Assessment of an Expanded Functional Disability Scale for Older Adults With Diabetes

Abstract

Although prior literature has shown the plausibility of combining the Activities of Daily Living (ADL) and Instrumental Activities of Daily Living (IADL) items to form an expanded scale for measuring the degree of functional decline, this has not been shown in older adults with diabetes who are disproportionately affected by functional disability. Using the 2009 Medicare Current Beneficiary Survey data, we evaluated the factor structure of the pooled ADL and IADL items. Based on our study comprising 2,158 community-dwelling older adults (≥65 years) with diabetes, the unidimensional model exhibited good fit. Despite well-fitting indices, high correlations were observed between the latent constructs (>.70) of the multi-factor models, suggesting a lack of discriminant validity. These findings provide empirical support for a combined scale that can comprehensively and efficiently characterize the extent of functional disability in older adults with diabetes for research, risk adjustment, and evaluation in patient-centered medical homes.

Keywords

diabetes functional disability confirmatory factor analysis Medicare

Diabetes mellitus is a highly prevalent chronic disease in the United States, affecting about 27% of the population aged 65 years and over (Centers for Disease Control and Prevention, 2011), and about 34% of the older adults in nursing homes (Coxe, Lennertz, & McCullough, 2013). Functional disability disproportionately affects older adults with diabetes, where approximately 30% to 55% of them experience some form of limitation in their usual activities or mobility (Bruce, Davis, & Davis, 2005; Chiu & Wray, 2011; Kalyani, Saudek, Brancati, & Selvin, 2010). The care and management of the elderly with diabetes is further complicated by the heterogeneous spectrum of functional disability, where some individuals might be physically and cognitively robust whereas others are fraught with multiple comorbidities (American Diabetes Association [ADA], 2014; Blaum, Ofstedal, Langa, & Wray, 2003; Durso, 2006; Sinclair et al., 2012). Recent epidemiologic studies have consistently found that diabetes is associated with disability, which affects quality of life and is a major risk factor for loss of independence, falls, injuries, infections, and institutionalization (ADA, 2008, 2013; Chau et al., 2011; Chiu & Wray, 2011; Salas, Bubolz, & Caro, 2000; Volpato, Maraldi, & Fellin, 2010). The primary concern of older adults with diabetes was also demonstrated to be disease-related complications, such as amputation and blindness, resulting in disabilities (Quandt et al., 2013). These findings highlight the need to effectively identify and assess the extent of functional disability among older adults with diabetes for both clinical research and screening purposes.

In current practice, the Activities of Daily Living (ADL; Katz, Ford, Moskowitz, Jackson, & Jaffe, 1963) and Instrumental Activities of Daily Living (IADL; Rosow & Breslau, 1966) are well-established scales conventionally used to ascertain functional disability. The ADL scale measures activities that are essential for self-care, whereas the IADL scale measures activities that are necessary for independent adaptation to living within the environment (Spector, Katz, Murphy, & Fulton, 1987). Psychometric properties of the ADL and IADL scales, such as test–retest reliability and internal consistency, have been previously evaluated and well established (Ivanova et al., 2013). Although both scales are well accepted and widely used for assessing different domains of functional disability, previously published research has suggested the plausibility of combining the ADL and IADL items to form an expanded scale, which is potentially a more sensitive assessment than the stand-alone scales to measure a wider range of functional decline (Fortinsky, Garcia, Joseph Sheehan, Madigan, & Tullai-McGuinness, 2003; Spector & Fleishman, 1998; Spector et al., 1987). This combined ADL-IADL scale can offer several practical advantages to clinicians and researchers. For one, it could provide clinicians with a more comprehensive assessment of the extent of functional disability that would aid in the tailoring of treatment regimens. In addition, in the context of research practice, a single scale confers the analytic advantage of avoiding problems with collinearity that typically occur when multiple, highly correlated scales are concurrently used.

To date, studies addressing the dimensionality of functional disability have yielded inconsistent findings in the general older adult population. Although a body of evidence suggests that the ADL and IADL can be combined into a unidimensional scale (Fortinsky et al., 2003; Kempen & Suurmeijer, 1990; Spector & Fleishman, 1998; Spector et al., 1987) (i.e., the ADL and IADL measure the same underlying factor or construct), others argue that functional disability is multidimensional and multiple scales are needed to provide a functional disability profile for patients (Clark, Stump, & Wolinsky, 1997; Thomas, Rockwood, & McDowell, 1998). Furthermore, the IADL scale has also been hypothesized to comprise two dimensions—the physical IADLs and the cognitive IADLs (Ng, Niti, Chiam, & Kua, 2006; Thomas et al., 1998), such that the combined ADL and IADL items could potentially represent three separate constructs accounting for the full spectrum of functional disability.

Clinical guidelines highlight the importance of recognizing and addressing functional disability among older adults with diabetes, for whom treatment goals should be tailored according to the degree and complexity of functional disability (ADA, 2014; A. F. Brown, Mangione, Saliba, & Sarkisian, 2003; Durso, 2006; Sue Kirkman et al., 2012). Consequently, there is a need to appropriately characterize functional disability within this population. Coupled with the debate over the dimensionality of functional disability measures, this study seeks to evaluate the dimensionality and to examine the factor structure of the pooled items from the individual ADL and IADL scales. This is the first study that examines whether the ADL and IADL items can be combined into a singular and unidimensional scale for an enhanced and practical assessment of functional disability in older adults with diabetes. It is particularly important to assess the unidimensionality of functional disability within the diabetes subpopulation of older adults as they have unique or varying needs from the general older adult population.

