• apathy;
  • apathy evaluation scale;
  • family distress;
  • reliability;
  • validation


  1. Top of page
  2. Abstract

Aim:  Apathy is a well-recognized symptom of Alzheimer's disease (AD). The aim of the present study was to validate the Taiwanese version of the Apathy Evaluation Scale, clinician version (AES-C) and assess the severity of apathy in patients with AD.

Methods:  Comprehensive evaluations were conducted in a total of 144 AD patients. This study used a cross-sectional comparative design. Data were collected from clinical interviews using the AES, the Mini-Mental Status Examination (MMSE), the Neuropsychiatric Inventory (NPI), and the Clinical Dementia Rating Scale (CDR).

Results:  Internal consistency was indicated by Cronbach's alphas of 0.85; test–retest reliability was 0.89 over a period of 3 days. Criterion-related validity was supported by the fact that AES-C significantly correlated with the apathy scores of the NPI. Factor analysis indicated a three-factor structure. Convergent validity was supported by a positive correlation between the AES-C score and the anxiety score of the NPI. Discriminant validity was supported by the fact that the AES-C scores did not correlate with the depression subscale of the NPI, and the correlation between the AES-C score and the euphoria score of the NPI score was negative. Known-group validity was supported by results showing that AD patients in a moderate stage of dementia (CDR = 2) had significantly higher scores on the AES-C than patients with mild-stage dementia (CDR = 1).

Conclusion:  The AES-C is a reliable and valid instrument for measuring symptoms of apathy among AD patients in Taiwan.

ALZHEIMER'S DISEASE (AD) is increasingly becoming a critical issue in countries with aging populations. The prevalence of dementia among the elderly is around 2–4% in Taiwan.1 This rate is lower than that in developed countries. Possible reasons to explain the low prevalence in Taiwan include that the elderly are relatively young, that the mortality rate among dementia patients is high, and that there may be underestimation of the extent of the problem due to the difficulty of AD diagnosis. AD is the leading cause of dementia,1 and 80–90% of AD patients have neuropsychiatric symptoms.2 The most frequently observed neuropsychiatric symptoms in AD patients are night-time behavior and apathy as measured on the Neuropsychiatric Inventory (NPI).3 Such symptoms interfere with the quality of life for both patients and caregivers, and may be an important factor in decision-making about how to manage an elderly person's disease.4

Apathy is a symptom defined as lack of motivation not attributable to diminished level of consciousness, cognitive impairment, or emotional distress related to daily function.5,6 Apathy is reported frequently in patients with Parkinson's disease, AD, vascular dementia, and psychosis.4,7–9 Apathy in AD has been consistently associated with relatively more severe cognitive deficits, more severe impairments in activities of daily living, higher levels of burden and distress in caregivers, and increased resource use.10,11 It is also associated with reduced functional level, decreased response to treatment, and poor illness outcome.12

Starkstein et al. noted that apathy in patients with AD is associated with poor insight into their apathy syndrome, and the low self-ratings of AD patients is related to their difficulty in remembering behavioral changes from the time before disease onset.10 Because AD patients have poor insight into their apathy syndrome and cognitive deficits, screening and early detection of the symptoms are important to facilitate differential diagnosis by clinical staff and to guide families in grappling with patients’ performance difficulties.

In recent years, many comprehensive apathy scales have been developed. They include the NPI,13 the Apathy Evaluation Scale (AES),5 the Apathy Scale14 and the Dementia Apathy Interview and Rating.15 Although all of these assessment instruments are well-established and have been shown to have good reliability and validity, the AES has several advantages over the others and has greatly facilitated and promoted research into apathy. The AES is the most widely used to characterize and quantify symptoms, and warrants scrutiny.16

The AES was developed based on a definition of apathy as a neuropsychiatric syndrome of primary motivational loss not attributed to emotional distress, intellectual impairment or diminished level of consciousness.5,6 The AES has three versions: the self-rated (the AES-S), the informant-rated (AES-I), and the clinician-administered (the AES-C). The psychometric properties of the AEC-C have been studied in different diagnosis groups, and apathy is a well-recognized AD symptom.5,6,10,17 An abbreviated version of the AES has been used in recent studies of nursing home residents with dementia.18

Only one study has reassessed the factor structure and AES psychometric properties.16 Clarke studied a more homogenous patient group of dementia patients, and found that the AES-C has fairly good psychometric traits.16 Two factors (apathy and interest) were identified on the AES-C. To our knowledge, no previous studies have examined the psychometric properties of the AES-C in an AD patient group in Taiwan. The AES has been translated into many languages,18,19 and its psychometric properties have been established. Validation of the Taiwan version of the AES may provide a reliable tool to rapidly screen for apathy in Taiwanese AD patients, and will allow study results to be compared with those from different countries.

