To determine the minimal clinically important difference and the patient acceptable symptom state for visual acuity and visual function, 2 key visual outcomes for patients undergoing cataract extraction, as an aid for evaluating the outcome of cataract extraction.
Prospective cohort study.
setting: Multicenter study of 17 hospitals. patients: A total of 4335 consecutive patients advised to undergo cataract extraction. main outcome measures: Sociodemographic and clinical data, including visual acuity (VA), patient satisfaction, and responses to the Visual Function Index 14 (VF-14) and transitional questions were collected before and after cataract extraction. The patient acceptable symptom state and minimal clinically important difference were estimated for the entire sample and for subgroups by preintervention status and presence of ocular comorbidities.
Among patients with simple cataract, postintervention patient acceptable symptom state values for VA ranged from 0.67 to 0.80, depending on preintervention VA, while VF-14 scores ranged from 88 to 90. For patients with any additional ocular comorbidity, VA ranged from 0.63 to 0.75, depending on preintervention VA, while VF-14 scores ranged from 86 to 92. For the entire sample, postintervention patient acceptable symptom state values were 0.75 for VA and 86.1 for VF-14 scores. Minimal clinically important difference for patients who reported being a little better ranged from 0.17 to 0.5 in VA and 0.41 to 37.46 in VF-14 scores, depending on preintervention status and presence of ocular comorbidities.
The minimal clinically important difference and patient acceptable symptom state values are complementary parameters that may help in the clinical decision-making process by providing more meaningful estimates of the impact of cataract extraction on 2 important outcomes, visual acuity and visual function.
Although cataract extraction is the most frequently performed surgical intervention in many developed countries, how best to interpret outcomes of this procedure is still an open question. Visual acuity and other clinical measures have traditionally been used to evaluate the success of cataract extraction. In recent years, patients’ perceptions of their conditions have been given increasing importance in defining effective outcomes, influencing clinical decision making, and evaluating health-related quality of life (HRQoL) outcomes. Patient-reported outcomes (PROs) address fundamental values of a treatment over and above its effect on measurable, objective parameters of health. Patients have a unique perspective on the impact of a treatment since they experience first-hand its benefits and burdens. Asking patients about their perspectives has the potential to yield clinically meaningful endpoints that are an integral part of clinical decision making. PROs are generally measured using questionnaires completed by the patient.
One widely used, disease-specific PRO is the Visual Function 14 (VF-14) score, which offers an estimate of the patient’s perspective of his or her visual function and how it can be affected by cataracts and their removal. To help interpret changes in individual patient scores, researchers have developed the minimal clinically important difference and the minimal clinically important improvement. Both represent the proportion of patients reporting improvement following an intervention. Another potentially clinically relevant measure is the patient acceptable symptom state, defined as the value beyond which patients consider themselves well. While the minimal clinically important difference encompasses the concept of improvement (feeling better), the patient acceptable symptom state measures patient perception of well-being or remission of symptoms (feeling good), which is undoubtedly a clinically relevant outcome. The minimal clinically important difference and patient acceptable symptom state concepts are complementary. To date, the patient acceptable symptom state term has been used primarily in rheumatology and orthopedics.
This prospective cohort study was designed to determine the patient acceptable symptom state and minimal clinically important difference for 2 key visual outcomes for patients undergoing cataract extraction: visual acuity (VA) and visual function (VF). Data are presented for the entire cohort and for subgroups based on baseline VA status and the presence of ocular comorbidities, factors that may influence outcomes of cataract extraction.
We recruited consecutive patients from 17 hospitals in different regions of Spain who were advised to have cataract removal by phacoemulsification between October 1, 2004 and July 30, 2005. All the hospitals participating in this study belong to the Spanish National Health Service; all are public hospitals and shared similar technical and human resources. To be included in the study, patients had to be older than 18 years of age and provide informed consent. Patients were excluded if they underwent additional intraocular procedures, had malignant processes or serious organic or psychiatric disorders that would prevent them from collaborating in the study, or had serious difficulties with reading or understanding that rendered them unable to participate or to complete the questionnaires. Personal data were collected only for tracking patients during the study period and all data were kept confidential.