Method

Data Source

This cross-sectional study analyzed data from the 2009 Medicare Current Beneficiary Survey (MCBS), which is a nationally representative survey of Medicare beneficiaries conducted by the Centers for Medicare and Medicaid Services (CMS; 2013). The survey dataset contains comprehensive information regarding individual demographic characteristics, health and functional status, and person-level health service utilization and expenditure. Survey responses were also supplemented and linked to administrative data on Medicare enrollment and all Medicare Part A (hospital insurance), Part B (medical insurance), and Part D (prescription drug coverage) claims records. This study was reviewed and approved by the University of Maryland Baltimore Institutional Review Board.

Sample

The study sample was selected according to the following inclusion criteria: (a) beneficiaries aged 65 years and older, (b) resided in the community, and (c) diagnosed with diabetes. Presence of diabetes was ascertained using both self-report and claims information. We identified presence of diabetes from claims using an algorithm adapted from the CMS for the Chronic Condition Data Warehouse (Hebert et al., 1999; Chronic Conditions Data Warehouse, 2014), where diabetes was considered present if a beneficiary has either (a) one or more inpatient hospital or home health care claims or (b) at least two outpatient hospital or physician claims with the following International Classification of Disease, Ninth revision (ICD-9) codes: 250.xx, 357.2, 362.01, 362.02, 366.41. Diabetes was also considered present if a beneficiary self-reported having diabetes with a positive response to the survey question, “Has a doctor ever told you that you had any type of diabetes, including: sugar diabetes, high blood sugar, (borderline diabetes, pre-diabetes, or pregnancy-related diabetes/borderline diabetes, or pre-diabetes)?”

Measures

To ascertain the presence of limitation in the ADL, beneficiaries in the MCBS were presented with a series of questions asking “Because of a health or physical problem, do you have any difficulty in (1) bathing or showering, (2) dressing, (3) eating, (4) getting in or out of bed or chairs, (5) walking, and (6) using the toilet?” The IADL scale measures higher order tasks that require greater neuropsychological organization than the basic ADL, and are necessary for independent living (Ng et al., 2006; Spector et al., 1987). To ascertain limitations in IADL, beneficiaries were asked whether they have difficulty in (a) using the telephone, (b) doing light housework, (c) doing heavy housework, (d) preparing meals, (e) shopping, and (f) managing money.

The available responses for all items in the ADL and IADL scales consisted of “yes,” “no,” “doesn’t do,” “not ascertained,” “don’t know,” or “refused.” Responses of “not ascertained,” “don’t know,” or “refused” were deemed as missing. When a beneficiary indicated “doesn’t do,” a follow-up question was asked to determine if they did not do the particular activity because of health problems. If they indicated yes to the follow-up question, they were recoded as having limitation for that particular activity. In contrast, those who indicated “doesn’t do” in the original question and “no” to the follow-up question would be recoded as having a missing response because we cannot determine with certainty if they did not perform that activity due to health constraints or their lifestyles do not include engagement in that particular activity (e.g., some older persons might not need to prepare their own meals on a regular basis). In this way, the final responses to every item on both scales were dichotomous (1 = yes, 0 = no), with allowances for missing responses.

Data Analysis

Descriptive analyses of the baseline characteristics were first conducted. Individuals who had missing responses on all 12 questions were excluded from the analysis. Internal consistency reliability of the ADL, IADL, and combined ADL-IADL scales was assessed using Cronbach’s alpha. A Cronbach’s alpha value between .70 and .95 is generally acceptable (Tavakol & Dennick, 2011). Concurrent validity of the combined ADL-IADL scale was further assessed by examining the proportion of beneficiaries with self-reported comorbidities at different score ranges along the ADL-IADL scale. It was hypothesized that a higher score on the ADL-IADL scale would reflect a higher burden of comorbidities. These analyses were conducted using SAS version 9.3 (SAS Institute Inc., Cary, NC, USA) or STATA/MP version 12 (StataCorp, 2011).

Confirmatory factor analyses (CFAs) were conducted using Mplus version 7.1 (Muthén & Muthén, 1998-2012) with weighted least squares with mean- and variance-adjusted estimation (WLSMV) for dichotomous data (Schmitt, 2011) to examine the factor structure of the pooled items from the ADL and IADL. CFA is appropriate in this context as the dimensionalities of these items from the two scales have been explored in previous studies (Clark et al., 1997; Fortinsky et al., 2003; Kempen & Suurmeijer, 1990; Ng et al., 2006; Spector & Fleishman, 1998; Spector et al., 1987; Thomas et al., 1998). Hence, instead of using an exploratory factor analysis to derive the factor structure for a new set of variables, our research interest lies in investigating whether existing established dimensionalities apply to a different population, older adults with diabetes.

Based on prior research as stated above, three competing models were tested: (a) a one-factor model combining all ADL and IADL items into one factor, (b) a two-factor model with the ADL items on one factor and the IADL items on a second factor, and (c) a three-factor model with the ADL items, the physical IADLs, and the cognitive IADLs each forming separate factors. The three competing models were evaluated and compared with multiple fit indices, including the chi-square goodness-of-fit index, comparative fit index (CFI), Tucker–Lewis index (TLI), and the root mean square error of approximation (RMSEA). A model has good or acceptable fit when it has a nonsignificant chi-square test, CFI and TLI above 0.95 (Hu & Bentler, 1999; Kline, 2011), and RMSEA of .06 or less (Hu & Bentler, 1999). In addition, both unstandardized and standardized coefficients were examined; ideally, the standardized coefficients should be close to or greater than 0.70 (Kline, 2011).

