Activities of Daily Living Are Associated With Older Adult Cognitive Status

Abstract

We compared the extent to which subjective report of activities of daily living (ADLs) by caregivers and older adults were associated with objective measures of older adults’ cognition. In independent studies (Study 1 N = 238; Study 2 N = 295), bivariate correlations and multiple regression analyses examined the association of caregiver and self-rated reports of older adult basic, instrumental, and total ADLs and older adult cognition. We examined the magnitude of the caregiver/self-report discrepancy and older adult cognition. In both studies, caregiver reports more accurately accounted for older adult cognitive differences. Older adult visuospatial/constructional deficits were uniquely related to caregiver basic ADL reports. Results indicate that caregiver reports of older adult ADLs are more reliable indicators of older adult cognition than self-reports, and this difference grows as older adult cognition decreases. Thus, older adult ADL assessment may be useful in providing information on potential cognitive decline.

Keywords

functional status informant report caregiver care recipient dyads cognitive status

Older adults adapt to declining physical and cognitive capabilities, in part, with informal assistance with daily activities from close relatives. A clear and objective method for quantifying needed assistance is critical for health care providers, particularly for older adults who may evidence some level of cognitive impairment. However, it can be a challenge to integrate objective measures of cognitive function with self- and caregiver reports of daily functioning. Integrating objective and subjective measures may prove particularly daunting for health care providers who are faced with discrepant reports regarding daily functioning, as this depends on the nature and validity of responses on these measures. Although functional issues may arise in the absence of cognitive difficulties (e.g., Bennett et al., 2006), functional impairment is frequently associated with cognitive decrements and is a key criteria for dementia diagnosis (American Psychiatric Association, 2000; Dodge et al., 2005; Freilich & Hyer, 2007). Thus, it is critical to understand the association between reports of daily functioning and cognitive status.

Functional impairment has frequently been assessed through use of self-report or caregiver report of activities of daily living (ADL). ADLs have been classified into two domains: Basic ADLs (BADLs) involve less complex, implicitly learned activities, such as bathing, dressing, and eating (Katz, Ford, Moskowitz, Jackson, & Jaffe, 1963). Instrumental ADLs (IADLs) involve more cognitively demanding tasks, such as managing money or medication (Lawton & Brody, 1979). Previous research has demonstrated a relationship between accuracy of self-reported functional ability and level of cognitive impairment in nondemented community-based older adults (Mitchell & Miller, 2008a) but has not evaluated that relationship directly against informant report relationships and cognition. This emphasizes the need for accurate report of functional status to attain associative validity.

A roadblock to accuracy in reporting involves potential bias within reports of functional status. Although obtaining self-report remains an important factor in identifying functional limitations (Elam et al., 1991), older adults may not self-report such problems to a physician or other health professional. Researchers have found that older adults frequently overestimate their functional abilities. This could be a result of denying difficulties as a coping mechanism, a desire to not appear as a burden to others, or a misperception of current abilities (Rubenstein et al., 1984).

However, caregiver reports of functional status may also have inaccuracies. Cotter, Burgio, Stevens, Roth, and Gitlin (2002) examined the correlation between caregiver report of assistance with ADL tasks and behavioral observation of this assistance within the home. Caregivers were found to be accurate reporters of the nature of assistance when compared with behavioral observation, although they tended to overestimate the amount of time it took to provide assistance to older adults. Other researchers have shown caregivers to report more pervasive exaggerations of older adult functional deficits, perhaps as a result of caregiver burden (Rubenstein et al., 1984). Some researchers have shown no effect of variables such as relationship of caregiver to older adult or amount of caregiver time spent with an older adult as related to the discrepancy between caregiver and self-report (Farias, Mungas, & Jagust, 2005). Others have shown that live-in spouse caregivers are associated with higher ratings of accuracy (Ready, Ott, & Grace, 2004). Such potential biases make the relationship between caregiver reports and objectively measured indices of older adult function more complex.

Accuracy of caregiver report has been related to decrements in older adult functional and cognitive domains. A study contrasting caregiver-reported ADLs and behaviorally observed ADLs in older adults exhibited strong correlations in the domain of motor functioning (walking) and moderately strong correlations in other basic activities (dressing). Correlations were weaker for IADLs, such as managing money (Zanetti, Geroldi, Frisoni, Bianchetti, & Trabucchi, 1999). Studies focusing on IADLs have similarly shown weak accuracy of self- and informant reports of financial abilities, consistent older adult overestimation of skills, and low stability of both types of reports (Wadley, Harrell, & Marson, 2003). Research reporting changes in cognitive status suggest that older adult decrements in memory predict a self-report–caregiver report discrepancy (Farias et al., 2005).

A growing literature has investigated the association between cognitive deficit and functional status. Dodge et al. (2005) estimated that cognition accounted for 18% to 36% loss in BADL abilities and 11% to 29% of loss in IADL abilities in a community-based sample of older adults. Longitudinally, cognitive impairment at baseline in participants experiencing no functional decrement was a risk factor for later decline (1 year post baseline) in BADLs and IADLs (Dodge et al., 2005). Bennett et al. (2006) suggested global cognitive functioning accounted for up to 22% of the variance in functional status, particularly within the domain of housework but found no specific cognitive domain that was uniquely related to functional status. However, several studies suggest that executive functioning plays a significant role in facilitating activities of daily living (Grigsby, Kaye, Baxter, Shetterly, & Hamman, 1998; Lewis & Miller, 2007; Mitchell & Miller, 2008a, 2008b). After controlling for demographic factors, Johnson and colleagues found that decrements in executive functioning, as measured by the Trails B test, were predictive of later decline in ADLs in female older adults (Johnson, Lui, & Yaffe, 2007). Others suggest that cognitive and physical processes independently contribute to an individual’s ability to complete ADLs (Gill, Williams, Richardson, & Tinetti, 1996; Petrella, Cress, & Miller, 2004).

The current investigation sought to examine caregiver versus self-report of functional status and their relationship to objective measures of cognitive function. In the current investigation, we compared caregiver and self reports of BADLs and IADLs as they relate to measures of cognition in older adults. Discrepancies within the current literature associating the role of caregiver and self-report with older adult cognition are difficult to reconcile due to variability in methodology and sample size across studies. To account for such issues, we used two samples of community-based dyads, with both older adult and caregiver in each sample completing nearly identical functional report measures.

Another aim of the current study was to use a large sample size of caregiver–older adult dyads and specifically employ common clinical measures of cognition to evaluate the relationship between reported functional status and cognition. We used the Cognistat: Neurobehavioral Cognitive Status Examination (Cognistat; Northern California Neurobehavioral Group, Inc., 1995; Kiernan, Mueller, Langston, & Van Dyke, 1987) and the Repeatable Battery for the Assessment of Neuropsychological Status (RBANS; Randolph, 1998) to compare the correlation of functional measures with neuropsychological measures commonly used in geriatric populations. We sought to identify the relationship between older adult cognitive status and reported functional status and also to examine what domains of cognition may hold the most associative power in caregiver report of older adult functional status. Domains of cognitive function associated with caregiver reports on older adult functional impairment would be helpful for clinicians to clarify what aspects of cognitive decline may be most readily observed when assessing older adult functional status by self- and collateral report.

Based on the extant literature, we hypothesized that (a) caregiver reports of older adult total ADLs would be more highly associated with older adult cognition than self-reports of total ADLs; (b) caregiver reports of older adult BADLs and IADLs would each be more strongly associated with cognition than self-report; and (c) IADLs would be more highly associated with cognition than BADLs, regardless of report method.

