1. Background
Globally, dementia is a growing public health problem, and due to the ageing population, the number of people with dementia has increased over the last decades (1-3). Approximately 10% of older adults over the age of 70 are suffering from dementia, and Alzheimer’s disease has been diagnosed in 50% of cases. In high-income countries, only 50% of people with dementia are diagnosed, and this figure is less than 10% in low- to middle-income countries (3). Iran will encounter explosive growth in the number of older adults. Based on the National Elderly Health Survey report in Iran, the prevalence of dementia in Iranian adults over 60 years old is 7.9% , and in adults over 80 years old, it reaches 13% (4).
Early detection of dementia helps to manage the disease more effectively and reduce the patient costs. It is estimated that about 10 - 15% of people with mild cognition impairment (MCI) will develop dementia per year compared to 1 - 2% of those with normal cognitive functioning (5, 6). Research has shown that increasing the score of the Mini-Mental State Examination (MMSE) test through treatment by 1 point can help to considerably reduce the related costs (7).
Many cognitive screening instruments have been developed for screening of cognitive impairment. The MMSE is one of the most widely used tools for screening, estimating the severity, and monitoring the cognitive problems. Due to the low difficulty of MMSE items, it is easy to distinguish healthy people from those with dementia (8). However, having a ‘ceiling effect’ in mild cognitive impairments, especially in people with higher levels of literacy or intelligence, limits the usefulness of this test for research purposes (7, 9-11). In both clinical and research settings, the follow-up of these patients is difficult due to old age and physical disabilities. Having a cognitive screening test similar to the MMSE, that does not require face-to-face visits, would make such follow-up, especially on a large scale, more practical and cost-effective. The Telephone Interview for Cognitive Status-modified (TICS-m) questionnaire is a convenient and useful tool developed for use in situations where in-person cognitive screening is impractical or inefficient. The TICS-m correlated highly with the MMSE (12-15). The psychometric properties of the TICS-m questionnaire among Iranian older adults have not been established yet. The Neyshabur Longitudinal Study on Ageing (NeLSA), which is an elderly component of the Prospective Epidemiological Research Studies in IrAN (PERSIAN) (16), includes a biennial evaluation of the cognitive status. Due to limitations in research resources, choosing an appropriate and valid tool for follow-up assessment is very important. Therefore, the present study aimed to validate the TICS-m questionnaire to be used in the telephone-based biennial follow-up of the NeLSA.
2. Objectives
The main objective of this study was to assess the face validity, content validity, and construct validity of the Persian version of TICS-m questionnaire.
3. Methods
3.1. Study Design
This cross-sectional study was conducted as a pilot in the city of Neyshabur in northeastern Iran, from January to May 2019 at the NeLSA Centre. We used simple random sampling based on the number of households with an elderly person. To determine the sample size for EFA, the researchers suggested a ratio of the number of observations to the number of variables from 3:1 to 20:1 (17). In this study the ratio of 20:1 was used. Therefore, a total of 210 individuals aged 50 years or older were enrolled in the study.
The inclusion criteria were enrolment with the NeLSA and willingness to participate in the study. The subjects were all community-dwelling, and none were from health facilities, such as hospitals or nursing homes. The exclusion criteria were as follows: hearing impairment; use of hearing aids; the presence of any psychiatric or neurological disease that causes cognitive disorders such as depression, epilepsy, mental retardation, and significant learning disability; history of brain surgery; addiction to alcohol; and a history of head trauma with loss of consciousness for more than two hours.
3.2. Measures
3.2.1. The 13-Item Telephone Interview for Cognitive Status-modified
The telephone interview was conducted one month after the in-person assessments with the MMSE. All research assistants who administered the TICS-m and MMSE held an MSc in clinical psychology and received training on the procedure. The 13-item TICS-m questionnaire of Brandt et al. consists of six cognitive dimensions, including orientation (7 points), registration/free recall (10 points), attention/calculation (6 points), comprehension/semantic/recent memory (5 points), language/repetition (1 points), and delayed recall (10 points). In this questionnaire, the highest score is allocated to memory; but unlike the MMSE test, which allocates 20% of its score to memory, in the TICS-m test, 56% of the total score is allocated to memory (8, 12). The total scores range from 0 to 39. Individuals who score 28 - 31 are considered as having ‘mild cognitive impairment’, and those who score ≤ 27 are considered as having ‘severe cognitive impairment’ (13) (Appendix 1 in the Supplementary File).
