- Email: [email protected]
Predictors of Outcome in Patients Who Underwent Cataract Surgery
Predictors of Outcome in Patients Who Underwent Cataract Surgery Oliver D. Schein, MD, MPH,1 Earl P. Steinberg, MD, MPP,1,2 Sandra D. Cassard, ScD,2 James M. Tielsch, PhD,1 Jonathan C. Javitt, MD, MPH, 3 Alfred Sommer, MD, MHS 1,2 Purpose: To identify preoperative patient characteristics associated with a lack of improvement on one or more measures 4 months after cataract surgery. Methods: The authors collected preoperative and 4-month postoperative information on 552 patients undergoing first-eye cataract surgery from the practices of 72 ophthalmologists in three cities. The principal outcomes assessed were (1) Snellen visual acuity, (2) a cataract-related symptom score (possible range: 0, 0 of 6 symptoms present or bothersome, to 18, all 6 symptoms very bothersome), and (3) a measure of functional impairment in patients with cataract-the VF-14 score (possible range: 0, inability to perform any of the applicable activities, to 100, no difficulty performing any of the applicable activities). Multiple logistic regression was used to assess the association between preoperative patient characteristics and failure to improve on one or more outcome measures. Multiple linear regression was used to estimate the adjusted rate of lack of improvement in one or more outcome measures for one group of patients compared with another. Results: Although 91 patients (16.5%) failed to improve on one or more of the outcome measures assessed, only 2 (0.4%) failed to improve on all three measures. The 91 patients who did not improve on at least one measure were approximately one sixth as likely to be satisfied with their vision postoperatively as the 461 patients who improved on all three outcome measures. Preoperative age of 75 years of age or older, VF-14 score of 90 or higher, cataract symptom score of 3 or lower, and ocular comorbidity (glaucoma, diabetic retinopathy, or age-related macular degeneration) were associated independently with increased likelihood of not improving on one or more measure (odds ratio: 3.57, 2.10, 3.29, and 2.16, respectively). The mean adjusted rate of failure to improve on at least one of the outcome measures ranged from 20.5% to 26.5% for patients with these preoperative characteristics compared with 8.8% to 13.8% for those patients without them. The preoperative level of Snellen visual acuity was not associated with the likelihood of not improving on one or more of the outcomes assessed. Conclusions: The authors conclude that specific preoperative characteristics (age, comorbidity, cataract symptom score, and VF-14 score) are independent predictors of patient outcome after cataract surgery. Ophthalmology 1995;102:817-823
Originally received: September 22. 1994. Revision accepted: December 31. 1994. I The Johns Hopkins University School of Medicine. Baltimore. 2 The Johns Hopkins University School of Hygiene and Public Health, Baltimore. 3 Georgetown University School of Medicine, The Worthen Center for Eye Care Research, Washington, DC. Presented at the American Academy of Ophthalmology Annual Meeting, San Francisco, October/November 1994.
Supported by grant HS06280 from the Agency for Health Care Policy and Research, Rockville, Maryland. The Johns Hopkins University holds a copyright on the VF-14 and the Cataract Symptom Score. Were the Johns Hopkins University to receive any royalties related to future licensing of these instruments, then, under University policy, the inventors (includes all of the authors of this article) of those instruments would receive a share of those royalties. Reprint requests to Oliver D. Schein, MD, MPH, 116 Wilmer Bldg, Johns Hopkins Hospital, 600 North WolfeSt, Baltimore, MD 21287-9091.
817
Ophthalmology
Volume 102, Number 5, May 1995
Cataract surgery is highly effective. Recent data from the Cataract Patient Outcomes Research Team (PORT) documented improvement in visual function and satisfaction with vision 4 months postoperatively in approximately 90% of patients undergoing first-eye cataract surgery.l ? With an annual rate of cataract surgery in the United States in excess of I million, however, the PORT data suggest that more than 100,000 patients undergoing cataract surgery each year do not show improvement in these measures postoperatively. Historically, success rates for cataract surgery have been measured in terms of Snellen visual acuity (for example, the percent of patients achieving 20/40 or better visual acuity). This measure ofsuccess derives from an eligibility criterion for a driver's license in many states, but by itself it is a poor indicator of patient benefit or outcome for at least two reasons. First, given the current levels of visual acuity at which cataract surgery is performed (a median preoperative visual acuity in the surgical eye of 20/60 was observed in our national cohort, and 12.3% had 20/40 or better visual acuity preoperatively'), a result of 20/40 or better may not necessarily represent a clinically significant change in visual acuity or functional capability. Second, and more importantly, it has become increasingly recognized that Snellen visual acuity is a poor measure of functional impairment and patients' satisfaction with their vision .!" Practicing cataract surgeons, for example, have had the experience of performing uncomplicated surgery , obtaining improved visual acuity, and yet being faced with a patient who is dissatisfied with his/her vision. Conversely, some patients have improvements in visual acuity after surgery that do not meet the expectations of their surgeons, yet the patient is satisfied with the result. It is likely that such disparities between surgeons' and patients' perspectives reflect not only different expectations but also different operational definitions of a "good outcome." Both patients and ophthalmologists would benefit if it were possible to predict, preoperatively, which patients are least likely to achieve improvement in vision after surgery. Such information would aid in the selection of patients most likely to benefit from surgery and assist cataract surgeons and their patients in setting realistic expectations. The Cataract PORT prospectively has collected preoperative and postoperative data on a large cohort of patients who underwent first eye cataract surgery. These data include information on patients' overall health status, functional impairment potentially related to vision , satisfaction with vision , and symptoms potentially related to cataract, as well as visual acuity and surgical complications. We seek to identify preoperative patient characteristics associated with a lack of improvement on one or more measures 4 months after surgery.
Methods The study design and data collection methods have been described in detail elsewhere.l v In summary, three cities were chosen based on their annual rate ofcataract surgery
818
of Medicare beneficiaries: Columbus, Ohio (low rate) ; St. Louis, Missouri (medium rate); and Houston, Texas (high rate). Within each site, a random sample of ophthalmologists, stratified by their reported annual volume of cataract surgery, was recruited. The 75 ophthalmologists recruited were asked to enroll 14 consecutive patients meeting the following eligibility criteria: (I) patient seen on or after July, 15, 1991; (2) patient scheduled for first eye cataract surgery within 3 months of that visit; (3) patient was 50 years of age or older; and (4) no simultaneous glaucoma, corneal, or vitreoretinal procedures planned.
