordinators received assignments from the online randomization system and informed patients and clinical staff of the treatment assignments. Preplanned stratification was by participating site and BCDVAbetter-eye (20/50 to 20/63 and worse than 20/63 to 20/200).

Interviewers who were certified and masked read a script to inform participants that questionnaire responses and treatment assignments were anonymous and confidential. Outcomes data were not shared with investigators or clinical staff until the study concluded.

The primary outcome measure was comparison of the changes in reading ability (measured with responses to 10 items in the 48-item VA LV VFQ) at baseline compared with 4 months later in the treatment groups. The questionnaire was validated previously in LV populations.^30,32,33,37 Patients rated their difficulty performing 48 daily activities using ordered response categories or they responded that they do not perform the activity for nonvisual reasons. Comparison of changes between groups for overall visual ability (from responses to all 48 items) and the other VA LV VFQ-48 visual ability domains (mobility, visual information processing, and visual motor skills from responses to different subsets of items) from baseline to 4 months, and changes in MNREAD measures of maximum reading speed, critical print size (smallest print that can be read at the maximum speed), and reading acuity from baseline to completion of treatment were secondary outcomes.

A 0.35-treatment effect, 5% type I error, and 85% power were selected. This effect size corresponds to a clinically significant change of 2.5 lines due to the strong linear trend between visual ability person measures (measured in dimensionless log odds units [logits]) and visual acuity (in logMAR units).²⁰ A 0.14-logit visual ability change corresponds to the ability change expected from a 1-line change in visual acuity.³⁸ With a 2-sided t test for 2 independent groups, a sample size of 300 (150 per group) was calculated.²⁰ A 10% withdrawal rate was estimated based on previous studies, yielding a sample size of 330 patients (165 per group). Statistical guidelines for early stopping were not used because both groups received treatment, the study duration was short, and the interventions were low risk. The data and safety monitoring committee reviewed the interim progress report biannually.

Rasch analyses³⁹ of responses to the 48 items and different subsets of items were used to estimate linear item and person measures in logits (with the origin arbitrarily set to the mean of the 48-item measures) for each of the 4 functional domains and overall visual ability for each participant. Comparisons of visual function person measures and subgroup analyses based on stratification were analyzed according to the intention-to-treat principle. The 2-sample t test was used to compare the differences in the primary and secondary outcomes between the treatment groups. Cohen d was used to calculate the magnitude of treatment effects as small (0.2), medium (0.5), or large (0.8).⁴⁰ Analysis of covariance was used to compare mean changes in the outcomes between the 2 arms adjusting for age, baseline measures (visual ability, BCDVAbetter-eye, contrast sensitivity, maximum reading speed, critical print size, and reading acuity), presence of visual fluctuations, history of receiving anti–vascular endothelial growth factor (VEGF) injections in the year before the study, and presence of central or juxtafixational scotomas.

The paired t test was used to test within-group changes. The differences between the treatment groups in quality of life from baseline to 4 months and changes in MNREAD measures of maximum reading speed, critical print size, and reading acuity, measured from baseline to completion of treatment, were also analyzed using 2-sample t tests.

Stepwise linear regression models were used to determine whether the mean changes in overall visual ability person measures and person measures for each functional domain from baseline to 4 months can be determined by baseline measures of visual impairment, baseline BCDVAbetter-eye, maximum reading speed, critical print size, reading acuity, presence of scotomas, visual fluctuations, treatment with anti-VEGF injections, age, and treatment group.

All analyses were 2-sided; P ≤ .05 was considered statistically significant. SAS software, version 9.4, was used to perform all analyses.⁴¹

Enrollment began September 27, 2010; accrual was completed July 31, 2014; and follow-up ended February 17, 2015. The Figure describes the flow of participants through the trial. A total of 2051 patients were screened, of whom 1728 were excluded (1706 ineligible per medical records review and 22 ineligible after screening). Randomization was completed for 323 patients, with 163 assigned to LV rehabilitation and 160 assigned to basic LV services. The most frequent reason for exclusion was that the BCDVAbetter-eye did not fall into the required range. Nineteen patients discontinued the study before completion. The actual withdrawal rate was 6%, lower than the 10% withdrawal estimate used for sample size calculation. Based on actual attrition rate, recruitment exceeded the required sample size of 320 patients. Final analyses included all 323 randomized patients.

