Patient Preferences for Diabetic Retinopathy Health States

Abstract

Purpose.: To develop standardized descriptions of health states that characterize vision-specific functional impacts of diabetic retinopathy (DR) according to levels of visual acuity and contrast sensitivity and to elicit preferences for these health states from persons with DR and assign weighted values to them.

Methods.: Vision-specific descriptions of health states were developed based on a literature review and patient and physician interviews. The content was based on items from the National Eye Institute Visual Functioning Questionnaire (VFQ) and reflected functional impacts experienced by DR patients. Values were assigned to the range of health states, anchored by the extremes full vision and death, by using the time-tradeoff method in a sample of 98 Canadian DR patients from three clinical centers.

Results.: The mean age of the sample was 60.4 years, and 56% were men. Mean preferences decreased from 0.98 (better-eye logMAR [Snellen equivalent] acuity, ≥20/40; worse-eye Snellen equivalent, ≥20/200) to 0.67 (Snellen equivalent visual acuity, ≤20/200, contrast sensitivity, ≤21 letters bilaterally). Preferences decreased with increasing severity of functional deficits and did not vary significantly by sex, age, VFQ quartile, or better- or worse-eye acuity.

Conclusions.: This is the first study that has been conducted to estimate preferences for standardized DR-specific health states, accounting for visual acuity and contrast sensitivity in both eyes. The results showed that the development and progression of DR are associated with substantial declines in preferences. In addition to the progressively greater impact from declining ETDRS visual acuity and contrast sensitivity, preference weights declined with increasing bilateral disparity. These preference values are useful for comparing the cost effectiveness of ophthalmic treatments.

The incidence of diabetes mellitus worldwide is increasing, and the World Health Organization's predicted 39% global increase from 2000 to 2030 ¹ is very likely a gross underestimate. ² As such, the incidence of retinopathy, a common microvascular complication of diabetes, is also anticipated to reach epidemic proportions. ³ Diabetic retinopathy (DR) is the leading cause of severe visual impairment among working-age people, ^4–6 compounding the already substantial burden ^7,8 associated with the progression of diabetes mellitus. ^9–13 Both DR and diabetes can have wide-ranging impacts on almost all aspects of health-related quality of life. ^9–19  

The impact of visual impairment on quality of life is most frequently estimated using patient-reported outcome measures (PROs). Although there are several vision-specific PROs, ^20,21 the National Eye Institute's Visual Function Questionnaire (NEI VFQ) ^22,23 is the most widely validated and frequently used. Although useful for quantifying the impact on quality of life of visual impairment, psychometric PRO questionnaires are not measures of individuals' preferences for the various health states. Preference measures can do exactly that: ask respondents about their preferences for a level of health against the reference health state of full vision. Weighting of patient preferences reflects the net impact of being at a particular level of health and can be used in adjusting life expectancy based on quality when assessing the cost effectiveness of new treatments for ophthalmic conditions. 

Although several preference studies have been conducted among DR patients, ^24–34 there are at least four limitations in the methodology that restrict the reliability and validity of the results: Small samples were recruited that included only patients who rated their own health states, there was no standardized scale of health states, and health states were based only on visual acuity in the better-seeing eye. The latter limitation is problematic when applying preference values in economic evaluations that are based on the results of randomized trials, because logMAR visual acuity (measured on the Early Treatment of Diabetic Retinopathy Study [ETDRS] chart) and Pelli-Robson contrast sensitivity are the standard measures used to monitor visual changes in trials in ophthalmology. 

A recent health technology assessment for the National Institute of Health and Clinical Excellence (NICE) in the United Kingdom reported that the measurement of preferences for health states in ophthalmic diseases, considering visual acuity and contrast sensitivity in both eyes, is one of the research priorities in the area. ³⁵ In this study, we therefore (1) developed standardized descriptions of vision-specific, quality-of-life impacts of DR that considered both visual acuity and contrast sensitivity in the better- and worse-seeing eyes and (2) weighted the preferences for these health states in a representative sample of Canadian patients with DR. As the health states are based on items from the NEI VFQ, investigators in the future will be able to estimate preference values for patients based on NEI VFQ responses if visual acuity data are missing. 

