Background: The Alzheimer's Disease Assessment Scale (ADAS) was designed as a rating scale for the severity of dysfunction in the cognitive and non-cognitive behaviours that are characteristic of persons with Alzheimer's disease. Its subscale, the ADAS-cog, is a cognitive testing instrument most widely used to measure the impact of the disease. However, the ADAS-cog takes more than 45 min to administer and requires a qualified clinical psychologist as the rater. A more comprehensive rating battery is therefore required. In the present study, we developed a computerized test battery named the Touch Panel-type Dementia Assessment Scale (TDAS), which was intended to substitute for the ADAS-Cog, and was specifically designed to rate cognitive dysfunction quickly and without the need of a specialist rater.
Methods: The hardware for the TDAS comprises a 14-inch touch panel display and computer devices built into one case. The TDAS runs on Windows OS and was bundled with a custom program made with reference to the ADAS-cog. Participants in the present study were 34 patients with Alzheimer's disease. Each participant was administered the ADAS-cog and the TDAS. The test scores for each patient were compared to determine whether the severity of cognitive dysfunction of the patients could be rated equally as well by both tests.
Results: Pearson's correlation coefficient showed a significant correlation between the total scores (r= 0.69, P < 0.01) on the two scales for each patient. The Kendall coefficients of concordance obtained for the three corresponding pairs of tasks (word recognition, orientation, and naming object and fingers) showed the three TDAS tasks can rate symptoms of cognitive decline equally as well as the corresponding items on the ADAS-cog.
Conclusions: The TDAS appears to be a sensitive and comprehensive assessment battery for rating the symptoms of Alzheimer's disease, and can be substituted for the ADAS-cog.
One of the most important public health issues in Japan at present concerns the country's rapidly ageing society, and how to prevent and manage an accompanying increase in the numbers of dementia patients. It is estimated that the number of persons with dementia in Japan will have reached more than 2 million in 2010. Detecting dementia in the early stages is important, because intervention programs are more effective when useful cognitive function can still be preserved.1,2 Furthermore, some medication can already improve symptoms and functioning, and might slow the progression of the basic disease process.3,4 However, regardless of whether dementia is addressed through intervention programs or medication, an appropriate assessment of its symptoms is necessary to identify early onset.
The Alzheimer's Disease Assessment Scale (ADAS)5 was designed as a rating scale for the severity of dysfunction in the cognitive and non-cognitive behaviours that are characteristic of persons with Alzheimer's disease. Its subscale, the ADAS-cog, consists of 11 tasks measuring disturbances of memory, language, praxis, attention and other cognitive abilities that are often referred to as the core symptoms of Alzheimer's disease, and is a cognitive testing instrument most widely used to measure the impact of the disease.6,7 At present, it is also used for evaluating medical therapies8–10 and intervention programs11,12 used with dementia patients. However, the ADAS-cog takes more than 45 min to administer and requires a qualified clinical psychologist as the rater. Although some studies13,14 have sought to improve the accuracy of the ADAS-cog, they did not address the difficulty of its administration. A more comprehensive rating battery is therefore required.
To this end, we developed a computerized test battery named the Touch Panel-type Dementia Assessment Scale (TDAS), which was intended to substitute for the ADAS-Cog, and was specifically designed to rate cognitive dysfunction quickly and without the need of a specialist rater. Here, we describe the TDAS and present a comparison of the rating capabilities of the TDAS and the ADAS-cog for assessing Alzheimer's disease.
MATERIALS AND METHODS
Participants in the present study were 34 patients (mean age 79.2 years) attending our memory disorder clinics. All patients were diagnosed with Alzheimer's disease by a neurological specialist, based on the results of several medical examinations as well as neuroimaging and the National Institute of Neurological and Communicative Disorders and Stroke Alzheimer's Disease and Related Disorders Association (NINCDS-ADRDA)15 criteria. Table 1 shows the demographics for the participants. All participants gave informed consent to be a part of this study.
Table 1. Demographic information for participants
Mean age ± SD
Mean MMSE ± SD
MMSE, Mini-Mental State Examination.
78.1 ± 8.1
22.0 ± 3.6
79.7 ± 7.7
21.5 ± 3.3
Each participant was administered the ADAS-cog followed by the TDAS 1–6 months later. The test scores for each patient were compared to determine whether the severity of cognitive dysfunction of persons with Alzheimer's disease could be rated equally as well by both tests. Statistical analysis was carried out using spss software (Version 17 for Windows; SPSS, Chicago, IL, USA). The level of statistical significance was set at 0.05 for all tests.
Description of the TDAS
The hardware for the TDAS comprises a 14-inch touch panel display and computer devices built into one case, making the unit compact (450 W × 300 L × 38 D mm), lightweight (2 kg) and portable (Fig. 1). The TDAS runs on Windows OS and was bundled with a custom program made with reference to the ADAS-cog. Although the ADAS-cog is composed of 11 test items, it was difficult to computerize all of these items, because they were created based on dialogical observation. Therefore, we decided to use seven of the ADAS-cog items, which were highly significant and easily modifiable for computerization. With the addition of two other tasks, the TDAS was arranged into nine computerized test tasks and can typically be administered within 30 min. All test questions are asked verbally by the computer and/or are presented visually. In this procedure, a perfect score is 0, and the score increases as the number and value of false answers increases, reaching a score of 101 when all answers are incorrect. A description of each of the nine TDAS tasks follows.
