Short- and Long-Term Effects of an Intensive Inpatient Vision Rehabilitation Program

Article Outline

• Abstract
• Methods
  • Instruments
  • Participants
  • Data Collection
  • Analysis
• Results
• Discussion
• Conclusions
• References
• Copyright

Abstract 

Stelmack JA, Moran D’A, Dean D, Massof RW. Short- and long-term effects of an intensive inpatient vision rehabilitation program.

Objective

To assess the effects of a visual rehabilitation program on visually impaired subjects’ visual ability and ability to perform activities.

Design

Prospective observational study.

Setting

Telephone interviews of respondents in their homes the week before admission to the rehabilitation center and 3 months and 1 year after discharge from the rehabilitation center.

Participants

A total of 178 consecutive patients from the Hines Blind Rehabilitation Center participated in development of the 48-item Veterans Affairs Low Vision Visual Functioning Questionnaire (VA LV VFQ-48). Data were analyzed for 95 who participated in all 3 administrations of the questionnaire.

Intervention

Comprehensive blind rehabilitation program (mean hospital admission, 40d).

Main Outcome Measure

The self-report ratings of patients’ difficulty performing 48 activities on the VA LV VFQ-48.

Results

The increase in visual ability ± standard deviation of .981±.482 logits (equivalent to an 8-line improvement in visual acuity on an Early Treatment of Diabetic Retinopathy Study chart) at 3 months postrehabilitation decreased to .682±.485 logits (equivalent to a loss of 2.5 lines of visual acuity on the same chart) 1 year postrehabilitation. The effect sizes measured at 3 months (2.035) and 1 year (1.495) indicate large treatment effects corresponding to statistically significant differences for the increase in visual ability at 3 months and 1 year postrehabilitation (paired 2-tailed t tests, P<.001) relative to pretreatment measures. The difference in visual abilities measured at 3 months and 1 year posttreatment also is statistically significant (P<.001).

Conclusions

Treatment effects decreased over the 12-month follow-up period. However, the group of patients whose data were analyzed was still statistically and clinically significantly better at their 1-year follow-up than before beginning treatment.

Key Words: Questionnaires, Rehabilitation, Vision, low

VISUAL IMPAIRMENT, that is, “chronic visual deficiencies that impair everyday function that are not correctable by ordinary glasses,” is included among the 10 most prevalent causes of disability in America.¹ People confronted with vision loss often feel frustrated performing everyday activities and may experience loss of self-esteem, social isolation, difficulty working, and reduced independence.¹ Vision rehabilitation has the potential to restore functional abilities and lower societal costs by reducing the need for services and increasing independence and quality of life. Although some clinical trials are underway, evidence from multicenter randomized clinical trials evaluating treatment strategies and their cost-effectiveness is not yet available. Reports from observational studies and case series confirm the efficacy of vision rehabilitation, but success rates reported vary from 23% to 97%.² These studies cannot be compared because different criteria were used to establish benefit, and a paucity of information was reported on the actual treatment provided in most of the studies.²

Comprehensive vision rehabilitation programs are offered in the 10 Veterans Affairs (VA) Blind Rehabilitation Centers (BRCs). The VA BRC teams consist of a physician, an optometrist, a psychologist, a social worker, nurses, and blind rehabilitation therapists. BRCs offer a variety of skill courses (manual skills, visual skills, daily living skills, adaptive computer training, orientation and mobility) that teach veterans to more effectively use their remaining senses, adjust to disability, and return to a contributing place at home, work, and in the community.³ Therapy includes vision substitution (using other senses and nonvisual devices, eg, use of touch and hearing while traveling with a white cane) as well as vision enhancement (strategies and devices to more effectively use remaining vision, eg, reading with magnifiers). A psychologist and social worker provide individual and group therapy. Veterans also participate in recreational activities and socialize at events sponsored by local service organizations. The mean length of stay at the Hines BRC is currently 40 days.