The aim of the present study was to establish the validity and reliability of a Taiwanese version of the AES-C and to assess severity of apathy in AD.


  1. Top of page
  2. Abstract

Subjects and study design

We used a cross-sectional and descriptive correlation design in this study. We recruited subjects from eight nursing homes and three outpatient clinics of general medical hospitals in Taiwan. The researcher group (geriatric psychiatrist, psychiatric nurse, and research nurse trained in psychometric measurement) did the comprehensive evaluations.

The evaluations included a history of the present illness, mental status examination, and standardized neurologic examination. In addition, we reviewed the subjects’ medical records and prescriptions (excluding selective serotonin re-uptake inhibitors), and, when available, we reviewed the laboratory tests (complete blood count and routine chemistry) over the previous 6 months. Physicians helped to make the diagnosis of dementia based on the DSM-IV criteria.20 The dementia subtypes (AD, vascular dementia and others) were based on standard clinical criteria currently in use.21 We also identified cognitive, neurological, and psychiatric disturbances. We included subjects who were at least 60 years of age. Those who had a diagnosis of delirium, a history of alcoholism or drug abuse, or who had received a new psychotropic treatment <15 days before the evaluations were excluded from the study. A total of 144 subjects was included in the study. All research subjects provided written informed consent in conformity with guidelines of the institutional review board at Taipei Medical University. The administrators of long-term care facilities approved this study.


Apathy Evaluation Scale, clinician version

We used the AES-C to measure indicators of apathy in the previous 4 weeks.6 It consists of an 18-item questionnaire to evaluate behavior, cognition and emotion subscales. The items are rated on a 4-point Likert scale from 1 (not at all characteristic) to 4 (very characteristic). The AES-C has multiple forms that have has been shown to have good reliability and validity.6,16,18 Higher scores indicate higher levels of apathy, reflecting a lower level of motivation to get things done during the day.

The AES-C was developed using a translation and back-translation process. The AES-C was first translated from English into Taiwanese by a bilingual person. A second bilingual person who had not seen the original English AES-C version, back-translated from Taiwanese into English. The two English translations were then compared for consistency. If the back-translated items and the originals did not agree, the first translator attempted a second translation, which was then compared to the original. This process was repeated until the back-translated items and the originals agreed.

Mini-Mental Status Examination

The Mini-Mental Status Examination (MMSE) is a brief 30-point questionnaire that is commonly used to screen for cognitive impairment in the elderly.22 The scale is used to assess cognitive function, including orientation, registration, attention and calculation, recall, and language, and a score of ≤23 indicates cognitive impairment.22 The MMSE has been validated and extensively used in both clinical practice and research.

Neuropsychiatric Inventory

The NPI was developed to assess psychopathology in patients with dementia and other neuropsychiatric disorders. The NPI has 12 sub-areas (delusions, hallucinations, agitation, apathy, anxiety, depression, euphoria, irritability, disinhibition, aberrant motor behavior, changes in appetite, and night-time behavior disturbances).13 The NPI is based on a structured interview with a caregiver who is familiar with the patient. Frequency is rated 1–4 (1, occasionally; 2, often; 3, frequently; 4, very frequently) and severity is scored 1–3 (1, mild; 2, moderate; 3, severe). The product of severity × frequency is calculated for each behavioral change during the previous month or since the last evaluation. Cronbach's α for overall reliability of the NPI was 0.88 and the concurrent validity was good, as shown by an acceptable correlation between the NPI scores and other validated measurements.13

Clinical Dementia Rating Scale

The Clinical Dementia Rating Scale (CDR) was used to evaluate stage of severity of dementia. The CDR was developed primarily for use in AD patients. This is a 5-point scale with scoring from 0 (no impairment), 0.5 (very mild dementia) to 3 (severe impairment). CDR-1 indicates a mild stage, CDR-2 a moderate stage, and CDR-3 a severe stage.23,24 The CDR score was based on information obtained about the patient's everyday performance in six domains.

