Purpose
To test the reliability and validity of questionnaires shortened from the National Eye Institute 25-item Vision Function Questionnaire (NEI VFQ-25).
Design
Cross-sectional, multicenter cohort study.
Methods
Reliability was assessed by Cronbach α coefficients. Validity was evaluated by studying the association of vision-targeted quality-of-life composite scores with objective visual function measurements. A total of 5482 women between the ages of 65 and 100 years participated in the year-10 clinic visit in the Study of Osteoporotic Fractures (SOF). A total of 3631 women with complete data were included in the visual acuity (VA) and visual field (VF) analyses of the 9-item NEI VFQ (NEI VFQ-9), which is defined for those who care to drive, and a total of 5311 women with complete data were included in the VA and VF in the analyses of the 8-item NEI VFQ (NEI VFQ-8). To assess differences in prevalent eye diseases, which were ascertained for a random sample of SOF participants, 853 and 1237 women were included in the NEI VFQ-9 and the NEI VFQ-8 analyses, respectively.
Results
The Cronbach α coefficient for the NEI VFQ-9 scale was 0.83, and that of the NEI VFQ-8 was 0.84. Using both questionnaires, women with VA worse than 20/40 had lower composite scores compared with those with VA of 20/40 or better ( P < .001). Participants with mild, moderate, and severe binocular VF loss had lower composite scores compared with those with no binocular VF loss ( P < .001). Compared with women without chronic eye diseases in both eyes, women with at least 1 chronic eye disease in at least 1 eye had lower composite scores.
Conclusions
Both questionnaires showed high reliability across items and validity with respect to clinical markers of eye disease. Future research should compare the properties of these shortened surveys with those of the NEI VFQ-25.
The National Eye Institute Visual Function Questionnaire (NEI VFQ) was devised to provide a self-reported measure of vision-targeted functioning. Originally, a 51-item questionnaire was developed, and the questionnaire later was shortened to 25 items (NEI VFQ-25). The instruments assess vision-related quality of life within different domains such as general, near, distance, and color vision; role limitations; dependency; mental health; and social function. They have been validated by a variety of studies showing they are useful tools in assessing vision-specific quality of life.
One advantage of these questionnaires is their comprehensiveness, because they capture several dimensions of vision-targeted quality of life. Subscales, usually consisting of several items, can be used to provide an independent measure of visual ability within each domain. The length of the surveys, however, can be a challenge, especially in large, multipurpose, population-based studies involving a variety of other questionnaires. The developers of the original NEI VFQ and the NEI VFQ-25 therefore devised a 9-item instrument (NEI VFQ-9) that has been used in the National Health and Nutrition Examination Survey since 1999. No published studies previously have assessed the reliability and validity of the NEI VFQ-9.
The NEI VFQ-9 was administered to 5482 women who participated in the year-10 clinic visit in the Study of Osteoporotic Fractures from 1997 through 1998. The purpose of this research was to evaluate the reliability and validity of the NEI VFQ-9, as well as of the 8-item NEI VFQ (NEI VFQ-8), in which the driving question is excluded because of a large proportion of nondrivers and statistically significant differences between drivers and nondrivers.
Methods
Subjects
From 1986 to 1988, 9704 ambulatory white female volunteers who were 65 years or older, with no history of osteoporosis or bilateral hip replacement, were enrolled in the SOF, a multicenter, prospective, longitudinal cohort study. These women were located in Baltimore, Maryland; Minneapolis, Minnesota; Portland, Oregon; and the Monongahela Valley, Pennsylvania. A cohort of 662 black women was recruited from population listings at each of the 4 clinic centers using the original recruitment criteria between 1996 and 1998. From January 1997 through September 1998, all surviving participants were invited to participate in a follow-up clinical examination (year-10 clinic visit), which included a comprehensive eye evaluation. All individuals in the study gave informed consent to participate after obtaining institutional review board approvals from all study institutions. A total of 5482 women, consisting of 4820 white participants (63% of the surviving cohort) and 662 black women, attended the year-10 clinic visit. These 5482 women completed the NEI VFQ-9 questionnaire.