In testing the above hypotheses, we separately analyzed data from two different studies. In Study 1, participants were drawn from the Family Relationships in Late Life (FRILL) study (described below), in which the Cognistat was used as a measure of older adult cognition. In Study 2, participants included caregiver–older adult dyads from the second FRILL study (FRILL2), including an oversampled set of African American dyads. In Study 2, the RBANS was used as a measure of older adult cognition.

Furthermore, in identifying an associative relationship between caregiver reports of older adult total ADL and older adult RBANS scores, we sought to isolate specific domains of cognitive function that could drive this relationship. In Study 2, we expanded our comparison of BADLs and IADLs tested in Study 1 by also examining specific domains of cognitive functioning as they relate to the maintenance of daily functioning. We were interested in whether the relationship between cognitive functioning and both BADLs and IADLs tested within Study 1 were being driven by particular cognitive processes. Specifically, we were interested in determining which components of cognition were essential for the maintenance of simple versus complex activities of daily living (i.e., BADLs vs. IADLs). To accomplish this goal, we reversed associated variables to directly compare the variance in functional status (caregiver reported older adult BADL and IADL) accounted for by five RBANS index scores. We hypothesized that (a) the older adult performance on RBANS indices of visuospatial and attentional tasks would be most highly associated with BADL decrements because these fundamental activities (e.g., dressing, grooming, bathing) require visuomotor coordination, speed, and spatial navigation. In contrast, we hypothesized that (b) older adult performance on RBANS indices of language and delayed memory would be the most highly associated with IADL decrements because these more complex activities (e.g., managing finances, managing medications, grocery shopping) require higher order cognitive abilities such as naming ability, verbal fluency, and memory functioning. In addition, language and delayed memory functioning are most likely to decline in preclinical Alzheimer’s disease (Randolph, 1998) and decrements in these domains may be more likely in individuals experiencing functional decline in IADLs. We chose to use caregiver report of older adult ADL functioning because Study 1 findings suggested that caregiver ratings of older adult functional status were more sensitive (i.e., larger range in scores and higher mean rating of functional decrements).

Study 1

Method

Sample and Procedures

In Study 1, participants were drawn from the FRILL study of informal caregiver–older adult dyads, recruited from three recruitment sites: Athens, Georgia; Pittsburgh, Pennsylvania; and Dallas, Texas. Caregivers were defined as those who provided unpaid assistance with at least two IADLs (e.g., managing money) or one BADL (e.g., bathing) to community-dwelling adults aged 60 years or older. The source of older adult impairment was not a factor in participant selection. Details of the FRILL protocol have been reported elsewhere (Beach et al., 2005; Miller et al., 2006).

The FRILL project was approved by all necessary institutional review boards. Participants were screened by telephone for eligibility and then interviewed in person within their homes. Caregivers and care recipients were interviewed concordantly and separately to prevent data contamination. Research staff were properly trained on interview procedures and assessment administration. Each individual received US$20 for their participation. Testing lasted 1.5 to 2 hr per participant.

Recruitment efforts yielded a total of 283 dyads. Because of the extremely small number of dyads being neither White nor African American (4 dyads in “Other” category), we excluded those dyads, creating a dichotomous category for older adult race (n = 279, coded as 1 = White, 2 = African American). Of the 279 remaining dyads, 41 were excluded from analyses due to missing data on the primary measures of interest (older adult demographics; caregiver and self reports of BADLs and IADLs; older adult composite Cognistat score) leaving 238 dyads.

Measures

Older adult cognitive status

Older adults completed the Cognistat (Northern California Neurobehavioral Group, Inc., 1995), which examines general cognitive functioning in 10 domains (orientation, attention, language comprehension, memory, design construction, language repetition, naming, calculations, reasoning, and judgment) and has been shown to have good sensitivity in distinguishing cognitively impaired adults in older adults (Drane & Osato, 1997). Due to limited sample sizes in published normative data for the Cognistat, we followed Miller and colleagues’ (2006) impairment rating procedure to create a composite severity score for each participant. A rating scale of 1 to 4 was created for each domain, with 1 representing lack of impairment, and 4 representing severe impairment in a given area. The summed composite rating for each individual domain was used as an estimate of older adult cognitive functioning (range = 10-40).

Older adult and caregiver reported amount of help provided

Both the caregiver and older adult completed an 18-item assessment, adapted from the ADL instrument (Older American Resources and Services, Duke University, 1978) to assess older adult level of functional status. The older adult’s performance of BADLs and IADLs were based on a metric of 0 to 5. Activities that older adults had never participated in were rated 0 (e.g., if an older adult had not participated in managing couple’s money prior to impairment), to differentiate preimpairment delegated activities from level of postimpairment assistance. Activities that older adults had engaged in preimpairment received a rating between 1 and 5: A score of 1 on the scale represented activities which the older adult completed with no help, whereas a 5 represented an activity requiring a great deal of help by the caregiver. Total ADL scores have a theoretical range of 0 to 90, BADLs range from 0 to 50, and IADL range from 0 to 40. In addition, discrepancy scores were calculated by subtracting self-rated ADLs from caregiver reported older adult functional status to have a single measure of the relative size differences between caregiver and self-reports of functional status.

Results

Sample Description

Mean caregiver age was 62 years (SD = 15, range = 20-88 years), and mean older adult age was 77 years (SD = 8.4, range = 60-98 years). Similar to national estimates (e.g., National Alliance for Caregiving & AARP, 2004), the majority of participants in the caregiver role were female (78%) and either a spouse (48%) or adult child (37%). Education and race for both studies are reported in Table 1.

Table 1.

Study 1 and 2 Demographic Information

	Study 1		Study 2
	n	% of total sample	n	% of total sample
Caregiver race
African American	39	16.4	113	40
White	194	81.5	164	58
Other racial category	5	2.1	5	2
Caregiver education
<7 years	1	0.4	5	1.8
Junior high	9	3.8	16	5.7
Some high school	19	8.0	36	12.8
High school graduate	52	21.8	100	35.5
Trade school	14	5.9	29	10.3
Some college	40	16.8	43	15.2
College graduate	55	23.1	27	9.6
Graduate degree	48	20.2	26	9.2
Older adult race
African American	41	17.2	111	39.4
White	197	82.8	171	60.6
Older adult education
<7 years	13	5.5	21	7.4
Junior high	28	11.8	37	13.1
Some high school	33	13.9	56	18.8
High school graduate	55	23.1	80	28.4
Trade school	14	5.9	20	7.1
Some college	41	17.2	32	11.3
College graduate	35	14.7	25	8.9
Graduate degree	19	8.0	14	5.0
	M	SD	M	SD
Older adult cognition^a	18.51	6.49	75.79	18.53
Older adult functional status
Self-reported BADL	18.87	9.58	2.31	2.08
Self-reported IADL	21.00	9.73	3.01	2.40
Caregiver-reported BADL	20.76	10.10	2.89	2.38
Caregiver-reported IADL	25.73	9.08	3.58	2.44

Note. BADL = basic activities of daily living; IADL = instrumental activities of daily living.

Study 1 = Cognistat, Study 2 = Repeatable Battery for the Assessment of Neuropsychological Status.

Cognitive and Functional Status

Older adults demonstrated a wide range of abilities in both cognitive domains, based on Cognistat summed composite rating, as well as within functional domains, based on self- and collateral reports of BADLs and IADLs (see Table 1).