Major Outcome Measures Before cataract surgery , and again at approximately 4 months after surgery, enrolled patients underwent a detailed interview. The interview included (1) the VF-14 test, a reliable and valid index of functional impairment in patients with cataract.P VF-14 scores potentially range from 0 (inability to perform any of the applicable activities) to 100 (no difficulty performing any of the applicable activities); (2) a question related to patients' overall satisfaction with their vision (with response options of"very dissatisfied," " dissatisfied," "satisfied," or " very satisfied" ; (3) a set of six questions directed at eliciting symptoms potentially related to cataract (double vision , halos or glare, blurry vision, disturbing brightness, color distortion, worsening vision)." Those reporting such symptoms preoperatively or postoperatively were asked the degree to which the y were bothered by the symptoms ("not at all" [0 points], "a little" [I point], "somewhat" [2 points], and "very" [3 points]) . Based on a patient's response, a cataract symptom score was constructed for the patient. Possible scores ranged from 0 (0 of the 6 symptoms present or bothersome) to 18 (all 6 symptoms very bothersome). The participating surgeons completed standardized forms regarding patients' ocular history and examination preoperatively and approximately 4 months postoperatively, as well as a form regarding the surgery itself (completed within 48 hours after surgery). The study was approved by The Johns Hopkins University School of Hygiene and Public Health Committee on Human Research.
Statistical Analysis Associations between preoperative patient characteristics and patient outcome measures were examined initially using contingency tables with chi-square statistics or univariate logistic regression models. Multiple logistic regression analysis was used to assess the independent association between preoperative characteristics and failure to improve in one or more outcome measures at 4 months postoperatively. No improvement in any outcome was defined as no change or a worsening in (I) VF-14 score, (2) Snellen visual acuity, and (3) the cataract symptom score . Change in Snellen visual acuity was assessed by comparing the logarithms of preoperative versus postoperative minimum angles of resolution (IogMAR). 5 To mitigate the possibility of a ceiling effect on our measures (i.e., those with a "perfect" score on a measure preoper-
Schein et al . Cataract Surgery atively cannot improve on that measure), the 17 patients with no preoperative or postoperative dysfunction as measured by the VF-14 score, the I patient with preoperative and postoperative visual acuity of 20/20, and the 44 patients with no preoperative or postoperative cataract symptoms were considered in our analyses as having had an improvement in the VF-14 score, visual acuity, and cataract symptom score, respectively. Multiple linear regression analysis was used to estimate the adjusted rate oflack of improvement in one or more outcome measures for one group of patients compared with another.
VF-14
(n-46)
Results Complete baseline, perioperative , and 4-month follow-up information was available on 699 (91%) of766 patients. Five hundred fifty-two (79%) of these 699 patients had not had second eye surgery by the 4-month postoperative data collection. Table I presents summary preoperative and 4-month postoperative characteristics of these 552 patients .'>' Macular degeneration refers to the characteristic retinal findings in this condition and the opinion of the surgeon that the retinal changes observed were visually significant. Diabetic retinopathy refers to the presence of any degree of diabetic retinopathy. Glaucoma refers to any degree of glaucoma as diagnosed by the participating ophthalmologist. No change or a worsening of the Snellen visual acuity was observed in 24 patients (14.3%), whereas no change or worsening in VF-14 score and cataract symptom score was observed in 46 (8.3%) and 43 patients (7.8%), respectively. Figure I, a Venn diagram depicting the intersection of the three measures of lack of improvement,
Visual Acuity (n-24)
Figure 1. Venn diagram illustrates number and overlap of patients with no change or worsening in VF-14 score, cataract symptom score, and visual acuity.
illustrates that these three measures of lack of improvement are not congruent. Of the 91 patients who failed to improve on at least one of the measures, vision in only 2 (2.2%) failed to improve on all three measures. One of these two patients had 20/30 Snellen visual acuity preoperatively, and a postoperative macular hole developed. The other patient had 20/50 visual acuity preoperatively and had an intraoperative choroidal hemorrhage. Twenty patients (3.6%) showed lack of improvement on two or more outcome measures. Twelve (60.0%) of these 20 patients had a postoperative complication, whereas the rate of an adverse postoperative event in the remainder of the cohort was 15.4% (P < 0.000 I). Overall,
Table 1. Preoperative and Four-month Postoperative Patient Characteristics (n = 552) Patient Characteristic
Baseline
Age (yrs) (mean ± SD) Sex (% female) Ocular comorbidity (%) Macular degeneration Diabetic retinopathy Glaucoma Visual acuity (median) (range) Surgical eye Better eye VF-14 (mean ± SD) Six-item cataract symptom score (mean ± SD) Satisfaction with vision (%) Very satisfied Satisfied Dissatisfied Very dissatisfied
7Z ± 8 63
4Mos
10.0
4.7
10.9
20/60 (20/20-LP) 20/40 (20/20-HM) 76.5 ± 16.3
20/25 (20/15-CF) 20/25 (20/15-CF)
5.6 ± 4.1 0.9 10.7 64.8
23.6
93.2 ± 9.4 0.8 ± 1.9
45.7 42.4 9.8 2.2
SD = standard deviation ; LP = light perception; CF = counting fingers; HM = hand motions.
