Trend, event analyses have distinct advantages

March 6, 2009

San Diego-When it comes to assessment of progression of glaucoma, trend analysis and event analysis have distinct advantages, said Joseph Caprioli, MD. He is the David May II Professor of Ophthalmology, chief of the glaucoma division, and director of the glaucoma basic science and clinical laboratories at the Jules Stein Eye Institute, Los Angeles.

San Diego—When it comes to assessment of progression of glaucoma, trend analysis and event analysis have distinct advantages, said Joseph Caprioli, MD. He is the David May II Professor of Ophthalmology, chief of the glaucoma division, and director of the glaucoma basic science and clinical laboratories at the Jules Stein Eye Institute, Los Angeles.

Generally speaking, “trend analysis requires more tests and takes longer to reach a conclusion, but it provides important information about rate,” Dr. Caprioli said. “Event analysis requires fewer tests. You can reach an endpoint sooner, but there’s no good information about rate.

“Both approaches require some evidence of persistence and confirmation,” he added.

“In event analysis, we develop limits for test/re-test variability in an effort to differentiate variability from true change,” Dr. Caprioli said. These limits subsequently are applied to an individual series of tests to determine the threshold of variability that needs to be exceeded to indicate that change has occurred, he said.

Examples of event analysis include glaucoma change probability analysis and glaucoma progression analysis, according to Dr. Caprioli.

“In trend analysis, we establish a statistical significance of change over the entire series of tests,” he said. “We then relate that magnitude of change to the variability observed within the data series for that individual.”

Examples of trend analysis include pointwise linear regression, clusterwise linear regression, and visual field indices, Dr. Caprioli said.

Both trend- and event-based methods can be applied to functional and structural changes from glaucoma, he said. Trend-based analysis is more appropriate for use with perimetry, however, and event-based analysis is more suited for use with optic nerve imaging, he said.

Qualitative estimates of progression rates for optic nerve images are useful in the clinical setting, Dr. Caprioli added.

Related Content:

News