Background: The Informant Questionnaire on Cognitive Decline in the Elderly (IQCODE) is a widely used screening tool for dementia. We aimed to determine the ability of the German version of the 16-item IQCODE with a two-year time frame to discriminate healthy mature control participants (NC) from mild cognitive impairment (MCI) and probable early Alzheimer's disease (AD) patients (all with Mini-mental State Examination (MMSE) scores ≥ 24/30) and to optimize diagnostic discriminability by shortening the IQCODE.
Methods: 453 NC (49.7% women, age = 69.5 years ± 8.2, education = 12.2 ± 2.9), 172 MCI patients (41.9% women, age = 71.5 years ± 8.8, education = 12.3 ± 3.1) and 208 AD patients (59.1% women, age = 76.0 years ± 6.4, education = 11.4 ± 2.9) participated. Stepwise binary logistic regression analyses (LR) were used to shorten the test. Receiver operating characteristic curves (ROC) determined sensitivities, specificities, and correct classification rates (CCRs) for (a) NC vs. all patients; (b) NC vs. MCI; and (c) NC vs. AD patients.
Results: The mean IQCODE was 3.00 for NC, 3.35 for MCI, and 3.73 for AD. CCRs were 85.5% (NC-patient group), 79.9% (NC-MCI), and 90.7% (NC-AD), respectively. The diagnostic discriminability of the shortened 7-item IQCODE (i.e. items 1, 2, 3, 5, 7, 10, 14) was comparable with the longer version (i.e. 7-item CCRs: NC-patient group: 85.3%; NC-MCI: 80.1%, NC-AD: 90.5%).
Conclusions: The German 16-item IQCODE with two-year time frame showed excellent screening properties for MCI and early AD patients. An abbreviated 7-item version demonstrated equally high diagnostic discriminability, thus allowing for more economical screening.