Results

After excluding 10 observations with missing responses on all 12 items, the final study sample comprised 2,158 elderly Medicare beneficiaries who resided in the community and had diabetes. Table 1 includes baseline sociodemographic characteristics, self-reported comorbidities, and self-rated health status of the study sample. More than half of the beneficiaries were between 65 and 80 years old, married, and had a high school education or less. The beneficiaries were also predominantly white (81.4%) and reported household income of US$50,000 or less (86.7%).

Table 1.

Baseline Characteristics of Beneficiaries With Diabetes (N = 2,158).

	All beneficiaries (N = 2,158)
Characteristics	n	%
Age
65-69	459	21.3
70-74	491	22.8
75-79	462	21.4
80-84	412	19.1
85+	334	15.5
Gender
Female	1,096	50.8
Male	1,062	49.2
Race (missing = 10)
White	1,748	81.4
African American	235	10.9
Others	165	7.7
Marital status (missing = 2)
Single	1,014	47.0
Married	1,142	53.0
Education (missing = 8)
Less than high school	589	27.4
High school or vocational diploma	813	37.8
Some college	332	15.4
College degree or higher	416	19.4
Income (missing = 1)
<US$25,000	1,079	50.0
US$25,000 to US$50,000	792	36.7
>US$50,000	286	13.3
Number of comorbidities^a
0-1	325	15.1
2-3	1,051	48.7
4+	782	36.2
Self-rated health status (missing = 5)
Excellent	181	8.4
Very good	548	25.5
Good	770	35.8
Fair	471	21.9
Poor	183	8.5

Data Source. Medicare Current Beneficiary Survey (2009).

Note. Ten observations were dropped as they had missing values on all items of the Activities of Living scale and the Instrumental Activities of Living scale.

Comorbidity categories are self-reported and include cardiac disease, hypertension, cerebrovascular disease (CVD), respiratory disease, cancer, arthritis, neurological conditions, psychiatric disorder, and osteoporosis or bone fracture.

The proportions of beneficiaries who indicated difficulties in any of the 12 functional activities are presented in Table 2. For the ADL items, the largest proportion of beneficiaries (34.5%) with diabetes faced difficulties in walking. Between 10% and 20% of the study population indicated difficulties with each of bathing/showering, dressing, and getting in and out of bed/chairs. Less than 10% reported difficulties with using the toilet or eating. With regard to IADL items, a substantial proportion of the diabetic elderly (40.4%) had difficulties with doing heavy housework. In contrast, approximately 1 in 10 beneficiaries reported problems using the telephone or managing money.

Table 2.

Proportion of Beneficiaries With Difficulties in ADL or IADL (N = 2,158).

	Beneficiaries with difficulty (N = 2,158)
ADL/IADL items	n	%
ADL scale
Do you have any difficulty in bathing or showering?
Yes	332	15.4
No	1,825	84.5
NA/missing	1	0.1
Do you have any difficulty in dressing?
Yes	232	10.8
No	1,926	89.2
NA/missing	—	—
Do you have any difficulty in eating?
Yes	49	2.3
No	2,109	97.7
NA/missing	—	—
Do you have any difficulty getting in or out of bed or chairs?
Yes	369	17.1
No	1,783	82.6
NA/missing	6	0.3
Do you have any difficulty in walking?
Yes	744	34.5
No	1,414	65.5
NA/missing	—	—
Do you have any difficulty in using the toilet?
Yes	164	7.6
No	1,994	92.4
NA/missing	—	—
IADL scale
Do you have any difficulty in doing light housework (such as washing dishes, straightening up, or light cleaning)?
Yes	349	16.2
No	1,725	79.9
NA/missing	84	3.9
Do you have any difficulty in doing heavy housework (such as scrubbing floors or washing windows)?
Yes	871	40.4
No	1,092	50.6
NA/missing	195	9.0
Do you have any difficulty in preparing (your/his/her) own meals?
Yes	280	13.0
No	1,771	82.0
NA/missing	107	5.0
Do you have any difficulty in shopping for personal items (such as toilet items or medicines)?
Yes	408	18.9
No	1,720	79.7
NA/missing	30	1.4
Do you have any difficulty in using the telephone?
Yes	172	8.0
No	1,979	91.7
NA/missing	7	0.3
Do you have any difficulty in managing money (such as keeping track of expenses or paying bills)?
Yes	203	9.4
No	1,886	87.4
NA/missing	69	3.2

Data Source. Medicare Current Beneficiary Survey (2009).

Note. Available responses were “yes,” “no,” “doesn’t do,” “not ascertained,” “don’t know,” or “refused.” If the latter three are selected, responses were treated as missing. If “doesn’t do” is indicated, a follow-up question was asked to determine whether they did not do the particular activity because of health problems. Recoding: If they indicated yes to the follow-up question, they would be recoded as “yes.” If they indicated no to the follow-up question, they would be recoded as missing. Ten observations were dropped as they were missing on all items of the ADL and IADL. ADL = Activities of Daily Living; IADL = Instrumental Activities of Daily Living; NA = not applicable.