Questionnaire for demographic and disease information

A demographic information sheet covered basic patient information including age, education years, sex, marital status, religious affiliation, recruitment sites, and the disease degree level.

Statistical analysis

We evaluated the reliability and validity of the AES-C. Internal consistency was established by calculating the Cronbach's alpha coefficient, which ranges from 0 to 1, with higher values indicating less measurement error. Test–retest reliability was evaluated by calculating the Pearson product moment correlation coefficient between pre-test and post-test assessments, with a 3-day interval in a sample of 12 patients.

Criterion-related validity was established using the Pearson product moment correlation coefficient between AES-C scores and the apathy scores of the NPI. Construct validity was established by principal-axis factor analysis with direct oblimin rotation. We chose the items of anxiety and depression for the convergent and discriminate validity analysis of AES-C towards the different negative symptoms on the NPI, because their definition most closely resembles the Marin et al. definition of apathy.6 We chose the item of euphoria because this word indicates the extreme opposite. We computed the Pearson product moment correlation coefficient between AES-C scores and the MMSE. Known-group validity was established by comparing the AES-C scores among patients with different stages of severity of dementia (CDR-1 = mild, CDR-2 = moderate). We hypothesized that AD patients with a moderate stage of cognitive dysfunction would obtain higher scores on the AES-C than those with a mild stage of cognitive dysfunction. We divided the total sample into four groups with different degrees of apathy. We used one-way analysis of variance to test for differences in the MMSE score of these four groups. We then used post-hoc tests (Bonferroni-corrected) to determine differences between the groups. We further hypothesized that patients would experience more severe apathy symptoms when they had poorer cognitive function.

All statistical procedures were performed using SPSS version 15.0 (SPSS, Chicago, IL, USA). The differences of the groups were considered significant for P < 0.05.


  1. Top of page
  2. Abstract

Participant characteristics

Comprehensive evaluations were conducted in a total of 144 AD patients. Table 1 lists demographic and disease-related characteristics of all valid subjects. The mean ± SD age was 74.5 ± 9.1 years and the mean years of education was 7.3 ± 4.9. Most subjects were widowed (57.6%), and the main religious affiliation was Buddhist/Taoist (61.8%). Nearly 82% of participants were recruited in nursing homes. Most were diagnosed with mild AD (CDR = 1).

Table 1.  Subject characteristics (n = 144)
Age (years)74.59.1
Education (years)7.34.9
  1. OPD, outpatient department.

Marital status  
Religious affiliation  
Recruitment sites  
 Nursing home11881.9
Stage of dementia  

Reliability of AES

Internal consistency was established by calculating Cronbach's alpha coefficients, which were 0.85 for the total scale. Test–retest reliability was evaluated by calculating the Pearson product moment correlation coefficient between pre-test and post-test assessments over a 3-day period, for a sample of 12 AD patients. The test–retest reliability was 0.89 for all items scored by the AES-C. The scale scores were significantly correlated with the apathy scores of the NPI (r = 0.61, P < 0.001).

Validity of AES

Factor analysis was used to determine the underlying constructs measured by the items in the AES-C. Principal axis factor analysis with oblimin rotation indicated six factors with an eigenvalue > 1.0, accounting for 68.23% of the variance. From the scree plot, both a three- and a four-factor model could fit. A three-factor model gave the most clinically meaningful results. The factor loading was high, ranging from 0.37 to 0.74, which indicated that the AES-C items were associated with three factors (Table 2). Three factors had eigenvalues ≥1.0, accounting for 49.32% of the variation in the current analysis.

Table 2.  Factor analysis and factor loadings of the AES-C (n = 144)
ItemsFactor loading
Factor 1Factor 2Factor3
  1. AES-C, Apathy Evaluation Scale, clinician version.