Methods

A cross-sectional study was conducted to elicit patient preferences ³⁶ for health states predicated on difficulties with activities of daily living due to DR. The patient population included Canadians 18 years of age or older with various severities of DR, but no other sight-threatening conditions. Participants were recruited to represent the distribution between the sexes of diabetes mellitus in Canada. ³⁷ Interviewing took place at three locations, two in Vancouver, British Columbia, and one in London, Ontario, between May and August 2008. All patients were compensated financially and provided informed consent. The target sample size of 100 patients was calculated to allow a 95% confidence interval (CI) around mean preference values of up to 0.04. ³⁸ This study was approved by the institutional review boards of the Universities of British Columbia and Western Ontario and the Western Institutional Review Board and adhered to the tenets of the Declaration of Helsinki. 

One-on-one interviews were conducted by trained interviewers to elicit time-tradeoff (TTO) ³⁹ preference values for the health states, as well as current vision status (Your Current Vision), by standard methodology. ⁴⁰ Briefly, participants were asked to imagine that they were in each of the standardized health states (described later). For each health state, the participants chose between (1) remaining in the health state without improvement for 10 years or (2) trading some of the 10 years for a hypothetical treatment that would restore full vision. Participant demographics, most recent Snellen visual acuity (within 3 months), and clinical information were recorded. All participants also completed the NEI VFQ. ²²  

Establishment of the Health States

Vision-specific health states were established to illustrate the functional impacts of DR at different levels of visual function (based on ETDRS and contrast sensitivity measures in the better- and worse-seeing eyes). Functional limitations due to visual impairment ^{14–19,22,23,41,42} were described by using items from the NEI-VFQ survey. ^22,23 We verified the content validity of the health states through qualitative interviews with 30 DR patients with various levels of visual impairment (measured by ETDRS acuity and Pelli-Robson contrast sensitivity) recruited from a retinal practice in Vancouver. A trained interviewer queried participants on limitations due to DR before the participants completed the NEI-VFQ. A thematic analysis ⁴³ was conducted to characterize visual impairment according to acuity level and to identify the key functional impacts of DR from the patient's perspective. As functional status varied considerably within each visual category, more than one health state was defined for some levels of vision. For example, within one visual category, the ability to drive could differ substantially, and thus within the same visual acuity, more than one health state were established based on the various levels of driving ability. Refinements were made to the health state descriptions based on patient feedback, mean NEI-VFQ scores by vision category, and review by three clinicians and four PRO researchers experienced with characterizing health states. 

The final set of 10 health states reflect 6 clinically and patient-relevant combinations of better- and worse-seeing eye ETDRS visual acuity and contrast sensitivity (1). Table 1 shows a classification scheme to assign specific health states to the patients based on visual acuity and contrast sensitivity. If such clinical data are lacking in some samples, but data are available from the NEI-VFQ, the responses to the VFQ items for driving, mobility, reading, and activity may be useful in determining the specific health state that a selected patient should be assigned. In addition to the 10 DR health states derived based on the VFQ, a health state for current health (your current vision) was included. 

Table 1.

 

Vignette Classification Scheme

Health State	Better-Seeing Eye ETDRS/logMAR	Better-Seeing Eye Snellen Equivalent	Worse-Seeing Eye ETDRS/logMAR	Worse-Seeing Eye Snellen Equivalent	VFQ Driving under Difficult Conditions Response*	VFQ Mobility Response*	VFQ Reading Response*	VFQ Activity Response*
1a	0.0–0.3	20/20–20/40	>1.0 or CS >21	>20/200 or CS ≥21	1	1	1	1
1b	0.0–0.3	20/20–20/40	>1.0 or CS >21	>20/200 or CS ≥21	2	1	2	2
2a	0.0–0.3	20/20–20/40	<1.0 or CS <21	<20/200 or CS <21	5	3	3	4
2b	0.0–0.3	20/20–20/40	<1.0 or CS <21	<20/200 or CS <21	4	3	3	3
3a	0.4–0.6	20/50–20/80	>1.0 or CS >21	>20/200 or CS ≥21	3	3	2	2
3b	0.4–0.6	20/50–20/80	>1.0 or CS >21	>20/200 or CS ≥21	4	3	3	3
4a	0.4–0.6	20/50–20/80	<1.0 or CS <21	<20/200 or CS <21	4	3	4	4
4b	0.4–0.6	20/50–20/80	<1.0 or CS <21	<20/200 or CS <21	4	3	3	3
5	0.7–0.9	20/100–20/160	<1.0	<20/200 or CS <21	5	4	4	4
6	1.0	20/200 or CS <21	<1.0 or CS <21	<20/200 or CS <21	5	4	4	5

*  Mean response values to each specific item on the NEI VFQ survey estimated from the patient sample and expert opinion. Potential NEI VFQ responses for each item range from 1 (no difficulty performing the activity) to 5 (stopped performing the activity because of eyesight). CS, contrast sensitivity.