Bivariate Analysis

Older adult education and race correlated with their Cognistat composite score (r = −.334, p < .001; r = .244, p < .001, respectively), with lower Cognistat scores (higher functioning) seen more frequently in higher educated individuals and in White samples. Caregiver education and ethnic background similarly correlated with older adult Cognistat score (r = −.149, p < .05, and r = .263, p < .001, respectively). Neither older adult nor caregiver gender showed significant correlations with measures of functional status and so were not included in further analyses.

Caregiver-reported older adult total ADL correlated positively with older adult Cognistat composite score (r = .296, p <.01), whereas self-reported total ADL showed no significant correlation (r = .093, p > .05). Older adult composite Cognistat score was significantly correlated with both self-reported and caregiver-reported IADLs (caregiver: r = .404, p < .001; older adult: r = .182, p <.01) as well as caregiver reported BADLs (r = .132, p = .04). The older adult Cognistat composite score showed a significantly higher correlation with caregiver-reported IADLs than with older adult IADL (dependent correlation comparison: t = 3.14, p < .01).

Caregiver and Self-Reports of Total ADLs

To test our hypothesis that caregiver reports of older adult total ADLs would be more closely associated with older adult Cognistat composite score than self-reports of total ADLs, we conducted a hierarchical regression analysis. Demographic variables of caregiver race and older adult education were identified as significant through a single simultaneous regression analyses of all demographic variables. These were entered into the first step of the ensuing models. As expected, they were found to be significantly associated with older adult Cognistat composite score, ΔR² = .163, F(2, 235) = 22.824, p < .001. In the second step of the model, self-reports of total ADLs were not incrementally associated with older adult cognition when compared to Step 1, ΔR² = .003, ΔF(1, 234) = .866, p > .05. In the third step of the model, caregiver reports of older adult total ADL scores were incrementally associated with older adult cognition when compared to Step 2, ΔR² = .068, ΔF(1, 233) = 20.747, p < .001. Regression details are presented in Table 2.

Table 2.

Study 1 Regression Analysis for Prediction of Cognistat Composite Score by Older Adult and Caregiver Demographic Variables and Total ADL Measures and Older Adult–Caregiver Total ADL Discrepancy Score

Step	Variables entered			β	t	p
Total ADL measures
1A	Caregiver race			.228	3.787	<.001
	Older adult education			−.308	−5.124	<.001
1B	Caregiver race			.231	3.832	<.001
	Older adult education			−.299	−4.914	<.001
	Self-reported total ADL			.056	0.930	.353
1C	Caregiver race			.200	3.428	.001
	Older adult education			−.295	−5.044	<.001
	Self-reported total ADL			−.127	−1.794	.074
	Caregiver-reported total ADL			.320	4.555	<.001
Caregiver–older adult combined BADL/IADL discrepancy score
2A	Caregiver race			.228	3.787	<.001
	Older adult education			−.308	−5.124	<.001
2B	Caregiver race			.201	3.397	<.001
	Older adult education			−.321	−5.447	.001
	Caregiver–older adult total ADL discrepancy score			.201	3.407	<.001
Step	R	SE	R² change	F change	df	p
1A/2A	.403	5.96	.163	22.824	2, 235	<.001
1B	.407	5.96	.003	0.866	1, 234	.353
2B	.450	5.83	.040	11.605	1, 234	.001
1C	.484	5.73	.068	20.747	1, 233	<.001

Note. ADL = activities of daily living; BADL = basic activities of daily living; IADL = instrumental activities of daily living. N = 238.

To compare the impact of the overall magnitude of differences between caregiver and self-reports of total ADLs, we ran a parallel regression analysis but replacing Steps 2 and 3 with our discrepancy score variable. As in the individual analyses, the size of the differences between caregiver and self-rated total ADL scores were significantly associated with poorer older adult cognition, ΔR² = .04, ΔF(1, 234) = 11.605, p < .01.

Caregiver and Self-Reports of BADLs and IADLs

To test our hypotheses that caregiver reports of both BADLs and IADLs would be more highly associated with older adult Cognistat composite score than self-report, we conducted separate regression analyses for BADLs and IADLs. Regression details are presented in Table 3. In the BADLs regression, the first step was identical to that done previously, controlling for older adult race and caregiver education. The second step added self-reports of BADLs and the third step added caregiver reported BADLs. Self-reports of BADLs alone were not associated with older adult Cognistat composite score (older adult BADL: β = −.044, t = −.728, p >.05) but were significant when caregiver reported BADL were added into the model (older adult BADL: β = −.195, t = −2.452, p < .05) in the third step. Caregiver reports of basic functional status did reach significance in the third step of the model (caregiver BADL: β = .226, t = 2.876, p < .01). In the IADLs contrast, only caregiver reports of older adult IADLs was positively associated with older adult Cognistat composite score in the final step of the model (older adult IADL: β = .013, t = .219, p >.05; caregiver IADL: β = .349, t = 5.766, p < .001).

Table 3.

Study 1 Regression Analysis for Prediction of Older Adult Cognistat Score by Older Adult and Caregiver Basic ADL/IADL Measures and Caregiver–Older Adult BADL and IADL Discrepancy Score

Step	Variables entered			B	t	p
Older adult and caregiver BADL
1A	Caregiver race			.228	3.787	<.001
	Older adult education			−.308	−5.124	<.001
1B	Caregiver race			.224	3.702	<.001
	Older adult education			−.315	−5.169	<.001
	Self-reported ADL			−.044	−0.728	.467
1C	Caregiver race			.209	3.504	<.01
	Older adult education			−.316	−5.271	<.001
	Self-reported basic ADL			−.195	−2.452	<.05
	Caregiver-reported basic ADL			.226	2.876	<.01
Older adult and caregiver IADL
2A	Caregiver race			.228	3.787	<.001
	Older adult education			−.308	−5.124	<.001
2B	Caregiver race			.229	3.858	<.001
	Older adult education			−.289	−4.817	<.001
	Self-reported IADL			.144	2.421	<.05
2C	Caregiver race			.187	3.326	<.01
	Older adult education			−.278	−4.931	<.001
	Self-reported IADL			.013	0.219	.827
	Caregiver-reported IADL			.349	5.766	<.001
Caregiver–older adult BADL/IADL discrepancy score
3A	Caregiver race			.228	3.787	<.001
	Older adult education			−.308	−5.124	<.001
3B	Caregiver race			.208	3.499	<.001
	Older adult education			−.319	−5.381	<.001
	Caregiver–older adult discrepancy: BADL			.175	2.959	.003
3C	Caregiver race			.201	3.388	.001
	Older adult education			−.321	−5.442	<.001
	Caregiver–older adult discrepancy: BADL			.122	1.825	.069
	Caregiver–older adult discrepancy score: IADL			.113	1.689	.093
Step	R	SE	R² change	F change	df	p
1A/2A/3A	.403	5.963	.163	22.824	2, 235	<.001
1B	.406	5.968	.002	0.531	1, 234	.467
2B	.428	5.902	.020	5.859	1, 234	<.05
3B	.439	5.87	.030	8.756	1, 234	<.01
1C	.440	5.878	.029	8.273	1, 233	<.01
2C	.534	5.533	.102	33.244	1, 233	<.001
3C	.450	5.84	.010	2.851	1, 233	.093

Note. ADL = activities of daily living; BADL = basic activities of daily living; IADL = instrumental activities of daily living. N = 238.

