Chilean version of the INECO Frontal Screening (IFS-Ch): psychometric properties and diagnostic accuracy

OBJECTIVE This study sought to analyze the psychometric properties and diagnostic accuracy of the Chilean version of the INECO Frontal Screening (IFS-Ch) in a sample of dementia patients and control subjects. METHODS After adapting the instrument to the Chilean context and obtaining content validity evidence through expert consultation, the IFS-Ch was administered to 31 dementia patients and 30 control subjects together with other executive assessments (Frontal Assessment Battery [FAB], Modified version of the Wisconsin Card Sorting Test [MCST], phonemic verbal fluencies [letters A and P] and semantic verbal fluency [animals]) and global cognitive efficiency tests (Mini mental State Examination [MMSE] and Addenbrooke's Cognitive Examination-Revised [ACE-R]). Caregivers of dementia patients and proxies of control subjects were interviewed with instruments measuring dysexecutive symptoms (Dysexecutive Questionnaire [DEX]), dementia severity (Clinical Dementia Rating Scale [CDR]) and functional status in activities of daily living (Activities of Daily Living Scale [IADL] and Technology-Activities of Daily Living Questionnaire [T-ADLQ]). Convergent and discriminant validity, internal consistency reliability, cut-off points, sensitivity and specificity for the IFS-Ch were estimated. RESULTS Evidence of content validity was obtained. Evidence of convergent validity was also found showing significant correlations (p<0.05) between the IFS-Ch and the other instruments measuring: executive functions (FAB, r=0.935; categories achieved in the MCST, r=0.791; perseverative errors in the MCST, r= -0.617; animal verbal fluency, r=0.728; "A" verbal fluency, r=0.681; and "P" verbal fluency, r=0.783), dysexecutive symptoms in daily living (DEX, r= -0.494), dementia severity (CDR, r= -0.75) and functional status in activities of daily living (T-ADLQ, r= -0.745; IADL, r=0.717). Regarding reliability, a Cronbach's alpha coefficient of 0.905 was obtained. For diagnostic accuracy, a cut-off point of 18 points (sensitivity=0.903; specificity=0.867) and an area under curve of 0.951 were estimated to distinguish between patients with dementia and control subjects. DISCUSSION The IFS-Ch showed acceptable psychometric properties, supported by evidence of validity and reliability for its use in the measurement of executive functions in patients with dementia. The diagnostic accuracy of the IFS-Ch for detecting dementia patients was also considered acceptable.


INTRODUCTION
E xecutive functions constitute a group of higher order abilities that coordinate basic cognitive processes in order to regulate, control and execute goal-oriented behaviors that require new and creative solutions. [1][2][3] These include a wide range of cognitive processes such as inhibition, working memory, shifting, verbal reasoning, multitasking and planning, 4,5 all of which involve significant activity of the frontal lobes and "frontal lobe systems", i. e. those areas with direct connections with the frontal lobes. 6 This cognitive domain is impaired in numerous neurological and neuropsychiatric pathologies, such as focal lesions involving the frontal lobes (abscesses, strokes or tumors), inflammatory diseases, neurodegenerative disorders, schizophrenia, obsessive compulsive disorder, etc. 7 Executive dysfunction has also been observed early in most types of dementia, to the point where some authors have defined it as its core symptom. 8 Accordingly, the assessment of executive functions contributes to an early diagnosis of dementia. Moreover, executive deficits are prominent symptoms of some dementia syndromes, such as frontotemporal dementia (behavioral variant) 9 and vascular dementia 10,11 Hence, the assessment of this cognitive domain also contributes to the differential diagnosis of the specific type of dementia.
The above-mentioned facts, together with the high and increasing prevalence of dementia, 12 have prompted the development of executive screening tests to be applied in neurological and general medical practice with elderly patients that can provide brief and quick assessment of this cognitive domain. The INECO Frontal Screening (IFS) is an executive screening test that assesses several executive processes using a few tasks. 13 It comprises three of the subtests included in the Frontal Assessment Battery (FAB) -another executive screening test that has shown good characteristics for assessing executive dysfunctions: 14,15 those which have shown the highest sensitivity according to the test author's everyday clinical experience 13 as well as empirical evidence 16 (Luria Motor Series, Conflicting Instructions and Go-no go). In addition, the IFS includes new subtests, most of them assessing various dimensions of working memory. Figure 1 shows the detailed structure of the IFS, describes the variables assessed by the test, its indicators and sub-indicators, and the subtests that measure each indicator or subindicator.
Since the IFS has only been validated in Argentina and to the best of our knowledge neither content validity nor correlation of the IFS with functionality and dysexecutive behaviors in daily living have been examined, it would be valuable to consider these aspects in order to complement the study of the instrument. Therefore, the aim of the present study was to adapt the IFS to the Chilean cultural context and evaluate its psychometric properties and diagnostic accuracy in a sample of control subjects and dementia patients.