819
Ophthalmology
Volume 102, Number 5, May 1995
13 of these 20 patients (65.0%) reported being dissatisfied or very dissatisfiedwith their vision 4 months after surgery, compared with 53 (10.0%) of 532 patients (P < 0.0001) in the remainder of the cohort. Of the 91 patients (16.5%) who failed to improve on at least one of our three outcome measures, 27 (29.7%) had one or more adverse events after surgery, approximately double the rate (14.5%) of the 461 patients who showed some improvement on all three measures (P = 0.001). Thirty-one (34.1%) of the 91 patients who did not improve on one or more measures reported being dissatisfied or very dissatisfied with their vision 4 months after cataract surgery. The 9 L patients who showed no improvement on at least one of the three outcome measures were approximately one sixth as likely (age-adjusted odds ratio, 0.17; 95% confidence interval [CI], 0.10-0.30) to be satisfied or very satisfied with their vision 4 months postoperatively as the 461 patients who showed some improvement on all three outcome measures. To identify factors present preoperatively that are associated with a lack of improvement in at least one of the three outcome measures 4 months after cataract surgery, we performed a series of bivariate analyses (Table 2). Preoperative characteristics associated with a lower probability of postoperative improvement in at least one outcome measure include a VF-14 score of 90 or greater, a cataract symptom score of 3 or less, age older than 64 years, and presence of ocular comorbidity. Conversely, the presence of posterior subcapsular cataract was found to be associated with an increased likelihood of improvement in at least one outcome measure. No statistically significant association was found between a lack of improvement in one or more outcome measures and preoperative visual acuity in the surgical eye, sex, or medical comorbidity. To assess the independent association between preoperative patient characteristics and lack of improvement in the specific outcomes we considered, a multiple logistic regression analysis was performed. Patients with a preoperative VF-14 score of90 or greater (little preoperative dysfunction) were twice as likely (odds ratio, 2.10; 95% CI, 1.19-3.71) to have a lack of improvement in at least one outcome measure as those with more preoperative dysfunction as measured by the VF-14 score. Similarly, those with a low preoperative cataract symptom score (03) were approximately three times more likely (odds ratio, 3.29; 95% CI, 1.93-5.64) to show no improvement in at least one of the three outcome measures we assessed compared with those with higher scores (4-18). Age also was independently associated with the likelihood of having a lack of improvement in at least one of the three outcomes. Compared with those 50 to 64 years of age, those 65 to 74 years of age had a 2A-fold higher risk (odds ratio, 2.38; 95% CI, 0.85-6.65), and those 75 to 95 years of age had a 3.6-fold higher risk (odds ratio, 3.57; 95% CI, 1.2810.00) of not improving in at least one outcome assessed. Finally, those with glaucoma, diabetic retinopathy, or agerelated macular degeneration were approximately twice as likely (odds ratio, 2.16; 95% CI, 1.29-3.64) not to improve in at least one outcome as those with none of these
820
Table 2. Bivariate Associations between Preoperative Cataract Surgery Patient Characteristics and Lack of Improvement on One or More Outcome Measures" Variable VF-14 score [referent = Q, (25-66)]
o, 67-79
80-89 Q4 9O-1OO Visual acuity of surgical eye (referent = 20/40 or better) 20/50-20/70 20/80-20/100 20/200 or worse logMAR visual acuity (surgical eye) Cataract symptom score (referent = 0-3 points) 4-7 points 8-18 points Age (yrs) (referent = 50-64) 65-74 75-95 Female Any ocular comorbidity'] PSC cataract Medical comorbidity score[ Satisfied or very satisfied with vision ~
Odds Ratio
95%CI
0.86 1.43 2.85
0.41, 1.82 0.73,2.80 1.51,5.38
1.17 0.68 1.28 0.96
0.56, 1.72 0.25, 1.84 0.56,2.92 0.53, 1.72
0.24 0.23
0.13,0.43 0.13,0.43
2.20 3.53 0.67 2.18 0.58 0.99 1.34
0.89,5.43 1.46,8.52 0.43, 1.05 1.35,3.52 0.35,0.96 0.95,1.03 0.70,2.55
CI = confidence interval; Q = quartile; PSC = posterior subcapsular cataract. • No change or worsening in one or more VF-14, visual acuity, or cataract symptom score.
t t
Macular degeneration, glaucoma, diabetic retinopathy. See reference 3.
conditions. Preoperative visual acuity and medical comorbidity were unrelated to the risk of having no improvement. Females and those with posterior subcapsular cataract tended to be more likely to improve in all three outcome measures, although these two associations were not statistically significant once other characteristics had been controlled. We performed identical analyses for patients who failed to improve on two or more of the outcomes (n = 20). Because of limited statistical power, however, no factors were associated with this more stringent definition oflack of improvement at a level that was statistically significant. However, the direction and magnitude of the associations were similar to those observed in the principal analysis. Table 3 presents the mean adjusted rates of failing to improve in at least one of the outcome measures. These rates provide estimates of the likelihood of lack of postoperative improvement in at least one outcome measure based on preoperative status. Those with either a preoperative VF-14 score of90 or greater, a cataract symptom score of 3 or less, or with ocular comorbidity have an
Schein et al Table 3. Mean Adjusted Rates of Failing to Improve on at Least One Outcome Measure* Preoperative Characteristic VF-14 ~90
<90 Visual acuity of the logMAR surgical eye" Cataract symptom score ::;3 >3 Age (yrs) 50-64 65-74 75-95 Sex F M Other ocular cornorbldlty] No other ocular comorbidity PSC cataract No PSC cataract Medical comorbidity score
Mean Adjusted Rate (%)
95%CI
24.5 13.8
18.0, 31.1 10.4,17.3
-0.013 (0.039)
-0.089, 0.063
26.5 10.4
21.4,31.6 6.6, 14.2
8.8 15.1 20.5
0.9, 16.7 10.8, 19.5 15.7,25.2
14.3 19.9 24.4 13.8 14.1 17.5 0.003 (0.003)
10.5, 18.0 15.0,24.7 18.3,30.4 10.4,17.2 9.1, 19.1 13.9,21.2 -0.003, 0.008
CI = confidence interval; PSC = posterior subcapsular cataract. • For continuous variables, parameter estimates and standard errors are provided rather than adjusted rates .
t
Macular degeneration, glaucoma, diabetic retin opathy.
estimated rate of between 20% and 25% of failing to improve on one or more of the outcome measures assessed.