Two national studies evaluated outcomes of VA blind rehabilitation programs. In a 12- to 24-month follow-up of veterans discharged from BRCs and Visual Impairment to Optimize Remaining Sight programs, Watson et al⁴ found that 85% of veterans continued to use their prescribed low vision devices 12 to 24 months after the intervention. Watson et al⁵ reported that closed-circuit television-viewing systems had the lowest abandonment rate among low-vision reading devices. De l’Aune et al⁶ reported that more than 90% of 2500 veterans surveyed were either satisfied or completely satisfied with VA BRC programs. Six weeks after completing VA blind rehabilitation programs, veterans reported large gains in functional ability including reading mail daily or weekly and crossing a street with a traffic light. In a study by Stelmack et al,⁷ veterans who completed the Hines BRC program reported fewer difficulties performing 7 items from the National Eye Institute Visual Functioning Questionnaire−25 (NEI VFQ-25) near and distance vision subscales (reading ordinary print in newspapers; going out to see movies, plays, or sports events; reading small print in a telephone book, on a medicine bottle, or a legal form; figuring out whether bills you receive are accurate; see well up close; reading street signs or the names of stores; seeing and enjoying programs on television) after rehabilitation.

These studies measured outcomes by using a posttest only design or a pretest and 1 posttest design. The question whether the effects of VA blind rehabilitation treatment persist over time has not been addressed. The objective of this study is to compare visual ability, patients’ ability to perform activities modulated by visual impairment, before rehabilitation at the Hines BRC to 2 postrehabilitation time points: 3 months and 1 year after rehabilitation.

Back to Article Outline

Methods 

Instruments 

The 48-item Veterans Affairs Low Vision Visual Functioning Questionnaire (VA LV VFQ-48) was used for this study rather than the NEI VFQ-25 because the VA LV VFQ-48 has a larger set of 48 items that are addressed and affected positively by rehabilitation strategies. Development of the VA LV VFQ-48 and its validation were described in prior publications.⁸, ⁹, ¹⁰, ¹¹

The VA LV VFQ-48 includes 4 questions that are administered orally. Question 1 (“Is it difficult to …?”) is asked about all 48 items. Response choices include the following: not difficult, slightly difficult, moderately difficult, extremely difficult, impossible, and do not do it for nonvisual reasons (which is scored as missing data). Question 1 is the subject of this study.⁸, ⁹, ¹⁰, ¹¹

Participants 

One hundred seventy-eight consecutive patients from the Hines BRC, who participated in development of the VA LV VFQ-48, were recruited, and data were analyzed for 95 participants who participated in all 3 administrations of the questionnaire. The 83 participants who did not complete all 3 questionnaires were not included in the analysis. These subjects were less than 1 year postdischarge when funding for the study ended (40 subjects); were lost to follow-up (15 subjects); experienced major depression, cognitive loss, or terminal illness during the rehabilitation program and were excluded based on the recommendations of the staff psychologist or nurse practitioner (12 subjects); did not complete their rehabilitation program (8 subjects); declined to participate in additional questionnaires (4 subjects); or were deceased (4 subjects).

The mean age of subjects ± standard deviation (SD) included in the analysis was 69.4±11.6 years (age range, 42−87y; median, 74y). The mean age of the subjects who were not included was 74.6±9.3 years (age range, 48−89y; median, 78.0y). Habitual distance visual acuity of subjects included in the analysis was a mean of 1.03±0.27 logarithm of minimal angle of resolution (log MAR) (range, 0.40−1.60 log MAR; median, 1.00 log MAR). The 1.03 log MAR corresponds to a 20/200 Snellen equivalent. Four subjects had no light perception, and 12 subjects showed severely constricted visual fields with visual acuity better than 20/200 Snellen. Habitual visual acuity of subjects not included in the analysis was 1.07±0.24 log MAR (range, 0.30−1.6 log MAR; median, 1.08 log MAR). One subject had no light perception, and 2 subjects had severely constricted visual fields. The presenting visual acuity of subjects’ better-seeing eye for both groups is presented in figure 1.

• 
  • View Large Image
  • Download to PowerPoint

• Fig 1. 

  Histograms of the visual acuity of subjects’ better eye at entry into the vision rehabilitation program.

