Background
The mastery of everyday practical capacities is essential for the elderly to maintain their independence. Lawton and Brody [
1] defined a set of everyday activities for the elderly, so-called activities of daily living (ADLs). They differentiated between basic ADLs (BADLs), which refer to self-maintenance skills such as feeding, dressing, and toileting, and instrumental ADLs (IADLs), which cover more complex behaviors of domestic functioning and enable independent living. IADLs include food/meal preparation, financial administration, housekeeping, laundry, use of the telephone, responsibility for one’s own medication, mode of transportation, and shopping.
With an aging population, the number of people with dementia has dramatically increased in recent decades, and dementia has become a public health challenge [
2]. Alzheimer’s disease is the most common form of dementia and begins years before the onset of clinical symptoms. Its pathology can be described on a continuum that ranges from a preclinical stage (changes in biomarkers) to a prodromal stage with minor cognitive symptoms/mild cognitive impairment, to a symptomatic stage that includes dementia [
3,
4]. At different stages of the disease, different assessments are needed. While a patient is in the preclinical stage, an assessment of biomarkers is most important, whereas functional assessments become more important in the prodromal and symptomatic stages [
5]. IADLs can be used for a functional assessment as early as in the prodromal stage because it has been shown that impairments in IADLs are associated with the diagnosis and development of dementia [
6‐
10] and, more important, deficits in BADLs and IADLs seem to occur at different stages of the dementing process [
10,
11]. Whereas BADLs have been found to be more strongly correlated with motor functioning and coordination [
12] and thus are more likely to remain preserved until the later stages of the disease, IADLs have been found to be more sensitive to the earlier stages of cognitive decline as these activities are more complex and require greater neuropsychological organization [
11]. Even more, IADL impairments have been shown to predict the progression to dementia and can be used to help distinguish between dementia and early forms of cognitive decline, such as mild cognitive impairment (MCI) [
6,
13]. MCI refers to a state that is defined by the presence of the first cognitive impairments that do not yet constitute dementia [
14] but have a high probability of progressing to dementia [
15]. Persons with MCI can experience subtle changes in everyday functional competence [
8]. There is scientific evidence showing that IADLs can be impaired in MCI [
8,
16‐
18]. In addition, in a systematic review, Jekel et al. [
17] reported that patients with MCI and IADL deficits seem to have a higher risk of developing dementia than patients with MCI without IADL deficits, again stressing the importance of IADLs.
Because there is ample evidence that the ability to perform IADLs plays a crucial role in identifying the development of the dementia syndrome, there is a need for assessment tools that have been specifically designed and validated for patients with the first signs of impairments in IADLs (i.e. persons with MCI or mild dementia). As one study showed that several informant-based IADL questionnaires were limited in their quality [
19], it remains important to identify an optimal way to measure IADLs. A promising approach is the use of performance tests as these tests provide standardized and more objective results [
17]. To move in this conceptual direction, the Erlangen Test for Activities of Daily Living (E-ADL) [
20] was developed in 2009 and can be characterized by its excellent economy. In contrast to other performance tests, it requires only about 10 min to be performed and does not require any tasks to be done outside the test room. The E-ADL was designed to assess BADL capabilities and can be used with persons with moderate or severe dementia [
21]. Because it is too easy for persons with less severe dementia, there is a need for a performance test that has been validated for persons with mild dementia or even MCI. For this reason, the Erlangen Test of Activities of Daily Living in Persons with Mild Dementia or Mild Cognitive Impairment (ETAM) was developed as a performance-based tool for the assessment of IADLs [
22]. The ETAM addresses some of the disadvantages of existing performance tests for ADL capabilities as some of these are very time-consuming (from 45 min, Functional Living Skills Assessment [FLSA] [
23], up to 1.5 h, Direct Assessment of Functional Abilities [DAFA] [
24]), cover only a limited range of relevant domains of IADLs, or include culture-specific items (e.g. “calling directory assistance” or “refilling a prescription” in the Revised Direct Assessment of Functional Status [DAFS-R] [
25]). Above all, the ETAM can be used with persons with MCI [
22]. In a first validation study of 107 study participants, including participants with normal cognition, persons with MCI, and persons with mild dementia, the ETAM was shown to be a feasible performance-based assessment tool with good psychometric parameters [
22]. In this first study, the final structure of the ETAM was developed, and the items were reduced from ten to six items on the basis of an exploratory factor analysis and other criteria.