Discussion As concerns relating to healthcare expenditure have increased over recent years, so has interest in the effectiveness and value of medical interventions. This interest is now shared by the government and other insurers, employers, providers (particularly those paid on a capitated basis), and patients. Increased attention to the outcomes of medical care has led to the realization that measures of success may vary, depending on the measure and the beholder, and that the perceptions and function of patients are vital to meaningful definitions of medical outcomes . We examined three measures of outcome of cataract surgery-change in a standard clinical measure (Snellen visual acuity), change in patient self-report of his/her ability to perform various activities requiring vision (the VF14), and change in bothersome symptoms that might be caused by cataract and which might or might not result in functional impairment (cataract symptom score). A priori, one might expect a significant overlap among those patients who fail to improve on one versus another of these measures. However, our finding that 91 patients
Cataract Surgery failed to improve on one or more outcome measures but that only 2 patients failed to improve on all three outcome measures indicates that that is not the case. Patients may have an improvement in visual acuity, for example, without an improvement in either functional impairment or symptoms. This observation is consistent with our earlier findings that the correlation between these measures preoperatively' and the correlation between changes in the these measures postoperatively' are moderate at best. In our cohort, lack of improvement on one or more outcome measures was strongly associated with the occurrence of a postoperative complication; 42.9% of patients who showed no improvement in one or more of the outcome measures we examined had a complication , and 23.7% of patients who had a complication showed no improvement on at least one of the outcome measures we assessed. This finding is of limited value to cataract surgeons and their patients preoperatively, since, except for a few relatively uncommon anatomic features (e.g., pseudoexfoliation and high myopia), few patient characteristics are known to be associated with a high risk of complications of cataract surgery. The identification of more common patient characteristics which can be determined easily before surgery and which are associated with an increased risk of not improving postoperatively would be of great interest, however. Our analysis identified several preoperative patient characteristics that are independently associated with failure to improve on one or more of the three outcome measures we examined . One of these characteristics, ocular comorbidity (i.e., glaucoma, diabetic retinopathy, or age-related macular degeneration) , was anticipated, because it is logical that other diseases that impair visual function would be associated with a reduced likelihood of improvement of visual acuity and functional status. Three additional characteristicswere as strongly associated with failure to improve on one or more outcome measure as ocular comorbidity. First, age was strongly associated with this outcome, wherein patients 65 years of age or older were two to four times less likely to improve on all three outcomes compared with those 50 to 64 years of age. While the adjusted rate of not improving on all three measures (i.e., after controlling for ocular comorbidity, VF-14 score, and cataract-related symptom score) is 8.8% for those 50 to 64 years of age, it is 15.1 % for those 65 to 74 years of age and 20.5% for those 75 to 95 years of age. We hypothesize that this association may be due to a combination of factors, including increased unrecognized ocular comorbidit y in older patients, an age effect on best potential visual acuity, and a tendency for older patients to have and/or perceive more functional impairment whether or not it is related to their vision. Second, we found that patients with a low level of preoperative functional impairment, specifically a VF-14 score of90 or greater, had almost a threefold higher likelihood of not impro ving on one or more outcome measures compared with patients having more dysfunction (VF-14 score of 89 or lower) preoperatively. After adjusting for age, ocular comorbidity, and cataract symptom score, patients with a preoperati ve VF- [4 score of 90 or
821
Ophthalmology
Volume 102, Number 5, May 1995
higher had a 24.5% rate of not improving on at least one outcome measure, compared with a rate of 13.8%for those with preoperative VF-14 scores less than 90. This finding indicates that those patients who perceive minimal functional impairment before surgery are at increased risk for failing to show improvement in VF-14 score, cataract symptom score, or visual acuity. Third, we found that patients with fewer or less bothersome symptoms before surgery were more likely not to improve on one or more outcome measures. Specifically, patients with a cataract symptom score ofless than 4 preoperatively had approximately a threefold higher likelihood of having no improvement in one or more outcomes compared with those With higher preoperative symptom scores. Thus, as with functional impairment, patients who are less symptomatic from their cataract preoperatively are more likely to show no improvement in at least one outcome measure. Interestingly, no difference in the likelihood of not improving was observed for those with symptom scores of 4 to 7 compared with those with scores of 8 to 18. This suggests that those who are somewhat or very bothered by two symptoms (e.g., glare and blurry vision) (symptom score, 4 or 6) are as likely to show some type of improvement after surgery as those who are at least somewhat bothered by four or more symptoms (symptom score, ~8). The level of preoperative Snellen visual acuity was not associated with lack of improvement in the measures we assessed. Those with relatively good visual acuity (e.g., better than 20/40) preoperatively were as likely to improve on all three outcome measures as those with worse visual acuity. This finding has important implications regarding the limited value of Snellen visual acuity as a sole or primary criterion for the need for or appropriateness of cataract surgery. This finding supports the approach adopted by the Agency for Health Care Policy and Research Guideline Panel on Cataract in Adults, which stipulated that there is no Snellen visual acuity that can be used as a threshold for appropriateness of cataract surgery. 6 There are several caveats that should be taken into account in the interpretation of our data and in potential extrapolations from them. Most importantly, although we have identified preoperative characteristics associated with lack of improvement in at least one of the outcome measures we examined (Snellen visual acuity, the VF-14, and cataract symptom score), it is clear that benefits from (and indications for) cataract surgery do not require an expectation of improvement in all three measures we considered. For example, in our analyses, a patient with no postoperative improvement in visual acuity but who had reductions in functional impairment and cataract-related symptoms (e.g., glare) was classified as having failed to improve in at least one outcome measure in our primary analysis . As shown in Figure I, patients who fail to improve in one of the outcomes frequently improve in one or more of the other two outcomes we assessed. There is value , nonetheless, in considering the three outcomes in aggregate , because patients may expect to improve in all
822
of them. We also examined characteristics associated with failure to improve on any two of the three outcomes we considered, because this approach might be more likely to identify those with truly poor surgical outcomes. The direction and magnitude of the associations were similar to those observed in our primary analysis, thus supporting the robustness of the findings presented. A second caveat is that the approach used to define the presence of major ocular comorbidities was not strictly standardized. Therefore, we cannot differentiate the varying effectsof different levels of disease (e.g., diabetic macular edema versus diabetic ischemic retinopathy). However, because many of the patients who met our definition of ocular comorbidity (e.g., minimal diabetic retinopathy) in fact had good visual acuity, the reported association between ocular comorbidity and lack of improvement on one or more measures likely represents a conservative estimate of the true association. Third, there are potentially important outcomes of cataract surgery other than the three measures we addressed. In light of this limitation, and the fact that the VF-14 and cataract symptom score may not be sufficiently sensitive to measure some clinically important improvement, we may have misclassified some patients as not having improved when, in fact, they did improve in some way. On the other hand, because we did not require any minimal amount of improvement in our outcome measures for a patient to be considered improved, it is possible that we misclassified some patients as improved who did not necessarily have any clinically significant improvement. However, we are reassured by the relation of our measures oflack of improvement to patients' overall satisfaction with their vision, wherein the unadjusted rate of being satisfied with one's vision 4 months after cataract surgery decreased from 92.4% to 74.6% to 35.0% for patients with lack of improvement on no, one, and two outcome measures, respectively. In summary, we have shown that specificpreoperative patient characteristics are predictive of lack of improvement in one or more pertinent outcomes of cataract surgery. Preoperative visual acuity was found to be unrelated to the likelihood of improvement, a finding that has important implications for constructing profiles of when cataract surgery is appropriate, as well as for patient education. Two measures derived from patient questionnaires, the VF-14, and an index based on cataract-related symptoms are independent predictors of outcome. Such measures, or their analogues, are important patient characteristics that are relevant to assessing outcomes from cataract surgery and may be of benefit to physicians and patients in decision making regarding the need for cataract surgery.