Internal Consistency and Concurrent Validity of the Combined ADL-IADL Scales

The Cronbach’s alpha for the ADL (.798), IADL (.834), and combined ADL-IADL (.890) scales displayed good internal consistency reliability. As the scores on the combined ADL-IADL scale increased, the proportion of beneficiaries with comorbidities increased correspondingly (Table 3). In particular, comparing a combined ADL-IADL score of 9–12 to 0, we observed a fivefold or more increase in cerebrovascular disease, neurological, psychiatric, and respiratory conditions.

Table 3.

Proportion of Beneficiaries With Comorbidities, Stratified by Scores on the Combined ADL-IADL Scale.

	Scale score on the combined ADL-IADL
	0 (%)	1-4 (%)	5-8 (%)	9-12 (%)
Heart	43.7	61.1	63.4	73.1
Hypertension	80.3	86.4	88.6	87.5
CVD	9.6	17.0	26.4	46.2
Cancer	17.6	26.0	24.0	27.9
Arthritis	51.8	72.7	78.1	84.6
Neurological	3.8	9.5	22.0	52.9
Psychiatric	8.3	15.9	27.2	57.7
Bone	15.0	21.7	31.2	55.8
Respiratory	14.2	24.0	36.2	73.1

Data Source. Medicare Current Beneficiary Survey (2009).

Note. Comorbidity categories are self-reported and include cardiac disease, hypertension, CVD, respiratory disease, cancer, arthritis, neurological conditions, psychiatric disorder, and osteoporosis or bone fracture. ADL = Activities of Daily Living; IADL = Instrumental Activities of Daily Living; CVD = cerebrovascular disease.

Confirmatory Factor Analysis

Estimated coefficients and model fit indices for the one-factor, two-factor, and three-factor models are presented in Tables 4 and 5, respectively. The unidimensional model is illustrated in Figure 1. Standardized coefficients for nine items in the one-factor model are 0.70 or greater; three items have coefficients approaching 0.70: eating (0.696), getting in or out of bed or chairs (0.668), and using the telephone (0.637) (see Table 4). Similarly, most of the standardized coefficients in the two-factor and three-factor models were at least 0.70 (see Table 4).

Table 4.

Coefficients of CFAs—One-Factor, Two-Factor, and Three-Factor Models (Unstandardized Coefficient [SE]; Standardized Coefficient [SE]).

		Two-factor		Three-factor
Item	One-factor	ADL	IADL	ADL	IADL cognitive	IADL physical
Bathing or showering	1.000 (0.000); 0.905 (0.011)	1.000 (0.000); 0.930 (0.010)		1.000 (0.000); 0.929 (0.010)
Dressing	0.989 (0.013); 0.895 (0.012)	0.987 (0.015); 0.918 (0.012)		0.988 (0.015); 0.918 (0.012)
Eating	0.770 (0.042); 0.696 (0.038)	0.762 (0.041); 0.709 (0.038)		0.763 (0.041); 0.709 (0.038)
Getting in or out of bed or chairs	0.738 (0.017); 0.668 (0.013)	0.743 (0.017); 0.691 (0.014)		0.744 (0.017); 0.692 (0.014)
Walking	0.957 (0.017); 0.866 (0.013)	0.966 (0.017); 0.898 (0.013)		0.966 (0.017); 0.897 (0.012)
Using the toilet	0.959 (0.019); 0.867 (0.016)	0.952 (0.019); 0.885 (0.016)		0.953 (0.019); 0.885 (0.016)
Using the telephone	0.704 (0.035); 0.637 (0.031)		1.000 (0.000); 0.654 (0.032)		1.000 (0.000); 0.726 (0.033)
Managing money	0.901 (0.021); 0.815 (0.017)		1.274 (0.061); 0.833 (0.017)		1.297 (0.064); 0.942 (0.023)
Doing light housework	1.051 (0.013); 0.951 (0.008)		1.479 (0.072); 0.966 (0.007)			1.000 (0.000); 0.963 (0.007)
Doing heavy housework	0.987 (0.017); 0.893 (0.012)		1.398 (0.071); 0.9914 (0.012)			0.944 (0.014); 0.910 (0.012)
Preparing own meals	1.039 (0.014); 0.939 (0.009)		1.457 (0.071); 0.952 (0.009)			0.985 (0.011); 0.948 (0.009)
Shopping for personal items	1.034 (0.014); 0.935 (0.008)		1.454 (0.071); 0.950 (0.008)			0.982 (0.010); 0.946 (0.008)

Data Source. Medicare Current Beneficiary Survey (2009).

Note. A total of 2,158 observations were included; 10 observations were excluded as they were missing on all items. The weighted least squares with mean and variance-adjusted estimator was used to derive unstandardized and standardized coefficients. The standardized coefficients reported are the STDYX coefficients from MPlus. All p values were highly significant at <.0001. CFAs = confirmatory factor analyses; SE = standard error; ADL = Activities of Daily Living; IADL = Instrumental Activities of Daily Living.

Table 5.

Fit Statistics Values for Confirmatory Factor Analysis—One-Factor, Two-Factor, and Three-Factor Models.

Index	One-factor	Two-factor	Three-factor
χ²	737.218	478.689	431.442
Degree of freedom	54	53	51
p value	<.001	<.001	<.001
RMSEA	.077	.061	.059
RMSEA 90% CI	[0.072, 0.082]	[0.056, 0.066]	[0.054, 0.064]
pRMSEA	p < .001	p < .001	p = .002
CFI	0.980	0.988	0.989
TLI	0.976	0.985	0.986

Data Source. Medicare Current Beneficiary Survey (2009).

