[Development of a German telephone version of the IQCODE (The Informant Questionnaire on Cognitive Decline in the Elderly) within a cohort of relatives of ischemic stroke patients].

Objectives: To date, no validated long or short telephone version of the German "Informant Questionnaire on Cognitive Decline in the Elderly" (IQCODE) exists for determining cognitive changes over a period of 10 years by interviewing relatives.

Method: Sixty relatives of 60 ischemic stroke patients were interviewed both face-to-face and by phone with the 26-item German IQCODE version in randomized order. Inter-method reliability was calculated as between-method agreement and quantified by a weighted kappa with Fleiss-Cohen weights. A short telephone version of the IQCODE was developed with the Variance Inflation Factor (VIF). Its reliability in regard to the 26-item telephone ICQODE version was calculated by Spearman's correlation coefficient.

Results: The weighted kappa between the telephone and face-to-face interview was 0.84 (95 % confidence interval: 0.72-0.97). The short version of the IQCODE consists of 10 items. Spearman's Correlation Coefficient in regard to the long and short telephone version of the IQCODE was 0.97 (95 % confidence interval 0.96-0.99).

Conclusions: A long and short telephone version of the IQ-CODE for detecting cognitive changes in the elderly within a period of 10 years was developed by interviewing relatives.