3.3. The Process of Validation
3.3.1. Translation of the TICS-m
After obtaining permission to translate and use the instrument, the English version of the questionnaire and its instructions were translated into Persian by two fluent Persian translators (forward translation). Then, the translated questionnaire was retranslated to English by two independent translators (backward translation). After these steps, a team of experts discussed and resolved the degree of discrepancy between the two Persian and English versions.
3.3.2. The Content and Face Validity of the Persian Version of TICS-m Questionnaire
The content and face validity of the translated questionnaires were examined quantitatively and qualitatively. The questionnaires were sent to ten experts in the field of psychology, neuropsychology, psychiatry, and community medicine to evaluate and provide feedback on the items in terms of relevancy, simplicity, clarity, necessity, and importance.
To examine the content validity, the content validity ratio (CVR) and the content validity index (CVI) were calculated. The acceptable value for CVR based on the Lawshe table was considered as ≥ 0.62 (18, 19). After calculating the CVR and eliminating unnecessary questions, the CVI was calculated for the remaining items. The acceptable values were as follows: (1) If I-CVI was higher than 0.79, the item was accepted; (2) If the value of I-CVI was between 0.70 and 0.79, the item needed to be reviewed; and (3) If the value of I-CVI was less than 0.70, the item was removed from the measurement tool (20, 21).
To examine the face validity of questionnaire, the Impact score (with acceptance value of > 1.5) was calculated (22). In assessing the content and face validity of the questionnaire qualitatively, the comments of the expert panels and three speech therapists were applied to replace with alternative words. After that, the revised questionnaire was completed experimentally by 30 healthy elderly subjects, and the questionnaire was finalized.
3.3.3. The Construct Validity of the Persian Version of TICS-m Questionnaire
For this purpose, the questionnaires were administered to 210 community-dwelling older adults aged 50 years and older. The collected data was then analyzed using the EFA.
3.4. Reliability
In order to assess the internal consistency of the Persian version of TICS-m questionnaire, the questionnaires were administrated to 30 volunteers aged ≥ 50 with normal cognition at the NeLSA Centre, and then Cronbach’s coefficient alpha was calculated. A value of 0.7 or above was considered as an acceptable internal consistency (23). To examine the external reliability of the questionnaire, test-retest was used, in which the same questionnaires were completed by the same 30 respondents after a two-week interval. The collected data were entered into SPSS software, and the intra-class correlation (ICC) was calculated. The criteria for interpretation of ICC values were as flows: (1) ICC value < 0.5: week reliability, (2) ICC value: 0.5 - 0.75: moderate reliability, (3) ICC value: 0.75 - 0.90: good reliability, and (4) ICC value > 0.9: great reliability (24).
3.5. Statistical Analysis
Quantitative data were presented as mean and standard deviation, and qualitative data were presented as frequency and percentage. The normality of data was examined using the Kolmogorov-Smirnov test. The Cronbach’s alpha and ICC tests were calculated to assess the internal and external reliability of the translated TICS-m questionnaire, respectively. To examine the construct validity of the instrument, the EFA with the extraction method of principal component analysis (PCA) and Oblimin rotation was used. Kaiser-Meyer-Olkin (KMO) and Bartlett’s tests were used to determine the sufficiency of sample size and its suitability for factor analysis (acceptable value for conducting EFA: KMO ≥ 0.6). The factors were retained based on whether or not the factor had an eigenvalue greater than 1. Factor loadings greater than 0.40 were considered relevant in interpreting the factors (25). The data were analyzed by SPSS statistical software version 21. CVR, CVI, and Impact scores were calculated in the Excel software version 2016.
4. Results
4.1. The Content and Face Validity of the Persian Version of TICS-m Questionnaire
The questionnaire was reviewed by ten experts to assess the content and face validity. The CVR values of all items were higher than 0.62, but the CVI values of three items (I-CVI) (items 6, 10, and 12) were lower than 0.78 (0.67, 0.77, and 0.67, respectively). Therefore, these items were revised and replaced with alternative words through consultation with a language and literature expert. The mean CVI of the instrument was higher than 0.90 (S-CVI / Ave = 0.91), so the content validity of the instrument was confirmed. In assessing the face validity of the instrument quantitatively, the Impact scores of all items were higher than 1.5. Therefore, the face validity of the instrument was confirmed (Table 1).
Table 1.