Analysis

For each health state, the preference weight was calculated by dividing the number of years the respondent would like to live in full vision by 10 (for example, 7 years in full vision is equivalent to a health state preference value of 0.70). ⁴⁰ Preference values were summarized by the mean, 95% CI, standard error (SE), median, range, and interquartile (IQR) range. Demographic data were summarized by the mean and standard deviation or proportion, as appropriate. Using the Student's t-test or analysis of variance as appropriate, we performed subgroup analyses comparing preference weights among the following groups: age, sex, NEI-VFQ score quartile, NEI-VFQ general health score, bilateral disparity of ETDRS acuity (in Snellen equivalents) between the eyes (better eye ≥20/40; worse eye ≤20/200), and better- or worse-seeing eye Snellen visual acuity (expressed using continuous, decimal value Snellen-equivalent scores). Statistical comparisons were two-tailed, with an α of 0.05 (Stata for Windows Version 10; StatSoft, College Station, TX). 

Results

A total of 100 participants with DR were recruited. Two participants' responses were eliminated because of illogical responses on the TTO exercise, resulting in an effective sample size of 98. Mean participant age was 60.4 years and 55 (56%) were men (Table 2). Forty (41%) participants reported always wearing glasses and seven (7%) wore none. Eighty-two (84%) participants had bilateral diabetic retinopathy and 76 (77.6%) were receiving treatment. Twenty-seven (28%) participants had diabetic macular edema without proliferative disease, 26 (26%) had proliferative DR without macular edema, 36 (37%) had both, and 9 (9%) had neither. Mean better- and worse-seeing eye Snellen acuities were 20/30 (range, 20/20–20/400) and 20/50 (range, 20/25–20/800), respectively. 

Table 2.

 

Demographic Characteristics of Canadian Participants from Whom TTO Utilities Were Elicited