Vision-Specific Quality of Life
The 9-item questionnaire was devised from the NEI VFQ by the developers of the NEI VFQ using similar quantitative and qualitative analysis when developing the NEI VFQ-25. With respect to the domains represented on the 51-item version of the questionnaire, the 9-item version includes one question about the person’s general vision, 3 questions about near vision (reading normal newsprint, seeing well up close, finding objects on a crowded shelf); 1 question about each of distance vision (going down stairs at night), driving (driving during the daytime in familiar places), peripheral vision (seeing objects off to the side), role limitation (limited in endurance), and well-being or mental health (amount of time worrying about eyesight); and no questions about ocular pain, vision-specific social functioning, expectations for visual function, vision-specific role-functioning, dependency because of vision, or color vision. Table 1 lists questionnaire items and the set of response options; other than a question asking whether the person is currently driving, each question includes either 5 or 6 response options. The NEI VFQ-9 takes approximately 3 to 4 minutes to complete. The questionnaire was administered using written forms that participants returned in the mail or during clinic visits. Subjects with low vision were able to seek assistance in completing the forms during their clinic visit.
| Nine-Item National Eye Institute Visual Function Questionnaire |
|---|
| Q1. General vision (6-level) |
| At the present time, would you say your eyesight (with glasses or contact lenses, if you wear them) is: |
| (1) excellent, (2) good, (3) fair, (4) poor, (5) very poor, or (6) are you completely blind? |
| Q2. Well being/mental health (5-level) |
| How much of the time do you worry about your eyesight? |
| (1) None of the time, (2) a little of time, (3) some of the time, (4) most of the time, or (5) all of the time. |
| Q3. Near vision, reading normal newsprint (6-level) |
| How much difficulty do you have reading ordinary print in newspapers? |
| (1) No difficulty at all, (2) a little difficulty, (3) moderate difficulty, (4) extreme difficulty, (5) stopped doing because of your eyesight, or (6) stopped doing this for other reasons or not interested in doing this. |
| Q4. Near vision, seeing well up close (6-level) |
| How much difficulty do you have doing work or hobbies that require you to see well up close, such as cooking, sewing, fixing things around the house, or using hand tools? |
| (1) No difficulty at all, (2) a little difficulty, (3) moderate difficulty, (4) extreme difficulty, (5) stopped doing because of your eyesight, or (6) stopped doing this for other reasons or not interested in doing this. |
| Q5. Distance vision, going down stairs at night (6-level) |
| Because of your eyesight, how much difficulty do you have going down steps, stairs, or curbs in dim light or at night? |
| (1) No difficulty at all, (2) a little difficulty, (3) moderate difficulty, (4) extreme difficulty, (5) stopped doing because of your eyesight, or (6) stopped doing this for other reasons or not interested in doing this. |
| Q6. Driving (6-level) a |
| How much difficulty do you have driving during the daytime in familiar places? |
| (1) No difficulty at all, (2) a little difficulty, (3) moderate difficulty, (4) extreme difficulty, stopped doing because of your eyesight, or (5) stopped doing this for other reasons or not interested in doing this. |
| Q7. Role limitation (5-level) |
| Are you limited in how long you can walk or do other activities such as housework, child care, school, or community activities because of your vision? |
| (1) All of the time, (2) most of the time, (3) some of the time, (4) a little of time, or (5) none of the time. |
| Q8. Peripheral vision (6-level) |
| Because of your eyesight, how much difficulty do you have noticing objects off to the side while you are walking along? |
| (1) No difficulty at all, (2) a little difficulty, (3) moderate difficulty, (4) extreme difficulty, (5) stopped doing because of your eyesight, or (6) stopped doing this for other reasons or not interested in doing this. |
| Q9. Near vision, finding objects on a crowded shelf (6-level) |
| Because of your eyesight, how much difficulty do you have finding something on a crowded shelf? |
| (1) No difficulty at all, (2) a little difficulty, (3) moderate difficulty, (4) extreme difficulty, (5) stopped doing because of your eyesight, or (6) stopped doing this for other reasons or not interested in doing this. |
a For Question 6, “not currently driving” also was an option; however, if selected, then the question was omitted and the remaining questions comprised the 8-item National Eye Institute Visual Function Questionnaire.