Clinical data were collected during the visit prior to cataract extraction and approximately 6 weeks after the surgery by ophthalmologists and nurses collaborating in the study. These data included sociodemographic information (sex, age, level of education, marital status, social support), preintervention clinical data (VA, laterality of cataract, contralateral VA, ocular pathology, visual function, and technical complexity of the cataract extraction), data related to surgical technique, and information regarding complications during and immediately after the cataract extraction as well as up to 6 weeks after surgery. VA in each eye and expected postoperative visual acuity were determined by the ophthalmologist at the preintervention visit when the patient was added to the waiting list for cataract extraction. VA was evaluated again at 6 weeks after surgery. In all cases, each eye was evaluated separately using Snellen eye charts; the best-corrected VA was recorded in decimal units.
The VF-14 questionnaire is a validated index that measures the difficulty related to performing 14 vision-dependent activities of daily living, such as daytime and nighttime driving, reading traffic signs or small print, and engaging in recreational activities. Scores on the VF-14 range from 0 (unable to do any activities) to 100 (able to do all activities without difficulty). The VF-14 has been translated and validated in Spanish. The questionnaire was mailed to patients at the time of the preintervention visit. To increase the response rate to the mailed questionnaire, up to 2 reminder letters were mailed at scheduled times to patients who had not returned their questionnaires; telephone calls were made when necessary to collect this information. Approximately 3 months after the intervention, the VF-14 was again sent to patients, along with additional questions on the clinical aspects of their disease and questions on satisfaction with the outcomes of their surgery. These included a global satisfaction question about the results of the intervention with 5 answer choices ranging from “not at all satisfied” to “totally satisfied.” At this time, patients were also asked a transitional question about the improvement in their vision after cataract removal: “How is your vision now compared with how it was before your cataract surgery?” There were 7 answer options, ranging from “a great deal better” to “a great deal worse.” Follow-up for patients who did not reply in a timely fashion was the same as previously described for the preintervention mailing.
The different categories of the satisfaction and of the transitional items were grouped. In the overall satisfaction question, categories “not satisfied,” “not at all satisfied,” and “somewhat satisfied” were regrouped into “not satisfied,” and “slightly satisfied” and “much satisfied” remained as before. With regard to the transitional item, the 7 answer choices were reduced into 5 options, summarized in the following way: “much worse” and “slightly worse” summarized as “much worse”; “a little worse”; “same”; “a little better”; and “slightly better” and “much better” grouped into “much better.”
Mean and standard deviations were calculated for the preintervention VA and VF-14 scores and for mean change in VA and VF-14 scores after cataract extraction, stratified by patient satisfaction with the results of the cataract extraction (quite satisfied/slightly satisfied/not satisfied) and response to the transitional question. Frequencies and percentages were computed for categorical variables.
To assess the relationship of the preintervention VA and VF-14 scores and mean change in these measures with patient satisfaction, t tests for independent samples (quite-slightly satisfied/not satisfied) or analysis of variance (ANOVA) analyses (quite satisfied/slightly satisfied/not satisfied) were used. The Scheffé post hoc test was used for pairwise mean comparisons. When categorical variables were compared, a χ 2 was developed. ANOVA was used to evaluate postintervention changes in vision by baseline visual status and response to the transitional question.
We computed the patient acceptable symptom state value to identify the postintervention VA or VF-14 score beyond which patients considered themselves well. Satisfaction state (quite-slightly satisfied/not satisfied) at 3 months after cataract extraction was chosen as the anchoring variable. An empirical cumulative probability curve was constructed as a function of the scores of interest (postintervention VA and VF-14 score) for patients in the satisfactory group. Patient acceptable symptom state was targeted as the 75th percentile of the obtained curve. This value was considered as the optimal patient acceptable symptom state cut point.
To determine the optimal cut points for the postintervention scores of the receiver operating characteristic (ROC) curves, logistic regression models were developed. Satisfaction was selected as the outcome variable and the postintervention VA or VF-14 values as independent. The values of the VA or VF-14 that showed the highest sensitivity and specificity were, respectively, considered as the optimal ROC cut-off point (the upper-left cut point in the graph) for those parameters. Sensitivity, specificity, and area under the curve (AUC) for the obtained value were assessed.
