The ICD-10 Glaucoma Severity Score Underestimates the Extent of Glaucomatous Optic Nerve Damage

Ari Leshno,Emmanouil Tsamis,Noga Harizman,George A Cioffi,Qing Wang,Sol La Bruna,Anvit Rai,Carlos Gustavo De Moraes,Jeffrey M Liebmann,Donald C Hood,George A. Cioffi,Jeffrey M. Liebmann,Donald C. Hood
DOI: https://doi.org/10.1016/j.ajo.2022.08.009
IF: 5.488
2022-08-29
American Journal of Ophthalmology
Abstract:Purpose The-International-Classification-of-Disease, 10 th revision (ICD-10) codes used for glaucoma severity classification are based on the location of visual-field (VF) defects. Given the known poor sensitivity of the 24-2 VF test to early disease and macular damage, we hypothesized that the ICD-10 codes will not accurately reflect the extent of glaucomatous damage. Design Validity and reliability analysis. Methods We evaluated 80 eyes with glaucomatous optic neuropathy (GON). Masked reviewers assigned an ICD-10 severity grade based on 24-2 VF. Two additional masked examiners determined the presence of optical coherence tomography (OCT) structural damage in each hemifield and/or central 5 degrees to define an OCT-based equivalent ICD-10 classification. Results 80 eyes with GON were classified as Mild, Moderate and Advanced in 15, 23 and 42 cases, respectively, based on the 24-2 VF and in 6, 7 and 67 cases, respectively based on the OCT. The OCT classifications were more severe in 29/80 (36%) of cases. In 33 cases (41.3%), macular damage detected by OCT was missed by the 24-2. In 4/80 (5%) cases, the VF overestimated the severity, likely due to variability of the 24-2 test. Conclusion The ICD-10 system relies solely on damage seen on the 24-2 and as such is a 24-2 functional score rather than a "glaucoma" severity score. OCT revealed wide variation of damage across grades, with a significant proportion of the eyes showing macular structural damage missed with the 24-2 VF. Adding OCT information to the ICD-10 system would help it to more accurately reflect the extent of glaucomatous damage.
ophthalmology
What problem does this paper attempt to address?