Primary eye diagnosis (better-seeing eye) of the subjects was 61% macular degenerations, 3% other retinal degenerations, 13% diabetic retinopathy, 12% glaucoma, 5% neurologic disorders, 1% cataract, 1% trauma, 1% retinal detachment, and 3% other miscellaneous diagnoses. Patients with active major depression, cognitive loss, terminal illness, and other serious health conditions were identified by the nurse practitioner and staff psychologist during their clinical evaluations. Patients with these conditions were excluded from further participation in the study based on the recommendations of these staff members. The study was conducted in compliance with the tenets of the Declaration of Helsinki for research in human subjects.¹² The study was approved by the institutional review board at Hines Hospital.

Data Collection 

Three interviewers, who were not rehabilitation service providers, administered the VA LV VFQ-48 by telephone before veterans were admitted to the rehabilitation center and again at 3 months and 1 year after discharge. The administration time varied from 25 to 35 minutes. Visual acuity was not measured at follow-up. The interviewers asked each patient if loss of vision since their admission to the rehabilitation program prevented the use of low-vision devices.

Analysis 

Frequently, visual function rating scale questionnaires are scored by linearly rescaling the average response rank across items. At best, this method of scoring could generate values that are monotonic with visual ability, but necessarily the relationship will be nonlinear. The nonlinearities arise because the first and last response categories are unbounded. Also, most visual function questionnaires, the VA LV VFQ-48 included, permit respondents to skip items or respond that they do not do the activity described by an item for reasons other than vision. Such responses result in missing data. Scales based on average response rank scores across items are instrument and sample specific; therefore, missing data can distort the scale. Because of the limitations and inaccuracies of traditional scoring methods,¹³ we have used Rasch models to estimate visual ability in logits (log odds units) for each subject from his/her responses to the VA LV VFQ-48.¹⁰, ¹¹ Visual ability estimates from Rasch models are on an interval scale and are not distorted by missing data.

Theoretically, each activity on the questionnaire requires some threshold level of visual ability.¹³, ¹⁴ Each patient possesses a level of visual ability, the ability to perform daily activities that is modulated by the type and extent of visual impairment. Activities are easy to perform when the patient’s visual ability exceeds that needed, hard to perform when the patient’s visual ability is only slightly greater than that needed, and impossible to perform when the patient’s visual ability is less than the amount needed. The difficulty ratings from the VA LV VFQ-48 are the patients’ self-report of the difference between their visual ability and the visual ability necessary to perform the activity.¹⁴ The effectiveness of the intervention is defined by the increase in visual ability for performing activities after rehabilitation.

Rasch analysis with the Andrich rating scale model was performed on the self-report ratings by using Winsteps.¹⁵, ^16a Rasch analysis provides estimates of visual ability for each person; required visual ability for each item; separation reliabilities for person and item measure distributions; response category thresholds; and fit statistics for person, item, and response threshold estimates.¹⁴, ¹⁷ As performed in our earlier validation study of the VA LV VFQ-48,¹¹ changes in visual ability over time were evaluated by treating the pre- and postrehabilitation ratings in the data matrix as if there were 3 different groups of respondents. This technique is called stacking.¹¹ The change in visual ability from prerehabilitation to 3 months postrehabilitation, the change in visual ability from prerehabilitation to 1 year postrehabilitation, and the change in visual ability from 3 months post- to 1 year postrehabilitation were compared. Normally, one would be concerned with local dependencies when stacking data collected from the same subjects at different time points. But in the case of vision rehabilitation, the effects of intervention can be item specific. For example, a magnifier might make reading items easier to perform but have no impact on mobility items. Therefore, to force all effects of intervention to appear in estimates of visual ability, we used the stacking technique to constrain the item measures to be the same for all time points. To evaluate dependencies in the data, principal components analysis was performed on the residual errors across items for each person-time and across person-time for each item. The effect size is a measure of treatment magnitude that is not dependent on sample size.¹⁸ The effect size¹⁸, ¹⁹ is the change in mean visual ability from time 1 to time 2 (μ₁−μ₂) divided by the SD of the change in visual ability, that is,

where ES is the effect size, σ₁ and σ₂ are the SDs of the distributions of person measures at time 1 and time 2, respectively, and r₁₂ is the product-moment correlation between-person measures for the 2 different times of measurement. The paired t test was used to test whether the differences between the means were different from 0 with statistical significance defined as α less than .05 with correction for multiple comparisons.