References 1. Schein 00, Steinberg EP, Javitt .IC, et a!. Variation in cataract surgery practice and clinical outcomes. Ophthalmology 1994; 10 I: 1142-52.
Schein et al . Cataract Surgery 2. Steinberg EP, Tielsch JM , Schein 00, et al. National study of cataract surgery outcomes. Variation in 4-month postoperative outcomes as reflected in multiple outcome measures. Ophthalmology 1994;I0 I:113I-4 I. 3. Steinberg EP, Tielsch JM , Schein 00, et al. The VF-14: an index of functional impairment in patients with cataract. Arch Ophthalmol 1994;112:630-8. 4. Mangione CM, Phillips RS, Seddon JM, et al. Development
of the Activities of Daily Vision Scale. A measure of visual functional status. Med Care 1992;30: J J J 1-26. 5. Westheimer G. Scaling of visual acuity measurements. Arch Ophthalmol 1979;97:327-30. 6. Cataract Management Guideline Panel. Cataract in Adults: Management of Functional Impairment. Rockville, MD: U.S. Department of Health and Human Services, 1993. (AHCPR Pub!. No. 93-0542).
823
Originally received: September 22. 1994. Revision accepted: December 31. 1994. I The Johns Hopkins University School of Medicine. Baltimore. 2 The Johns Hopkins University School of Hygiene and Public Health, Baltimore. 3 Georgetown University School of Medicine, The Worthen Center for Eye Care Research, Washington, DC. Presented at the American Academy of Ophthalmology Annual Meeting, San Francisco, October/November 1994.
Supported by grant HS06280 from the Agency for Health Care Policy and Research, Rockville, Maryland. The Johns Hopkins University holds a copyright on the VF-14 and the Cataract Symptom Score. Were the Johns Hopkins University to receive any royalties related to future licensing of these instruments, then, under University policy, the inventors (includes all of the authors of this article) of those instruments would receive a share of those royalties. Reprint requests to Oliver D. Schein, MD, MPH, 116 Wilmer Bldg, Johns Hopkins Hospital, 600 North WolfeSt, Baltimore, MD 21287-9091.
817
Ophthalmology
Volume 102, Number 5, May 1995
Cataract surgery is highly effective. Recent data from the Cataract Patient Outcomes Research Team (PORT) documented improvement in visual function and satisfaction with vision 4 months postoperatively in approximately 90% of patients undergoing first-eye cataract surgery.l ? With an annual rate of cataract surgery in the United States in excess of I million, however, the PORT data suggest that more than 100,000 patients undergoing cataract surgery each year do not show improvement in these measures postoperatively. Historically, success rates for cataract surgery have been measured in terms of Snellen visual acuity (for example, the percent of patients achieving 20/40 or better visual acuity). This measure ofsuccess derives from an eligibility criterion for a driver's license in many states, but by itself it is a poor indicator of patient benefit or outcome for at least two reasons. First, given the current levels of visual acuity at which cataract surgery is performed (a median preoperative visual acuity in the surgical eye of 20/60 was observed in our national cohort, and 12.3% had 20/40 or better visual acuity preoperatively'), a result of 20/40 or better may not necessarily represent a clinically significant change in visual acuity or functional capability. Second, and more importantly, it has become increasingly recognized that Snellen visual acuity is a poor measure of functional impairment and patients' satisfaction with their vision .!" Practicing cataract surgeons, for example, have had the experience of performing uncomplicated surgery , obtaining improved visual acuity, and yet being faced with a patient who is dissatisfied with his/her vision. Conversely, some patients have improvements in visual acuity after surgery that do not meet the expectations of their surgeons, yet the patient is satisfied with the result. It is likely that such disparities between surgeons' and patients' perspectives reflect not only different expectations but also different operational definitions of a "good outcome." Both patients and ophthalmologists would benefit if it were possible to predict, preoperatively, which patients are least likely to achieve improvement in vision after surgery. Such information would aid in the selection of patients most likely to benefit from surgery and assist cataract surgeons and their patients in setting realistic expectations. The Cataract PORT prospectively has collected preoperative and postoperative data on a large cohort of patients who underwent first eye cataract surgery. These data include information on patients' overall health status, functional impairment potentially related to vision , satisfaction with vision , and symptoms potentially related to cataract, as well as visual acuity and surgical complications. We seek to identify preoperative patient characteristics associated with a lack of improvement on one or more measures 4 months after surgery.
Methods The study design and data collection methods have been described in detail elsewhere.l v In summary, three cities were chosen based on their annual rate ofcataract surgery
818
of Medicare beneficiaries: Columbus, Ohio (low rate) ; St. Louis, Missouri (medium rate); and Houston, Texas (high rate). Within each site, a random sample of ophthalmologists, stratified by their reported annual volume of cataract surgery, was recruited. The 75 ophthalmologists recruited were asked to enroll 14 consecutive patients meeting the following eligibility criteria: (I) patient seen on or after July, 15, 1991; (2) patient scheduled for first eye cataract surgery within 3 months of that visit; (3) patient was 50 years of age or older; and (4) no simultaneous glaucoma, corneal, or vitreoretinal procedures planned.
Major Outcome Measures Before cataract surgery , and again at approximately 4 months after surgery, enrolled patients underwent a detailed interview. The interview included (1) the VF-14 test, a reliable and valid index of functional impairment in patients with cataract.P VF-14 scores potentially range from 0 (inability to perform any of the applicable activities) to 100 (no difficulty performing any of the applicable activities); (2) a question related to patients' overall satisfaction with their vision (with response options of"very dissatisfied," " dissatisfied," "satisfied," or " very satisfied" ; (3) a set of six questions directed at eliciting symptoms potentially related to cataract (double vision , halos or glare, blurry vision, disturbing brightness, color distortion, worsening vision)." Those reporting such symptoms preoperatively or postoperatively were asked the degree to which the y were bothered by the symptoms ("not at all" [0 points], "a little" [I point], "somewhat" [2 points], and "very" [3 points]) . Based on a patient's response, a cataract symptom score was constructed for the patient. Possible scores ranged from 0 (0 of the 6 symptoms present or bothersome) to 18 (all 6 symptoms very bothersome). The participating surgeons completed standardized forms regarding patients' ocular history and examination preoperatively and approximately 4 months postoperatively, as well as a form regarding the surgery itself (completed within 48 hours after surgery). The study was approved by The Johns Hopkins University School of Hygiene and Public Health Committee on Human Research.