This task was made by computerizing the word-recognition task of the ADAS-cog. The task begins with the presentation of 12 target words. Individual target words are presented onscreen for 3 s at 2-s intervals. After each of the 12 words have been shown, the computer displays, in turn, 24 words composed of a random mixture of the 12 previously presented words and 12 previously non-viewed words, and the subject responds to the question shown about whether the word was shown previously. The subject answers by touching the ‘yes’, ‘no’ or ‘unknown’ button shown on the computer display. This task is repeated three times using the same target words and another 36 previously non-viewed words. The score equals the total number of incorrect responses for three trials (maximum = 72).
Following a command
This task was made by modifying the following commands task of the ADAS-cog. The task examines whether a subject can follow what the computer says. The computer presents 10 choice icons labelled from 0 to 9 and then directs the subject to touch a number specified. An incorrect response is recorded as one point by the computer. This trial is repeated twice using different numbers (maximum = 2).
This task is similar to the orientation task of the ADAS-cog. The computer presents four screens in turn. On each screen, the computer shows icons displaying choices and asks the subject what year, month, day and weekday it is, respectively. An incorrect response is scored as one point (maximum = 4).
This task, which was made by modifying the constructional praxis task of the ADAS-cog, evaluates the visual–spatial perception of the subject. The ADAS-cog requires subjects to copy the geometric forms presented. As it is difficult for a computer to assess the accuracy of a drawn form automatically, this test examines whether the subject can remember geometric forms correctly. The computer presents four screens in turn. On each screen, the computer presents one geometric form (i.e., a square, lozenge, cube or triangular prism) for 5 s. Two seconds after the target form disappears, five geometric forms are displayed consisting of the target form and four other forms, and the subject is asked which form was presented previously. An incorrect response is scored as one point (maximum = 4).
This task examines whether the subject can name the fingers correctly, in the same manner as for the ADAS-cog. The computer presents five screens in turn. Each screen shows a picture of the hand on which one finger is marked by a red circle on the left side of the screen, and five choice icons labelled with the five finger names on the right side. The subject is asked to select the correct icon. An incorrect response is scored as 1 point (maximum = 5).
This task, made by modifying the naming objects task of the ADAS-cog for computerization, examines whether the subject can recognize the utilization of tools correctly. For this task, the computer presents three screens in turn. On each screen, the computer presents one common tool (i.e., a pair of scissors, a comb or a broom) and five choice icons each labelled with the purpose of usage. The subject is asked to select the icon explaining the correct usage of the tool. An incorrect response is scored as one point (maximum = 3).
Accuracy of the order of a process
This task, a modification of the ideational praxis of the ADAS-cog, examines whether the subject can recognize the process of writing a letter through to posting it. The screen displays seven icons on which seven actions in the process are labelled randomly. The subject is instructed to select the icons in the correct order. If the order is incorrect in any way, five points are recorded.
This task examines whether the subject can calculate money correctly. The computer presents three screens in turn, each of which presents an amount of money and various denominations of coins. The subject must combine the coins displayed to equal the amount shown. An incorrect response is scored as one point (maximum = 3).
Clock time recognition (non-digital)
This task examines whether the subject can recognize clock time correctly. The computer asks three kinds of questions. First, the subject is asked to select one clock from a choice of six that shows the same time presented on the screen. Second, the subject is asked to select one icon from a choice of six that labels the time corresponding to a clock presented on the screen. Third, the subject is asked to select one icon from a choice of six that labels the difference in time between a clock presented on the screen (shows 10:40) and 11 o'clock. An incorrect response is scored as one point (maximum = 3).
Table 2 shows the score for each of the test items on the TDAS and the ADAS-cog. Figure 2 shows the relationship between total scores on the two scales for each patient. Pearson's correlation coefficient showed a significant correlation between the total scores (r= 0.69, P < 0.01). The TDAS tasks of word recognition, following command, orientation, visual–spatial perception, naming fingers, object recognition and accuracy of the orders of process were created with reference to the tasks on the ADAS-cog. To evaluate whether the TDAS ratings of these six tasks are equal to those of the corresponding original six tasks on the ADAS-cog, the Kendall coefficients of concordance were obtained for the corresponding pairs of tasks (Table 3). Word recognition has the highest coefficient of 0.57, followed by orientation and naming object and fingers with 0.41 and 0.32, respectively. Thus, these three TDAS tasks can rate symptoms of cognitive decline equally as well as the corresponding items on the ADAS-cog.