Assessment of Vision
Visual acuity (VA) and visual fields (VFs) were assessed among all women who participated in the year-10 clinic visit. The presence of eye diseases, specifically age-related macular degeneration (AMD), glaucoma, and visually significant cataract, was assessed among women who were selected randomly to have their eye photographs graded.
Visual Acuity and Visual Field
Distance VA was measured in each eye separately with habitual correction using Bailey-Lovie charts, and VFs were assessed in each eye separately using the Humphrey Field Analyzer suprathreshold 76-point 30-degree VF program (Humphrey Field Analyzer; Zeiss, Oberkochen, Germany), as described in a previous SOF publication.
Eye Diseases
The presence of AMD was determined using 45-degree stereoscopic fundus photographs. Early AMD was characterized by the presence of soft drusen (drusen area ≥ 95 μm but < 960 μm in diameter) with retinal pigment epithelium depigmentation or soft drusen (drusen area ≥ 960 μm in diameter) with or without pigmentary abnormalities. The presence of geographic atrophy or subfoveal choroidal neovascularization indicated the existence of late AMD.
Glaucoma was diagnosed based on the appearance of the optic nerve head in the optic disc photographs and 76-point VF screening results (at least 1 point missed). Glaucomatous optic nerves were those having diffuse and localized thinning of the neuroretinal rim, loss of retinal fiber layer, increased cupping, asymmetry of the optic nerve cup-to-disc ratios of 0.2 or more, or a combination thereof.
Aphakia or pseudophakia was determined for each eye using external or lens photographs, or both. Visually significant cataract was defined as the presence of nuclear, posterior subcapsular, or cortical cataract and VA worse than 20/40.
The presence of a chronic eye disease was defined as (1) the presence of late AMD, or the presence of early AMD with VA worse than 20/40; (2) the presence of glaucoma; or (3) the presence of visually significant cataract, or aphakia, or pseudophakia because of cataract surgery.
Women Included in the National Eye Institute Visual Function Questionnaire 9-Item and 8-Item Analyses
The distinction between the NEI VFQ-9 and NEI VFQ-8 is in their treatment of driving, which raises conceptual issues relevant to quality of life, because not everyone wants to or cares to drive. Rather than artificially assigning average quality of life, the possible quality of life, or some other quality of life to those who do not care to drive, the approach taken here is to have 2 alternate scales: the NEI VFQ-9 for those who care to drive, and the NEI VFQ-8, which is applicable to all. For those who do not care to drive, we consider their NEI VFQ-9 score to be undefined. Accordingly, in the NEI VFQ-9 instrument, subjects are asked first whether they are currently driving. For those whose response is “not currently driving,” 2 additional options are provided on the reasons of not driving: stopped because of eyesight or stopped for other reasons. A person responding with the former option was coded as a nondriver and was included in the calculation of the NEI VFQ-9 composite score with the lowest possible item score (0), whereas a person responding with the latter option also was coded as a nondriver, but was considered to have an undefined NEI VFQ-9 composite score. Subjects who responded that they were completely blind were coded as nondrivers and were included in the calculation of the NEI VFQ-9 composite score with the lowest possible item score (0).
Among the 5482 women attending the SOF year-10 clinic visit in this study, 9 did not respond to the driving question. Therefore, 1793 subjects, including the 9 who did not respond to the driving question, 1563 who did not provide the reason for not driving, and 221 women who stopped for other reasons, were excluded from the calculation of the NEI VFQ-9 composite score. They were included in the calculation of the NEI VFQ-8 composite score when they responded to the other 8 NEI VFQ questions.