Back to Article Outline

Results 

One hundred seventy-eight consecutive subjects from the Hines BRC were recruited from an earlier validation study.¹¹ The prerehabilitation and 3-month postrehabilitation data for the 95 used in the present study are a subset of this data. The 83 who did not participate in all 3 administrations of the questionnaires were not included. Mean visual acuity, Snellen equivalent 20/200, was the same in the 2 groups. Those who did not complete all 3 questionnaires were, on average, 5 years older (mean age, 74.6y) than those who were included in the analysis.

For the 95 who were included in the analysis, the mean time from discharge to administration of the postrehabilitation questionnaires was 4.2 months (mode, 3.5mo) and 15.4 months (mode, 12.0mo).

The measures of required ability for each item estimated from the stacked data analyzed for the present study, which also included the 1-year postrehabilitation data, were compared with measures of required visual ability for each item estimated from the stacked prerehabilitation and 3-month postrehabilitation data analyzed in the earlier study.¹¹ The results are in strong agreement with an intraclass correlation coefficient of .93. Fit statistics and separation reliabilities for both person and item-measure estimates also were in strong agreement with those reported earlier. Thus, adding the 1-year postrehabilitation data to the analysis did not change the instrument calibration.

The estimated measure accounts for 87% of the person variance and 95% of the item variance. The first principal component of the residuals accounts for 14% of the variance in the person measure residual error and 9% of the item measure residual error. The eigenvalues of all principal components are in the scree, suggesting random errors. The correlations of residuals between time points for each person were low when comparing prerehabilitation to 6-month postrehabilitation residual errors (average r=.03) and when comparing prerehabilitation to 1-year postrehabilitation residual errors (average r=.25). As might be expected, because there was no intervention between the follow-up measures, the correlation between factor loadings was much stronger when comparing 6-month and 1-year postrehabilitation residuals (average r=.66). These results combined with the high intraclass correlation coefficient on item measure estimates suggest that local dependencies from repeated measures had little influence on estimates from the Rasch analysis.

Statistically significant increases in visual ability were reported from prerehabilitation to 3 months postrehabilitation (.981 logit, which is equivalent to an 8-line improvement in visual acuity on the Early Treatment of Diabetic Retinopathy Study [ETDRS] chart; P<.001) and from prerehabilitation to 1 year postrehabilitation (.682 logit, which is equivalent to a 5.5-line improvement in visual acuity on the ETDRS chart; P<.001). A significant decrease in visual ability occurred from 3 months postrehabilitation to 1 year postrehabilitation (.299 logit, which is equivalent to a loss of 2.5 lines of visual acuity on the ETDRS chart; P<.001).

The magnitude of the effect size for the increase in visual ability of subjects from prerehabilitation to 3 months postrehabilitation was 2.035, and the change from prerehabilitation to 1 year postrehabilitation was 1.405. The effect size of the change from 3 months postrehabilitation to 1 year postrehabilitation was 0.311. Boxplots of patients’ visual ability for the 3 time points of measurement are plotted in figure 2.

• 
  • View Large Image
  • Download to PowerPoint

• Fig 2. 

  The boxplot shows the distribution of patients’ visual ability in logits prerehabilitation and at 3 months and 1 year postrehabilitation. The mean visual ability is set at the origin. Visual abilities above the mean are positive, and visual abilities below the mean are negative. In the boxes, the lower line is the 25th percentile, the central line is the median, and the top line is the 75th percentile. Ninety-five percent of the data is included in the lines extending from either side of the central boxes. *Outlying points.

There is considerable overlap between the distribution of visual ability at 3 months postrehabilitation and the distribution of visual ability at 1 year postrehabilitation but minimal overlap between the distribution of visual ability prerehabilitation and the distribution of visual ability either at 3 months postrehabilitation or 1 year postrehabilitation.