To compare the impact of the overall magnitude of differences between caregiver and self-reports of both BADLs and IADLs, we ran a regression analysis using our discrepancy score variable. The discrepancy between caregiver and self-rated BADL scores were significantly associated with older adult cognition, ΔR² = .03, ΔF(1, 234) = 8.756, p < .01. Interestingly, when the discrepancy between caregiver and self-rated IADL scores were added to the model to identify unique variance accounted for by IADLs above and beyond that accounted for by BADLs, the discrepancy scores of BADL and IADL functioning appeared to share variance. This step did not display significant additional variance accounted for in older adult cognition ΔR² = .010, ΔF(1, 233) = 2.851, p > .05. This is further supported by results from a separate regression analysis of caregiver minus self-rated IADLs significantly accounting for variance in older adult cognition, ΔR² = .029, ΔF(1, 234) = 8.263, p < .01, when not including BADLs.

Tolerance

Tolerance was calculated to examine collinearity of all regression models. A cutoff of a tolerance of .1 or greater was used to rule out collinearity (Lin, 2008). All tolerance levels were well within the specified range.

Study 2

Method

Sample and Procedures

In Study 2, a total of 450 caregiver–older adult dyads were recruited for the second FRILL (FRILL2) study at three recruitment sites: Athens, Georgia; Pittsburgh, Pennsylvania; and Tuscaloosa, Alabama. Recruitment was similar to FRILL with some notable exceptions. To be eligible for the study, older adults and their caregivers had to be coresiding community-dwelling individuals. Inclusion criteria for caregiver–older adult dyads required that the caregiver was primarily responsible for the care of a cognitively or physically impaired older adult above the age of 60 and had to provide unpaid help for at least two IADLs or one BADL.

A primary goal of FRILL2 was to oversample African American caregiving dyads to obtain data sufficient to address research questions missing from previous research on the quality of informal older adult care (e.g., longitudinal comparisons between White and African American caregivers). Within these constraints, we attempted to obtain as representative a sample as possible, employing the services of the Survey Research Center at the University of Georgia. These efforts began with random digit dialing (RDD) in the areas including and surrounding the data-collection sites. We then narrowed our search to age-targeted RDD (e.g., individuals 60 years of age and older, according to U.S. Census data). These methods produced more eligible White than African American dyads. To increase the number of African American participants, we used community-based snowball referral methods at the Georgia site in which completed African American dyads were recontacted and asked to provide the names and telephone numbers of other potentially eligible dyads. Project staff then contacted these individuals. In initial screening, RDD methods identified 877 potential dyads. Of these, 35% refused to be interviewed, 5.6% could not be reached due to technical phone problems, and 18% were subsequently determined to be ineligible based on study criteria. Snowballing methods produced 95 potential dyads, of which 14.7% refused participation. Together, these methods resulted in a sample that was 57% White, 42% African American, and 1% in another racial category. Overall, recruitment efforts resulted in 765 eligible dyads, 321 (42%) of which declined participation, resulting in a sample of 450 caregiver–older adult dyads (58% participation rate). For the purposes of simplicity in interpreting the influence of race on our observed relationships between older adult cognitive status and reported functional ability, we again excluded dyads in which older adult race was reported as “other” for the analyses described below. Of these 450 dyads of participants, 7 dyads were excluded because the older adult was not African American or White, 20 were excluded because the older adult was younger than 60 years of age, and 128 were excluded due to missing data on the main measures used in the analyses below (i.e., demographic data, older adult RBANS scores). Thus, the dyads represented in the results below included older adults 60 years of age or older identifying as African American or White who provided sufficient data on the main measures used in our analyses (n = 295). Some participants met more than one of the above-described exclusionary criteria (e.g., missing data and the older adult was in the “Other” racial category). Furthermore, some participants retained for the analyses below had partially missing functional data and were included in analyses if data were sufficient.

Similar to Study 1, face-to-face structured interviews lasting between 1.5 and 2 hr, for which participants were paid US$25, were conducted in respondents’ homes by pairs of carefully trained interviewers, and caregivers and older adults were interviewed separately and simultaneously. The study was approved by the institutional review boards of the Universities of Alabama, Georgia, and Pittsburgh.

Measures

Older adult cognitive status

Cognitive functioning was measured with the RBANS (Randolph, 1998). The RBANS is a 30-min neuropsychological test designed to assess cognitive decline in older adults and serve as a screening tool for cognitive functioning in younger adults (Randolph, Tierney, Mohr, & Chase, 1998). The RBANS consists of 12 subtests that generate 5 index scores: Visuospatial/Constructional, Attention, Language, Immediate Memory, and Delayed Memory. The RBANS additionally generates a global score (the Total Scale score) derived from the raw scores on all 12 subtests. The 12 subtests of the RBANS are Figure Copy, Line Orientation, Digit Span, Coding, Picture Naming, Semantic Fluency, List Learning, Story Memory, List Recall, List Recognition, Story Recall, and Figure Recall.

Older adult and caregiver reported amount of help provided

The same 18-item ADL assessment from Study 1 was used in Study 2 (ADL; Older American Resources and Services, Duke University, 1978). However, in this study, each item on the functional measure was asked and coded on a dichotomous scale (i.e., 0 = no help needed, 1 = yes, help needed). Thus, total score range for ADL was from 0 to 18. As in Study 1, discrepancy scores were calculated by subtracting self-rated ADLs from caregiver-reported functional status to have a single measure of the relative size differences between caregiver and self-reports of functional status.

Results

Sample Description

As in Study 1, the majority of caregivers were women (69%). Mean caregiver age was 61 years (SD = 14.6, range = 18-91). Within Study 2, 54% of caregivers were a older adult’s spouse. Older adults (55% female) were, on average, 76 years of age (SD = 9.2). Education and race for both studies are reported in Table 1.

Cognitive and Functional Status

Please refer to Table 1 for further information regarding cognitive and functional status scores.

Bivariate Analysis

Older adult education correlated with older adult RBANS total score (r = .467, p < .001), as did race (r = .353, p < .001), indicating that at the group level, higher RBANS scores correlated with higher education and White racial status. Caregiver reported older adult total ADL correlated with older adult RBANS total score (r = −.330, p <.001), as did self-reported total ADL (r = −.185, p < .001). When examining the breakdown of the total ADL score in a BADL/IADL analysis, older adult RBANS total score was significantly correlated with both caregiver and self-reports of older adult BADLs (caregiver: r = −.265, p < .001; older adult: r = −.116, p <.05) as well as with caregiver and self-reported IADLs (caregiver: r = −.347, p < .001; older adult: r = −.221, p < .001). These relationships suggested that lower cognitive impairments were associated with lower functional difficulties reported by both caregiver and older adult.

Caregiver and older adult education, older adult age, and older adult gender accounted for significant variance among all demographic variables when entered into a single simultaneous regression of demographic variables. These were entered into the first step of the ensuing models.

Caregiver and Self-Reports of Total ADLs

Regression details are presented in Table 4. To again test our hypothesis that caregiver reports of older adult total ADLs would be more highly associated with older adult RBANS total score than self-reports, we conducted a hierarchical regression analysis entering caregiver education, older adult education, older adult age, and older adult gender in the first step of the model, which was found to be significantly associated with older adult RBANS performance, ΔR² = .253, F(4, 273) = 23.175, p < .001. In the second step, we added self-reported total ADLs and found that, contrary to results found in Study 1, older reports of total ADLs were modestly associated with older adult RBANS total score, ΔR² = .017, F(1, 272) = 6.385, p < .05, when caregiver reports were not included. In the final step, we added caregiver reports of total ADLs. As in Study 1, the overall model accounted for a significant amount of the variance in older adult RBANS total score, ΔR² = .040, F(1, 271) = 15.655, p < .001. Caregiver report was significantly associated with RBANS (β = −.265, t = −3.957, p < .001) as hypothesized, whereas self-report was not (p > .05), once caregiver reports were included.