Table 2. Mean score for each task of the Alzheimer's Disease Assessment Scale-Cognitive and the Touch Panel-type Dementia Assessment Scale
Table 3. Kendall's coefficient of concordance showing an agreement among two ratings obtained by the Alzheimer's Disease Assessment Scale-Cognitive and the Touch Panel-type Dementia Assessment Scale
Test items corresponding pair
Kendall's coefficient of concordance
In the case where Kendall's coefficient of concordance is more than 0.05 (when the level of statistical significance was set at 0.05), two raters can be considered the same. †Naming object and fingers task of the Alzheimer's Disease Assessment Scale-Cognitive (ADAS-cog) were computerized separately as naming fingers and object recognition for the Touch Panel-type Dementia Assessment Scale (TDAS). Therefore, sum score of two tasks was compared with score obtained by naming object and fingers task of the ADAS-cog.
Constructional praxis of the ADAS-cog and visual–spatial perception of the TDAS
Naming object and fingers of the ADAS-cog and sum of naming fingers and Object recognition of the TDAS†
Ideational praxis of the ADAS-cog and accuracy of the order of a process of the TDAS
With the increase in the number of aged persons who are suspected to be in cognitive decline, an assessment of cognition is becoming a more crucial part of many medical consultations for this population. This is increasing the burden on medical staff who must administer the cognitive tests. Thus, it is anticipated that there will be a greater need in the future for a quick and sensitive screening test that is suitable for use by primary healthcare workers as well as by non-specialists.16,17 Some computerized test batteries have recently been developed and their usefulness reported.18–20 Using a computerized cognitive test battery offers some clear advantages; for example, it can provide quick, objective and accurate results based on the same standards for all subjects examined. Because the TDAS was intended for use with the elderly, the design of the computer and user interface was important. We used a touch panel display as an input device so that elderly people could operate it easily by touching the icon shown on the display, without needing to use a keyboard or mouse. In the present study, almost of all of the subjects could operate the TDAS by themselves. However, some subjects with poor eyesight or hearing required help to operate the TDAS. In such cases, support from medical staff was required for them to complete the assessment. There were some severe AD patients who ended the TDAS halfway through. Almost all of such patients could not complete the ADAS-cog either. The TDAS can't respond flexibly according to the condition of human subjects, which was a weak point of the computerized battery. It was reported that the ADAS-cog subscale was most accurate for measuring AD-related cognitive dysfunction at a moderate level of impairment.13 We believe that patients with moderate impairment could also operate the TDAS easily without any support.
We intended to develop a computerized test battery to substitute for the ADAS-cog and developed the TDAS as a result. The ADAS-cog assesses not only memory, but also speaking fluency and behaviour characteristics of subjects, and the scale items were determined based on face-to-face interviews. However, because the computer used in the TDAS is simply a personal computer, it is difficult for it to recognize what the subject says and how the subject behaves. During the entire test process, the TDAS asks the test questions verbally and/or through text prompts and the subject is required to select the correct icon from many shown on the screen. The TDAS decides if the answer provided is correct or not from the icon selected. Such interaction does not allow for the TDAS to introduce those test items on the ADAS-cog, which are assessed through spoken responses or other subject behaviours. Therefore, out of the 11 tasks on the ADAS-cog, only six tasks (following commands, naming objects and fingers, orientation, word-recognition, constructional praxis and ideational praxis) could be computerized and incorporated into the TDAS.
Kendall's coefficient of concordance can be used for assessing agreement between two raters and it ranges from 0 (no agreement) to 1 (complete agreement). In the case in which Kendall's coefficient of concordance is more than 0.05 (when the level of statistical significance is set at 0.05), two raters can be considered the same. The coefficients for the naming objects and fingers, orientation and word-recognition tasks were more than 0.05, indicating that these three TDAS tasks have the same assessment capability as the corresponding ADAS-cog tasks. However, the following command task, constructional praxis and ideational praxis did not yield the same ratings on both scales and the probable reasons are as follows. For the following command task, the mean score on the TDAS was less than that on the ADAS-cog and this was because it was difficult to computerize complex commands, so the commands on the TDAS were easier than those on the ADAS-cog. For ideational praxis, although the ADAS-cog gives a score from 0 to 5 according to the accuracy of the subject's performance, the TDAS can only judge the subject's answer as correct or incorrect and allot 0 or 5 points (not the range). For constructional praxis on the ADAS-cog, the subject is required to copy figures presented on paper and the rater assesses the similarity of the copy. This was difficult to computerize, so we modified the test to become a visual–spatial perception test for the TDAS.
Instead of using tasks that assess capabilities through spoken responses or behaviours, we introduced two tasks into the TDAS that could be scored automatically by computer: one was money calculation and the other was time recognition. Handling money and clock reading are common daily activities, and impairments of these abilities are indicators of dementia.21,22 Although not all of the tasks on the TDAS could provide the same rating outcome as the ADAS-cog tasks, the total scores for the TDAS and the ADAS had a large correlation coefficient. Thus, the TDAS appears to be a sensitive and comprehensive assessment battery for rating the symptoms of Alzheimer's disease and can be substituted for the ADAS-cog.
The ADAS-cog requires a qualified clinical psychologist as the rater for an administration. The TDAS does not require a specialist rater. Furthermore, the TDAS can be completed in a short time compared with the ADAS-cog. We believe these are advantages of the TDAS.