Reliability and validity of A Quick Test of Cognitive Speed (AQT) in Iranian older adults

Background: Cognitive disorders are one of the most important issues in old age. They may remain hidden in the early stages. There are many cognitive tests, but some variables affect their results (e.g., age and education.) This study aimed to evaluate the reliability and validity of A Quick Test of Cognitive Speed (AQT) in Iranian older adults. Methods: This study aimed to test the psychometric properties of AQT. 114 older adults participated in the study and were divided into three groups (46 with mild cognitive impairments (MCI), 24 with dementia, and 45 without MCI and dementia) based on the diagnosis of two geriatric psychiatrists. Participants were assessed by AQT and Mini-Mental State Examination (MMSE). Data were analyzed using Pearson correlation, independent t-test, and ROC curve by SPSS v.23. Results: There was no signicant correlation between AQT subscales and age and no signicant difference between the AQT subscales in male and female, educational levels, and marital status. The test-retest correlations (r) were signicant for Color (C) 0.84, Form (F) 0.91 and Color-Form (CF) 0.94. Convergent validity was signicant between MMSE and AQT. Its correlation was with Color -0.78, Form -0.71, and Color-Form -0.72. The cut-off point for Color was 43.50 s, Form 52 s, and Color-Form 89 s were based on sensitivity and specicity for differentiating older patients with MCI with controls. The cut-off point for Color was 62.50 s, for Form 111 s, and Color-Form 197.50 s based on sensitivity and specicity measures for differentiating older patients with dementia and MCI. Conclusion: The ndings of this study showed that A Quick Test of Cognitive Speed (AQT) is a suitable tool for assessing cognitive function in older adults.


Participants
We recruited 114 participants divided into three clinical groups: controls (n = 44), mild cognitive impairment (MCI) (n = 46), and dementia (n = 24) based on cognitive status. The sampling method was census which was performed from May 2018 to Feb 2019. Inclusion criteria in the control group include age over 60 years, no complaints of memory, or any other cognitive symptoms, normal cognitive function, no cognitive disorders as approved by two geriatric psychiatrists. Inclusion criteria in the MCI group include age over 60 years, the complaint of memory problems, or any other cognitive symptoms, and con rmation by two geriatric psychiatrists of the presence of MCI. Inclusion criteria in the dementia group include age over 60 years, mild to moderate dementia (any type) based on DSM-V criteria and approved by two geriatric psychiatrists. Exclusion criteria included visual problems. We obtained written consent from all study participants, if they were able, or from their families after explaining the objectives and methods of study.

Measurements
We collected data using the AQT and MMSE screening tests. We used a stopwatch to record the time used for completing each of the three AQT tests. The time was recorded in seconds from the beginning of the test to its end. We used the test-retest method to evaluate the reliability of AQT after one month.
A Quick Test of Cognitive Speed (AQT) is a screening tool for identifying cognitive impairments. It consists of three subtests: Color (C), Form (F) and Color-Form (CF). The time used for rapid automatized naming of the forty visual stimuli in each subtest is measured in seconds. The Color (C) and Form (F) tests measure reaction, retrieval and response time (perceptual processing) and the Color-Form (CF) combination test assesses visual working memory and active attention [19,20]. The Color-Form (CF) test is appropriate for examining changes in cognitive function related to neurological or psychiatric disorders and the effectiveness of pharmacological therapies (9). Naming the combinations activates the bilateral parietotemporal regions and the subcortical region of the brain, including the hippocampus, and examines central executive functions [15,21]. Administering the three tests in succession takes from 3-5 minutes. The visual stimuli for AQT are presented on three test plates. The rst features eight lines of colored squares (black, blue, red, and yellow) and the second eight lines of black forms (circle, line, square, and triangle) that are repeated randomly. The third page consists of eight lines of color and form combinations (Fig. 1). The patient is allowed to use other names to describe colors and forms, and time is recorded in seconds according to the test instructions [14]. In previous studies, the sensitivity and speci city of this test have been 0.78 and 0.67, respectively and was higher than for MMSE (0.61) and CDT (0.46) [10]. The test-retest reliability ranges from r = 0.84 to 0.96. Naming times were not dependent on sex or formal education after establishing literacy. The cut-off point (in seconds) for average-normal performance was set at one standard deviation above the mean (+ 1SD), for slower than normal between + 1 and + 2 standard deviations, and for abnormal performance at more than + 2 standard deviations for English and Swedish [16,17].
Mini-Mental State Examination (MMSE): This examination was developed by Flocsetin in 1975. Its maximum score is 30 and a score below 25 is a sign of cognitive impairments. MMSE examines the cognitive state within ve areas: time and place orientation, memory, attention, calculation, and language.
Its test-retest reliability was reported as 0.89 [22]. In Iran the MMSE psychometric properties showed that the test-retest reliability is 0.78 and its cut-off point is 21. With a sensitivity of 0.90 and a speci city of 0.84, this tool should be interpreted according to age and education [23].

Data analysis
The results were analyzed with descriptive statistics (including mean and standard deviation, sensitivity and speci city) and data analysis used Pearson correlation (r), paired and independent t-tests and ANOVA, and Receiver Operating Characteristics (ROC curve). SPSS v.23 software was used for all data analyses.