Variance %28.9111.578.84
 1. S/he is interested in things.0.74  
 2. S/he gets things done during the day.0.52  
 3. Getting things started on his/her own is important to him/her.0.41  
 5. S/he is interested in learning new things.0.44  
 6. S/he puts little effort into anything.0.46  
 8. Seeing a job through to the end is important to her/him.0.49  
 9. S/he spends time doing things that interest her/him.0.51  
10. Someone has to tell her/him what to do.0.67  
11. S/he is less concerned about her problems than s/he should be.0.66  
16. Getting things done during the day is important to her/him.0.53  
17. S/he has initiative.0.47  
 4. S/he is interested in having new experiences. 0.48 
 7. S/he approaches life with intensity. 0.67 
15. S/he has an accurate understanding of her/his problem. 0.56 
18. S/he has motivation. 0.54 
12. S/he has friends.  0.39
13. Getting together with friends is important to her/him  0.37
14. When something good happens, s/he gets excited.  0.52

We verified the convergent and discriminate validity as we hypothesized. The Pearson product moment correlation between the AES-C and the anxiety score of the NPI showed significant correlation (r = 0.35, P < 0.001). The AES-C scores were not significantly correlated with the depression scores of the NPI (r = 0.16, P > 0.05). The Pearson product moment correlation between the AES-C and the euphoria score was negative and statistically significant (r = −0.46, P < 0.001).

The Pearson product moment correlation between the AES-C and the MMSE was negative and statistically significant (r = −0.48, P < 0.001). The results supported the hypothesis that the AES-C scores would correlate negatively with the MMES score. We also examined the relation between cognitive function and apathy symptoms, and found that this patient subgroup differed significantly on MMSE score (F3.140 = 15.41, P < 0.001; Figure 1). As can be seen in Figure 1, on Bonferroni-corrected t-test, the patient group in the lowest quartile of apathy symptoms (first quartile of the AES-C distribution) had the highest significant MMSE scores of all groups. Results supported the research hypothesis that patients with poor cognitive function would experience the most severe apathy symptoms.


Figure 1. Mean Mini-Mental Status Examination (MMSE) score vs Apathy Evaluation Scale-clinician version (AES-C) score. AES-C quartiles: first, 37–51 points (n = 35); second quartile, 52–56 points (n = 35); third quartile, 57–61 points (n = 35); fourth quartile, >62 points (n = 39). Cognitive function in dementia decreased with rising apathy. *P < 0.05 (Bonferroni corrected t-tests).

Download figure to PowerPoint

Known-group validity, according to our hypothesis that AD patients with moderate dementia (CDR = 2) would report significantly higher scores on the AES-C than patients with mild dementia (CDR = 1), was confirmed (t = −2.98; P < 0.005).

The symptom of apathy in the past month was present for 61.7% of the present patients. The apathy score on the NPI-D was 1.10 ± 1.36 (mean ± SD; maximum, 4; minimum, 0).


  1. Top of page
  2. Abstract

In the present study the AES-C was found to have satisfactory internal consistency. This finding agrees with that of the original study by Marin et al.6 and that of another study by Clarke et al.25 The test–retest reliability of the present study (r = 0.89) was as high as that of the Marin et al. study.6

In the present study, apathy symptoms were measured according to three factors, which were identified as a single predominant factor and two minor factors by the original scale constructor.6 In another study reanalyzing the factor structure and psychometric properties of the AES-C in a wide range of clinic outpatients, two factors (apathy and interest) were identified.25 In another study, Faerden et al.'s factor analysis of the AES-C in patients with a first episode of psychosis identified three subscales: apathy, insight and social.9 Like the present study (Table 2), these previous studies showed that corresponding factor solutions shared six items in the main factor (item 2, 6, 8, 9, 10, 17).8,9 But the present findings are also similar to those of the original study, in identifying apathy and interest as factors. Compared with the results of the three studies by Marin et al.,6 both the present study and Faerden et al. study included two items about social contact (items 12 and 13).9 In examining AES-C psychometric properties for measuring apathy in dementia patients, we found that one recent study also identified one main factor of the AES-C as including 11 items.25 This agrees with the present results. The major difference between the present study and the other studies relates to a minor factor, the items of which belong to another factor in the other studies. One reason for this discrepancy could be the heterogeneity of research subjects in each study.

In the Lueken study on developing a short version of the AES, the investigator found that the short-version AES is a valuable and time-efficient instrument for assessing apathy.18 It would therefore be worthwhile to develop a short version of the AES-C.

The present study thus shows that the AES-C has acceptable criterion-related validity, and that apathy syndrome was found to have a positive significant correlation with apathy scores of the NPI. Similar to the Clarke et al. study,25 the present study found fair validity for the AES-C, with its significant correlation with the apathy subscale of the NPI.