The Scores for Relevancy, Clarity, Simplicity, CVI, CVR, and the Impact Score of the 13-item TICS-m Questionnaire
| Dimensions of the Memory, Questions | I-CVI a | CVRb | Impact score | Evaluation c | |||
|---|---|---|---|---|---|---|---|
| Simplicity | Relevancy | Clarity | I-CVI/Ave | ||||
| Orientation | |||||||
| Q1 | |||||||
| (i) What day of the week is it? | 1.00 | 1.00 | 1.00 | 1.00 | 1 | 5 | Accept |
| (ii) What is today’s date? | 0.9 | 1.00 | 0.9 | 0.93 | 1 | 4.8 | Accept |
| (iii) What season are we in? | 1.00 | 1.00 | 1.00 | 1.00 | 1 | 4.9 | Accept |
| Q2 | |||||||
| What is your age? | 0.9 | 0.9 | 0.9 | 0.9 | 1 | 4.9 | Accept |
| Q3 | |||||||
| What is your telephone number? | 0.9 | 0.9 | 1.00 | 0.93 | 0.8 | 4.5 | Accept |
| Registration/free recall | |||||||
| Q4 | |||||||
| I’m going to read you a list containing ten words (cabin, pipe, elephant, chest, silk, theatre, watch, whip, pillow, and giant). Please listen carefully and try to remember them. When I am done, tell me as many as you can in any order. Ready? Now, tell me all the words you can remember. | 0.6 | 1.00 | 0.4 | 0.67 | 1 | 4.9 | Accept |
| Attention/calculation | |||||||
| Q5 | |||||||
| Please take 7 away from 100. Now continue to take 7 away from what you have left over until I ask you to stop. | 0.8 | 1.00 | 1.00 | 0.93 | 0.8 | 4.7 | Accept |
| Q6 | |||||||
| Please count backwards from 20 to 1 | 1.00 | 1.00 | 1.00 | 1.00 | 1 | 4.9 | Accept |
| Comprehension, semantic, and recent memory | |||||||
| Q7 | |||||||
| What do people usually use to cut paper? | 1.00 | 1.00 | 1.00 | 1.00 | 1 | 5 | Accept |
| Q8 | |||||||
| What is the prickly green plant found in the desert? | 0.7 | 1.00 | 0.6 | 0.77 | 0.8 | 4.8 | Accept |
| Q9 | |||||||
| Who is the reigning monarch now? | 0.9 | 0.9 | 0.9 | 0.9 | 0.8 | 4.6 | Accept |
| Q10 | |||||||
| Who is the prime minister now? | 0.9 | 0.9 | 1.00 | 0.93 | 0.8 | 4.7 | Accept |
| Q11 | |||||||
| What is the opposite of east? | 1.00 | 1.00 | 1.00 | 1.00 | 0.8 | 5 | Accept |
| Language/repetition | |||||||
| Q12 | |||||||
| Please say this: “Methodist Episcopal”. | 0.4 | 1.00 | 0.6 | 0.67 | 0.8 | 4.7 | Accept |
| Delayed recall | |||||||
| Q13 | |||||||
| Please repeat the list of 10 words I read earlier. | 1.00 | 1.00 | 0.9 | 0.97 | 1 | 5 | Accept |
| S-CVI/Ave = 0.91 d | |||||||
a I-CVI=item - level content validity index
b CVR= Content Validity Ratio
c Acceptance was based on CVR ≥ 0.62.
d S-CVI/Ave = Scale- level of content validity index/Average = mean of I-CVIs
According to the experts’ ideas, all three items were related to the concept of the questionnaire, but they were difficult in terms of simplicity and transparency. Therefore, to revise these three items, the opinions of experts in the fields of linguistics, speech therapy, clinical psychology, and psychiatry were applied and these three items were replaced by linguistic and cultural phrases that were appropriate to our target community. Therefore, instead of the three words in item 4 (cabin, theater, and giant), which were associated with the free recall, the words ‘home (Khaneh in Persian)’, ‘cinema’, and ‘demon (Div in Persian)’ were replaced, respectively. Also, the item number 8, ‘What is the prickly green plant found in the desert?’, with the answer of cactus, was replaced by the phrase ‘What is the thorny plant found in the desert?’, with the answer of camels- thorn (Khar Shotor in Persian), and the item number 12, ‘Please say this: ‘Methodist Episcopal’, was replaced by the phrase ‘Please say this: Samsum Al-Saltaneh’.
4.2. The Construct Validity of the Persian Version of TICS-m Questionnaire
A total of 210 community-dwelling older adults aged ≥ 50 were included in the study. The sample size for conducting factor analysis was suitable, and data was inter-related and ideal for factor analysis according to the values of KMO and Bartlett’s Test of Sphericity (KMO= 0.737 and Approx, χ2 = 590.92, P < 0.0001). The socio-demographic characteristics of the participants are shown in Table 2.
Table 2.