Subjects.
The study was carried out in a convenience sample, which included participants of both sexes, Spanish speakers, aged 52 or older, with at least three years of formal education. All subjects had a proxy that gave relevant information about their everyday activities and behavior. Subjects were divided into two groups: A clinical sample, including 31 patients recruited from the Cognitive Neurology and Dementias Unit (Unidad de Neurología Cognitiva y Demencias) of the Neurology Service at the Hospital del Salvador in Santiago, Chile. The diagnosis of dementia was provided by a Neurologist based on detailed neurological, neuropsychological, laboratory, and neuroimaging data from each participant. The first step in the diagnostic process was to determine the presence of dementia using the DSM-IV-TR criteria. 17 When these criteria were met, the Neurologist determined the specific type of dementia using multiple diagnostic criteria for AD (i. e., NINCDS-ADRDA), vascular dementia (i. e., AD-DTC, NINDS-AIREN), Dementia with Lewy Bodies (i. e., third report of the DLB Consortium) or frontotem-Ihnen J, et al.
The control sample comprised 30 subjects with similar socio-demographic characteristics (age, sex and years of education) to those of the clinical sample. All participants included in this group had CDR=0 and presented no symptoms or history of neurological or psychiatric diseases.
Finally, exclusion criteria for both groups were: [1] presence of depression as measured by the Geriatric Depression Scale (score ≥5 points); [2] presence of Anxiety Disorder as measured by the Zung Scale (score ≥51 points); and [3] presence of severe sensory deficits (loss of vision and/or hearing) that could impede test administration.
IFS and other neuropsychological tests. As outlined above, the IFS is a screening test for executive dysfunctions. The tasks included in the IFS are: Luria motor series (3 points), Conflicting instructions (3 points), Go-no go (3 points), Months backwards (2 points), Backwards digit span (6 points), Modified Corsi tapping test (4 points), Proverb interpretation (3 points) and Modified Hayling Test (6 points). Thus, the IFS has a maximum possible score of 30 points. High scores indicate preservation of the executive functions. In this study, the IFS was adapted to the Chilean cultural context (IFS-Ch) and then administered to all subjects.
All subjects were assessed with the following executive tests to estimate convergent validity. [1] The Modified version of the Wisconsin Card Sorting Test (MCST), 22 a brief version of the widely known Wisconsin Card Sorting Test 23,24 designed originally to study "abstract behavior" and "set-shifting ability" and later proposed as being sensitive for assessing frontal damage. 24 The MCST is a classification task in which the subject must find the sorting criteria and maintain it for a number of trials. 14 This particular version was used as it simplifies and reduces ambiguity in administration, making it more suitable for elderly patients. 25 [2] Verbal fluency tasks, or controlled oral word-association, in which subjects have to generate words following a given criteria. This test is sensitive for assessing executive dysfunction 24,26 and semantic memory impairment. 24