Statistical Analysis Associations between preoperative patient characteristics and patient outcome measures were examined initially using contingency tables with chi-square statistics or univariate logistic regression models. Multiple logistic regression analysis was used to assess the independent association between preoperative characteristics and failure to improve in one or more outcome measures at 4 months postoperatively. No improvement in any outcome was defined as no change or a worsening in (I) VF-14 score, (2) Snellen visual acuity, and (3) the cataract symptom score . Change in Snellen visual acuity was assessed by comparing the logarithms of preoperative versus postoperative minimum angles of resolution (IogMAR). 5 To mitigate the possibility of a ceiling effect on our measures (i.e., those with a "perfect" score on a measure preoper-
Schein et al . Cataract Surgery atively cannot improve on that measure), the 17 patients with no preoperative or postoperative dysfunction as measured by the VF-14 score, the I patient with preoperative and postoperative visual acuity of 20/20, and the 44 patients with no preoperative or postoperative cataract symptoms were considered in our analyses as having had an improvement in the VF-14 score, visual acuity, and cataract symptom score, respectively. Multiple linear regression analysis was used to estimate the adjusted rate oflack of improvement in one or more outcome measures for one group of patients compared with another.
VF-14
(n-46)
Results Complete baseline, perioperative , and 4-month follow-up information was available on 699 (91%) of766 patients. Five hundred fifty-two (79%) of these 699 patients had not had second eye surgery by the 4-month postoperative data collection. Table I presents summary preoperative and 4-month postoperative characteristics of these 552 patients .'>' Macular degeneration refers to the characteristic retinal findings in this condition and the opinion of the surgeon that the retinal changes observed were visually significant. Diabetic retinopathy refers to the presence of any degree of diabetic retinopathy. Glaucoma refers to any degree of glaucoma as diagnosed by the participating ophthalmologist. No change or a worsening of the Snellen visual acuity was observed in 24 patients (14.3%), whereas no change or worsening in VF-14 score and cataract symptom score was observed in 46 (8.3%) and 43 patients (7.8%), respectively. Figure I, a Venn diagram depicting the intersection of the three measures of lack of improvement,
Visual Acuity (n-24)
Figure 1. Venn diagram illustrates number and overlap of patients with no change or worsening in VF-14 score, cataract symptom score, and visual acuity.
illustrates that these three measures of lack of improvement are not congruent. Of the 91 patients who failed to improve on at least one of the measures, vision in only 2 (2.2%) failed to improve on all three measures. One of these two patients had 20/30 Snellen visual acuity preoperatively, and a postoperative macular hole developed. The other patient had 20/50 visual acuity preoperatively and had an intraoperative choroidal hemorrhage. Twenty patients (3.6%) showed lack of improvement on two or more outcome measures. Twelve (60.0%) of these 20 patients had a postoperative complication, whereas the rate of an adverse postoperative event in the remainder of the cohort was 15.4% (P < 0.000 I). Overall,
Table 1. Preoperative and Four-month Postoperative Patient Characteristics (n = 552) Patient Characteristic
Baseline
Age (yrs) (mean ± SD) Sex (% female) Ocular comorbidity (%) Macular degeneration Diabetic retinopathy Glaucoma Visual acuity (median) (range) Surgical eye Better eye VF-14 (mean ± SD) Six-item cataract symptom score (mean ± SD) Satisfaction with vision (%) Very satisfied Satisfied Dissatisfied Very dissatisfied
7Z ± 8 63
4Mos
10.0
4.7
10.9
20/60 (20/20-LP) 20/40 (20/20-HM) 76.5 ± 16.3
20/25 (20/15-CF) 20/25 (20/15-CF)
5.6 ± 4.1 0.9 10.7 64.8
23.6
93.2 ± 9.4 0.8 ± 1.9
45.7 42.4 9.8 2.2
SD = standard deviation ; LP = light perception; CF = counting fingers; HM = hand motions.
819
Ophthalmology
Volume 102, Number 5, May 1995
13 of these 20 patients (65.0%) reported being dissatisfied or very dissatisfiedwith their vision 4 months after surgery, compared with 53 (10.0%) of 532 patients (P < 0.0001) in the remainder of the cohort. Of the 91 patients (16.5%) who failed to improve on at least one of our three outcome measures, 27 (29.7%) had one or more adverse events after surgery, approximately double the rate (14.5%) of the 461 patients who showed some improvement on all three measures (P = 0.001). Thirty-one (34.1%) of the 91 patients who did not improve on one or more measures reported being dissatisfied or very dissatisfied with their vision 4 months after cataract surgery. The 9 L patients who showed no improvement on at least one of the three outcome measures were approximately one sixth as likely (age-adjusted odds ratio, 0.17; 95% confidence interval [CI], 0.10-0.30) to be satisfied or very satisfied with their vision 4 months postoperatively as the 461 patients who showed some improvement on all three outcome measures. To identify factors present preoperatively that are associated with a lack of improvement in at least one of the three outcome measures 4 months after cataract surgery, we performed a series of bivariate analyses (Table 2). Preoperative characteristics associated with a lower probability of postoperative improvement in at least one outcome measure include a VF-14 score of 90 or greater, a cataract symptom score of 3 or less, age older than 64 years, and presence of ocular comorbidity. Conversely, the presence of posterior subcapsular cataract was found to be associated with an increased likelihood of improvement in at least one outcome measure. No statistically significant association was found between a lack of improvement in one or more outcome measures and preoperative visual acuity in the surgical eye, sex, or medical comorbidity. To assess the independent association between preoperative patient characteristics and lack of improvement in the specific outcomes we considered, a multiple logistic regression analysis was performed. Patients with a preoperative VF-14 score of90 or greater (little preoperative dysfunction) were twice as likely (odds ratio, 2.10; 95% CI, 1.19-3.71) to have a lack of improvement in at least one outcome measure as those with more preoperative dysfunction as measured by the VF-14 score. Similarly, those with a low preoperative cataract symptom score (03) were approximately three times more likely (odds ratio, 3.29; 95% CI, 1.93-5.64) to show no improvement in at least one of the three outcome measures we assessed compared with those with higher scores (4-18). Age also was independently associated with the likelihood of having a lack of improvement in at least one of the three outcomes. Compared with those 50 to 64 years of age, those 65 to 74 years of age had a 2A-fold higher risk (odds ratio, 2.38; 95% CI, 0.85-6.65), and those 75 to 95 years of age had a 3.6-fold higher risk (odds ratio, 3.57; 95% CI, 1.2810.00) of not improving in at least one outcome assessed. Finally, those with glaucoma, diabetic retinopathy, or agerelated macular degeneration were approximately twice as likely (odds ratio, 2.16; 95% CI, 1.29-3.64) not to improve in at least one outcome as those with none of these
820
Table 2. Bivariate Associations between Preoperative Cataract Surgery Patient Characteristics and Lack of Improvement on One or More Outcome Measures" Variable VF-14 score [referent = Q, (25-66)]
o, 67-79
80-89 Q4 9O-1OO Visual acuity of surgical eye (referent = 20/40 or better) 20/50-20/70 20/80-20/100 20/200 or worse logMAR visual acuity (surgical eye) Cataract symptom score (referent = 0-3 points) 4-7 points 8-18 points Age (yrs) (referent = 50-64) 65-74 75-95 Female Any ocular comorbidity'] PSC cataract Medical comorbidity score[ Satisfied or very satisfied with vision ~
Odds Ratio
95%CI
0.86 1.43 2.85
0.41, 1.82 0.73,2.80 1.51,5.38
1.17 0.68 1.28 0.96
0.56, 1.72 0.25, 1.84 0.56,2.92 0.53, 1.72
0.24 0.23
0.13,0.43 0.13,0.43
2.20 3.53 0.67 2.18 0.58 0.99 1.34
0.89,5.43 1.46,8.52 0.43, 1.05 1.35,3.52 0.35,0.96 0.95,1.03 0.70,2.55
CI = confidence interval; Q = quartile; PSC = posterior subcapsular cataract. • No change or worsening in one or more VF-14, visual acuity, or cataract symptom score.