Note. 2,158 observations were included; 10 observations were excluded as they were missing on all items. Weighted least squares estimation was used for the confirmatory factor analysis models.

RMSEA = root mean square error of approximation; CFI = comparative fit index; TLI = Tucker–Lewis index.

Figure 1.

Confirmatory factor analysis of one-factor model with standardized coefficients (standard errors).

All three models had significant chi-square goodness-of-fit values (see Table 5). The three-factor model had a RMSEA value that was less than .06, and the RMSEA for the one-factor and two-factor models were above .06 but below .08. Incremental fit indices for all three models were excellent, with CFI and TLI values above 0.95. Because all three models exhibited reasonable fit based on the CFI, TLI, and RMSEA indices, no modifications were made to avoid increasing the complexity of the models.

In the two-factor model, we observed a very strong correlation of .898 between the ADL and IADL factors. Likewise, the correlations among the latent constructs in the three-factor model were also strong to very strong: ADL with physical IADL (r = .910), physical IADL with cognitive IADL (r = .908), and cognitive IADL with ADL (r = .745).

When we excluded beneficiaries reporting pre-diabetes, which represented only 3% of the study population, the findings did not change (see the online appendix).

Discussion

Assessing and quantifying functional disability among older adults with diabetes is extremely important in the management of diabetes because many diabetes-related complications are associated with functional limitations (ADA, 2014; Durso, 2006; Sinclair et al., 2012). For example, nerve problems arising from diabetes may limit mobility, and uncontrolled diabetes can interfere with cognition and affects the ability to manage money (Chiu & Wray, 2011). The ADL and IADL scales are traditionally used as separate scales, and to the best of our knowledge, this is the first study examining the factor structure of the ADL and IADL items among a large representative sample of older adults with diabetes living in the community.

Our incremental fit indices (CFI and TLI) and RMSEA values were satisfactory for the one-factor model, indicating that the ADL and IADL can be combined to form a single measure of functional disability for older adults with diabetes. Good fit indices were also obtained with the two-factor and three-factor models; however, very strong correlations among the factors were observed in these multi-factor models. The very high correlation (>.85) between the ADL and IADL in the two-factor model implied that both factors have a high degree of overlap and lack discriminant validity (T. A. Brown, 2006). This suggests that functional disability can be characterized as a single dimension among older adults with diabetes, and further demonstrated the plausibility of a combined ADL-IADL scale, which would also be more parsimonious. Parallel conclusions can be drawn with the observed high correlations among ADL, physical IADL, and cognitive IADL factors in the three-factor model.

Our findings are consistent with earlier work by Spector et al. (1987) and Kempen & Suurmeijer, 1990). Employing Item Response Theory, Spector and colleagues (1987) demonstrated the feasibility and validity of combining the ADL and IADL into one scale for community-dwelling older adults. In the same study, the authors noted that the two-factor solution did not perform well and displayed very high inter-factor correlations. Similar to this study, Kempden & Suurmeijer, (1990) also provided evidence for combining the ADL and IADL scales, with the use of Guttman scaling and principal components analysis.

Although the individual ADL and IADL scales of the MCBS have been widely used in various disability-related epidemiology and outcomes studies (Salas et al., 2000; Saliba et al., 2001; Yang & Hall, 2008), a unidimensional ADL-IADL scale may offer several key advantages. For one, an expanded scale would have practical implications as the broader scale range may better quantify the severity of functional disability within community-dwelling older adults with diabetes and thus, help identify the distribution of disabling problems, service needs, and service utilization. It has been suggested that among older adults with limitations in ADL, a substantial portion of them reported receiving inadequate assistance (Desai, Lentzner, & Weeks, 2001). Therefore, screening for such persons requiring assistance can help to identify any unmet needs among this population and facilitate provision of more targeted efforts toward improving care management for these individuals (Desai et al., 2001).

Both the American Geriatrics Society and ADA have highlighted the importance of regular assessment of functional status among older adults with diabetes as treatment goals are tailored according to the individual’s functional status and frailty (ADA, 2014; A. F. Brown et al., 2003; Sue Kirkman et al., 2012). Functional status is heterogeneous among older adults with diabetes; some experience fairly dire complications resulting in significant functional disability whereas others are able to maintain reasonably good health and independence (ADA, 2014; Blaum et al., 2003; Sinclair et al., 2012). In older adults who are frail and fraught with major comorbidities and functional impairment, it is recommended that diabetes therapeutic targets be less stringent as these individuals are more vulnerable to the negative consequences of hypoglycemia arising from aggressive treatment. In such cases, it is acceptable for higher A1C targets to be established (<8% as compared with guidelines for all adults at <6.5% or 7%) (ADA, 2014; Sue Kirkman et al., 2012). For such older adults with diabetes, functional status may be a more appropriate outcome to assess as compared with relying solely on conventional numerical outcomes (e.g., A1C, blood pressure, low density lipoprotein levels). Practical and valid assessment tools that could improve the evaluation of the degree of functional disability within this vulnerable population are needed to identify needs and plan appropriate treatment. Such tools can efficiently stratify older adults according to their likelihood of adverse outcomes (Durso, 2006), such as injurious falls due to functional decline. Our study provides empirical evidence for an enhanced ADL-IADL scale for use among older adults with diabetes to quantify the extent to which they face functional disability.