Characteristic	Qualitative Interviewees (n = 30)	Preference Interviewees (A = 98)
Mean age, y (SD)	64.8 (12.25)	60.38 (12.59)
Male sex, n (%)	17 (56.67)	55 (56.12)
Race, n (%)
White	22 (73.33)	81 (82.65)
Native Canadian	0 (0.00)	2 (2.04)
Asian	5 (16.67)	7 (7.14)
Other	1 (3.33)	6 (6.12)
Did not specify	2 (6.67)	2 (2.04)
Education category, n (%)
Grade school	1 (3.33)	1 (1.02)
Some high school	3 (10.00)	13 (13.27)
Graduated high school	17 (56.67)	46 (46.94)
Completed degree/post graduate degree	9 (30.00)	38 (38.78)
Occupation category, n (%)
Full time employed/self-employed	8 (26.67)	27 (27.55)
Part time employed/self-employed	0 (0.00)	5 (5.10)
Retired	16 (53.33)	45 (45.92)
Stay at home	3 (10.00)	7 (7.14)
Unemployed	2 (6.67)	3 (3.06)
Other	1 (3.33)	11 (11.22)
Marital status, n (%)
Single	4 (13.33)	20 (20.41)
Married/Partnership	23 (76.67)	63 (64.29)
Divorced	2 (6.67)	8 (8.16)
Widowed	1 (3.33)	7 (7.14)
Chronic medical conditions, n (%)
Cancer	2 (6.67)	6 (6.12)
Headaches	3 (10.00)	10 (10.20)
Respiratory	2 (6.67)	10 (10.20)
Heart disease	7 (23.33)	41 (41.84)
Gastrointestinal	5 (16.67)	16 (16.33)
Psychological problem	3 (10.00)	20 (20.40)
Other	5 (16.67)	22 (22.45)
None	14 (46.67)
Usage of corrective lenses, n (%)
All the time	13 (43.33)	40 (40.82)
For seeing up close	11 (36.67)	33 (33.67)
For seeing at distance	8 (26.67)	17 (17.35)
Hardly ever wear	1 (3.33)	2 (2.04)
Never wear	0 (0.00)	1 (1.02)
Do not have glasses	2 (6.67)	7 (7.14)
Does wearing glasses improve vision?	(n = 27)	(n = 91)
Yes, I see a lot better, n (%)	15 (53.57)	58 (63.74)
Yes, I see a little better, n (%)	11 (39.29)	26 (28.57)
Glasses do not help, n (%)	1 (3.57)	7 (7.69)
Receiving treatment for DR, n (%)
Yes	27 (90.00)	76 (77.55)
No	3 (10.00)	21 (21.43)
Did not specify	0 (0.00)	1 (1.02)
DR in both eyes, n (%)	19 (67.86)	82 (83.67)
Median better eye habitual Snellen VA	20/40	20/30
Mean better eye habitual Snellen VA	20/40	20/30
20/20–20/25, n (%)	5 (16.67)	46 (46.94)
20/30–20/40, n (%)	13 (43.33)	26 (26.53)
20/50–20/80, n (%)	11 (36.67)	17 (17.35)
20/100–>20/200, n (%)	0 (0.00)	1 (1.02)
20/200–>20/400, n (%)	0 (0.00)	6 (6.12)
<20/400, n (%)	0 (0.00)	1 (1.02)
Unknown, n (%)	1 (3.33)	1 (1.02)
Median worse eye habitual Snellen VA	20/70	20/60
Mean worse eye habitual Snellen VA	20/60	20/50
20/20–20/25, n (%)	2 (6.67)	23 (23.47)
20/30–20/40, n (%)	6 (20.00)	17 (17.35)
20/50–20/80, n (%)	11 (36.67)	25 (25.51)
20/100–>20/200, n (%)	8 (26.67)	7 (7.14)
20/200–>20/400, n (%)	1 (3.33)	17 (17.35)
<20/400, n (%)	1 (3.33)	8 (8.16)
Unknown, n (%)	1 (3.33)	1 (1.02)
Mean (SD) better eye habitual ETDRS VA	0.76 (0.27)	NA
Mean (SD) worse eye habitual ETDRS VA	0.45 (0.30)	NA
Mean (SD) better eye contrast sensitivity (letters)	31.43 (3.82)	NA
Mean (SD) worse eye contrast sensitivity (letters)	26.60 (8.43)	NA

Mean preference values decreased from 0.98 for the most favorable health state (better-seeing eye ETDRS acuity 20/20–20/40, worse-seeing eye ETDRS acuity >20/200 or contrast sensitivity >21) to 0.67 for the least favorable health state (ETDRS acuity <20/200 or contrast sensitivity ≤21; Table 3). A large, bilateral disparity in visual acuity or contrast sensitivity between the eyes affected the preference weights. For example, the preference value for a health state with good visual acuity in the better-seeing eye (ETDRS acuity 20/20–20/40) and poor vision in the worse-seeing eye (ETDRS acuity ≤20/200 or contrast sensitivity ≤21 letters) was lower than a second health state with similar visual function in both eyes: 0.81 vs. 0.85, respectively. A fundamental difference between these two health states is the loss of the ability to drive in the former health state, with a wide disparity in acuity. 

Table 3.

 

TTO Preference Values for All Health States: Total Study Population

Health State	Better-Seeing Eye*	Worse-Seeing Eye*	Mean (SE)	95% CI	Median	Range	IQR
1a	20/20–20/40	>20/200 or CS ≥21	0.98 (0.01)	0.96–0.99	1.00	0.55–1.00	0.00
1b	20/20–20/40	>20/200 or CS ≥21	0.94 (0.01)	0.92–0.96	1.00	0.35–1.00	0.10
3a	20/50–20/80	>20/200 or CS ≥21	0.91 (0.01)	0.88–0.94	0.98	0.20–1.00	0.12
	Current vision	Current vision	0.88 (0.02)	0.84–0.93	1.00	0.03–1.00	0.15
2b	20/20–20/40	<20/200 or CS <21	0.85 (0.02)	0.82–0.89	0.90	0.03–1.00	0.20
3b	20/50–20/80	>20/200 or CS ≥21	0.85 (0.02)	0.82–0.89	0.93	0.03–1.00	0.19
4b	20/50–20/80	<20/200 or CS <21	0.85 (0.02)	0.81–0.89	0.90	0.03–1.00	0.19
2a	20/20–20/40	<20/200 or CS <21	0.81 (0.02)	0.77–0.85	0.85	0.03–1.00	0.28
4a	20/50–20/80	<20/200 or CS <21	0.76 (0.02)	0.71–0.81	0.81	0.03–1.00	0.28
5	20/100–20/160	<20/200 or CS <21	0.70 (0.03)	0.64–0.75	0.76	0.03–1.00	0.33
6	20/200 or CS <21	<20/200 or CS <21	0.67 (0.03)	0.62–0.73	0.70	0.03–1.00	0.40

 Total study population, N = 98. CS, contrast sensitivity.