However, because this study was only cross-sectional, there is currently no longitudinal data on the ETAM’s sensitivity to change. This is essential because sensitivity to change or responsiveness is an essential aspect of validity. It provides important information about the ETAM’s ability to measure change over time, and consequently, it determines whether the ETAM can be used in intervention studies. At this time, there are currently no performance tests for assessing IADLs in persons with MCI that can be used in intervention studies. Thus, one aim of the present study was to analyze the ETAM’s sensitivity to change. In addition, we wanted to investigate whether the original target group of persons with MCI or mild dementia could also be extended to include persons with moderate dementia. This would extend the application of the ETAM enormously because dementia is a progressive disease. Other aims of the present study involve other test construction criteria. The exploratory factor analysis in the validation study supported a one-factor structure for the ETAM. In the current study, we conducted a confirmatory factor analysis and investigated whether this structure could be supported. This was important to do in order to determine whether actual data were consistent with the hypothesis that the ETAM consists of a single IADL factor. Other test construction criteria included analyzing discriminant validity with additional instruments and determining criterion-related validity.
Discussion
In this study, we examined the reliability and validity of the ETAM and confirmed that the ETAM can be used not only with people with MCI and mild dementia but also with people with moderate dementia. We showed that ETAM scores differed between the level of cognitive impairment with people with MCI achieving the best results, people with mild dementia second best, and people with moderate dementia the worst. In addition, we confirmed that the ETAM is able to detect change over time. Also, a confirmatory factor analysis supported the postulated single factor structure of IADLs.
The present study supports the application of the ETAM for persons with MCI or mild dementia. In addition, the ETAM can also be recommended for assessing the subgroup of persons with moderate dementia. This is meaningful because functional assessment becomes more important when the degree of cognitive impairment increases [
5]. Our analyses showed that persons with MCI achieved the best results, persons with mild dementia scored on average five points lower, and persons with moderate dementia scored another six points lower. Thus, these results show that as the dementing disease progresses, participants find it increasingly difficult to carry out the IADL-oriented tasks of the ETAM, thus providing support for the ETAM’s reliability and validity.
Further support for the validity of the ETAM was provided by care level, which is primarily related to BADL capacities. We found that participants who had not yet qualified for a care level achieved the most points (i.e. they showed a better performance on the ETAM), and with a higher care level, participants achieved fewer points on the ETAM. Persons with no care level and persons with care level 1 did not show significantly different ETAM scores, which might be due to the different sample sizes that were used or the fact that care level is more strongly related to BADL capacities than to IADL capacities. This finding is especially interesting because care level is an external criterion that was rated by independent testers who were not involved in the study.
We were able to confirm the discriminant validity of the ETAM as predicted in our hypotheses (moderate overall correlation with the MMSE; low correlations with all other tests). Whereas the ETAM scores of people with MCI were barely correlated with the MMSE, the correlation increased when we analyzed the subgroup of persons with mild or moderate dementia. This finding is consistent with Giebel et al.’s [
46] results in suggesting that with the progression of the dementing disease, cognition is increasingly affected, and people have more trouble mastering IADLs. Further support for the association between cognitive levels and functional abilities such as IADLs was found, for example, by Njegovan [
47], who showed that progressive cognitive decline is associated with a specific pattern of loss of functional tasks. All in all, these findings appear to suggest that activities of daily living and cognitive tasks are increasingly associated as cognitive impairment progresses. This means that the relationship between IADL capacities and performance on cognitive tasks increases as cognitive impairment progresses. A similar yet weaker pattern was found for the correlation between the ETAM and the Self-Care item from the EQ-5D, which can be applied to assess BADLs to a certain extent. Again, as the dementing condition progressed, the correlation with the ETAM increased.