Groups of tasks (executive processes) Indicators Subindicators Subtests
Executive functions  Semantic verbal fluency (animals) and phonemic verbal fluencies (letters A and P) 27 were specifically used. [3] The FAB, a screening test for executive dysfunction that assesses conceptualization, mental flexibility, motor programming, resistance to interference, inhibitory control and environmental autonomy. 14 All participants were also tested with global cognitive efficiency measures: [1] the Mini Mental State Examination (MMSE), 28 the most commonly used cognitive screening test internationally; 29  Procedure. The IFS was first adapted to the Chilean cultural context and its content validity was assessed by consultation with experts through a content validity questionnaire. All subjects were assessed by the modified IFS (IFS-Ch) and the other instruments previously described.
Statistical analysis. All statistical analyses were performed with significance level set at 0.05. Data analysis was performed with PASW Statistics 18 software. Differences in sex were analyzed using the χ² test. Differences in age, years of education and test scores between groups were analyzed using the t test for independent samples. A one-way MANOVA analysis was conducted to compare results across subtests of the IFS-Ch by diagnostic category. The correlations between scores of two tests were evaluated using the Pearson coefficient, with the exception of the association between CDR and IFS-Ch scores, for which the Spearman rank correlation test was employed. Reliability was assessed using the Cronbach's alpha coefficient. The sensitivity and specificity of the IFS-Ch for detecting the presence of dementia were evaluated using the receiver operating characteristic (ROC) analysis.
Ethical concerns. The study was approved by the Ethics Committee at the Servicio de Salud Metropolitano Oriente. Informed consent was obtained from control subjects, dementia patients and their closest relatives.

RESULTS
Adaptation. Given its sociocultural nature, the proverb interpretation subtest of the IFS was adapted to the Chilean cultural context. Using a four-point Likert scale, six experts in the neuropsychological field were consulted about the capacity of the three proverbs included in the original test and three proverbs proposed as relatively common in Chile to assess executive function and their level of familiarity in the Chilean cultural context. The three proverbs that presented the highest means and the lowest standard deviations were selected. Table 1 summarizes the statistical parameters for the experts' responses. Only minor modifications were made to the rest of the test administration procedure and scoring instructions in order to standardize the assessment procedure as much as possible.
Demographic and neuropsychological data. Table 2 shows demographic and neuropsychological data for the clinical and control samples. No significant differences in sex, age or years of education were found among the groups (p>0.05). In contrast, the scores of all the instruments administered to subjects and their informants differed significantly between the studied groups (p<0.05).

Influence of socio-demographic variables on IFS-Ch perfor-
mance. In order to determine the influence of demo- graphic variables on IFS-Ch performance, the correlation between demographic characteristics and IFS-Ch total scores was estimated. No significant association was found between IFS-Ch total scores and age (r= -0.197; p>0.05), whereas a significant correlation was found between IFS total scores and years of education (r=0.48; p<0.001). Regarding sex, no significant gender differences were found on IFS performance (t= -0.25; p>0.05). In summary, only years of education showed an influence on IFS performance.

Evidence of validity. Content validity.
Five experts with at least two years of experience in the field of neuropsychology answered a content validity questionnaire designed for the IFS-Ch. In this questionnaire, the conceptual and operational definitions of executive functions and its indicators were presented. The definition of each indicator was followed by the administration and scoring instructions for the corresponding subtest. Subsequently, the experts were asked about the capacity of each subtest to assess executive function, its capacity to measure the corresponding indicator, and the clarity of the administration and scoring instructions, leaving a space for any other observations. All the experts agreed that each of the subtests measured executive functions and that each subtest assessed its respective indicator.
For 5 of the 8 subtests, all the experts considered that the instructions were formulated clearly, while for the 3 remaining subtests, one expert considered that the instructions were formulated poorly. The latter expert suggested changes to clarify the instructions, which were later incorporated into the test. A new version of the IFS-Ch was then devised according to these observations. This new version had only minor differences compared with the original test.
Discriminant validity. The performance of the two groups differed significantly (p<0.05). Average total scores on the IFS-Ch and each of its subtests were significantly lower in the clinical sample (   ency with letters A and P, semantic verbal fluency of animals and the FAB); global cognitive efficiency (ACE-R-Ch and MMSE); dysexecutive symptoms (DEX); dementia severity (CDR); and functionality (IADL and T-ADLQ). The coefficients estimated for each association are given in Table 4. The association between IFS-Ch and measures of global cognitive efficiency indicates no evidence of divergent validity.
Evidence of reliability. The Cronbach's alpha coefficient calculated for the total test was 0.901. Regarding the subtests that included more than one item, the Cronbach's alpha coefficient was 0.577 for the Modified Corsi tapping test, 0.781 for the Proverb interpretation task, and 0.836 for the Modified Hayling test.

