Abstract:
Binocular data typically arise in ophthalmology where pairs of eyes are
evaluated, through some diagnostic procedure, for the presence of certain
diseases or pathologies. Treating eyes as independent and adopting the usual
approach in estimating the sensitivity and specificity of a diagnostic test
ignores the correlation between fellow eyes. This may consequently yield
incorrect estimates, especially of the standard errors. This research is
concerned with diagnostic studies wherein several diagnostic tests, or the
same test read by several readers, are administered to identify one or more
diseases. A likelihood-based method of estimating disease-specific
sensitivities and specificities via hierarchical generalized linear mixed
models (HGLMMs) is proposed to meaningfully delineate the various
correlations in the data. The efficiency of the estimates is assessed in a
simulation study. Data from a study on diabetic retinopathy are analyzed to
illustrate the methodology.