t t
Macular degeneration, glaucoma, diabetic retinopathy. See reference 3.
conditions. Preoperative visual acuity and medical comorbidity were unrelated to the risk of having no improvement. Females and those with posterior subcapsular cataract tended to be more likely to improve in all three outcome measures, although these two associations were not statistically significant once other characteristics had been controlled. We performed identical analyses for patients who failed to improve on two or more of the outcomes (n = 20). Because of limited statistical power, however, no factors were associated with this more stringent definition oflack of improvement at a level that was statistically significant. However, the direction and magnitude of the associations were similar to those observed in the principal analysis. Table 3 presents the mean adjusted rates of failing to improve in at least one of the outcome measures. These rates provide estimates of the likelihood of lack of postoperative improvement in at least one outcome measure based on preoperative status. Those with either a preoperative VF-14 score of90 or greater, a cataract symptom score of 3 or less, or with ocular comorbidity have an
Schein et al Table 3. Mean Adjusted Rates of Failing to Improve on at Least One Outcome Measure* Preoperative Characteristic VF-14 ~90
<90 Visual acuity of the logMAR surgical eye" Cataract symptom score ::;3 >3 Age (yrs) 50-64 65-74 75-95 Sex F M Other ocular cornorbldlty] No other ocular comorbidity PSC cataract No PSC cataract Medical comorbidity score
Mean Adjusted Rate (%)
95%CI
24.5 13.8
18.0, 31.1 10.4,17.3
-0.013 (0.039)
-0.089, 0.063
26.5 10.4
21.4,31.6 6.6, 14.2
8.8 15.1 20.5
0.9, 16.7 10.8, 19.5 15.7,25.2
14.3 19.9 24.4 13.8 14.1 17.5 0.003 (0.003)
10.5, 18.0 15.0,24.7 18.3,30.4 10.4,17.2 9.1, 19.1 13.9,21.2 -0.003, 0.008
CI = confidence interval; PSC = posterior subcapsular cataract. • For continuous variables, parameter estimates and standard errors are provided rather than adjusted rates .
t
Macular degeneration, glaucoma, diabetic retin opathy.
estimated rate of between 20% and 25% of failing to improve on one or more of the outcome measures assessed.
Discussion As concerns relating to healthcare expenditure have increased over recent years, so has interest in the effectiveness and value of medical interventions. This interest is now shared by the government and other insurers, employers, providers (particularly those paid on a capitated basis), and patients. Increased attention to the outcomes of medical care has led to the realization that measures of success may vary, depending on the measure and the beholder, and that the perceptions and function of patients are vital to meaningful definitions of medical outcomes . We examined three measures of outcome of cataract surgery-change in a standard clinical measure (Snellen visual acuity), change in patient self-report of his/her ability to perform various activities requiring vision (the VF14), and change in bothersome symptoms that might be caused by cataract and which might or might not result in functional impairment (cataract symptom score). A priori, one might expect a significant overlap among those patients who fail to improve on one versus another of these measures. However, our finding that 91 patients
Cataract Surgery failed to improve on one or more outcome measures but that only 2 patients failed to improve on all three outcome measures indicates that that is not the case. Patients may have an improvement in visual acuity, for example, without an improvement in either functional impairment or symptoms. This observation is consistent with our earlier findings that the correlation between these measures preoperatively' and the correlation between changes in the these measures postoperatively' are moderate at best. In our cohort, lack of improvement on one or more outcome measures was strongly associated with the occurrence of a postoperative complication; 42.9% of patients who showed no improvement in one or more of the outcome measures we examined had a complication , and 23.7% of patients who had a complication showed no improvement on at least one of the outcome measures we assessed. This finding is of limited value to cataract surgeons and their patients preoperatively, since, except for a few relatively uncommon anatomic features (e.g., pseudoexfoliation and high myopia), few patient characteristics are known to be associated with a high risk of complications of cataract surgery. The identification of more common patient characteristics which can be determined easily before surgery and which are associated with an increased risk of not improving postoperatively would be of great interest, however. Our analysis identified several preoperative patient characteristics that are independently associated with failure to improve on one or more of the three outcome measures we examined . One of these characteristics, ocular comorbidity (i.e., glaucoma, diabetic retinopathy, or age-related macular degeneration) , was anticipated, because it is logical that other diseases that impair visual function would be associated with a reduced likelihood of improvement of visual acuity and functional status. Three additional characteristicswere as strongly associated with failure to improve on one or more outcome measure as ocular comorbidity. First, age was strongly associated with this outcome, wherein patients 65 years of age or older were two to four times less likely to improve on all three outcomes compared with those 50 to 64 years of age. While the adjusted rate of not improving on all three measures (i.e., after controlling for ocular comorbidity, VF-14 score, and cataract-related symptom score) is 8.8% for those 50 to 64 years of age, it is 15.1 % for those 65 to 74 years of age and 20.5% for those 75 to 95 years of age. We hypothesize that this association may be due to a combination of factors, including increased unrecognized ocular comorbidit y in older patients, an age effect on best potential visual acuity, and a tendency for older patients to have and/or perceive more functional impairment whether or not it is related to their vision. Second, we found that patients with a low level of preoperative functional impairment, specifically a VF-14 score of90 or greater, had almost a threefold higher likelihood of not impro ving on one or more outcome measures compared with patients having more dysfunction (VF-14 score of 89 or lower) preoperatively. After adjusting for age, ocular comorbidity, and cataract symptom score, patients with a preoperati ve VF- [4 score of 90 or
821
Ophthalmology
Volume 102, Number 5, May 1995