Among the 5482 women attending the SOF year-10 clinic visit, 8 did not return the questionnaire and were excluded from the statistical analyses, resulting in a sample of 5474 women. One additional individual did not respond to the driving question, and 1886 (34%) were coded as nondrivers. Of the 1886 women, 1563 were not currently driving without stating the reason for not driving, 221 had stopped for other reasons, 96 had stopped because of eyesight, and 6 were completely blind. Approximately 97% (5311/5474) provided valid responses to the 8 questions applicable to the NEI VFQ-8; of those who did not, 105 were drivers and 58 were nondrivers. To assess the differences in composite scores based on VA and VF, a total of 3631 women were included in the analyses of the NEI VFQ-9, and 5311 were part of the analyses of the NEI VFQ-8.
Eye diseases were diagnosed in a random sample of 1274 women of all participants in the year-10 visit. Forty-two (3.5%) women had late AMD in at least 1 eye, 513 (43.1%) had early AMD in at least 1 eye, and 634 (53.3%) women did not have AMD in either eye. A total of 105 (8.5%) had glaucoma in at least 1 eye. A total of 376 (31.7%) women had aphakia or pseudophakia in both eyes, 332 (28.0%) had visually significant cataract in at least 1 eye, and 479 (40.4%) women did not have visually significant cataract in both eyes. To assess the differences in composite scores based on prevalent eye diseases, 853 women (of 1274, or 67%) had complete data for the 9-item analysis and 1237 (of 1274, or 97%) had complete data for the 8-item analysis.
Statistical Analysis
Distributions of demographic and health characteristics are presented for all women who participated in the year-10 clinic visit. Self-rated health status was measured by asking participants to rate their health relative to that of others as excellent, good, fair, poor, or very poor. These responses were grouped into: (1) fair, poor, or very poor; or (2) excellent or good. Participants were asked about having been told by a doctor that they have 1 or more of 17 comorbidities such as heart attack, coronary, or myocardial infarction, stroke, diabetes, high blood pressure, Parkinson disease, and dementia or Alzheimer disease.
The percentages of women with missing NEI VFQ-9 item-level responses at the missing level, or with responses at the ceiling (best possible) or floor (worst possible) levels, were calculated for the sample of women who had valid responses to all 9 questions. Individuals characterizing their vision as excellent (question 1, general vision), have a response at the ceiling level. The calculation of composite scores requires no missing values in any items.
To assess reliability, the internal consistencies of the NEI VFQ-9 and the NEI VFQ-8 were evaluated using Cronbach α coefficients. The Cronbach α coefficient varies from 0 to 1; higher values indicate greater internal consistency.
Spearman correlations between VA or binocular VF and both multi-item instruments were used to assess whether people with poorer vision had lower vision-specific quality of life. Number of points missed in binocular VF loss and logarithm of the minimal angle of resolution score of VA in the better and worse eye were analyzed as continuous variables. To support validity, Spearman correlation coefficients should be negative, indicating a higher logarithm of the minimal angle of resolution score (signifying worse VA), and more points missed in binocular VF are related to lower vision-specific quality of life.
The comparison of mean NEI VFQ-9 and NEI VFQ-8 scores by VA groups (20/40 or better and worse than 20/40) and by VF groups (no binocular visual field [BVF] loss, mild BVF loss [1 to 9 points missed], moderate BVF loss [10 to 19 points missed], and severe BVF loss [20 or more points missed]) were used to assess whether statistically significant differences existed in the composite scores between people with worse vision and those with good vision.
The validity of both multi-item questionnaires using comparisons between questionnaire composite scores also was assessed based on prevalent eye diseases. To support validity, women in the reference groups comprising women with no chronic eye disease should have higher composite scores compared with those with an eye disease.
The statistical comparisons of composite scores included the following: (1) scores of women with VA 20/40 or better versus women with VA worse than 20/40 in the better eye and worse eye; (2) scores of women with no binocular VF loss versus those with mild, moderate, and severe binocular VF loss; (3) scores of women without any chronic eye diseases in both eyes versus those with at least 1 eye disease in at least 1 eye; (4) scores of women without any chronic eye diseases in both eyes versus those with early or late AMD in at least 1 eye; (5) scores of women without any chronic eye diseases in both eyes versus those with glaucoma in at least 1 eye; and (6) scores of women without any chronic eye diseases in both eyes compared with phakic women with visually significant cataract in at least 1 eye and those with aphakia or pseudophakia in both eyes. The composite scores range from 0 to 100, with a score of 100 showing the highest quality of life. All statistical analyses were performed using SAS software version 9.1 (SAS Institute, Inc, Cary, North Carolina, USA). Statistical significance was defined as P < .05.