Baseline characteristics, patients’ health status, changes in visual function (baseline to 4 months), and changes in reading performance measures (baseline to end of treatment) are presented in Tables 1, 2, and 3. Overall, 97.2% of the participants were male and 90.4% were white; mean (SD) age was 80 (10.5) years. The most frequent eye diagnoses (better-seeing eye) were nonexudative age-related macular degeneration (AMD) (LV rehabilitation group, 65 [39.9%]; basic LV group, 70 [43.8%]) and exudative AMD (LV rehabilitation group, 37 [22.7%]; basic LV group, 41 [25.6%]). Anti-VEGF injections were received in the year before the study by 53 patients (32.5%) in the LV rehabilitation group and 60 (37.5%) in the basic LV group. Vision fluctuations were experienced by 50 patients (30.7%) in the LV rehabilitation group and 59 (36.9%) in the basic LV group. Mean BCDVAbetter-eye was 0.6 (0.2) logMAR (20/80 Snellen equivalent) for both groups; mean contrast sensitivity (better-seeing eye) was 1.1 (0.8) for both groups. There were no differences in baseline health status between the groups. Mean MREAD reading performance measures at baseline were reading acuity, 0.9 (0.4) logMAR (20/160 Snellen equivalent) for both groups; maximum reading speed was 109.1 (68.6) words per minute for the LV rehabilitation group and 120.2 (72.2) words per minute for the basic LV group. Critical print size was 1.2 (0.3) logMAR (20/300 Snellen equivalent) for both groups.

The LV rehabilitation group received a mean (SD) of 1.9 (0.3) therapy sessions and completed 9.8 (4.8) homework assignments. Mean therapy time was 234.2 (145.9) minutes. Therapy was longest for eccentric viewing skills at 64.1 (48.8) minutes; the time varied for LV devices, but was longest for portable electronic magnifiers (62.3 [28.2] minutes) and desktop video magnifiers (68.6 [39.2] minutes). Dispensing time for the basic LV group varied based on LV devices prescribed (54.2 [11.1] minutes). There was no difference in the percentages of LV devices prescribed for both groups. The LV rehabilitation group (n = 163) received 56 (34.4%) monocular telescopes, 82 (50.3%) teleloupes, 102 (62.6%) pocket magnifiers, 76 (46.6%) stand magnifiers, 48 (29.4%) intermediate distance devices, 51 (31.3%) reading glasses, 52 (31.9%) desktop electronic magnifiers, 40 (24.5%) portable electronic magnifiers, 95 (58.3%) filters or prescription sunglasses, and 18 (11.0%) filters to control glare indoors. The basic LV services group (n = 160) received 73 (45.6%) monocular telescopes, 90 (56.3%) teleloupes, 122 (76.3%) pocket magnifiers, 91 (56.9%) stand magnifiers, 44 (27.5%) intermediate distance devices, 54 (33.8%) reading glasses, 55 (34.4%) desktop electronic magnifiers, 56 (35.0%) portable electronic magnifiers, 101 (63.1%) filters or prescription sunglasses, and 25 (15.6%) filters to control glare indoors.

Quiz Ref IDTable 2 presents the comparison of the mean changes in primary and secondary outcomes in logits from baseline to 4 months between the treatment groups. A 0.14-logit change in visual ability corresponds to the ability change expected from a 1-line change in visual acuity. Compared with those in the LV group who received basic services, patients in the LV rehabilitation group who received basic LV services plus LV rehabilitation reported greater improvement in visual ability (reading, visual information processing, visual motor skills, and overall). Within groups, improvement was found in all functional domains and overall visual ability in the LV rehabilitation group and for all functional domains and overall visual ability except mobility in the basic LV group. The outcomes comparisons between the treatment groups were not altered after adjusting for all covariates (n = 272).