Table 4.

Study 2 Regression Analysis for Prediction of Older Adult RBANS Total Scale Score by Older Adult and Caregiver Total ADL Measures and Caregiver–Older Adult Combined BADL/IADL Discrepancy Score

Step	Variables entered			β	t	p
Total ADL measures
1A	Caregiver education			.101	1.652	.100
	Older adult education			.410	6.829	<.001
	Older adult age			−.183	−3.198	.002
	Older adult gender			−.065	−1.230	.220
1B	Caregiver education			.093	1.532	.127
	Older adult education			.418	7.018	<.001
	Older adult age			−.152	−2.620	.009
	Older adult gender			−.075	−1.433	.153
	Self-reported total ADL			−.134	−2.527	.012
1C	Caregiver education			.137	2.285	.023
	Older adult education			.392	6.713	<.001
	Older adult age			−.132	−2.324	.021
	Older adult gender			−.044	−0.858	.391
	Self-reported total ADL			.026	0.401	.689
	Caregiver-reported total ADL			−.265	−3.957	<.001
Caregiver–older adult total ADL discrepancy score
2A	Caregiver education			.101	1.652	.100
	Older adult education			.410	6.829	<.001
	Older adult age			−.183	−3.198	.002
	Older adult gender			−.065	−1.230	.220
2B	Caregiver education			.132	2.132	.034
	Older adult education			.391	6.528	<.001
	Older adult age			−.183	−3.230	.001
	Older adult gender			−.042	−0.793	.429
	Caregiver–older adult total ADL discrepancy			−.137	−2.567	.011
Step	R	SE	R² change	F change	df	p
1A/2A	.503	15.559	.253	23.175	4, 273	<.001
1B	.520	15.408	.017	6.385	1, 272	<.05
2B	.521	15.402	.018	6.589	1, 272	<.05
1C	.557	15.009	.040	15.655	1, 271	<.001

Note. RBANS = Repeatable Battery for the Assessment of Neuropsychological Status; ADL = activities of daily living; BADL = basic activities of daily living; IADL = instrumental activities of daily living. N = 295.

To compare the impact of the overall magnitude of differences between caregiver and self-report ADLs on cognition, significant caregiver and older adult demographics (caregiver and older adult education, older adult age, older adult gender) were entered into the first step of a separate regression, followed by the discrepancy score for total ADL. The discrepancy of caregiver–self-rated total ADL scores were significantly associated with cognitive performance, ΔR² = .018, ΔF(1, 272) = 6.589, p < .05.

Caregiver and Self-Reports of BADLs and IADLs

To replicate findings from Study 1 and test our hypotheses that caregiver reports of both BADLs and IADLs would be more highly associated with cognitive performance than self-report, we conducted separate regression analyses for BADLs and IADLs. Regression details are presented in Table 5. As per earlier regressions, appropriate demographics (caregiver education, older adult education, older adult age, and older adult gender) were entered into the first step of the model, followed by self-reported BADL in the second step and caregiver reported BADL in the third step. Self-reported BADL did not account for a significant amount of the variance in the second step (B = −.075, t = −1.407, p > .05) and did not attain significance within the third step of the model (p > .05). Conversely, caregiver reports accounted for a significant amount of the variance, above self-reports, ΔR² = .037, F(2, 275) = 14.483, p < .001. In an identically run analyses of IADL relationships, self-report IADLs accounted for a significant amount of variance following demographics (B = −.182, t = −3.471, p > .001). However, this was not maintained into the third step (p > .05) whereas caregiver reports of IADLs attained significance (β = −.256, t = −4.163, p < .001).

Table 5.

Study 2 Regression Analysis for Prediction of Older Adult Cognistat Score by Older Adult and Caregiver Basic ADL/IADL Measures and Caregiver–Older Adult BADL and IADL Discrepancy Score

Step	Variables entered			B	t	p
Older adult and caregiver BADL
1A	Caregiver education			.100	1.629	.104
	Older adult education			.409	6.834	<.001
	Older adult age			−.176	−3.077	.002
	Older adult gender			−.066	−1.261	.208
1B	Caregiver education			.091	1.478	.141
	Older adult education			.415	6.933	<.001
	Older adult age			−.163	−2.817	.005
	Older adult gender			−.075	−1.412	.159
	Older adult–reported BADL			−.075	−1.407	.159
1C	Caregiver education			.124	2.039	.042
	Older adult education			.399	6.800	<.001
	Older adult age			−.143	−2.520	.012
	Older adult gender			−.057	−1.107	.269
	Older adult–reported BADL			.088	1.312	.190
	Caregiver-reported BADL			−.256	−3.806	<.001
Older adult and caregiver IADL
2A	Caregiver education			.102	1.707	.089
	Older adult education			.409	6.991	<.001
	Older adult age			−.190	−3.354	.001
	Older adult gender			−.063	−1.210	.227
2B	Caregiver education			.103	1.754	.081
	Older adult education			.412	7.171	<.001
	Older adult age			−.145	−2.542	.012
	Older adult gender			−.068	−1.326	.186
	Older adult–reported IADL			−.182	−3.471	.001
2C	Caregiver education			.151	2.607	.010
	Older adult education			.386	6.883	<.001
	Older adult age			−.121	−2.169	.031
	Older adult gender			−.033	−0.653	.514
	Older adult–reported IADL			−.053	−0.893	.373
	Caregiver–reported IADL			−.256	−4.163	<.001
Caregiver–older adult BADL/IADL discrepancy score
3A	Caregiver education			.101	1.652	.100
	Older adult education			.410	6.829	<.001
	Older adult age			−.183	−3.198	.002
	Older adult gender			−.065	−1.230	.220
3B	Caregiver education			.129	2.102	.036
	Older adult education			.391	6.548	<.001
	Older adult age			−.178	−3.142	.002
	Older adult gender			−.045	−0.855	.393
	Caregiver–older adult discrepancy: BADL			−.148	−2.796	.006
3C	Caregiver education			.132	2.135	.034
	Older adult education			.389	6.506	<.001
	Older adult age			−.179	−3.156	.002
	Older adult gender			−.043	−0.804	.422
	Caregiver–older adult discrepancy: BADL			−.135	−2.211	.028
	Caregiver–older adult discrepancy: IADL			−.028	−0.452	.651
Step	R	SE	R² change	F change	df	p
1A	.498	15.682	.248	22.854	4, 277	<.001
2A	.505	15.607	.255	23.664	4, 277	<.001
3A	.503	15.559	.253	23.175	4, 273	<.001
1B	.503	15.654	.005	1.980	1, 276	.161
2B	.535	15.304	.031	12.051	1, 276	<.01
3B	.524	15.368	.021	7.817	1, 272	<.01
1C	.539	15.285	.037	14.483	1, 275	<.001
2C	.573	14.871	.042	17.331	1, 275	<.001
3C	.524	15.391	.001	0.205	1, 271	.651

Note. ADL = activities of daily living; BADL = basic activities of daily living; IADL = instrumental activities of daily living. N = 295.