Profile of Participants in Exploratory Factor Analysis (N = 210)
| Variables | No. (%) | P-TICS-m, Median Score (P25 to P75) a |
|---|---|---|
| Gender | ||
| Male | 108 (51.4) | 29 (26 - 30) |
| Female | 102 (48.6) | 27 (24 - 30) |
| Age | ||
| 50 - 59 | 114 (54.3) | 28 (26 - 31) |
| 60 - 69 | 78 (37.1) | 28 (25 - 29) |
| ≥ 70 | 18 (8.6) | 21 (15.5 - 27.25) |
| Education | ||
| Illiterate | 17 (8.1) | 18 (13.50 - 20.5) |
| Elementary | 49 (23.3) | 27 (24 - 29) |
| Secondary | 24 (11.4) | 28 (26 - 30) |
| Tertiary | 4 (1.9) | 24.5 (21.5 - 29) |
| Diploma | 54 (25.7) | 29 (26 - 30.25) |
| Academic education | 62 (29.5) | 29 (27 - 31) |
a P-TICS-m: Persian version of the telephone interview for cognitive status-modified
As displayed in Table 3, factor analysis with Oblimin rotation method identified six factorial components with eigenvalues of greater than 1, which explained 68.77% of the total variance. All of the 15 items of the instrument remained in the extraction table. Only the location of the items related to each factor changed compared to the English version. For example, in the English version, ‘free recall’ and ‘delayed recall’ were in separate components, but in this study, they were placed under the same component. The Scree plot shows the number of extracted factors (Figure 1).
Table 3.
Factor Analysis After Oblimin with Kaiser Normalization Rotation for the Persian Version of the TICS-m in Middle- aged Adults (N = 210)
| Items | Components | |||||
|---|---|---|---|---|---|---|
| Orientation, Language/Repetition, Semantic | Orientation, Recent Memory | Comprehension, Recent Memory | Registration/ Free Recall, Delayed Recall | Comprehension, Attention/Calculation | Attention/Calculation, Orientation | |
| Q1 (iii) | 0.800 | |||||
| Q3 | 0.617 | -0.483 | ||||
| Q12 | 0.608 | -0.473 | -0.422 | |||
| Q11 | 0.565 | -0.428 | -0.545 | -0.430 | ||
| Q2 | -0.859 | |||||
| Q1 (i) | -0.718 | |||||
| Q9 | -0.657 | -0.511 | ||||
| Q7 | -0.868 | |||||
| Q10 | -0.728 | -0.494 | ||||
| Q4 | -0.900 | |||||
| Q13 | -0.878 | |||||
| Q8 | 0.840 | |||||
| Q5 | 0.709 | |||||
| Q6 | -0.818 | |||||
| Q1 (ii) | -0.456 | -0.430 | -0.505 | -0.700 | ||
| Eigenvalues | 4.209 | 1.648 | 1.273 | 1.098 | 1.084 | 1.002 |
| Variance explained (%) | 28.063 | 10.988 | 8.488 | 7.322 | 7.228 | 6.683 |
| Cumulative % | 28.063 | 39.050 | 47.538 | 54.860 | 62.088 | 68.771 |
a Extraction method: principal component analysis; rotation method: Oblimin with kaiser normalization; factor loadings < 0.4 removed.
Figure 1.
Scree plot diagram for the Persian version of the TICS-m in middle-aged adults
The Cronbach’s alpha for the Persian version of the TICS-m was 0.712. The item-to-total correlations ranged from zero to 0.688 (Table 4). The internal consistency of this questionnaire was moderate (Cronbach’s alpha= 0.712, P < 0.001) (Table 4). Also, the test-retest reliability of the questionnaire was excellent [ICC (95 % CI) = 0.918 (0.828 to 0.961)].
Table 4.
Corrected Item-to-total Correlation of the Persian Version of the TICS-m in Middle-aged Adults (N = 30) a
| Items | Scale Mean if Item Deleted | Scale Variance if Item Deleted | Corrected Item-Total Correlation | Cronbach’s Alpha if Item Deleted |
|---|---|---|---|---|
| Q1 (i) | 21.4333 | 32.323 | 0.214 | 0.711 |
| Q1 (ii) | 19.9333 | 26.892 | 0.547 | 0.669 |
| Q1 (iii) | 21.4333 | 31.702 | 0.518 | 0.704 |
| Q2 | 21.4000 | 32.800 | 0.000 | 0.716 |
| Q3 | 21.4667 | 31.430 | 0.459 | 0.702 |
| Q4 | 18.3333 | 20.023 | 0.688 | 0.630 |
| Q5 | 19.7000 | 20.907 | 0.524 | 0.678 |
| Q6 | 21.5000 | 31.293 | 0.414 | 0.701 |
| Q7 | 21.4000 | 32.800 | 0.000 | 0.716 |
| Q8 | 21.8333 | 32.420 | 0.022 | 0.720 |
| Q9 | 21.5000 | 31.155 | 0.456 | 0.700 |
| Q10 | 21.5000 | 32.121 | 0.169 | 0.711 |
| Q11 | 21.6667 | 30.230 | 0.479 | 0.692 |
| Q12 | 21.6000 | 31.766 | 0.189 | 0.710 |
| Q13 | 18.9000 | 19.541 | 0.632 | 0.648 |
a Overall Cronbach’s alpha of the Persian version of the TICS-m = 0.712.