Table 2 reports subgroup analyses based on the preplanned stratification by BCDVAbetter-eye. Patients with BCDVAbetter-eye worse than 20/63 to 20/200 assigned to the LV rehabilitation group who received basic LV plus LV rehabilitation experienced more improvement in visual ability (reading, visual motor, and overall) than those assigned to basic LV services. There was no difference in outcomes between treatment groups for patients with BCDVAbetter-eye 20/50 to 20/63. Compared with patients with worse visual acuity, those with better BCDVAbetter-eye received fewer anti-VEGF injections in the year before the study (35 [26.5%] vs 78 [40.8%]; P = .005); they had fewer central or juxtafixational scotomas (48 [44.9%] vs 125 [74.0%]; P < .001), higher contrast sensitivity (1.2 [0.8%] vs 1.0 [0.7%]; P = .04), and better MNREAD reading performance measures (reading acuity, 0.65 logMAR vs 1.03 logMAR; P < .001; maximum reading speed, 142.6 vs 94.8 words/min; P < .001; critical print size, 1.09 logMAR vs 1.31 logMAR; P = .03). In addition, patients with better BCDVAbetter-eye were prescribed fewer desktop electronic magnifiers (20 [15.2%] vs 87 [45.5%]; P < .001) or portable video magnifiers (28 [21.2%] vs 68 [35.6%]; P = .004).

Improvements in visual ability domains and overall visual ability were indicated by lower baseline scores (B coefficients from stepwise linear regression models: reading, −0.52; mobility, −0.41; visual information, −0.42; visual motor, −0.32; overall, −0.30; P < .001). Improvement in all domains except visual information processing was indicated by LV rehabilitation group assignment (B coefficients from stepwise linear regression models: reading, 0.36; mobility, 0.25; visual motor, 0.44; overall, 0.29; P < .05).

Table 3 presents the mean changes in MNREAD reading performance measures for all patients. The LV rehabilitation group demonstrated more improvement in reading acuity (P < .001) and maximum reading speed (differences: −0.11 logMAR reading acuity [equivalent to 1 line]; 95% CI, −0.15 to −0.07; P < .001; mean increase of 21.0 words/min in maximum reading speed, 95% CI, 6.4 to 35.5; P = .005). There were no changes in critical print size within or between groups.

Comparison of changes in quality-of-life scores from baseline to 4 months between treatment groups found no differences in Short Form-36 subscale scores for physical functioning, physical role limitations, bodily pain, vitality, social functioning, emotional role limitations, mental health, general health, and physical or mental components. There was also no change in the EuroQol-5D scores from baseline to 4 months between the treatment groups.

A total of 11 adverse events were reported from inception of the study to the 4-month follow-up. None of these adverse events was related to the study intervention.

This RCT demonstrated that both basic LV services and basic LV services plus LV rehabilitation provided for veterans with macular diseases, most of whom were white, male, and covered by Medicare,⁴² had improved visual ability (reading, visual information processing, visual motor skills, and overall) at 4-month follow-up. Basic LV services plus LV rehabilitation also improved mobility. In preplanned stratified analyses, visual ability (reading, visual motor skills, and overall) improved more in the LV rehabilitation group than in the basic LV services group for patients with BCDVAbetter-eye worse than 20/63 to 20/200; there were no differences between treatment groups for those with BCDVAbetter-eye 20/50 to 20/63. There were differences in the number of anti-VEGF injections received in the year before the study, contrast sensitivity, presence of central or juxtafixational scotomas, and LV devices prescribed between the stratified groups. These results led us to conclude that patients with mild LV (20/50 to 20/63) benefit from basic LV services except for mobility but gain no additional benefit from LV rehabilitation, whereas patients with moderate LV (worse than 20/63 to 20/200), also with the exception of mobility, benefit from basic LV services but gain even greater benefit with the addition of LV rehabilitation.

As expected from the relationship of visual ability with visual acuity, patients with better BCDVAbetter-eye had more overall visual ability at baseline than did patients with lower BCDVAbetter-eye. The patients with more visual ability were closer to the measurement ceiling of the VA LV VFQ-48; therefore, they had less room for improvement compared with those with lower BCDVAbetter-eye and less visual ability. Confirming this observation, the stepwise linear regression showed that improvement in all visual ability domains and overall visual ability is indicated by lower baseline scores and with the addition of LV rehabilitation.