The combined scale may also have practical implications and analytical advantages in research, such that this single combined scale would avoid problems of multicollinearity that potentially occur when the ADL and IADL are used concurrently (Spector & Fleishman, 1998). Our proposed combined ADL-IADL scale also demonstrated good internal consistency and concurrent validity. Specifically, we showed that a higher score on the ADL-IADL scale is associated with substantially more self-reported comorbidities, including cerebrovascular disease, cardiac diseases, arthritis, and neurologic and psychiatric conditions, which are associated with poorer outcomes and higher costs (Vogeli et al., 2007). Other studies have shown that almost half the older adults with diabetes have three other comorbidities, further complicating the management of their diabetes due to competing treatment demands (Kerr et al., 2007; Piette & Kerr, 2006). Patients also highlighted that presence of comorbidities presents a considerable barrier toward self-management of diabetes (Yankeelov, Faul, D’Ambrosio, Collins, & Gordon, 2013). Hence, when screening for functional disability, clinicians should be alerted to the fact that their patients are likely to have a greater comorbidity burden if they score high on the ADL-IADL scale. Nevertheless, even though our results supported the use of a unidimensional ADL-IADL scale, there may also be meaningful uses for the two- and three-factor models. Because our results also indicated good fit indices for the two-factor and three-factor models, this implied that both models might also be valid among older adults with diabetes. These two- or three-factor scales may be useful in instances where different domains of functional impairment are of particular importance. However, further research is warranted to further assess the applicability of these multi-factor scales in this population. For example, in the stroke population, it has been shown that a two-factor structure is valid and could discriminate stroke patients with or without dementia, depending on the responses to the cognitive subscale (Ng et al., 2006). Likewise, using our multi-factor models, clinicians can potentially distinguish diabetes patients with and without cognitive impairment. This is important in the older diabetes population, as special attention would have to be paid to patients with cognitive issues because these can impede the patients’ ability to exercise, perform glucose monitoring, check their feet for sores, or remember to take their medications (Kerr et al., 2007).

Recent guidelines from the International Diabetes Federation also emphasize the need for a holistic, multi-factorial approach toward the management of diabetes (International Diabetes Federation, 2013). Therefore, we also acknowledge that the unidimensional scale may not be applicable to all clinical settings. While the unidimensional tool provides an efficient screen for the extent of general functional decline, it does not preclude the need for clinicians to carefully characterize the types of disabilities faced by their patients so that individualized therapeutic regimens and care management plans can be made.

Some limitations of this study should be acknowledged. Although the study examines the factor structure of the ADL and IADL items and the internal consistency of the scales, other psychometric properties such as test–retest reliability and predictive validity require additional assessment. This should be a basis for future validation studies. The study population is also limited to Medicare beneficiaries with diabetes and our findings may not be generalizable to nonelderly adults with diabetes or older adults without diabetes. Finally, literature has also shown the possibility of a hierarchical relationship between ADL and IADL (Kempen & Suurmeijer, 1990; LaPlante, 2010; Spector et al., 1987), which might necessitate the use of scaling techniques to verify whether a hierarchical, combined scale applies within our study population. That is beyond the scope of this study, and future research is needed to explore whether a singular, hierarchical measure of functional disability has more discriminative power.

In conclusion, these findings provide empirical evidence for the plausibility of a combined ADL-IADL scale. Such a combined scale can comprehensively and efficiently characterize the extent of functional disability in older adults with diabetes for clinical screening, research, and risk adjustment purposes.

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 received no financial support for the research, authorship, and/or publication of this article.

Author Biographies

Xinyi Ng, BSPharm, is a doctoral candidate in the Department of Pharmaceutical Health Services Research at the University of Maryland, Baltimore. Her research interests are mainly in the areas of patient-reported outcomes and the application of conjoint analysis to elicit patients’ preferences.

Charlene C. Quinn, RN, PhD, is an associate professor at the University of Maryland School of Medicine, Department of Epidemiology and Public Health, Division of Gerontology. She is trained as a geriatric nurse and health services researcher. Her clinical and research expertise is in type 2 diabetes management of older adults. Her current research studies evaluate diabetes mobile health interventions. Working with other clinical investigators, she applies her knowledge of epidemiologic and randomized clinical trial design and methods to studies of diabetes and geriatrics.

Mehmet Burcu, MS, is a doctoral candidate in the Department of Pharmaceutical Health Services Research at the University of Maryland, Baltimore, with interest in the areas of pharmacoepidemiology and outcomes research, comparative effectiveness and safety research, and methods used in validating patient-reported outcomes.

Donna Harrington, PhD, is professor and associate dean for doctoral and post-doctoral education in the University of Maryland, Baltimore School of Social Work. She teaches doctoral courses in advanced data analysis and the integration of theory and research methods. Her research focuses on measurement, ethics, and older adults.

References

American Diabetes Association. (2008). Economic costs of diabetes in the U.S. in 2007. Diabetes Care, 31, 596-615. doi:10.2337/dc08-9017

American Diabetes Association. (2013). Economic costs of diabetes in the U.S. in 2012. Diabetes Care, 36, 1033-1046. doi:10.2337/dc12-2625

American Diabetes Association. (2014). Standards of medical care in diabetes—2014. Diabetes Care, 37(Suppl. 1), S14-S80. doi:10.2337/dc14-S014

Blaum

C. S.

Ofstedal

M. B.

Langa

K. M.

Wray

L. A.

(2003). Functional status and health outcomes in older Americans with diabetes mellitus. Journal of the American Geriatrics Society, 51, 745-753. doi:10.1046/j.1365-2389.2003.51256.x

Brown

A. F.

Mangione

C. M.

Saliba

Sarkisian

C. A.

(2003). Guidelines for improving the care of the older person with diabetes mellitus. Journal of the American Geriatrics Society, 51(5 Suppl. Guidelines), S265-S280. doi:10.1046/j.1532-5415.51.5s.1.x

Brown

T. A.

(2006). Confirmatory factor analysis for applied research. New York, NY: Guilford Press.