Patient acceptable symptom state values and optimal cut points were estimated for the overall sample and stratified by the presence or absence of ocular comorbidities and the baseline preintervention VA (≤0.1, 0.2–04, ≥0.5) and VF-14 score (<45, 45–70, >70).
The mean change score for patients whose response to the transitional question was “somewhat better” was used to estimate the minimal clinically important difference for VA and for the VF-14 score. Optimal cut points for the minimal clinically important difference, based on the ROC curves and patient acceptable symptom state values, were also estimated.
Statistical analyses were performed using SAS for Windows statistical software, version 9.2 (SAS Institute Inc, Cary, North Carolina, USA). Figures were developed with R release 2.10 statistical software (available at www.r-project.org/ ).
A total of 4335 patients were recruited for the study ( Supplemental Figure and Supplemental Table , available at AJO.com ). The mean age was 73.4 years, and 2520 (58.13%) of the participants were women. Simple cataract was present among 3321 patients (76.61%). The preintervention mean VA was 0.28 and the preintervention mean VF-14 was 61.02. 3843 of 4207 patients reported being satisfied with the outcomes of the intervention.
Table 1 presents the changes in VA and visual function (as measured by the VF-14) by preintervention values and the presence or absence of other ocular comorbidities across 3 levels of satisfaction with the outcomes of the intervention—quite satisfied, slightly satisfied, and not satisfied. Patients who reported being quite satisfied experienced substantial gains in VA (from 0.20 to 0.73) and VF-14 scores (from 8.22 points to 53.12 points), but the improvements varied according to preintervention status and presence of ocular comorbidities.
|Quite Satisfied (n = 3295) a||Slightly Satisfied (n = 548) b||Not Satisfied (n = 364) c||P Value|
|Mean VA Change|
|≤0.1||0.68 (0.24)||0.59 (0.29)||0.38 (0.32)||<.001|
|0.2–0.4||0.53 (0.20)||0.45 (0.22)||0.39 (0.28)||<.001|
|≥0.5||0.34 (0.19) c||0.31 (0.18)||0.17 (0.32) a||<.001|
|Cataract with other ocular comorbidity|
|≤0.1||0.51 (0.28)||0.27 (0.31)||0.12 (0.18)||<.001|
|0.2–0.4||0.43 (0.23)||0.35 (0.21)||0.23 (0.23)||<.001|
|≥0.5 d||0.21 (0.18)||0.22 (0.29)||0.20 (0.20)||.95|
|Mean Change in VF-14 Score|
|<45||56.25 (17.86)||36.50 (19.78)||20.15 (26.71)||<.0001|
|45–70||31.62 (13.26)||19.47 (19.16)||3.34 (23.94)||<.0001|
|>70||9.17 (11.83)||1.21 (14.74)||−5.66 (18.17)||<.0001|
|Cataract with other ocular comorbidity|
|<45||53.00 (18.65)||23.39 (22.02)||10.51 (19.72)||<.0001|
|45–70||28.23 (16.21)||12.91 (16.96)||2.61 (20.13)||<.0001|
|>70||8.71 (11.47)||−1.38 (17.37)||−22.13 (26.06)||<.0001|
The expected postintervention patient acceptable symptom state for VA and VF-14 scores for patients who reported being quite or slightly satisfied at the 75th percentile level are presented in Figure 1 . Postintervention VA among patients with simple cataract ranged from 0.67 to 0.80 across preintervention categories, while VF-14 scores ranged from 88 to 90. Among patients with any ocular comorbidity, VA ranged from 0.63 to 0.75 across preintervention categories, while VF-14 scores ranged from 86 to 92. For the entire sample, the expected postintervention VA was 0.75 and the expected postintervention VF-14 score was 86.1.