In the Andersson et al. study, the investigators found that apathy was correlated with the negative symptoms of anxiety, indicating good convergent validity, and that the AES-C was not significantly correlated with depressed mood symptoms, indicating good discriminant validity.8 This agrees with the present findings. Marin et al. also observed that the AES-C shows discriminant validity by its low correlation with depressed mood.17 Apathy, anxiety, and depression represent disorders with overlapping symptoms, but are in different constructs. Therefore, Marin et al. emphasized the importance of the AES-C not correlating very strongly with measures of depression and anxiety, and hypothesized a priori that this would be confirmed in their study.6 More conclusive evidence of the discriminant validity of the AES-C was found in subsequent studies, showing that the AES-C is poorly correlated with depressed mood, not reaching statistical significance.17 The evidence for the discriminant validity of the AES-C is thus satisfactory. Levy et al. emphasized that apathy is not depression, and that they should be expressed separately in different dementia syndromes and should co-occur at varying levels of disease severity if depression is associated with apathy.26

We also found that the AES-C had a significant inverse correlation (P < 0.05) with MMSE scores. This agrees with the Starkstein et al. study of apathy in dementia patients.10 Resnick et al. reported that a significant, inverse correlation exists between cognitive impairment (MMSE) and AES-C.27 This indicates that more severe apathy is significantly correlated with more severe cognitive deficits.10,25

Apathy is the only neuropsychiatric symptom that is significantly different between the groups when categorized according to CDR.28 Furthermore; apathy is more frequently seen in subjects with moderate to severe dementia. In the present study, patients with higher AES-C scores were significantly more likely to have moderate rather than mild AD, and more than half of moderate AD patients experienced apathy symptoms.

In addition, apathy is more evident than depression in AD patients, and may be more common among patients with a later age of dementia onset.17 From a diagnosis viewpoint, we suggest that the AES-C has the greatest validity.

In the Reekum et al. study they reviewed data derived from an extensive literature search of Medline and combined all of the study findings for AD outpatients, providing a point prevalence of 60.3%.12 In another study, severe apathy symptoms were present in 61.7% of patients.11 This indicates that AD patients have poor awareness of their emotions, blunting and lack of initiative,29 and that apathy is the most frequent neuropsychiatric syndrome in AD.26 In one longitudinal study, initial cognitive status was controlled for, and it was found that loss of interest contributes significantly to predicting AD over time.30 The authors therefore suggested that the need of discriminating the treatment effects on apathy versus depression carefully.30

Severe apathy symptoms in AD patients cause greater caregiver distress, and the pattern of behavioral disturbances produces varying caregiver responses.31 Thus, a critical need exists for research to explore more complex interventions (e.g. combination of pharmacological and other strategies), and to focus on a broader range of outcomes (e.g. caregiver burden). Clinicians who treat AD patients, particularly those in outpatient clinics, where most care for AD patients is carried out, or in nursing home settings or residential settings involving dementia, should pay increased attention to identifying and treating apathy disturbances, to relieve the distress of caregivers.

The present study had three limitations. First, we calculated test–retest reliability with a small sample size. Second, we did not use receiver operating characteristic analysis. Third, we did not examine the psychometric properties of the informant and subject versions for AD patients. In future studies on this topic, these issues should be addressed.

In this validation study of the Taiwanese version of the AES-C, we found that the scale had good reliability, validity, and sensitivity and agreed with the psychometrically validated English and Japanese versions. We also found that reliability was supported by high internal consistency (Cronbach's alpha) and test–retest coefficients. Further, the overall validity of the AES-C was supported by good concurrent validity as indicated by significant correlation between the AES-C and the apathy subscale of the NPI, and by good known-group validity as indicated by the higher AES-C scores reported for those with poor cognitive dementia ratings (CDR = 2). The patient group in the lowest quartile of apathy symptoms (the first quartile of the AES-C distribution) had significantly higher MMSE scores than all other groups.

The AES-C is one of the few instruments measuring apathy in Taiwanese patients with AD that has shown good reliability and validity. The use of the AES-C allows study results to be compared with those from different countries. Using reliable, valid samples, an easily used tool can improve communication between patients and clinicians about apathy symptoms, and thus hopefully contribute to further evaluation of prevalence, determinants, and treatment options for managing apathy in AD.


  1. Top of page
  2. Abstract

This study was supported by grants from Department of Health, Executive Yuan, R.O.C. (Taiwan), DOH95-TD-M-113-062-(2)(2/2).


  1. Top of page
  2. Abstract