We read with interest the article by Vianya-Estopa et al.,
1 “An Evaluation of the Amblyopia and Strabismus Questionnaire Using Rasch Analysis.” It criticized our Amblyopia and Strabismus Questionnaire (A&SQ) by highlighting “the limitations of the A&SQ instrument in the assessment of VR-QoL in subjects with strabismus and especially in those with amblyopia alone.” We would like to comment on the selection of subjects used in their appraisal.
The A&SQ was presented to 102 patients (mean age, 48 years) from ophthalmic, orthoptic, and optometric clinics and practices. Eleven had strabismus without amblyopia, 33 had anisometropic amblyopia, 39 had strabismic amblyopia, and 19 had combined-mechanism amblyopia. The sample seems representative of amblyopia of various causes in patients age 48. However, average visual acuity is worse than would be expected in 48-year-old U.K. citizens with amblyopia. Visual acuity of the amblyopic eye was 0.38 logMAR in anisometropic amblyopes, 0.6 logMAR in strabismic amblyopes, and 0.70 logMAR in combined-mechanism amblyopes. Accordingly, almost all had amblyopia with a visual acuity worse than 0.3 logMAR.
The prevalence of adults with a visual acuity of worse than 0.3 logMAR is approximately 2.5% in untreated populations,
2 –4 but 0.8% to 1.1% in treated populations.
5 –8 Treatment of amblyopia was well established in the United Kingdom the 1960s. Hence, the sample of patients consisted primarily of severe or insufficiently treated amblyopia, whereas there were only 11 patients with strabismus without amblyopia. The large category in the middle of the spectrum that runs from strabismus without amblyopia, through strabismus with mild or treated amblyopia, to more severe amblyopia is underrepresented in the sample of Vianya-Estopa et al.
1
Although the Rasch model does not assume a population distribution from a certain class, in practice, a sample that extends over the whole range of the spectrum provides more balanced statistical information about the functioning of the different response categories. This may well be the reason that they found that merging five-point scales into three-point scales did not make the A&SQ less informative for their sample. It could also explain the ceiling effect that they found. The relatively small sample size, 102, may have aggravated the problem.
In previous studies,
9 –11 in an almost nonselect sample of amblyopia and/or strabismus patients treated in the Waterland area 35 years earlier, we found that the two conditions strabismus and amblyopia are heavily interwoven. We agree that, in principle, separate instruments would be preferable to assess VR-QoL in individuals with strabismus, individuals with amblyopia, individuals with both amblyopia and strabismus, and individuals with loss of binocular vision. However, this conclusion cannot be drawn from the study by Vianya et al.
1 in a sample dominated by amblyopia with very low visual acuity.