To compare the impact of the overall magnitude of differences between caregiver and self-reports of both BADLs and IADLs in our second sample, we again conducted a regression analysis using our discrepancy score variable. Previously identified demographics (caregiver and older adult education, older adult age, older adult gender) were entered into the first step of a regression, followed by the discrepancy score for the BADL in the second step, and by the IADL discrepancy score in the third step. The discrepancy of caregiver–self-rated BADL scores were significantly associated with older adult cognition, ΔR² = .021, ΔF(1, 272) = 7.817, p < .01. However, the discrepancy of caregiver–self-rated combined IADL scores were not significantly associated with older adult cognition (p > .05). When entered in reverse order, the caregiver–older adult discrepancy scores of BADL maintained significance within the third step of the model, ΔR² = .013, ΔF(1, 271) = 4.890, p < .05, and IADL functioning did not display a significant association with cognition in either the second or third step (p > .05).

Cognitive Domains and Caregiver Reports of Older Adult ADLs

To explore domains of cognitive function that could help explain the relationship between measures of functional status and overall cognitive status, we examined whether specific RBANS cognitive indices were associated with reported difficulties with ADLs. Given our findings above, we restricted analyses only to caregiver reports of BADLs and IADLs in relation to older adult RBANS.

To test our hypothesis that older adult performance on RBANS indices of visuospatial and attentional tasks would be associated with caregiver reports of older adult ADL performance, we conducted a hierarchical regression analysis entering caregiver education, older adult education, older adult age, and older adult gender in the first step of the model (Table 6). This step accounted for a significant amount of BADL score variance, adjusted R² = .049, F(4, 279) = 3.581, p = .007. In the second step, we simultaneously added the five RBANS indices and found that the model accounted for a significant additional amount of the variance in caregiver reports of older adult BADL status, ΔR² = .070, F(5, 274) = 4.322, p = .001. RBANS performance on the visuospatial/constructional index (β = −.185, t = −2.553, p = .011) was the only index accounting for significant, unique variance, with higher visuospatial/constructional scores relating to lower functional difficulties. All other RBANS indices were nonsignificant (p > .05).

Table 6.

Study 2 Regression Analyses for Prediction of Caregiver Report of Older Adult Basic ADL Functioning by Older Adult RBANS Performance and Caregiver–Older Adult BADL Discrepancy Score

Step	Variables entered			β	t	p
Caregiver BADL report
1A	Caregiver education			.044	0.645	.519
	Older adult education			−.007	−0.106	.915
	Older adult age			.203	3.181	.002
	Older adult gender			−.009	−0.157	.875
1B	Caregiver education			.076	1.128	.260
	Older adult education			.110	1.540	.125
	Older adult age			.170	2.617	.009
	Older adult gender			−.009	−0.154	.878
	RBANS immediate memory			−.056	−0.609	.543
	RBANS visuospatial constructional			−.185	−2.553	.011
	RBANS language			.026	0.329	.742
	RBANS attention			−.119	−1.494	.136
	RBANS delayed memory			−.017	−0.174	.862
Caregiver–older adult BADL discrepancy score
2A	Caregiver education			.194	2.809	.005
	Older adult education			−.105	−1.560	.120
	Older adult age			.044	0.692	.490
	Older adult gender			.116	1.960	.051
2B	Caregiver education			.240	3.492	.001
	Older adult education			−.052	−0.716	.474
	Older adult age			.001	0.009	.993
	Older adult gender			.137	2.278	.024
	RBANS immediate memory			−.074	−0.798	.426
	RBANS visuospatial constructional			−.176	−2.418	.016
	RBANS language			−.192	−2.416	.016
	RBANS attention			.128	1.607	.109
	RBANS delayed memory			.045	0.469	.640
Step	R	SE	R² change	F change	df	p
1A	.221	2.314	.049	3.581	4, 279	.007
2A	.222	1.864	.049	3.590	4, 277	.007
1B	.344	2.248	.070	4.322	5, 274	.001
2B	.333	1.182	.062	3.786	5, 272	.002

Note. RBANS = Repeatable Battery for the Assessment of Neuropsychological Status; ADL = activities of daily living; BADL = basic activities of daily living. N = 295.

Discrepancy score analyses yielded conflicting results. Significant caregiver and older adult demographics (caregiver and older adult education, older adult age, older adult gender) were entered into the first step of a regression, In the second step, we simultaneously added the five RBANS indices and found that the model accounted for a significant additional amount of the variance in caregiver–older adult BADL discrepancy status, ΔR² = .062, F(5, 272) = 3.786, p < .01. In this analysis, language (β = −.192, t = −2.416, p < .05) and visuospatial/constructional performance (β = −.176, t = −2.418, p < .05) were identified as significantly associated with BADL discrepancy scores, above and beyond the aforementioned demographic variables.

To test our hypothesis that older adult performance on RBANS indices of language and delayed memory tasks would be associated with caregiver reports of older adult IADL performance, we conducted a hierarchical regression analysis entering all demographic variables accounting for significant variance in caregiver reported older adult IADLs as identified in a separate simultaneous regression. Caregiver education, older adult education, older adult age, and older adult gender were thus used in the first step of the model (Table 7). Caregiver education, older adult education, older adult age, and older adult gender were associated with caregiver reports of older adult IADL status, ΔR² = .123, F(4, 280) = 9.791, p < .001. In the second step, we simultaneously entered the five RBANS indices and found that the model accounted for a significant additional amount of the variance, ΔR² = .096, F(5, 275) = 6.796, p < .001. Contrary to our hypothesis, neither delayed memory or language nor any other RBANS index accounted for unique variance in caregiver-reported IADL status (all p > .05).

Table 7.

Study 2 Regression Analysis for Prediction of Caregiver Report of Older Adult IADL Functioning by Older Adult RBANS Performance and Caregiver–Older Adult IADL Discrepancy Score

Step	Variables entered			β	t	p
Caregiver report of older adult IADL
1A	Caregiver education			.189	2.934	.004
	Older adult education			−.084	−1.322	.187
	Older adult age			.237	3.885	<.001
	Older adult gender			.111	1.959	.051
1B	Caregiver education			.224	3.588	<.001
	Older adult education			.058	0.882	.379
	Older adult age			.149	2.149	.016
	Older adult gender			.078	1.394	.164
	RBANS immediate memory			−.122	−1.402	.162
	RBANS visuospatial constructional			−.052	−0.757	.450
	RBANS language			−.059	−0.795	.427
	RBANS attention			−.068	−0.917	.360
	RBANS delayed memory			−.142	−1.565	.119
Caregiver–older adult IADL discrepancy score
2A	Caregiver education			.197	2.925	.004
	Older adult education			−.111	−1.678	.095
	Older adult age			−.026	−0.403	.687
	Older adult gender			.157	2.655	.008
2B	Caregiver education			.210	3.093	.002
	Older adult education			−.063	−0.872	.384
	Older adult age			−.070	−1.046	.296
	Older adult gender			.134	2.202	.029
	RBANS immediate memory			−.132	−1.399	.163
	RBANS visuospatial constructional			.007	0.090	.928
	RBANS language			−.063	−0.775	.439
	RBANS attention			.082	1.009	.314
	RBANS delayed memory			−.047	−0.482	.630
Step	R	SE	R² change	F change	df	p
1A	.350	2.286	.123	9.791	4, 280	<.001
2A	.228	2.248	.052	3.797	4, 277	.005
1B	.468	2.176	.096	6.796	5, 275	<.001
2B	.279	2.237	.026	1.534	5, 272	.179

Note. RBANS = Repeatable Battery for the Assessment of Neuropsychological Status; ADL = activities of daily living; IADL = instrumental activities of daily living. N = 295.