LOVIT was a multicenter RCT²¹ that evaluated the effectiveness of an LV program rehabilitation for legally blind veterans with macular diseases similar to but more intense than the LV rehabilitation provided in LOVIT II. LOVIT demonstrated that LV rehabilitation significantly improved the visual ability of veterans compared with patients similarly impaired in the waiting-list control group who lost visual ability during the same 4-month interval.²¹ Patients in the LOVIT treatment group demonstrated more improvement in reading ability at 4-month follow-up than did patients in the LV rehabilitation group in the present study. Consistent with their more severe visual impairment, baseline reading ability for patients in LOVIT was less than that of patients in LOVIT II, so participants in LOVIT had more room for improvement. Another difference is that patients in LOVIT II received less therapy and homework and fewer desktop electronic magnifiers were prescribed compared with patients in LOVIT.

Changes in reading acuity and maximum reading speed from baseline to completion of treatment in the LV rehabilitation group compared with the basic LV services group (Table 3) mirrored changes in self-reported functional reading ability. Instruction in eccentric viewing, word and letter recognition skills, scanning techniques, and homework to practice skills may have contributed to these changes. There were no significant changes in critical print size between groups or within groups.

Three previous studies reported in the literature did not find differences in outcomes between basic and multidisciplinary or enhanced LV service delivery models in the United Kingdom, the Netherlands, or New Zealand.^43-47 It is a challenge to compare these LV effectiveness studies because (1) different outcome measures were used; (2) protocols differed with regard to inclusion criteria for severity of impairment, diagnosis, and follow-up time; and (3) access to LV devices and therapies was based on health services policies that vary among countries.³

Quiz Ref IDBoth LOVIT studies had many strengths: a RCT design, a well-defined treatment protocol guided by therapy and homework manuals that was consistently followed at all sites, the same validated questionnaires used to assess outcomes and health status, and scientific oversight and monitoring provided by a coordinating center and data and safety monitoring committee.^20,21 The LOVIT studies were conducted in the VA system where veterans are eligible for LV services and provided LV devices without charge. The primary weakness is that study results cannot be generalized to the US private sector where Medicare covers LV therapy prescribed by physicians and provided by occupational therapy, but does not cover the cost of LV devices.⁴²

Basic LV services alone or with the addition of LV rehabilitation was effective in this trial, but basic LV services plus LV rehabilitation was more effective than basic LV services alone only for patients with BCDVAbetter-eye worse than 20/63 to 20/200. These findings suggest that basic LV services are sufficient for most patients with LV who have mild visual impairment.

Corresponding Author: Joan Stelmack, OD, MPH, Blind Rehabilitation Center, Edward Hines Jr. Veterans Affairs Hospital (Mail Stop 124), 5000 S Fifth Ave, Hines, IL 60141 (joan.stelmack@va.gov).

Accepted for Publication: October 11, 2016.

Published Online: December 15, 2016. doi:10.1001/jamaophthalmol.2016.4742

Author Contributions: Drs Stelmack and Tang had full access to all the data in the study and take responsibility for the integrity of the data and accuracy of the data analyses.

Study concept and design: Stelmack, Tang, Massof.

Acquisition, analysis, or interpretation of data: All authors.

Drafting of the manuscript: Stelmack, Wei, Wilcox, Brahm, Sayers.

Critical revision of the manuscript for important intellectual content: Stelmack, Tang, Wilcox, Morand, Brahm, Sayers, Massof.

Statistical analysis: Tang, Wei, Morand, Massof.

Administrative, technical, or material support: Stelmack, Tang, Brahm, Sayers.

Conflict of Interest Disclosures: All authors have completed and submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest and none were reported.

Funding/Support: Funding for this research was provided by Department of Veterans Affairs (VA) Rehabilitation Research and Development grant C6958R. Funding for the low-vision devices prescribed and dispensed to veteran participants was provided by the Department of Veterans Affairs Prosthetics Service.

