COMBINING TWO DUAL-TASK TESTS TO DISCRIMINATE BETWEEN PEOPLE WITH DEMENTIA DISORDER, MILD COGNITIVE IMPAIRMENT, SUBJECTIVE COGNITIVE IMPAIRMENT, AND HEALTHY CONTROLS

Bozkurt Åhman H1, Cedervall Y1, Giedraitis V1, Berglund L1, McKee K2, Kilander L1, Rosendahl E3, Ingelsson M1, Åberg AC1,21Uppsala University, Department of Public Health and Caring Sciences, Uppsala, Sweden, 2Dalarna University, Health and Social Studies, Falun, Sweden, 3Umeå University, Department of Community Medicine and Rehabilitation, Umeå, SwedenBackground: According to the World Health Organization, dementia disorders are a global health challenge. In order to enable early interventions, there is a need for tools to identify individuals in the preclinical stages. Mild cognitive impairment (MCI) and subjective cognitive impairment (SCI) imply an increased risk of later progression to dementia disorders. Research has suggested that dual-task performance may be associated with cognitive decline.Purpose: To investigate if outcomes of dual-task performance, i. e. verbal tasks performed simultaneously with the mobility test Timed Up-and-Go (TUGdt), can be used in combination to discriminate individuals with dementia disorder, MCI, SCI, and healthy controls.Methods: Participants were consecutively recruited when going through assessment at specialist memory clinics. Individuals diagnosed with dementia disorder (n=86; mean age 76 years, range 55-94, 43% female), MCI (n=135; mean age 73 years, range 49-91; 44% female), and SCI (n=77; mean age 67 years, range 39-85; 47% female) were included. Controls (n=127; mean age 72 years, range 50-91, 56% female) were recruited through advertising. Data collection included established cognitive and motor function tests. Two types of TUGdt tests were performed; TUGdt “reciting animals”, and TUGdt “reciting months in reverse order”. The task “reciting animals” primarily reflects frontal lobe function, whereas “reciting months in reverse order” also places demands on parietal lobe functions. The TUGdt tests were registered by video cameras capturing mobility and verbal performance. The number of correct words (animals and months respectively) per 10 seconds for each TUGdt test was calculated. Associations between diagnoses and TUGdt variables were examined in logistic regression models adjusted for age, gender, and education. The areas under the receiver operating characteristics (ROC) curves (c-stat) were calculated for models combining outcomes from the TUGdt tests.Results: When combining the outcome variables “animals/10s” and “months/10s”, ROC curve analyses indicated significant differences between groups: Dementia vs MCI: c-stat=.77; Dementia vs SCI: c-stat=.91; Dementia vs Controls: c-stat=.96; MCI vs SCI: c-stat=.73; MCI vs Controls: c-stat=.87; SCI vs Controls: c-stat=.71; (all p< 0.0001). Furthermore, all the combined models showed significantly higher c-statistics than models based on the variables separated (p< 0.05).Conclusion(s): The results indicate that dual-task performance may reflect characteristic features of dementia disorder, MCI, and SCI. When combining outcomes from two different TUGdt tests, it is possible to discriminate these diagnostic groups from each other with an accuracy of 73 to 91%, and to discriminate them from the controls with an accuracy of 71 to 96%. Because of the heterogeneity of the diagnostic groups, these results call for further research with longitudinal design.Implications: Dual-task performance may be a useful tool to identify individuals in the preclinical stages of dementia disorders. TUGdt testing is easy to carry out, does not require expensive equipment or special knowledge, and could be used by physiotherapists or other health care professionals in different care settings.Keywords: Dementia, Dual-task, Cognitive impairmentFunding acknowledgements: The Swedish Research Council, Alzheimer Foundation, Promobilia Foundation, Swedish Society of Medicine, Uppsala-Örebro Regional Research Council, Geriatric Research Foundation, etc. Topic: Neurology; Older peopleEthics approval required: YesInstitution: The Regional Ethical Review Board in UppsalaEthics committee: The Regional Ethical Review Board in UppsalaEthics number: 2014/068All authors, affiliations and abstracts have been published as submitted.