Discrepancy score analyses were similar. Significant caregiver and older adult demographics (caregiver and older adult education, older adult age) were entered into the first step of a regression. In the second step, we simultaneously added the five RBANS indices and found that the model did not account for a significant proportion of additional variance. In this analysis, no specific RBANS indices were significantly associated with IADL discrepancy scores.

Discussion

In Study 1, we found that caregiver reports of older adult total ADLs was incrementally associated with older adults’ total Cognistat composite scores over self-reports of total ADLs, with cognitive impairment relating to reports of functional impairment. In Study 2, we found a similar pattern. However, unlike Study 1, self-reported functional impairment was associated with cognitive impairment when caregiver report was not taken into account. In both studies, caregiver reports of older adults’ daily functioning were significantly associated with older adult cognition beyond that accounted for by self-report, as hypothesized.

In comparing the association of cognition on BADLs and IADLs, BADL reports by caregiver of older adult functional difficulty was associated with greater cognitive impairment in Study 1, a trend in Study 2. In contrast, older adult BADL reports of functional difficulty were negatively associated with their own cognitive impairment such that those with higher cognitive impairment reported fewer difficulties. This pattern was not repeated in Study 2. These findings are an example of the complexity of reporter bias but suggest that with greater cognitive impairment there is less insight in self-report of one’s own functional difficulties, perhaps to the point of detriment. Trends for IADL reports showed similar patterns between Study 1 and Study 2. In both cases, cognitive impairment positively related to functional impairment for both older adult and caregiver, and in both cases, older adult report only showed significance when caregiver report was not taken into account. When discrepancy scores were used as the unit of measurement for functional status, Study 1 showed no unique variance of either BADLs or IADLs. In Study 2, a larger discrepancy between caregiver- and self-reported BADLs was associated with older adult cognitive impairment. Again, this supports the relatively greater accuracy of caregiver reports as well as the decreased accuracy of self-report when older adult cognitive impairment is high.

In Study 2, the use of the RBANS with its domain-specific cognitive indices allowed an additional exploration of the relationship between specific cognitive domains and BADLs and IADLs. The purpose of this analysis was to identify specific cognitive domains which affect the report of functional status. We hypothesized that older adult RBANS indices of visuospatial and attentional tasks would be associated with ADL decrements because they require visuomotor coordination, speed, and spatial navigation, and our findings were partially supported in that visual-spatial abilities were clearly associated with higher caregiver report of BADLs. However, the expected significant pattern was not observed for the attention index. When caregiver–self-report discrepancy scores were used, visual-spatial abilities and language were found to be strongly associated with BADL, indicating that impairments in these areas related to larger discrepancies between caregiver and self-reports. This finding could be clinically relevant in identifying areas to include in further assessment when self-report and collateral report discrepancy are high.

We further expected that older adults’ performance on the RBANS indices of language and delayed memory would be associated with IADL decrements because of their greater complexity and hence greater requirements of higher order cognitive abilities as well as their association in the literature with preclinical Alzheimer’s disease. However, neither delayed memory nor language processes significantly accounted for additional variance in our measure of IADLs. No one index was significantly associated with caregiver reported IADL original or discrepancy scores. The finding may be indicative of the complexity of skills used within IADLs in that the variance accounted for is not easily restricted to any single cognitive domain.

Cognitive abilities of older adults have long been informally associated with their functional independence. However, there is surprisingly little data evaluating the best functional measures to use in testing these associations, primarily surrounding the validity of caregiver versus self-reports. Our data support the use of caregiver report. These reports can play an important role in tailoring appropriate recommendations to an individual in an effort to maximize their functional independence in light of current cognitive abilities. Specifically, functional reports play a major role in decisions regarding the impact of cognitive impairment on an individual older adult’s functional independence. The current studies used large community samples of older adult dyads to directly compare self-report and caregiver report of older adult functional abilities with two commonly used measures of cognitive functioning in older adults to examine the association of older adult and caregiver reports of functional status and cognition. Our results indicate that caregiver reports of older adult functional ability consistently result in greater association of older adult functional status and cognition when compared with self-reports. This overall finding held true in both studies when ADLs were examined together as well as separated into basic and instrumental domains. A significant relationship linking size of discrepancy scores between self- and caregiver reports of basic functional abilities and care recipient cognitive performance was found within both studies; however, in Study 1, this pattern did not hold when variance due to IADLs was also included in the model. These findings suggest that although self-report of basic functional abilities may be less useful as it relates to cognitive functioning, it likely remains informative at times, framed by the discrepancy with caregiver report. Nevertheless, understanding of the cognitive/functional independence relationship in older adults will be stronger if it includes caregiver reports of overall functional status as this accounts for variance in older adult cognitive status beyond self-reports of overall functional status.

Furthermore, discrepancies in reported abilities for the older adult to successfully complete BADLs in Study 2 were more strongly correlated with cognition than were discrepancies within reported abilities to complete IADLs. We typically think of IADLs necessitating greater cognitive resources and ability. This finding highlights the importance of accurate assessment of BADLs as they may also have significant impact on functional independence. Our findings suggests this discrepancy could be clinically relevant in identifying the need for more in-depth cognitive testing.

The current findings add to the debate regarding the association between cognition and IADLs, an idea gaining interest within the literature through potential diagnostic implications for stages of cognitive decline. Research has implicated IADLs deficits within those suffering from mild cognitive impairment (Perneczky et al., 2006). However, this may be specific to only a subset of MCI sufferers, those individuals with amnestic MCI, a potential precursor to dementia, when compared with noncognitively impaired older adults (Mariani et al., 2008). In this case, only amnestic MCI individuals suffered from IADL deficits, showing no distinction from controls in the domain of BADLs. Some have suggested that initial cognitive difficulties creating IADL impairments may be an indicator of mild cognitive impairment, and current clinical diagnoses for the disorder ignoring the impact of functional status ought to therefore be reexamined (Nygard, 2003). Our findings suggest collateral–self-report discrepancies within BADLs as a potentially relevant variable to add to this debate.

Finally, it should be noted that, although our measures of ADLs are nearly identical in content from Study 1 to Study 2, the way information was collected and coded represent two different approaches reflecting the particular missions of the studies from which they derive. They should be viewed as similar but not identical measurement tools and thus not a complete replication from one to the other.

Footnotes

Declaration of Conflicting Interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This study was supported by the National Institute on Aging (Grant No. AG15321), Gail M. Williamson, Principle Investigator.

Bios

L. Stephen Miller (PhD, University of Mississippi, 1990) is professor of psychology and director of the UGA Bioimaging Research Center. His current research focuses on neurocognitive function and its neural substrates in older adults with and without pathology.

Courtney L. Brown (BA, Tufts University, 2003) is a clinical psychology graduate student in the Psychology Department at the University of Georgia. Her current research focuses on using neuroimaging techniques to study the role of aging within neuropsychological processes such as verbal memory.

Meghan B. Mitchell, PhD, is a postdoctoral fellow in clinical neuropsychology at the Bedford VA Medical Center and a research fellow in neurology at Massachusetts General Hospital/Harvard Medical School. Her research focuses on early cognitive and biological markers of decline in Alzheimer’s disease.

Gail M. Williamson, PhD, is professor emerita, Psychology Department, University of Georgia. As principal investigator of the FRILL projects, her research interests focus on the psychosocial factors that influence various outcomes of the caregiving experience, including caregiver depression and resentment and their impact on the quality of informal care provided.

References

American Psychiatric Association. (2000). Diagnostic and statistical manual of mental disorders (4th ed., text revision). Washington, DC: Author.