Bruce

D. G.

Davis

W. A.

Davis

T. M. E.

(2005). Longitudinal predictors of reduced mobility and physical disability in patients with type 2 diabetes: The Fremantle diabetes study. Diabetes Care, 28, 2441-2447. doi:10.2337/diacare.28.10.2441

Centers for Disease Control and Prevention. (2011). National diabetes fact sheet. Retrieved from http://www.cdc.gov/diabetes/pubs/pdf/ndfs_2011.pdf

Centers for Medicare & Medicaid Services. (2014). Medicare Current Beneficiary Survey (MCBS). Retrieved from http://www.cms.gov/Research-Statistics-Data-and-Systems/Research/MCBS/index.html?redirect=/mcbs/

10.

Chau

P. H.

Woo

Lee

Cheung

Chen

Chan

. . . McGhee

(2011). Older people with diabetes have higher risk of depression, cognitive and functional impairments: Implications for diabetes services. The Journal of Nutrition, Health & Aging, 15, 751-755. doi:10.1007/s12603-011-0071-z

11.

Chiu

C.-J.

Wray

L. A.

(2011). Physical disability trajectories in older Americans with and without diabetes: The role of age, gender, race or ethnicity, and education. The Gerontologist, 51, 51-63. doi:10.1093/geront/gnq069

12.

Chronic Conditions Data Warehouse. (2014). Chronic Condition Reference List. Retrieved from https://www.ccwdata.org/cs/groups/public/documents/document/ccw_conditionreferencelist2011.pdf

13.

Clark

D. O.

Stump

T. E.

Wolinsky

F. D.

(1997). A race- and gender-specific replication of five dimensions of functional limitation and disability. Journal of Aging and Health, 9, 28-42.

14.

Coxe

L. M.

Lennertz

McCullough

(2013). Changing prevalence of diabetes for Texas nursing home residents, 1999-2009. Journal of Applied Gerontology. doi:10.1177/0733464813512180

15.

Desai

M. M.

Lentzner

H. R.

Weeks

J. D.

(2001). Unmet need for personal assistance with activities of daily living among older adults. The Gerontologist, 41, 82-88. doi:10.1093/geront/41.1.82

16.

Durso

S. C.

(2006). Using clinical guidelines designed for older adults with diabetes mellitus and complex health status. Journal of the American Medical Association, 295, 1935-1940. doi:10.1001/jama.295.16.1935

17.

Fortinsky

R. H.

Garcia

R. I.

Joseph Sheehan

Madigan

E. A.

Tullai-McGuinness

(2003). Measuring disability in Medicare home care patients: Application of Rasch modeling to the outcome and assessment information set. Medical Care, 41, 601-615. doi:10.1097/01.mlr.0000062553.63745.7a

18.

Hebert

P. L.

Geiss

L. S.

Tierney

E. F.

Engelgau

M. M.

Yawn

B. P.

McBean

A. M.

(1999). Identifying persons with diabetes using Medicare claims data. American Journal of Medical Quality, 14, 270-277. doi:10.1177/10628606-9901400607

19.

Bentler

P. M.

(1999). Cutoff criteria for fit indexes in covariance structure analysis: Conventional criteria versus new alternatives. Structural Equation Modeling: A Multidisciplinary Journal, 6, 1-55. doi:10.1080/1070551-9909540118

20.

International Diabetes Federation. (2013). International diabetes federation managing older people with type 2 diabetes global guideline. Retrieved from http://www.idf.org/sites/default/files/IDF-Guideline-for-older-people-T2D.pdf

21.

Ivanova

J. I.

Mytelka

D. S.

Duh

M. S.

Birnbaum

H. G.

Cummings

A. K.

San Roman

A. M.

. . . Swindle

R. W.

(2013). Evaluating the utility of existing patient-reported outcome scales in novel patient populations with pancreatic cancer, lung cancer, and myeloproliferative neoplasms using Medicare current beneficiary survey data. Patient, 6, 189-200. doi:10.1007/s40271-013-0018-z

22.

Kalyani

R. R.

Saudek

C. D.

Brancati

F. L.

Selvin

(2010). Association of diabetes, comorbidities, and A1C with functional disability in older adults: Results from the National Health and Nutrition Examination Survey (NHANES), 1999-2006. Diabetes Care, 33, 1055-1060. doi:10.2337/dc09-1597

23.

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.

24.

Kempen

G. I.

Suurmeijer

T. P.

(1990). The development of a hierarchical polychotomous ADL-IADL scale for noninstitutionalized elders. Gerontologist, 30, 497-502. doi:10.1093/geront/30.4.497

25.

Kerr

E. A.

Heisler

Krein

S. L.

Kabeto

Langa

K. M.

Weir

Piette

J. D.

(2007). Beyond comorbidity counts: How do comorbidity type and severity influence diabetes patients’ treatment priorities and self-management? Journal of General Internal Medicine, 22, 1635-1640. doi:10.1007/s11606-007-0313-2

26.