higher had a 24.5% rate of not improving on at least one outcome measure, compared with a rate of 13.8%for those with preoperative VF-14 scores less than 90. This finding indicates that those patients who perceive minimal functional impairment before surgery are at increased risk for failing to show improvement in VF-14 score, cataract symptom score, or visual acuity. Third, we found that patients with fewer or less bothersome symptoms before surgery were more likely not to improve on one or more outcome measures. Specifically, patients with a cataract symptom score ofless than 4 preoperatively had approximately a threefold higher likelihood of having no improvement in one or more outcomes compared with those With higher preoperative symptom scores. Thus, as with functional impairment, patients who are less symptomatic from their cataract preoperatively are more likely to show no improvement in at least one outcome measure. Interestingly, no difference in the likelihood of not improving was observed for those with symptom scores of 4 to 7 compared with those with scores of 8 to 18. This suggests that those who are somewhat or very bothered by two symptoms (e.g., glare and blurry vision) (symptom score, 4 or 6) are as likely to show some type of improvement after surgery as those who are at least somewhat bothered by four or more symptoms (symptom score, ~8). The level of preoperative Snellen visual acuity was not associated with lack of improvement in the measures we assessed. Those with relatively good visual acuity (e.g., better than 20/40) preoperatively were as likely to improve on all three outcome measures as those with worse visual acuity. This finding has important implications regarding the limited value of Snellen visual acuity as a sole or primary criterion for the need for or appropriateness of cataract surgery. This finding supports the approach adopted by the Agency for Health Care Policy and Research Guideline Panel on Cataract in Adults, which stipulated that there is no Snellen visual acuity that can be used as a threshold for appropriateness of cataract surgery. 6 There are several caveats that should be taken into account in the interpretation of our data and in potential extrapolations from them. Most importantly, although we have identified preoperative characteristics associated with lack of improvement in at least one of the outcome measures we examined (Snellen visual acuity, the VF-14, and cataract symptom score), it is clear that benefits from (and indications for) cataract surgery do not require an expectation of improvement in all three measures we considered. For example, in our analyses, a patient with no postoperative improvement in visual acuity but who had reductions in functional impairment and cataract-related symptoms (e.g., glare) was classified as having failed to improve in at least one outcome measure in our primary analysis . As shown in Figure I, patients who fail to improve in one of the outcomes frequently improve in one or more of the other two outcomes we assessed. There is value , nonetheless, in considering the three outcomes in aggregate , because patients may expect to improve in all
822
of them. We also examined characteristics associated with failure to improve on any two of the three outcomes we considered, because this approach might be more likely to identify those with truly poor surgical outcomes. The direction and magnitude of the associations were similar to those observed in our primary analysis, thus supporting the robustness of the findings presented. A second caveat is that the approach used to define the presence of major ocular comorbidities was not strictly standardized. Therefore, we cannot differentiate the varying effectsof different levels of disease (e.g., diabetic macular edema versus diabetic ischemic retinopathy). However, because many of the patients who met our definition of ocular comorbidity (e.g., minimal diabetic retinopathy) in fact had good visual acuity, the reported association between ocular comorbidity and lack of improvement on one or more measures likely represents a conservative estimate of the true association. Third, there are potentially important outcomes of cataract surgery other than the three measures we addressed. In light of this limitation, and the fact that the VF-14 and cataract symptom score may not be sufficiently sensitive to measure some clinically important improvement, we may have misclassified some patients as not having improved when, in fact, they did improve in some way. On the other hand, because we did not require any minimal amount of improvement in our outcome measures for a patient to be considered improved, it is possible that we misclassified some patients as improved who did not necessarily have any clinically significant improvement. However, we are reassured by the relation of our measures oflack of improvement to patients' overall satisfaction with their vision, wherein the unadjusted rate of being satisfied with one's vision 4 months after cataract surgery decreased from 92.4% to 74.6% to 35.0% for patients with lack of improvement on no, one, and two outcome measures, respectively. In summary, we have shown that specificpreoperative patient characteristics are predictive of lack of improvement in one or more pertinent outcomes of cataract surgery. Preoperative visual acuity was found to be unrelated to the likelihood of improvement, a finding that has important implications for constructing profiles of when cataract surgery is appropriate, as well as for patient education. Two measures derived from patient questionnaires, the VF-14, and an index based on cataract-related symptoms are independent predictors of outcome. Such measures, or their analogues, are important patient characteristics that are relevant to assessing outcomes from cataract surgery and may be of benefit to physicians and patients in decision making regarding the need for cataract surgery.
References 1. Schein 00, Steinberg EP, Javitt .IC, et a!. Variation in cataract surgery practice and clinical outcomes. Ophthalmology 1994; 10 I: 1142-52.
Schein et al . Cataract Surgery 2. Steinberg EP, Tielsch JM , Schein 00, et al. National study of cataract surgery outcomes. Variation in 4-month postoperative outcomes as reflected in multiple outcome measures. Ophthalmology 1994;I0 I:113I-4 I. 3. Steinberg EP, Tielsch JM , Schein 00, et al. The VF-14: an index of functional impairment in patients with cataract. Arch Ophthalmol 1994;112:630-8. 4. Mangione CM, Phillips RS, Seddon JM, et al. Development
of the Activities of Daily Vision Scale. A measure of visual functional status. Med Care 1992;30: J J J 1-26. 5. Westheimer G. Scaling of visual acuity measurements. Arch Ophthalmol 1979;97:327-30. 6. Cataract Management Guideline Panel. Cataract in Adults: Management of Functional Impairment. Rockville, MD: U.S. Department of Health and Human Services, 1993. (AHCPR Pub!. No. 93-0542).
823