Beach

S. R.

Schulz

Williamson

G. M.

Miller

L. S.

Weiner

M. F.

Lance

c. e.

(2005). Risk factors for potentially harmful informal caregiver behavior. Journal of the American Geriatric Society, 53, 255-261.

Bennett

H. P.

Piguet

Grayson

D. A.

Creasey

Waite

L. M.

Lye

. . .Halliday

G. M.

(2006). Cognitive, extrapyramidal, and magnetic resonance imaging predictors of functional impairment in nondemented older community dwellers: The Sydney Older Person Study. Journal of the American Geriatrics Society, 54, 3-10.

Cotter

E. M.

Burgio

L. D.

Stevens

A. B.

Roth

D. L.

Gitlin

L. N.

(2002). Correspondence of the Functional Independence Measure (FIM) Self-Care subscale with real-time observations of dementia patients’ ADL performance in the home. Clinical Rehabilitation, 16, 36-45.

Dodge

H. H.

Kadowaki

Hayakawa

Yamakawa

Sekikawa

Ueshima

(2005). Cognitive impairment as a strong predictor of incident disability in specific ADL–IADL tasks among community-dwelling elders: The Azuchi Study. The Gerontologist, 45, 222-230.

Drane

D. L.

Osato

S. S.

(1997). Using the Neurobehavioral Cognitive Status Examination as a screening measure for older adults. Archives of Clinical Neuropsychology, 12, 139-143.

Elam

J. T.

Graney

M. J.

Beaver

Derwi

D. E.

Applegate

W. B.

Miller

S. T.

(1991). Comparison of subjective ratings of function with observed functional ability of frail older persons. American Journal of Public Health, 81, 1127-1130.

Farias

S. T.

Mungas

Jagust

(2005). Degree of discrepancy between self and other-reported everyday functioning by cognitive status: Dementia, mild cognitive impairment, and healthy elders. International Journal of Geriatric Psychiatry, 20, 827-834.

Freilich

B. M.

Hyer

L. A.

(2007). Relation of the Repeatable Battery for Assessment of Neuropsychological Status to measures of daily functioning in dementia. Psychological Reports, 101, 119-129.

10.

Gill

T. M.

Williams

C. S.

Richardson

E. D.

Tinetti

M. E.

(1996). Impairments in physical performance and cognitive status as predisposing factors for functional dependence among nondisabled older persons. Journal of Gerontology, 51A, M283-M288.

11.

Grigsby

Kaye

Baxter

Shetterly

S. M.

Hamman

R. F.

(1998). Executive cognitive abilities and functional status among community-dwelling older persons in the San Luis Valley Health and Aging Study. Journal of the American Geriatrics Society, 46, 590-596.

12.

Johnson

J. K.

Lui

Yaffe

(2007). Executive function, more than global cognition, predicts functional decline and mortality in elderly women. Journal of Gerontology, 62A, 1134-1141.

13.

Katz

Ford

A. B.

Moskowitz

R. W.

Jackson

B. A.

Jaffe

M. W.

(1963). Studies of illness in the aged. The index of ADL: A standardized measure of biological and psychosocial function. Journal of the American Medical Association, 185, 914-919.

14.

Kiernan

R. J.

Mueller

Langston

J. W.

Van Dyke

(1987). The Neurobehavioral Cognitive Status Examination: A brief but differentiated approach to cognitive assessment. Annals of Internal Medicine, 107, 481-485.

15.

Lawton

M. P.

Brody

E. M.

(1979). Assessment of older people: Self maintaining and instrumental activities of daily living. Gerontologist, 9, 179-186.

16.

Lewis

M. S.

Miller

L. S.

(2007). Executive control functioning and functional ability in older adults. Clinical Neuropsychologist, 21, 274-285.

17.

Lin

(2008). Solving multicollinearity in the process of fitting regression model using the nested estimate procedure. Quality and Quantity, 42, 417-426.

18.

Mariani

Monastero

Ercolani

Rinaldi

Mangialasche

Costanzi

. . .Mecocci

(2008). Influence of comorbidity and cognitive status on instrumental activities of daily living in amnestic mild cognitive impairment: Results for the ReGAI project. International Journal of Geriatric Psychiatry, 23, 523-530.

19.

Miller

L. S.

Lewis

M. S.

Williamson

G. M.

Lance

C. E.

Dooley

W. K.

Schulz

Weiner

M. F.

(2006). Caregiver cognitive status and potentially harmful caregiver behavior. Aging & Mental Health, 10, 125-133.

20.

Mitchell

M. B.

Miller

L. S.

(2008a). Executive functioning and observed versus self-reported measures of functional ability. Clinical Neuropsychologist, 22, 471-479.

21.

Mitchell

M. B.

Miller

L. S.

(2008b). Prediction of functional status in older adults: The ecological validity of four Delis-Kaplan Executive Function System tests. Journal of Clinical and Experimental Neuropsychology, 30, 683-690.

22.

National Alliance for Caregiving & AARP. (2004, April). Caregiving in the U.S. (Funded by MetLife Foundation). Retrieved from http://www.caregiving.org/data/04finalreport.pdf

23.

Northern California Neurobehavioral Group, Inc. (1995). Manual for the Neurological Cognitive Status Exam. Fairfax: Author.

24.

Nygard

(2003). Instrumental activities of daily living: A stepping-stone towards Alzheimer’s disease diagnosis in subjects with mild cognitive impairment? ACTA Neurologica Scandinavica, 107(Suppl. 179), 42-46.

25.

Older American Resources and Services, Duke University. (1978). Multidimensional functional assessment, the OARS methodology: A manual (2nd ed.). Durham, NC: Duke University, Center for the Study of Aging and Human Development.

26.

Perneczky

Pohl

Sorg

Hartmann

Tosic

Grimmer

. . .Kurz

(2006). Impairment of activities of daily living requiring memory or complex reasoning as part of the MCI syndrome. International Journal of Geriatric Psychiatry, 21, 158-162.

27.

Petrella

J. K.

Cress

M. E.

Miller

L. S.

(2004). Leg extensor power, cognition, and functional performance in independent and marginally dependent older adults. Age and Ageing, 33, 342-348.

28.

Randolph

(1998). RBANS manual: Repeatable Battery for the Assessment of Neuropsychological Status. San Antonio, TX: Psychological Corp (Harcourt).

29.

Randolph

Tierney

Mohr

Chase

(1998). The Repeatable Battery for the Assessment of Neuropsychological Status (RBANS): Preliminary clinical validity. Journal of Clinical and Experimental Neuropsychology, 20, 310-319.

30.

Ready

Ott

Grace

(2004). Validity of informant reports about AD and MCI patients’ memory. Alzheimer Disease & Associated Disorders, 18, 11-16.

31.

Rubenstein

L. Z.

Schairer

Wieland

G. D.

Kane

(1984). Systematic biases in functional status assessment of elderly adults: Effects of different data sources. Journal of Gerontology, 39, 686-691.

32.

Wadley

Harrell

Marson

(2003). Self- and informant report of financial abilities in patients with Alzheimer’s disease: Reliable and valid? Journal of the American Geriatric Society, 51, 1621-1626.

33.

Zanetti

Geroldi

Frisoni

G. B.

Bianchetti

Trabucchi

(1999). Contrasting results between caregiver’s report and direct assessment of activities of daily living in patients affected by mild and very mild dementia: The contribution of the caregiver’s personal characteristics. Journal of the American Geriatrics Society, 47, 196-202.