Role of the Funder/Sponsor: The Department of Veteran Affairs had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

Group Information: The LOVIT II Study Group: Chair’s Office: Joan Stelmack OD, MPH, principal investigator (Edward Hines Jr. VA Hospital); Robert W. Massof, PhD, coinvestigator (Wilmer Eye Institute, Johns Hopkins University School of Medicine); Scott Sayers, PhD, national coordinator, interviewer (Edward Hines Jr. VA Hospital); Nancy Ellis, MS,^a national coordinator, interviewer (Clement J. Zablocki VA Medical Center [VAMC]); Stephan Rinne, MS, interviewer, and Timothy Korwin, BS, Interviewer (Edward Hines Jr. VA Hospital). Hines Cooperative Studies Program Coordinating Center: Domenic J. Reda, PhD, director; X. Charlene Tang, MD, PhD, MPH, biostatistician; Kevin T. Stroupe, PhD, health economist; Dan Lippe, MA, project manager; Yongliang Wei, MS, statistical programmer; Kelly Tir, BA, data management programmer; Maria Rachelle, in-house monitor/site contact (data coordinator).

Participating Clinical Sites: Baltimore VAMC–Maryland Health Care System: Rex Ballinger, OD, site investigator, low-vision optometrist; Olga Whitman, OD, assistant site investigator, site coordinator, low-vision optometrist; Chana Hurvitz, MA, low-vision therapist; Sheila Davis, MA,^a low-vision therapist. Cincinnati VAMC: Timothy Morand, OD, site investigator, low-vision optometrist; Mary Colleen Rogge, RN, BSN, site coordinator; Brittany Swedelius, MA,^a low-vision therapist. Dayton VAMC: Timothy Morand, OD, site investigator, low-vision optometrist; Cynthia Thompson, site coordinator; Brian Joos, MS, low-vision therapist. Edward Hines, Jr. VA Hospital: Joan Stelmack, OD, MPH, site investigator, low-vision optometrist; Scott Sayers, PhD, site coordinator; Stephen Rinne, MA, site coordinator; Timothy Korwin, BS, site coordinator; Jack Houston, MSEd, MBA, low-vision therapist. William S. Middleton Memorial Veterans Hospital: Karen Brahm, OD, site coordinator, low-vision optometrist; David LaCrosse, BS, site coordinator; Amy Wurf, MSEd, low-vision therapist. Clement J. Zablocki Veterans Affairs Medical Center: Kenneth Rose, OD, site investigator, low-vision optometrist; Joseph Berman, PT, MHS, site coordinator; Nancy Ellis, MS,^a site coordinator; Claire Seefeldt, OTR, low-vision therapist. Philadelphia VAMC: Denise Wilcox, OD, PhD, site coordinator, low-vision optometrist; Connie Chronister, OD, assistant site coordinator, low-vision optometrist; Rajkaran Sachdej, BS, site coordinator; Janet Meyers, MS, OTR, low-vision therapist. Washington, DC, VAMC: Ellen Kwon, OD, site coordinator, low-vision optometrist; Andrew Pierce, BS,^a site coordinator; LaShandra Holmes-Russell, MS, low-vision therapist; W. G. (Bill) Hefner VA Medical Center: Roger W. Cummings, OD^a site coordinator, low-vision optometrist; Almeda Ruger, MS,^a site coordinator, low-vision therapist; Kimberly B. Gordon, MSN, RN,^a site coordinator; Brandy Carroll, OD, MPH,^a low-vision optometrist; Gary Mancil, OD,^a low-vision optometrist; Philip Roels, OD,^a low-vision optometrist. Consultants: Donald Fletcher, MD (California Pacific Medical Center); Janet Sunness, MD (Greater Baltimore Medical Center); Gislin Dagnelie, PhD (Wilmer Eye Institute, Johns Hopkins University School of Medicine). Data Monitoring Committee: Thomas W. Raasch, OD, PhD, chair, Data Safety and Monitoring Committee (The Ohio State University); Mae O. Gordon, PhD, biostatistician (Washington University School of Medicine); Leslie G. Hyman, PhD (Stony Brook University Medical Center); Patti S. Wimbs-Fuhr, OD, PhD, low-vision optometrist (Birmingham VA Medical Center). ^aFormer participant.