*  LogMAR acuities expressed as approximate Snellen equivalents.

Preferences for the your-current-vision health state were highly variable in the sample, ranging from 0.03 to 1.0. The mean preference value for that health state was 0.88 and varied significantly by NEI-VFQ score quartile and better- and worse-seeing eye visual acuity (P < 0.05). Mean health state preferences for current vision ranged from 0.20 to 1.0 among patients with (ETDRS acuity) better-seeing eye acuity, 20/20 to 20/40; from 0.20 to 1.0 among patients with better-seeing eye acuity, 20/50 to 20/100; and from 0.03 to 0.93 among patients with better-seeing eye acuity, ≤20/200 (Fig. 1). 

Figure 1.

 

Preference values according to baseline Snellen visual acuity in the better-seeing eye.

With the exception of the your-current-vision state, no consistent patterns were observed in mean preference values for health states by age category, sex, NEI-VFQ-score quartile, NEI-VFQ general health score, better- or worse-seeing eye Snellen visual acuity, or bilateral disparity (better-seeing eye, ≥20/40; worse-seeing eye, ≤20/200). Mean preference values for the current-vision health state varied significantly by NEI-VFQ score quartile and by bilateral disparity in visual acuity. 

Discussion

This study fills two gaps in the literature: the lack of both standardized DR health states and their respective patient-based utility weights. The health states in the present study are categorized according to visual acuity, contrast sensitivity, and bilateral disparity. These states enable the assignment of utilities that can be used to quality adjust life life-years associated with DR treatment outcomes. In addition, as the health states are based on content of important items in the NEI-VFQ, if those data are available and clinical efficacy data are not, responses to the VFQ items may be used to map patients into health state categories. This circumstance may occur in studies in which collecting repeated clinical measurements is not feasible, but PRO responses are. 

Patient preferences for health states ranged from 0.98 for good vision in both eyes (ETDRS acuity, 20/20–20/40 in the better-seeing eye with preserved vision in the worse-seeing eye) to 0.67 for poor vision in both eyes (ETDRS acuity, <20/200). Bilateral disparities in visual function led to lower preferences. This finding has been observed in some quality-of-life studies in ophthalmology (Pleil AM, et al. IOVS 2009;50:ARVO E-Abstract 3775), but not in others. ^25,44  

The preference values in this study contrast with those estimated by Lloyd et al., ²⁴ who reported patient utilities ranging from 0.81 (worse-seeing eye, 20/20–20/25) to 0.67 (worse-seeing eye, 20/100–20/200). That study is not directly comparable to ours, for two reasons. First, the preference elicitation method differed, with TTO being used in the present study and the standard gamble in Lloyd et al. TTO preferences are systematically higher than standard gamble preferences because there is no element of uncertainty in the former. ⁴⁰ Second, the health states were linked to worse-seeing eye visual acuity and included the systemic effects of diabetes. As such, it is not possible to isolate the impacts of DR from those of diabetes. In the present study, we focused only on the impacts of impaired vision caused by DR. 

The estimated preferences are slightly higher than those measured in other DR studies in which the TTO was also used. ^{9,27,28,30,33} The earlier studies had small samples of participants in each visual function category who rated only their present visual state (rather than evaluating all possible health states), and thus preference values estimated for the same levels of visual function in the different studies varied widely. For example, preferences for perfect vision or no retinopathy range from 0.85 ^25,28 to 0.94. ³² Consistent with these findings, we noted that preferences for the your-current-vision state varied substantially, even among patients at the same acuity level or within the same NEI-VFQ quartile. For example, participants in the worst health state (better-seeing eye ETDRS acuity ≤20/200) reported preferences for current vision ranging from 0.03 to 0.93. The fact that participants in the present study valued all the standardized health states is advantageous, in that we were able to control for potential idiosyncratic variation arising when individuals value only their present vision. 

