A new tool for quantifying visual function after cataract surgery is now available. The Cataract Outcomes Questionnaire is a valid, reliable instrument that can be completed in 3 minutes.
The COQ was developed using Rasch analysis to reengineer the 18-item visual disability assessment (VDA). The result is a 12-item instrument with rescaled response choices for several items. On average, patients can complete the COQ in about 3 minutes.
To develop the COQ, the VDA was administered to a heterogeneous cohort of 270 consecutive patients presenting before or after cataract surgery. Of that group, 99 patients underwent bilateral cataract surgery, 115 were having second eye procedures, and 56 were pseudophakic patients. Almost half of the study group (46%) had co-morbid eye disease.
The analyses of patient responses showed that several items on the VDA did not contribute well to the measurement of disability because they were redundant or did not fit with the model. Overall, those items also were identified as being too easy for the subjects.
"Plotting of patient ability and item difficulty showed both had normal distributions, but their distributions did not match as they should for a well-targeted questionnaire for assessing disability," Dr. Pesudovs said.
The newly created COQ was subjected to validity and reliability testing in an independent group of 122 patients who had cataract surgery. Analyses of those results established that the COQ functioned well in terms of precision, reliability, and internal consistency.
"We found the targeting of the items to the patients was improved with the COQ relative to the VDA, although it could be further enhanced with incorporation of some more difficult items that are geared to the more able patients," Dr. Pesudovs said.
The COQ also was found to have very good retest reliability, and its results correlated well with measures of vision. In addition, the minimal clinically important difference relative to a self-reported improvement in vision was low (about 3 units on a scale of 0 to 100).
Results of a double asymptotic nonlinear regression analysis examining the relationship between the raw item score and the Rasch measures showed that it would be possible to use a simple scoring conversion to translate a raw COQ score to a Rasch score in a clinical trial or clinical setting.
"Therefore, users could derive the benefits of Rasch scaling without having to perform Rasch analysis. As a limitation, however, that conversion does not work well if there are missing data," Dr. Pesudovs said.