The scatterplot in figure 3 shows that some patients (those above the solid line) had greater visual ability after 1 year, whereas other patients (those below the solid line) experienced a decline in visual ability from 3 months to 1 year postrehabilitation. Many patients (those on the solid line) showed no change in visual ability from 3 months to 1 year.

• 
  • View Large Image
  • Download to PowerPoint

• Fig 3. 

  The scatterplot of visual ability, ranging from a low of −1 to a high of 3 logits, shows that some patients benefit more from the rehabilitation program than others and some patients lose more visual ability at 1 year. The solid line is the line of equality where there is no change in visual ability from 3 months to 1 year postrehabilitation. Legend: ----, ±3 standard errors for significant gains (points above the solid line) or losses (points below the solid identity line).

Vision loss severe enough to prevent continued use of prescribed low-vision devices was reported by 16% of veterans at 3 months postrehabilitation and by 21% of veterans at 1 year postrehabilitation.

Back to Article Outline

Discussion 

This study was conducted to determine if the increase in visual ability measured 3 months after vision rehabilitation was maintained after 1 year. Results indicated that a large treatment effect continued 1 year after vision rehabilitation, even though there was a small but statistically significant decline in visual ability from 3 months to 1 year postrehabilitation. Some patients benefited more from the rehabilitation program than others, and some patients lost more visual ability over time than others.

One question that is raised by this study is what is the appropriate timeframe for administration of the self-report questionnaires that will be used to measure outcomes in vision rehabilitation clinical trials? The 3-month and 1-year assessment points were arbitrarily chosen for this study. Most researchers and clinicians would agree that patients should receive all prescribed low-vision and assistive devices and be allowed sufficient time to integrate devices and strategies into daily living before outcomes are measured. Visual function questionnaires may be susceptible to halo effects (ie, patients may overestimate their visual abilities at discharge, before they have the opportunity to integrate use of low-vision devices and adaptive techniques into daily living).²⁰ Although VA policy states that all discharged veterans should receive their prosthetic devices within 30 days of discharge, delays may occur. The 3-month timeline was chosen to allow 30 to 45 days for the veterans to receive their equipment and 30 to 45 days to integrate the equipment into daily living activities. The second outcome measurement was conducted at 1 year because outcomes in eye disease treatment trials are frequently reported at 1 year.

When outcomes were measured at 1 year, treatment effects were still large and statistically significant, but outcomes had declined compared with measurements at 3 months, probably as a result of changes in health associated with aging and progressive vision loss.

Questionnaires to measure comorbidities were not administered to all subjects with the 3-month and 1-year follow-up outcome measures. It was also not possible to measure visual-acuity changes along with administration of the telephone questionnaire 3 months and 1 year after rehabilitation in this study because the Hines VA BRC is a regional program and veterans do not return to the center for follow-up examinations.

The patients who did not complete 3 questionnaires (those who were not included in the analysis) were 5 years older, on average (mean age, 74.6y), than the veterans whose data were analyzed. These older veterans were more likely to experience a decline in health status. Fifteen of these subjects were excluded because of major depression, cognitive loss, or terminal illness during the rehabilitation program, and 4 were deceased before the 1-year follow-up.

The changes in rehabilitation outcomes over time should be addressed in future studies. Information from questionnaires to assess health status, mental health, and clinical measures of visual acuity postdischarge should be included to enable investigators to adjust for age, comorbidities, and visual acuity to improve generalizability of results to other patient populations. Previously published outcomes studies have not included this information. A control group should also be included to help explain changes in health status or mental health that might occur in the absence of treatment. The effect sizes reported in this study can be used to calculate sample sizes for observational studies and multicenter randomized clinical trials.

Back to Article Outline

Conclusions 

Although there was a significant decline in visual ability from 3 months to 1 year for the patients whose data were analyzed, treatment effects from the Hines VA BRC program were clinically and statistically significant 3 months and 1 year after rehabilitation.