Results
General Characteristics of Participants
Table 2 presents the characteristics of the 5482 women who participated in the SOF year-10 clinic visit. The sample consisted of primarily older women; their mean age was 79.5 ± 4.5 years with a range of 65 to 100 years. Most women (4359/5482, or 79%) were between the ages of 75 and 84 years. Approximately 7% (382/5482) were in the age group 65 to 74 years, and 14% (741/5,482) were 85 years of age or older. Black women represented 12% of the sample (662/5482); the remaining were white women. Approximately 80% (4376/5482) completed high school or had higher education. The same percentage of women reported excellent or good health status. The distribution of comorbidities was skewed; the median number of comorbidities was 1, with a range of 0 to 10 (of 17 possible). Approximately 17% (920/5482) had VA worse than 20/40. In sum, the sample consisted of women who primarily were older, white, high school graduates or more, and in excellent or good self-rated health status.
| Characteristics | Summary Statistics |
|---|---|
| No. of study sites, n (%) | |
| Baltimore | 1279 (23%) |
| Minneapolis | 1635 (30%) |
| Pittsburgh | 1409 (26%) |
| Portland | 1159 (21%) |
| Age (yrs) , n (%) | |
| Mean ± SD (range) | 79.5 ± 4.5 (65 to 100) |
| 65 to 74 | 382 (7%) |
| 75 to 79 | 2704 (49%) |
| 80 to 84 | 1655 (30%) |
| ≥85 | 741 (14%) |
| Race/ethnicity, n (%) | |
| White | 4820 (88%) |
| Black | 662 (12%) |
| Education (yrs) , n (%) | |
| <12 | 1096 (20%) |
| ≥12 (HS graduate or more) | 4380 (80%) |
| Self-rated health status, n (%) | |
| Fair/poor/very poor | 1097 (20%) |
| Excellent/good | 4376 (80%) |
| No. of comorbidity (0 to 17) | |
| Median (range) | 1 (0 to 10) |
| Habitual distance visual acuity in the better eye (logMAR) | |
| Mean ± SD (range) | 0.19 ± 0.18 (−0.18 to 1.36) |
| Worse than 20/40 | 920 (17%) |
Item Responses
Table 3 shows that many women performed at the optimal level for near, distance, and peripheral vision. The percentage of women reporting responses at the ceiling level exceeded 60% for all but 1 question, reaching 92% for the role limitation question. A relatively low percentage (15%) of women, however, reported that their eyesight was excellent ( Table 3 ).
| 9-Item National Eye Institute Visual Function Questionnaire Items | Missing, No. (%) | Floor, No. (%) | Ceiling, No. (%) |
|---|---|---|---|
| General vision | |||
| 6-level general vision | 0 | 7 (0.1%) | 817 (15%) |
| Near vision | |||
| 6-level reading normal newsprint | 18 (0.3%) | 121 (2%) | 3998 (73%) |
| 6-level seeing well up close | 59 (1%) | 114 (2%) | 4018 (73%) |
| 6-level finding objects on crowded shelf | 13 (0.2%) | 19 (0.3%) | 4615 (84%) |
| Distance vision | |||
| 6-level going down stairs at night | 94 (2%) | 39 (0.7%) | 3922 (72%) |
| Driving | |||
| 6-level daylight familiar places | 1785 (33%) | 102 (2%) | 3484 (64%) |
| Peripheral vision | |||
| 6-level seeing objects off to side | 9 (0.2%) | 15 (0.3%) | 4776 (87%) |
| Role limitation | |||
| 5-level limited in endurance | 4 (0.1%) | 74 (1%) | 5030 (92%) |
| Well-being/mental health | |||
| 5-level amount time: worry | 1 (0.02%) | 234 (4%) | 3644 (67%) |
a Excluding 8 women who did not have valid responses to all 9 questions.