Our study is the first to consider the impact of the degree of involvement of both eyes on preferences. Historically, most studies in ophthalmology (and specifically DR) ^{27,28,30,31,33} link preferences to best corrected visual acuity in the better-seeing eye only, leading to several limitations. Measuring best corrected vision in the better-seeing eye fails to consider the binocular nature of visual function, the impact of decreased contrast sensitivity, and whether the better-seeing eye is the treated eye. In the present study, health states describing poorer levels of contrast sensitivity in the worse-seeing eye, or disparate measures of visual function between the eyes, were preferred less than states where the vision in the worse-seeing eye remained relatively well-preserved. Measuring preferences for health states based only on better-seeing eye visual acuity ^9,25–34 therefore has the potential to overestimate the utility of a given health state. The importance of both contrast sensitivity and binocular visual acuity in overall function may help to explain some of the variability in published preference estimates where these were not accounted for, ^{9,24–28,30–34} an important consideration for studies reporting the impact of new treatments by visual acuity levels only. 

We based the content of our health states on NEI-VFQ items describing functional impairments of key importance to DR patients, which may differ from those that are of importance to patients with, for example, age-related macular degeneration (Pleil AM, et al. IOVS 2009;50:ARVO E-Abstract 3775). As other investigators have observed, ^{15,16,19,41,42} DR patients report impairments in driving ability, participation in leisure and work activities, reading ability, and activities related to diabetic management (specifically, reading glucometers and food labels). Functional impacts unique to DR challenge the commonly held belief that it only the level of visual acuity that affects quality of life. ^25,44 This study provides evidence that the cause of visual impairment also has an impact. 

There are two upper anchors presently in use in ophthalmology: Vision-specific health states may be compared to a full-health state, or a full-vision state (as we did in the present study). ³⁴ We chose to use full vision to allow a more precise comparison of the effect of ophthalmic treatments on quality of life while controlling for comorbid conditions, in line with the recommendation of the developer of the TTO method that “the utility of a health state should be unconfounded by utilities for other states that may or may not follow this one.” ⁴⁰ This approach is consistent with how utility values are measured for other conditions. Regardless of the nature of the health condition under consideration, in utilities-elicitation exercises, subjects are comparing target health states to an upper anchor state that describes the absence of that condition. In that sense, comparing visual impairment states to full vision (as we did here) could be considered analogous to a subject's comparing a cardiac health state to full health (the absence of any cardiac problems). Some investigators have found that, when they used different utilities-elicitation methods, the choice of the upper anchor affected the utility values. ³⁴ Although the findings of Lee et al. ³⁴ may affect the use of the present utility values for across-disease comparisons, it does not affect the use of the utility values presented herein for evaluations within ophthalmology. 

The health state classification scheme derived in this study is novel in the area of vision functioning and the results provide a larger number of discreet vision states with finer gradation than described previously. By applying these preference values, investigators can use patient responses to the selected items of the NEI-VFQ to complement the measurement of visual acuity or contrast sensitivity. This method may help to reduce the respondent burden in clinical trials by allowing investigators to collect some, but not all, of these (PRO, visual function, and preference) measurements. 

This study has limitations. We attempted to recruit participants with DR with various levels of visual impairment; however, consistent with the experience of Lloyd et al. ²⁴ in the United Kingdom, we had difficulty in identifying potential participants with poorer visual function. Contrast sensitivity and ETDRS acuity measures were obtained from participants in the qualitative phase, as they were thought to be fundamental to valid classification of responses during characterization of the health states. 

To better estimate the impact of DR on preferences and quality of life, we established standardized DR health states and used them in a rigorously conducted preferences-elicitation exercise among DR patients. The health state classification system provides a useful tool for categorizing DR patients in future studies. The patient-derived preference values recorded in this study provide a means of making comparisons within ophthalmology and of quality-adjusting life expectancy in economic evaluations of treatments for DR. 

Footnotes

 This study was carried out by Oxford Outcomes Ltd., a consultancy specializing in contract research for a wide range of clients in the life sciences industry, including public sector organizations as well as pharmaceutical and other private companies. The study described in this manuscript was conducted by Oxford Outcomes under contract to, and with funding provided by, Pfizer US, Inc.