Supplier

Back to Article Outline

References 

1. US Department of Health and Human Services. Vision research—a national plan 1999-2003: a report of the National Eye Advisory Council. Bethesda: National Eye Institute; 1998;NIH Publication No. 98-4120.
   • View In Article
2. Raasch TW, Leat SJ, Kleinstein RN, Bullimore MA, Cutter GR. Evaluating the value of low-vision services. J Am Optom Assoc. 1997;68:287–295
   • View In Article
   • MEDLINE
3. Department of Veterans Affairs, Veterans Health Administration, Blind Rehabilitation Service. Coordinated services for blinded veterans IB 11-59 (revised) P87250. Washington (DC): Department of Veterans Affairs; 1996;
   • View In Article
4. Watson GR, De l’Aune W, Stelmack J, Maino J, Long S. National survey of the impact of low vision device use among veterans. Optom Vis Sci. 1997;74:249–259
   • View In Article
   • MEDLINE
   • CrossRef
5. Watson GR, De l’Aune W, Long S, Maino J, Stelmack J. Veterans’ use of low vision devices for reading. Optom Vis Sci. 1997;74:260–265
   • View In Article
   • MEDLINE
   • CrossRef
6. De l’Aune WR, Welsh RL, Williams M. A national outcomes assessment of the rehabilitation of adults with visual impairments. J Vis Impair Blind. 2002;94:281–291
   • View In Article
7. Stelmack JA, Stelmack TR, Massof RW. Measuring low-vision rehabilitation outcomes with the NEI VFQ-25. Invest Ophthalmol Vis Sci. 2002;43:2859–2868
   • View In Article
   • MEDLINE
8. Stelmack J, Szlyk J, Stelmack T, et al. Use of Rasch person item map in exploratory data analysis: a clinical perspective. J Rehabil Res Dev. 2004;41:233–241
   • View In Article
   • MEDLINE
   • CrossRef
9. Szlyk JP, Stelmack J, Massof RW, et al. Performance of the veterans affairs low vision visual functioning questionnaire. J Vis Impair Blind. 2004;98:261–275
   • View In Article
10. Stelmack JA, Szlyk JP, Stelmack TR, et al. Psychometric properties of the Veterans Affairs Low-Vision Visual Functioning Questionnaire. Invest Ophthalmol Vis Sci. 2004;45:3919–3928
    • View In Article
    • MEDLINE
    • CrossRef
11. Stelmack JA, Szlyk JP, Stelmack TR, et al. Measuring outcomes of vision rehabilitation with the Veterans Affairs Low Vision Visual Functioning Questionnaire. Invest Ophthalmol Vis Sci. 2006;47:3253–3261
    • View In Article
    • MEDLINE
    • CrossRef
12. Dunn CM, Chadwick G. Code of Federal Regulations. In:  Allen W editors. Protecting study volunteers in research. Boston: CenterWatch; 1999;p. 203
    • View In Article
13. Massof RW. Likert and Guttman scaling of visual function rating scale questionnaires. Ophthalmic Epidemiol. 2004;11:381–399
    • View In Article
    • MEDLINE
    • CrossRef
14. Massof RW. The measurement of vision disability. Optom Vis Sci. 2002;79:515–552
    • View In Article
15. Linacre JM, Wright B. A user’s guide to Winsteps Rasch-model computer program. Chicago: Mesa Pr; 2001;
    • View In Article
16. Andrich D. A rating formulation for ordered response categories. Psychometrika. 1978;43:561–573
    • View In Article
    • CrossRef
17. Bond TG, Fox CM. Applying the Rasch model: fundamental measurement in the human sciences. Mahwah: Erlbaum; 2001;
    • View In Article
18. Cohen J. In: Statistical power analysis for the behavioral sciences. 2nd ed.. Hillsdale: Erlbaum; 1988;
    • View In Article
19. Rosnow RL, Rosenthal R. Computing contrasts, effect sizes and counternulls on other people’s published data: general procedures for research consumers. Psychol Methods. 1996;1:331–340
    • View In Article
    • CrossRef
20. Stelmack J, Babcock-Parziale J, Head DN, et al. Timing and directions for administration of questionnaires affect outcomes measurement. J Rehabil Res Dev. 2006;41:809–816
    • View In Article

• a Version 3.16; Winsteps, PO Box 811322, Chicago IL 60681-1322.