1. Introduction
The recent advances in early detection and treatment of cancer have led to a significant reduction in poor cancer-specific outcomes [
1]. Thus, more women are living cancer free with five-year relative survival rates for breast cancer exceeding 95% [
2]. However, the survival rates are not distributed uniformly among all cancer survivors, with the rates among African American women trailing behind those of other racial and ethnic groups significantly [
2]. While poor screening habits and socio-demographic characteristics predispose minority women to lower survival rates [
1], lifestyle characteristics (i.e., obesity, physical inactivity), and deficits in health-related quality of life may contribute equally to disparities in survival [
3,
4,
5,
6,
7].
Health-related quality of life (HRQOL) consists mainly of mental and physical health domains, which are important determinants of cancer survival. According to the Medical Outcomes Study 36-Item Short-Form Health Survey (SF-36) [
8], physical health is operationalized as physical functioning, bodily pain, fatigue (or vitality), and physical role functioning (SF-36). Similarly, mental health is operationalized as social roles, mental health, emotional role functioning, and general health perceptions (SF-36). Of the two domains, physical health, and specifically physical functioning, have been shown to be related to cancer survival [
9,
10]. Physical functioning includes limitations in mobility, decreases in strength, and limited range of motion in limbs [
11,
12]. The study of physical functioning is critical not only because it is an important indicator of cancer survival [
10], but cancer survivors are also twice as likely to have deficits in functional status when compared to age-matched controls [
13]. Of importance, studies have indicated that physical function scores are significantly lower among African American cancer survivors when compared to survivors of other ethnic groups [
14].
Physical function can be measured using performance-based or self-report measures [
15,
16]. Short physical performance batteries have been shown to have high reliability and are predictive of disability, institutionalization, and mortality [
17,
18,
19]. However, the objective assessment of physical function is not always feasible and practical as a survey and is more costly [
20]. It should also be noted that self-report measures, despite their ease of implementation, are associated with limitations as well. The limitations of self-report instruments are their lack of sensitivity to change, concerns over reproducibility, and inability to capture a broad range of functioning [
20]. Moreover, several studies have shown moderate to poor associations between performance-based measures and self-reported functional status. To address the limitations of existing instruments used to assess physical function, the Late-Life Function and Disability (LLFD) Instrument was developed [
21,
22]. The LLFD instrument is comprised of two components. The functional component assesses advanced lower extremity function (e.g., running 1/2 mile), basic lower extremity function (e.g., getting in and out of a car), and upper extremity function (e.g., holding a full glass of water in one hand) [
21]. The disability component assesses the frequency of performing social (e.g., visiting friends or family in their home) and personal (e.g., taking care of household business) role activities and the limitation in the capability of performing instrumental and management role activities [
22]. The LLFD instrument is a comprehensive measure of functional status. The advanced lower extremity function subscale represents items requiring reasonable effort, whereas basic lower extremity function represents activities of daily living that are essential to normal functioning [
21,
22]. Upper body functioning is of relevance to cancer survivors because approximately one-third of the breast cancer survivors suffer from upper extremity disability that could progress to chronic arm disability [
23]. More recently, an abbreviated version of the LLFD instrument was proposed as a more parsimonious model and found to have adequate construct, convergent, and predictive validity in an ethnically diverse sample of older adults [
24].
To date, the LLFD instrument and its abbreviated version have been assessed only in non-cancer samples. We do not know of any studies that have assessed the construct validity of this instrument in adult samples with chronic health conditions. Minority breast cancer survivors experience a greater burden of functional limitations and disability due to poor lifestyle characteristics and late stage of presentation when compared to non-Hispanic white breast cancer survivors. Thus, our sample of underserved African American breast cancer survivors represents a unique sample to examine the validity of this instrument. Therefore, the purpose of this study was to examine the construct validity of the abbreviated LLFD instrument in a sample of African-American breast cancer survivors and examine its association with sociodemographic and lifestyle characteristics.
4. Discussion
The primary purpose of this study was to examine the construct validity of the abbreviated LLFD instrument in a sample of African-American breast cancer survivors. We found that the LLFD instrument fit the data with minor modifications. These data support prior studies examining the LLFD instrument in an older adult population and potentially extends the validity of the instrument to cancer survivors. In addition, we observed that education and number of comorbid conditions were associated significantly with all measures of function. Overall, these data demonstrate adequate construct validity of the LLFD instrument in a vulnerable population of cancer survivors but also sheds some light on sub-populations within African American breast cancer survivors who deserve special attention for interventions to remedy functional limitations.
Minor modifications were made to improve the fit of the functional component of the LLFD instrument. In particular, we added a residual variance correlation among two items representing upper extremity functioning (i.e., Item #3: using common utensils for preparing meals; Item #4 holding a full glass of water in one hand). These items have similar conceptual meaning as both of them relate to working with one’s hands. Despite the minor modifications, the functional subscales demonstrated simple structure with high factor loadings indicative of strong construct validity. Thus, we feel that our results support those observed in prior studies using this abbreviated instrument [
24]. The brevity of functional components and its demonstrated reliability and validity is of importance given prior studies indicating that functional status was associated with cancer-specific and overall survival [
10,
30]. Importantly, this instrument has relevance for cancer survivors due to the variability in the scores and low frequency of floor and ceiling effects. Prior studies have demonstrated that the LLFD instrument was sensitive to change and may serve as a self-report outcome of functional status in intervention studies [
20].
Our results also confirm the construct validity of the social and personal components of the disability instrument. The instrument had adequate factor loadings and was correlated significantly with the functional components, similar to a prior study [
24]. In addition, there a sizable portion (>50%) of the sample who reported scores at the lowest end indicating a large floor effect. This may also suggest that most of the survivors participating here had reasonable levels of function and therefore few disabilities. Of note, the sociodemographic, medical, and lifestyle characteristics were not associated with the disability subscales. This could be due to the lack of variance in the responses as a sizable portion of women reporting the lowest possible scores on the subscales. More research is needed to determine how well these items correlate with objective measures of function and disability in survivor populations.
Further evidence to support the construct validity of the abbreviated LLFD instrument were the associations we observed between the sociodemographic, medical, and lifestyle characteristics with the functional components. In particular, less education was associated with worse upper and lower extremity function. The associations we observed among our survivors are similar to those found in prior studies because education is a proxy for socioeconomic status and thus indicates earning power or lack thereof [
31,
32,
33]. Becker et al. [
34] observed that comorbidities were associated with not only physical and functional health but also social and emotional health as well. The associations we observed here may point to high-risk groups that can be targeted for interventions. African American breast cancer survivors who are less educated may be more likely to have co-occurring conditions such as high blood pressure and diabetes and may be more vulnerable to functional limitations due to their disease state.
The results from this study provide noteworthy and unique information about a vulnerable population of cancer survivors. Strengths of our study include a modest sample size of underrepresented cancer survivors as well as the use of an instrument not commonly examined in the field of cancer survivorship. Despite the strengths, several limitations should be noted. Our data are cross-sectional and therefore can not imply causality. Our sample size is modest and consists of mostly affluent African American breast cancer survivors, limiting the generalizability of our study. In addition, our data on functional limitations are self-reported and subject to reporting biases. Further studies will be needed to compare the abbreviated LLFD instrument in cancer survivors to objective measures of functional status.