BACKGROUND: The characteristics and associated risks of spontaneously reported cognitive complaints have not been investigated due to the lack of a classification instrument. METHODS: In phase 1, a classification system with descriptive categories and cognitive domains was developed by experts through a modified Delphi technique. In phase 2, 180 elderly patients seeking medical attention for cognitive complaints provided free reports of their cognitive difficulties and each complaint was recorded verbatim. Three observers were asked to classify each complaint into a descriptive category. Perceived cognitive function was further characterized using the Memory Complaint Questionnaire (MAC-Q). RESULTS: The patients reported 493 spontaneous complaints, with a range of 1-6 complaints per patient and a mean of 2.7 (±1.3). The proportion of complaints that could be classified into a category by each of the three observers varied from 91.9% to 95.7%. Inter-observer agreement assessed using the κ statistic varied from 0.79 to 1 for descriptive categories and 0.83 to 0.97 for domains. Compared with the MAC-Q, spontaneously reported complaints provided complementary information by avoiding the cueing effect provoked by the questionnaire. The total number of complaints and their occurrences in specific domains were associated with important sociodemographic and clinical factors, indicating that their meaning and associated risks need to be further investigated. CONCLUSION: The instrument developed in this study proved to be a practical tool for classifying the majority of spontaneously reported cognitive complaints with high reliability. Further studies are needed to investigate clinical usefulness of this approach.
BACKGROUND: The characteristics and associated risks of spontaneously reported cognitive complaints have not been investigated due to the lack of a classification instrument. METHODS: In phase 1, a classification system with descriptive categories and cognitive domains was developed by experts through a modified Delphi technique. In phase 2, 180 elderly patients seeking medical attention for cognitive complaints provided free reports of their cognitive difficulties and each complaint was recorded verbatim. Three observers were asked to classify each complaint into a descriptive category. Perceived cognitive function was further characterized using the Memory Complaint Questionnaire (MAC-Q). RESULTS: The patients reported 493 spontaneous complaints, with a range of 1-6 complaints per patient and a mean of 2.7 (±1.3). The proportion of complaints that could be classified into a category by each of the three observers varied from 91.9% to 95.7%. Inter-observer agreement assessed using the κ statistic varied from 0.79 to 1 for descriptive categories and 0.83 to 0.97 for domains. Compared with the MAC-Q, spontaneously reported complaints provided complementary information by avoiding the cueing effect provoked by the questionnaire. The total number of complaints and their occurrences in specific domains were associated with important sociodemographic and clinical factors, indicating that their meaning and associated risks need to be further investigated. CONCLUSION: The instrument developed in this study proved to be a practical tool for classifying the majority of spontaneously reported cognitive complaints with high reliability. Further studies are needed to investigate clinical usefulness of this approach.