Footnotes

 Disclosure: S.M. Szabo, Oxford Outcomes (F, E, C); K.M. Beusterien, Oxford Outcomes (F, E, C); A.M. Pleil, Pfizer (E); B. Wirostko, Pfizer (E); M.J. Potter (C); H. Tildesley (C); J. Gonder (C); A. Barsdorf, Pfizer (E); A.R. Levy, Oxford Outcomes (F, E, C)

Appendix

Health State Descriptions (All Visual Acuity Measures Are Expressed as Snellen Equivalents of ETDRS Acuity Scores)

Better eye, 20/20 to 20/40; worse eye, better than 20/200; contrast sensitivity, ≥21:

•  
  No difficulty in reading ordinary print in newspapers, with glasses as necessary.
•  
  No difficulty in driving in difficult conditions such as in bad weather, during rush hour, on the freeway, or in city traffic, because of your eyesight.
•  
  No difficulty in doing work or hobbies that require seeing well up close, such as cooking, sewing, fixing things around the house, or using hand tools.
•  
  No difficulty in taking part in active sports or other outdoor activities that you enjoy (e.g., golf, camping, exercising, biking, jogging, walking, or taking pictures), because of your eyesight.
•  
  No difficulty in reading street signs or the names of shops from a car or bus.
•  
  No difficulty in going down steps, stairs, or curbs in dim light or at night, because of your eyesight.

Better eye, 20/20 to 20/40; worse eye, better than 20/200; contrast sensitivity, ≥21:

•  
  A little difficulty in reading ordinary print in newspapers; with glasses or magnifier as necessary.
•  
  A little difficulty in driving in difficult conditions such as in bad weather, during rush hour, on the freeway, or in city traffic, because of your eyesight.
•  
  A little difficulty in doing work or hobbies that require seeing well up close, such as cooking, sewing, fixing things around the house, or using hand tools.
•  
  A little difficulty in taking part in active sports or other outdoor activities that you enjoy (e.g., golf, camping, exercising, biking, jogging, walking, or taking pictures), because of your eyesight.
•  
  A little difficulty in reading street signs or the names of shops from a car or bus.
•  
  No difficulty in going down steps, stairs, or curbs in dim light or at night, because of your eyesight.

Better eye, 20/20 to 20/40; worse eye, worse than 20/200; contrast sensitivity, <21:

•  
  Moderate difficulty in reading ordinary print in newspapers; may need magnifier.
•  
  Do not drive in difficult conditions such as in bad weather, during rush hour, on the freeway, or in city traffic, because of your eyesight.
•  
  Moderate difficulty in doing work or hobbies that require seeing well up close, such as cooking, sewing, fixing things around the house, or using hand tools.
•  
  Extreme difficulty in taking part in active sports or other outdoor activities that you enjoy (e.g., golf, camping, exercising, biking, jogging, walking, or taking pictures), because of your eyesight.
•  
  Moderate difficulty in reading street signs or the names of shops from a car or bus.
•  
  Moderate difficulty in going down steps, stairs, or curbs in dim light or at night, because of your eyesight.

Better eye, 20/20 to 20/40; worse eye, 20/200; contrast sensitivity, <21:

•  
  Moderate difficulty in reading ordinary print in newspapers; may need magnifier.
•  
  Extreme difficulty in driving in difficult conditions such as in bad weather, during rush hour, on the freeway, or in city traffic, because of your eyesight.
•  
  Moderate difficulty in doing work or hobbies that require seeing well up close, such as cooking, sewing, fixing things around the house, or using hand tools.
•  
  Moderate difficulty in taking part in active sports or other outdoor activities that you enjoy (e.g., golf, camping, exercising, biking, jogging, walking, or taking pictures), because of your eyesight.
•  
  No difficulty in reading street signs or the names of shops from a car or bus.
•  
  Moderate difficulty in going down steps, stairs, or curbs in dim light or at night, because of your eyesight.

Better eye, 20/50 to 20/80; worse eye, better than 20/200; contrast sensitivity, ≥21:

•  
  A little difficulty in reading ordinary print in newspapers; must use a magnifier.
•  
  Moderate difficulty in driving in difficult conditions such as in bad weather, during rush hour, on the freeway, or in city traffic, because of your eyesight.
•  
  No difficulty in doing work or hobbies that require seeing well up close, such as cooking, sewing, fixing things around the house, or using hand tools.
•  
  A little difficulty in taking part in active sports or other outdoor activities that you enjoy (e.g., golf, camping, exercising, biking, jogging, walking, or taking pictures), because of your eyesight.
•  
  Moderate difficulty in reading street signs or the names of shops from a car or bus.
•  
  Moderate difficulty in going down steps, stairs, or curbs in dim light or at night, because of your eyesight.