Results
As shown in Table 1, the number of neurotypical controls was 44 (mean age: 69.24 ± 7.34), with mild cognitive impairment 46 (mean age 74.22 ± 6.21), and with dementia 24 (mean age: 78.54 ± 5.38). There was a signi cant difference between the ages of the control group and the elderly with MCI (P = 0.001) and the old patients with dementia (P < 0.001). There was no signi cant correlation between age and AQT time for all three subscales in the control group and among elderly with MCI (P > 0.01). For the older patients with dementia, a signi cant weak negative correlation was shown between age and shape subscale (r = 0.40, P = 0.05). Regression analysis resulted in the equation: y = 801.34-7.59 (age). Table 1 also shows comparative data for the variables gender, marital status, and levels of education. University education --11 23.9 9 37.5 Table 2 shows the time differences between the subscales of AQT based on demographic characteristics. There is a signi cant difference in the Color subscale measures between the two sexes in the elderly group with mild cognitive impairment. In other groups, there is no signi cant difference in any subscale of AQT with demographic variables.  Table 3. Shows the correlation between the test-retest measures for AQT after two weeks and the convergent validity of AQT with MMSE. The correlations for all subscales of AQT after two weeks are above 0.80 and signi cant (P < 0.01). The correlation between AQT and MMSE is also signi cant, but the correlation is negative because the scores of AQT and MMSE are opposite. The cut-off points for performance on the AQT subtests were determined with the gold standard (i.e., diagnosis by two geriatric psychiatrists) for the participants in the control, MCI and dementia groups (Tables 4 and 5).

Discussion
The ndings showed that AQT has suitable levels of reliability and validity for screening for cognitive impairments among elderly Iranians. Test-retest reliability showed that the correlation of all subscales after two weeks is above 0.80, which indicates an appropriate level of reliability for AQT. As a comparison, the test-retest reliability of AQT for detecting early-stage dementia in elderly Japanese was found to be 0.88, which was similar to this study [20].
The three groups of controls, elderly with MCI and dementia were signi cantly different in the demographic variables of age, sex, educational level, and marital status. However, these differences were not observed for all three subscales Color, Form, and Color-Form. The only signi cant correlation between age and AQT measures occurred for Form naming among old patients with dementia. In comparison, a study by Nielsen et al. (2006) indicated no difference in time measures between the two sexes, but the AQT time measures were shorter for literate than for illiterate old people [18]. In a study that evaluated the relationship between the AQT measures and neuropsychological test scores, no relationship was found between age and AQT naming time [8]. In this study, there were no signi cant associations between age, education level and AQT naming times, indicating that AQT, like MMSE, was not affected by age and education. A psychometric study of MSSE scores among the elderly reported signi cant correlations between MMSE scores and age and education [23]. The Abbreviated Mental Test Score (AMTS) also shows signi cant associations with the variables education and sex, and therefore, the role of these factors should be considered in the interpretation [24]. Montreal Cognitive Assessment (MoCA) is another tool for screening for MCI, but it uses classi ed scores and interpretations are based on age and level of education [25,26]. These ndings suggest that AQT was more effective than other screening tests (e.g. MMSE, AMTs, and MoCA).
Convergent validity for AQT was assessed with MMSE, which is a standard questionnaire used to assess cognitive status in its various domains. Our ndings showed that all AQT subscale measures had a signi cant correlation with MMSE (r > -0.70). Because the scoring for these two tests are opposite in value, less time on AQT indicates better cognitive status, whereas higher scores in MMSE indicate optimal cognitive status. In comparison, Nielsen (2007) assessed the relationship between AQT and MMSE and found signi cant negative correlations between tests that ranged from -0.60 to -0.72 (P = 0.01) [8]. Similar ndings were obtained in a study of Italian adults by Petrazzuoli et al [7].
Means and standard deviations for AQT Color, Form, and Color-Form naming times indicate a signi cant difference between the control, MCI, and dementia groups. Andersson (2007) also found signi cant differences between naming times for healthy participants and groups with dementia (18). Moreover, for participants with dementia with Lewy Bodies, the AQT times were longer than for patients with AD [14]. The group differences in this study proved greater than those reported by Andersson. One reason could be that we did not distinguish between different types of dementia. A study by Takahashi (2012) of elderly Japanese found that the mean AQT times for the healthy control group were two times shorter than for the group with MCI and three times shorter than for the group with dementia [20]. These differences can be related to many factors such as characteristics of the Japanese language or different levels of severity of the disease.
We used the gold standard (i.e., diagnosis by two geriatric psychiatrists) to determine the cut-off point based on the ROC curve. screening tests for dementia in primary care settings reported that for MMSE the sensitivity was 0.58 and speci city 0.91, for the CDT sensitivity was 0.26 and speci city 0.88, and for AQT the sensitivity was 0.78 and speci city 0.67 [10]. The authors concluded that the ndings supported the usefulness of AQT in Swedish primary care centers.

Conclusion
The ndings of this study indicate that A Quick Test of Cognitive Speed (AQT) is a suitable tool for assessing the cognitive status of older adults in primary care settings in Iran. AQT does not require literacy and is not language dependent for speakers of dialects and languages belonging to the same family. Therefore, AQT can be used for the initial assessment of the cognitive status of the elderly in all care centers.

Declarations Funding
This study is not funded by a speci c project grant.

Availability of data and materials
The datasets used during the current study are available from the corresponding author on reasonable request.

Ethics approval and consent to participate
This study has been approved by the Research Ethics Committee of the Iran University of Medical Sciences. We rst explained the study objectives to the participants and then obtained informed written consent from them.

Consent for publication
Not applicable.