Validation of Fuld object memory evaluation for the detection of dementia in nursing home residents

To examine the psychometric properties of the Fuld Object Memory Evaluation (FOME) as an instrument to detect dementia in nursing home residents.

Ninety-six elderly participants were recruited into a dementia group (n = 30) and a normal control group (n = 66). Forty participants (12 dementia, 28 normal controls) had visual impairment.

The test-retest reliability and parallel-form reliability of FOME were excellent, with intraclass correlation coefficients of 0.92 and 0.96, respectively. Satisfactory convergent validity of FOME was established with the Cantonese version of Mini-Mental State Examination, and the Memory subscale and the Initiation/Perseveration subscale of the Chinese version of Dementia Rating Scale (r = 0.43 - 0.68; p < 0.01). The FOME total retrieval (TR) score and delayed recall (DR) score showed good discriminative power to screen for dementia. Optimal cutoff scores for TR and DR were suggested as 33 (93% sensitivity, 82% specificity) and 7 (87% sensitivity, 76% specificity) respectively. The performance of FOME was not influenced by age, educational level and visual impairment.

The findings suggest that FOME is a valid assessment to screen for dementia in older nursing home residents and can be used with older individuals with limited education and those with visual impairments.