Diagnostic accuracy.
A ROC curve analysis on the IFS-Ch total score between control subjects and dementia patients generated several cut-off points, with 18 points being the best balance between sensitivity and specificity (sensitivity=0.903; specificity=0.867). The area under the curve (AUC) was 0.951 ( Figure 2). There were no significant differences among the areas under the curve of the IFS-Ch, FAB, categories completed on the MCST, Animals verbal fluency, A verbal fluency, and P verbal fluency (p>0.05). 35

DISCUSSION
In this paper, the IFS-Ch has shown good psychometric properties and diagnostic accuracy. First, it has shown validity evidence from multiple sources: content validity Table 3. Performance of dementia patients and control subjects in the IFS-Ch and its subtests.  through expert consultation, discriminant validity by comparing the means of IFS-Ch scores between groups, and convergent validity through associations between IFS-Ch scores and other executive and related measures. Second, the IFS-Ch demonstrated evidence of reliability, exhibiting a good internal consistency coefficient. This is relevant given that reliability is a common weakness of executive tests. 36,37 With regard to diagnostic accuracy, the selected cut-off point produced an excellent AUC as well as a very good balance between sensitivity and specificity for detecting dementia.

Dementia patients (n=31) Control subjects (n=30) t Significance
Although the test and two subtests (Modified Hayling test and Proverb interpretation) showed evidence of very good reliability, the Modified Corsi tapping subtest had a poor internal consistency coefficient. Further studies are needed to determine whether this subtest provides an accurate measure of spatial working memory.
It is noteworthy that the cut-off point found in this study (18) was much lower than that found in the original publication (25). This fact is probably due to the socio-demographic differences between both samples, particularly in relation to years of education. In our study, the mean of this parameter was 11.93 and 9.65 years for the control and clinical groups, respectively, whereas in the Argentinian investigation the mean for bvFTD, AD and control subjects was 16.3 years, 14.5 years and 14.5 years, respectively. These differences are coherent with our finding that years of education exhibited a significant association with IFS-Ch total scores and with results of studies showing that education is an important variable in executive test performance in general. [38][39][40] Overall, these data suggest that it is important to formulate local norms in order to interpret IFS-Ch scores accurately.
One of the main findings of this study was that the IFS-Ch showed a good association with functionality measures such as the IADL and T-ADLQ. This is coherent with the findings of previous studies which suggest that executive tests predict functional impairment more accurately than tests that assess other cognitive domains, [41][42][43] which is expected given that daily life activities are mainly goal-oriented behaviors. Thus, the described association contributes with evidence of convergent validity for the IFS-Ch.
Similarly, a good association was found between IFS-Ch total scores and the DEX, a questionnaire that assesses dysexecutive symptoms. In other words, the IFS-Ch, despite the fact that it is a non-ecological measure -i. e. it is a standardized test administered in a laboratory type setting -correlates significantly with the presence of dysexecutive behaviors in everyday life. The latter constitutes not only evidence of convergent validity for the test, but also suggests that it presents good ecological validity, a relevant and highly desirable feature for an executive assessment instrument. 4,44 One limitation of our study is the small number of subjects by category of dementia, a situation precluding proper assessment of the capacity of the IFS-Ch to discriminate between different types within the pathology. Moreover, the greater number of patients with AD compared with patients with bvFTD or VD, meant that most of our clinical group presented a multideficit clinical profile, a situation that could explain the significant correlation found between IFS-Ch total scores and measures of global cognitive efficiency (ACE-R-Ch and MMSE). Evidence of divergent validity for the IFS-Ch should be studied with patients presenting deficits mainly in the executive domain. Further research is needed to determine whether the IFS-Ch can differentiate between different forms of dementia and to obtain further evidence of divergent validity.
Acknowledgements. This study was supported by grants: Project FONDECYT No. 1100975 CONICYT and PIA-CONICYT Project CIE-05. We thank Constanza Ihnen for her English review of the manuscript.

-Specificity
Source of the curve Diagonal segments are produced by ties.