ROC curves for postintervention VA and VF-14 scores are presented in Figure 2 . These were generated to help identify the point that best balances sensitivity and specificity for both outcomes across patient satisfaction groups. Expected postintervention VA for patients with simple cataract was 0.50 for those with preintervention VA ≤0.1, 0.75 for those with VA between 0.2 and 0.4, and 0.85 for those with VA ≥0.5. Expected postintervention VF-14 scores were 75 points for patients with preintervention VF-14 scores less than 45 points, 80 points for those with VF-14 values between 45 and 70, and 90 points for those with VF-14 higher than 70 points. Postintervention VA values for patients with any other ocular comorbidity were 0.40 for patients with preintervention VA ≤0.1, 0.60 for those with VA between 0.2 and 0.4, and 0.80 for those with VA ≥0.5, while postintervention VF-14 scores were 45 points for patients with preintervention VF-14 scores <45 points, 65 for those with VF-14 values between 45 and 70, and 77 points for those with VF-14 >70 points. Globally, the expected postintervention VA was 0.70 (AUC, 0.69; sensitivity, 74%; specificity, 56%) and the expected postintervention VF-14 score was 75 points (AUC, 0.83; sensitivity, 84.16%, specificity, 64.36%).
Changes in VA and visual function across the responses to the transitional question are presented in Table 2 . The minimal clinically important difference for patients who reported they were “a little better than before the intervention” ranged from 0.17 to 0.57 in VA and from 0.41 points to 37.46 points in VF-14 scores; both were influenced by preintervention status and presence of ocular comorbidities.
|Much Better (n = 2999) a||A Little Better (n = 806) b||Same (n = 188) c||A Little Worse (n = 95) d||Much Worse (n = 78) e||P Value||Optimal ROC Cut Point||Optimal PASS Cut Point|
|Mean (SD) VA Change|
|Simple cataract (n = 3321)|
|≤0.1||0.69 (0.23) b , c , e||0.57 (0.28) a , e||0.42 (0.39) a||0.51 (0.29)||0.20 (0.21) a , b||<.001||0.69||0.69|
|0.2–0.4||0.53 (0.20) f||0.46 (0.21) a , d , e||0.40 (0.25) a||0.32 (0.30) a , b||0.34 (0.30) a , b||<.001||0.50||0.50|
|≥0.5||0.34 (0.18) e||0.31 (0.19) e||0.21 (0.32)||0.21 (0.20)||−0.10 (0.46) a , b||<.001||0.30||0.30|
|Cataract with other ocular comorbidity (n = 1014)|
|≤0.1||0.54 (0.28) f||0.32 (0.28) a , c||0.09 (0.14) a , b||0.08 (0.21) a||0.11 (0.23) a||<.001||0.40||0.48|
|0.2–0.4||0.44 (0.22) f||0.34 (0.21) a||0.22 (0.21) a||0.26 (0.25) a||0.14 (0.28) a||<.001||0.38||0.50|
|≥0.5||0.22 (0.19)||0.17 (0.26)||0.20 (0.18)||0.28 (0.17)||−0.05 (0.07)||.34||0.13||0.20|
|Mean (SD) Change in VF-14 Score|
|Simple cataract (n = 3321)|
|<45||56.76 (17.53) f||37.46 (20.54) a , d , e||34.49 (26.67) a , d , e||2.71 (16.26) a , b , c||1.98 (21.16) a , b , c||<.001||49.81||55.16|
|45–70||32.07 (12.95) f||20.86 (17.37) f||9.52 (24.15) a , b , e||1.55 (19.37) a , b||−10.04 (28.59) a , b , c||<.001||27.38||31.55|
|>70||9.75 (11.74) f||2.80 (13.30) a , d , e||−0.17 (12.97) a , e||−6.98 (14.23) a , b , e||−25.99 (26.00) f||<.001||6.25||12.27|
|Cataract with other ocular comorbidity (n = 1014)|
|<45||56.14 (16.31) f||26.28 (21.34) a , c , e||14.64 (16.78) a , b||8.28 (22.39) a||−1.54 (25.59) a , e||<.001||34.72||40.15|
|45–70||29.81 (15.38) f||13.64 (15.28) a , e||5.39 (22.06) a , e||2.60 (19.88) a||−24.37 (14.08) a , b , c||<.001||20.83||22.72|
|>70||9.54 (11.09) f||0.41 (16.27) a , e||−8.13 (17.70) a , e||−14.31 (15.22) a , e||−43.35 (26.17) f||<.001||4.17||9.85|