Selecting optimal screening items for delirium: an application of item response theory

Table 1 Summary of results from dimensionality assessment models

				Single factor model			Two-factor model		Bifactor model
Feature	Number of proposed indicators	Number of modeled indicators	Number of significant eigenvalues	Marginal reliability*	CFI^†	RMSEA^‡	CFI	RMSEA	Large secondary loadings	Retained dimensions
Feature 1-Acute Change and Fluctuating Course-Direct Interview	17	15	1	0.32	0.94	0.03	0.98	0.02	n/a	1
Feature 1-Acute Change and Fluctuating Course-Observational	12	11	1	0.81	0.99	0.05	n/a	n/a	n/a	1
Feature 2-Inattention-Direct Interview	15	15	2	0.80	0.96	0.06	0.99	0.03	no	2
Featured 2-Inattention-Observation	8	8	1	0.72	0.98	0.04	n/a	n/a	n/a	1
Featured 3-Disorganized Thinking-Direct Interview	13	13	2	0.64	0.98	0.05	0.99	0.03	no	1
Featured 3-Disorganized Thinking-Observational	13	13	2	0.89	0.90	0.08	0.98	0.04	yes	2
Featured 4-Altered Level of Consciousness-Observational ^§	25	14	2	0.94	0.89	0.05	0.98	0.03	yes	2

Note: *Marginal reliability coefficient is based on each feature at θ ₅₀ ⁺, ^†Comparative Fit Index, ^‡Root Mean Squared Error of Approximation. n/a implies the result is not applicable because the relevant models were not estimated in a bi-factor model. ^§Feature 4 –Altered Level of Consciousness-Direct Interview was dropped because items were redundant with Feature 2—Inattention-Direct Interview.

ISSN: 1471-2288