Kline

R. B.

(Ed.). (2011). Principles and practice of structural equation modeling (3rd ed.). New York, NY: Guildford Press.

27.

LaPlante

M. P.

(2010). The classic measure of disability in activities of daily living is biased by age but an expanded IADL/ADL measure is not. Journals of Gerontology Series B: Psychological Science & Social Science, 65, 720-732. doi:10.1093/geronb/gbp129

28.

Muthén

B.O.

Muthén

(1998-2012). Mplus Version 7.1 [Software]. Los Angeles, CA: Muthén & Muthén.

29.

T. P.

Niti

Chiam

P. C.

Kua

E. H.

(2006). Physical and cognitive domains of the Instrumental Activities of Daily Living: Validation in a multiethnic population of Asian older adults. Journals of Gerontology Series A: Biological Sciences & Medical Sciences, 61, 726-735. doi:10.1093/gerona/61.7.726

30.

Piette

J. D.

Kerr

E. A.

(2006). The impact of comorbid chronic conditions on diabetes care. Diabetes Care, 29, 725-731. doi:10.2337/diacare.29.03.06.dc05-2078

31.

Quandt

S. A.

Reynolds

Chapman

Bell

R. A.

Grzywacz

J. G.

E. H.

. . . Arcury

T. A.

(2013). Older adults’ fears about diabetes using common sense models of disease to understand fear origins and implications for self-management. Journal of Applied Gerontology, 32, 783-803.

32.

Rosow

Breslau

(1966). A Guttman health scale for the aged. Journal of Gerontology, 21, 556-559. doi:10.1093/geronj/21.4.556

33.

Salas

Bubolz

Caro

J. J.

(2000). Impact of physical functioning of health status on hospitalizations, physician visits, and costs in diabetic patients. Archives of Medical Research, 31, 223-227. doi:10.1016/s0188-4409(00)00048-5

34.

Saliba

Elliott

Rubenstein

L. Z.

Solomon

D. H.

Young

R. T.

Kamberg

C. J.

. . . Wenger

N. S.

(2001). The Vulnerable Elders Survey: A tool for identifying vulnerable older people in the community. Journal of the American Geriatrics Society, 49, 1691-1699. doi:10.1046/j.1532-5415.2001.49281.x

35.

Schmitt

T. A.

(2011). Current methodological considerations in exploratory and confirmatory factor analysis. Journal of Psychoeducational Assessment, 29, 304-321. doi:10.1177/0734282911406653

36.

Sinclair

Morley

J. E.

Rodriguez-Mañas

Paolisso

Bayer

Zeyfang

. . . Lorig

(2012). Diabetes mellitus in older people: Position statement on behalf of the International Association of Gerontology and Geriatrics (IAGG), the European Diabetes Working Party for Older People (EDWPOP), and the International Task Force of Experts in Diabetes. Journal of the American Medical Directors Association, 13, 497-502. doi:10.1016/j.jamda.2012.04.012

37.

Spector

W. D.

Fleishman

J. A.

(1998). Combining activities of daily living with instrumental activities of daily living to measure functional disability. Journals of Gerontology Series B: Psychological Science & Social Science, 53, S46-S57. doi:10.1093/geronb/53b.1.s46

38.

Spector

W. D.

Katz

Murphy

J. B.

Fulton

J. P.

(1987). The hierarchical relationship between activities of daily living and instrumental activities of daily living. Journal of Chronic Diseases, 40, 481-489. doi:10.1016/0021-9681(87)90004-x

39.

StataCorp. (2011). Stata statistical software: Release 12. College Station, TX: Author.

40.

Sue Kirkman

Briscoe

V. J.

Clark

Florez

Haas

L. B.

Halter

J. B.

. . . Odegard

P. S.

(2012). Diabetes in older adults: A consensus report. Journal of the American Geriatrics Society, 60, 2342-2356. doi:10.1111/jgs.12035

41.

Tavakol

Dennick

(2011). Making sense of Cronbach’s alpha. International Journal of Medical Education, 2, 53-55.

42.

Thomas

V. S.

Rockwood

McDowell

(1998). Multidimensionality in instrumental and basic activities of daily living. Journal of Clinical Epidemiology, 51, 315-321. doi:10.1016/S0895-4356(97)00292-8

43.

Vogeli

Shields

A. E.

Lee

T. A.

Gibson

T. B.

Marder

W. D.

Weiss

K. B.

Blumenthal

(2007). Multiple chronic conditions: Prevalence, health consequences, and implications for quality, care management, and costs. Journal of General Internal Medicine, 22(Suppl. 3), 391-395. doi:10.1007/s11606-007-0322-1

44.

Volpato

Maraldi

Fellin

(2010). Type 2 diabetes and risk for functional decline and disability in older persons. Current Diabetes Reviews, 6, 134-143.

45.

Yang

Hall

A. G.

(2008). The financial burden of overweight and obesity among elderly Americans: The dynamics of weight, longevity, and health care cost. Health Services Research, 43, 849-868. doi:10.1111/j.1475-6773.2007.00801.x

46.

Yankeelov

P. A.

Faul

A. C.

D’Ambrosio

J. G.

Collins

W. L.

Gordon

(2013). “Another day in paradise”: A photovoice journey of rural older adults living with diabetes. Journal of Applied Gerontology. doi:10.1177/0733464813493136