5. Discussion
This study aimed to provide a questionnaire for the Neyshabur elderly cohort to screen cognitive impairment in community-dwelling middle-aged and older adults using the phone without the need for a face-to-face interview. Since a Persian version of the TICS-m questionnaire has not been revised in Iran so far, this study was conducted to translate the questionnaire into Persian and investigate its reliability and validity in an Iranian adult sample.
After translating the English version of the TICS-m questionnaire into Persian, its content and face validity was evaluated by a panel of ten experts. For the Persian version of the TICS-m questionnaire, the CVR, S-CVI, and Impact scores of each item were above the defined criteria, suggesting good content and face validity. However, the I-CVI value of three items in the questionnaire was less than the acceptable value; so these items were revised.
In order to examine the reliability of the revised questionnaire, it was administered to 30 cognitively healthy older adults. The Cronbach’s alpha (0.712) indicated that the Persian version of the TICS-m questionnaire had satisfactory internal consistency and the value of ICC (0.918) suggested that the questionnaire also had an excellent test-retest reliability. The reported ICC value for the original version of the TICS-m questionnaire was high (ICC=0.99) (12). In the Korean version of TICS-m, the internal consistency (Cronbach’s alpha = 0.87) and ICC (0.95) among cognitively normal individuals aged 60-90 were also high (26). Similar findings have been reported for the Dutch (ICC=0.90) (27) and Japanese (ICC=0.94) versions of TICS-m (28); however, the ICC value in the Italian version was modest (ICC=0.73), since some cases were re-evaluated by a different examiner (29). One reason for the high value of the Cronbach’s alpha in the original version of the TICS-m questionnaire could be that there were a greater number of items in this questionnaire than the modified TICS-m questionnaire. For clinical applications, an ICC value of at least 0.90 is often recommended (30). Our results showed that the corrected item-to-total correlation for items Q1 (i) (What day of the week is it?), Q2 (What is your age?), Q7 (What do people usually use to cut paper?), Q8 (What is the prickly green plant found in the desert?), Q10 (Who is the prime minister now?), and Q12 (Please say this: ‘Methodist Episcopal’) was lower than 0.3, suggesting that these items may not belong to the scale and Cronbach’s alpha (0.712) increased only slightly when the items Q1 (i), Q7, and especially Q8 (α= 0.720) were removed from the scale.
Factor analysis of the TICS-m items in the present study yielded six factors: ‘orientation’, ‘registration/free recall’, ‘attention/calculation’, ‘comprehension, semantic, and recent memory’, ‘language/repetition’ , and ‘delayed recall’, which were consistent with the original version of the TICS-m questionnaire (12). However, van den Berg et al. performed factor analysis on the TICS-m to examine the underlying latent constructs; they extracted four factors including ‘verbal memory’, ‘orientation/mental tracking’, ‘language/reasoning’, and ‘attention/working memory’ (31). These differences may be justified by differences in the populations.
In summary, the TICS-m questionnaire, which is used for screening of dementia in older adults, especially when in-person interviews are not possible, had good internal consistency and excellent test-retest reliability in its Persian version, and the six extracted factors accounted for 68.8% of the total variance.
5.1. Limitations
In content validity studies, sampling bias may occur because the selection of experts is purposive. Also, our sampling population was selected from one center, the Neyshabur cohort population, which restricted the generalizability of the results. Due to the financial problems, for assessing the construct validity of translated TICS-m, only EFA was used and the confirmatory factor analysis was not performed. The concurrent validity of the TICS-m with the MMSE questionnaire was performed, but its results were not presented in this article.
5.2. Ethical Consideration
This study was approved by the Regional Ethics Committee at Tabriz University of Medical Sciences (IR.TBZMED.REC.1397.569). The project was implemented completely in the NeLSA Centre in Neyshabur. Participants were invited to take part in the study by telephone. Participation was voluntary and oral consent was taken. All principles of confidentiality for patients’ information were considered.
