Explicit scoring criteria for the Rey-Osterrieth and Taylor complex figures

Explicit scoring criteria based on the Taylor-Osterrieth method were developed for the Rey-Osterrieth and Taylor Complex Figure Tests. Without such criteria, clinicians and researchers are left to develop their own idiographic scoring systems, leading to low interrater reliability, especially in the middle range of scores. Criteria were applied to the copy and 30-min recall protocols of 101 subjects infected with HIV. Results indicate that the criteria presented are reliable across a wide range of scores, making them useful for tracking subtle changes over time. Independent raters were within two points of each other on 90.1 % of the copy and 88.1% of recall scores, with a maximum disagreement of five points on two copy protocols. The Taylor figure was also found to be more easily remembered than the Rey-Osterrieth, calling into question their interchangeability. Explicit criteria are particularly useful in research settings. They ensure interrater reliability over time and across examiners in longitudinal studies, enhance the detection of subtle differences within the normal range, and allow comparison of results obtained by different investigators.