In addition, we used the five EQ-5D items to compute correlation coefficients with dimensions such as pain, anxiety, etc. Aside from Self-Care, we found no meaningful correlation or pattern of correlations across the three subgroups of participants with MCI, mild dementia, or moderate dementia, thus providing support for the discriminant validity of the ETAM.
Another important relationship between the ETAM and the MMSE concerns sensitivity to change. For this purpose, we analyzed whether the ETAM was sensitive to other (cognitive) changes over a period of six months. We found that the change in MMSE over a period of six months turned out to be a significant predictor of the ETAM score after six months: When a person’s MMSE score had declined after six months, the person also achieved fewer points on the ETAM after six months. This is an important aspect of validity and it demonstrates that the ETAM is able to measure change over time. Thus, we recommend its use in intervention studies.
Similar to the first validation study, the item “phone call” turned out to be the most difficult item by far. The authors of the first validation study argued that how a person handles the phone is an important and sensitive indicator of incipient dementia processes [
22]. The item “traffic situations” was the second most difficult item. Apart from these findings, there were some differences in the order of items in comparison with the first validation study. This was most likely due to a smaller sample size in the previous study as well as less variation (the difficulties of the remaining four items ranged only from .47 to .67). In the current study, there was a consistent pattern of difficulty indices with one small exception. For the MCI subgroup, “pill organizer” and “alarm clock” were the easiest items (both .83), and “making tea” was the third easiest item (.77). Because it is common practice to arrange the items on a test in order of increasing difficulty, we propose that the order of the ETAM items be rearranged and adjusted to reflect the difficulties found in the current study. Specifically, we suggest the following order when carrying out the ETAM: 1) “making tea,” 2) “alarm clock,” 3) “pill organizer,” 4) “finances,” 5) “traffic situations,” 6) “phone call.” When the items are administered in this order, the participant is encouraged to continue the test, and this will also ensure that weaker candidates will not become discouraged.
Limitations
Some limitations of the current study should be mentioned. Because the lack of high-quality performance-based assessments for measuring IADL capabilities was the reason we developed the ETAM, we cannot provide convergent validity with other instruments that measure IADL capacities. To date, there is no gold standard for measuring IADL capacities especially by means of a performance-based assessment. Existing performance-based assessments are very time-consuming, taking up to 1.5 h [
48] in only very small groups [
24], or they seem to measure cognition rather than IADL functioning [
49] (for an overview, see [
22]). Because the ETAM already showed acceptable convergent validity with the informant-based Bayer Activities of Daily Living Scale [
50] in the first validation study [
22], we decided to focus on discriminant validity and sensitivity to change.
In addition, one should consider that differentiating between MCI, mild dementia, and moderate dementia can be performed only with the mean ETAM scores. This is because there is high between-subject variability on the ETAM in the three levels of cognitive impairment. Thus, ETAM scores should not be used to diagnose MCI, mild dementia, or moderate dementia.
Another limitation of the present study was that we used the NOSGER subscale Social Behavior and the EQ-5D items to analyze discriminant validity. However, a measure of mood would have been desirable because depressive mood is associated with a decline in cognitive abilities.
Future research perspectives
In our study, the categorization of MCI, mild dementia, and moderate dementia was solely based on the cognitive tests of the MMSE and the MoCA, which can be influenced by age and education [
51,
52]. Thereby, we defined cognitive impairment psychometrically and assessed clinical symptoms. In practice, the MMSE is one of the most commonly used screening tools for cognitive impairment [
32], and our analyses also showed that this categorization was successful. For a more accurate categorization for persons with MCI and different stages of dementia, future studies could focus on the use of other instruments besides the MMSE as well (e.g. the Consortium to Establish a Registry for Alzheimer’s Disease [CERAD], neuroimaging, and biomarkers). Especially in the preclinical and prodromal stages (MCI) of Alzheimer’s disease, biomarker assessments are very informative [
5].