Better eye, 20/50 to 20/80; worse eye, better than 20/200; contrast sensitivity, ≥21:

•  
  Moderate difficulty in reading ordinary print in newspapers; must use a magnifier.
•  
  Extreme difficulty in driving in difficult conditions such as in bad weather, during rush hour, on the freeway, or in city traffic, because of your eyesight.
•  
  A little difficulty in doing work or hobbies that require seeing well up close, such as cooking, sewing, fixing things around the house, or using hand tools.
•  
  Moderate difficulty in taking part in active sports or other outdoor activities that you enjoy (e.g., golf, camping, exercising, biking, jogging, walking, or taking pictures), because of your eyesight.
•  
  Moderate difficulty in reading street signs or the names of shops from a car or bus.
•  
  Moderate difficulty in going down steps, stairs, or curbs in dim light or at night, because of your eyesight.

Better eye, 20/50 to 20/80; worse eye, worse than 20/200; contrast sensitivity, <21:

•  
  Extreme difficulty in reading ordinary print in newspapers; must use a magnifier.
•  
  Extreme difficulty in driving in difficult conditions such as in bad weather, during rush hour, on the freeway, or in city traffic, because of your eyesight.
•  
  Moderate difficulty in doing work or hobbies that require seeing well up close, such as cooking, sewing, fixing things around the house, or using hand tools.
•  
  Extreme difficulty in taking part in active sports or other outdoor activities that you enjoy (e.g., golf, camping, exercising, biking, jogging, walking, or taking pictures), because of your eyesight.
•  
  Extreme difficulty in reading street signs or the names of shops from a car or bus.
•  
  Moderate difficulty in going down steps, stairs, or curbs in dim light or at night, because of your eyesight.

Better eye, 20/50 to 20/80; worse eye, worse than 20/200; contrast sensitivity, <21:

•  
  Moderate difficulty in reading ordinary print in newspapers; must use a magnifier.
•  
  Extreme difficulty in driving in difficult conditions such as in bad weather, during rush hour, on the freeway, or in city traffic, because of your eyesight.
•  
  No difficulty in doing work or hobbies that require seeing well up close, such as cooking, sewing, fixing things around the house, or using hand tools.
•  
  Moderate difficulty in taking part in active sports or other outdoor activities that you enjoy (e.g., golf, camping, exercising, biking, jogging, walking, or taking pictures), because of your eyesight.
•  
  Extreme difficulty in reading street signs or the names of shops from a car or bus.
•  
  Moderate difficulty in going down steps, stairs, or curbs in dim light or at night, because of your eyesight.

Better eye, 20/100 to 20/160; worse eye, worse than 20/100; contrast sensitivity, <21:

•  
  Extreme difficulty in reading ordinary print in newspapers; must use a magnifier.
•  
  Cannot drive in difficult conditions such as in bad weather, during rush hour, on the freeway, or in city traffic, because of your eyesight.
•  
  Extreme difficulty in doing work or hobbies that require seeing well up close, such as cooking, sewing, fixing things around the house, or using hand tools.
•  
  Extreme difficulty in taking part in active sports or other outdoor activities that you enjoy (e.g., golf, camping, exercising, biking, jogging, walking, or taking pictures), because of your eyesight.
•  
  Cannot read street signs or the names of shops from a car or bus.
•  
  Extreme difficulty in going down steps, stairs, or curbs in dim light or at night, because of your eyesight.

Better eye, 20/200; worse eye, worse than 20/200; OR both eyes have contrast sensitivity <21:

•  
  Extreme difficulty in reading ordinary print in newspapers.
•  
  Cannot drive in difficult conditions such as in bad weather, during rush hour, on the freeway, or in city traffic, because of your eyesight.
•  
  Extreme difficulty in doing work or hobbies that require seeing well up close, such as cooking, sewing, fixing things around the house, or using hand tools.
•  
  Cannot take part in active sports or other outdoor activities that you enjoy (like golf, camping, exercising, biking, jogging, walking, or taking pictures), because of your eyesight.
•  
  Cannot read street signs or the names of shops from a car or bus.
•  
  Extreme difficulty in going down steps, stairs, or curbs in dim light or at night, because of your eyesight.