Reliability
The Cronbach α coefficient for the 9-item scale was 0. 83 and for the 8-item was 0.84 ( Table 4 ). When removing questions from the scale, the α values remained relatively high, decreasing slightly for all items except for the well-being and mental health question, which if removed, resulted in Cronbach α coefficients of 0.85 for the NEI VFQ-9 and 0.86 for the NEI VFQ-8.
| 9-Item National Eye Institute Visual Function Questionnaire Item | National Eye Institute Visual Function Questionnaire: Cronbach α Coefficient | |
|---|---|---|
| 9-Item Composite Score (n = 3631) a | 8-Item Composite Score (n = 5311) b | |
| Overall | 0.83 | 0.84 |
| Cronbach coefficient α for deleted 9-item National Eye Institute Visual Function Questionnaire item | ||
| General vision | ||
| 6-level general vision | 0.81 | 0.82 |
| Near vision | ||
| 6-level reading normal newsprint | 0.80 | 0.80 |
| 6-level seeing well up close | 0.80 | 0.80 |
| 6-level finding objects on crowded shelf | 0.81 | 0.81 |
| Distance vision | ||
| 6-level going down stairs at night | 0.82 | 0.81 |
| Driving | ||
| 6-level daylight familiar places | 0.82 | NA |
| Peripheral vision | ||
| 6-level seeing objects off to side | 0.82 | 0.82 |
| Role limitation | ||
| 5-level limited in endurance | 0.82 | 0.83 |
| Well-being/mental health | ||
| 5-level amount time: worry | 0.85 | 0.86 |
a Nine-item National Eye Institute Visual Function Questionnaire composite score was calculated among women who had valid responses to all 9 items.
b Eight-item National Eye Institute Visual Function Questionnaire composite score was calculated among women who had valid responses to all 8 items, except for the driving item.
Validity
Table 5 presents the validity of both multiquestion instruments using VA and binocular VF assessments. Table 6 contains information regarding differences in prevalent eye diseases, specifically any chronic eye disease, AMD, glaucoma, and visually significant cataract.
| Clinical Measure of Visual Function | National Eye Institute Visual Function Questionnaire | |
|---|---|---|
| 9-Item Composite Score (n = 3631) a | 8-Item Composite Score (n = 5311) b | |
| Overall: mean ± SD (median) | 91.3 ± 11.5 (95) | 88.8 ± 13.7 (94) |
| Visual acuity (logMAR) | ||
| Better eye | ||
| Spearman correlation coefficient | −0.21 | −0.24 |
| P value | P < .001 | P < .001 |
| Mean ± SD (median) | ||
| 20/40 or better | 92.6 ± 8.5 (95) | 90.8 ± 10.4 (94) |
| Worse than 20/40 | 85.1 ± 17.3 (92) | 80.6 ± 19.7 (89) |
| P value c | P < .001 | P < .001 |
| Worse eye | ||
| Spearman correlation coefficient | −0.23 | −0.26 |
| P value | P < .001 | P < .001 |
| Mean ± SD (Median) | ||
| 20/40 or better | 93.3 ± 8.1 (96) | 91.7 ± 9.8 (95) |
| Worse than 20/40 | 89.1 ± 12.8 (94) | 85.9 ± 15.5 (91) |
| P value c | P < .001 | P < .001 |
| Binocular visual field loss (number of points missed) d | ||
| Spearman correlation coefficient | −0.14 | −0.16 |
| P value | P < .001 | P < .001 |
| Mean ± SD (median) | ||
| None (no point missed) | 93.6 ± 7.0 (96) | 92.4 ± 8.3 (95) |
| Mild (1 to 9 points missed) | 92.1 ± 9.2 (95) | 90.9 ± 10.2 (94) |
| Moderate (10 to 19 points missed) | 90.1 ± 11.1 (95) | 88.2 ± 12.1 (93) |
| Severe (20 or more points missed) | 87.0 ± 14.8 (92) | 83.0 ± 17.6 (89) |
| P value c | P < .001 | P < .001 |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
