| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
| ||||||||||||||||||||||||||||||||
Correspondence: Address correspondence to Joseph R. Sharkey, 1103 University Drive, Suite 203, College Station, TX 77840. E-mail: jrsharkey{at}srph.tamhsc.edu
 |
Abstract |
|---|
 TOP Abstract MethodsStatistical AnalysisResultsDiscussionReferences |
|---|
Key Words: Chronic care • Diabetes management • Dietary adherence • Food insufficiency • Home-delivered meals • Nutrition
Individuals with diabetes are expected to improve glycemic control through daily self-management of a complex regimen of multiple activities (e.g., glucose monitoring, appropriate food selection, physical activity, and medication use); thus, diabetes is one of the most psychologically and behaviorally demanding of all chronic medical conditions (Cox & Gonder-Frederick, 1992; Glasgow, Hampson, Strycker, & Ruggiero, 1997; Jack, Liburd, Spencer, & Airhihenbuwa, 2004). Poorly managed diabetes in later life may increase the risk for such diabetes-related complications as heart disease or stroke, diminished vision, high blood pressure, kidney disease, nervous system damage, or nontraumatic amputations (Conn et al., 2002).
Healthful eating (i.e., dietary regulation and meal planning) in accordance with dietary recommendations is a key factor of diabetes management and favorable clinical outcomes (American Diabetes Association [ADA], 2003; Nelson, Cunningham, Andersen, Harrison, & Gelberg, 2001; Wen, Shepherd, & Parchman, 2004). Dietary adherence requires the adoption and maintenance of major changes in eating behavior; these changes are influenced by psychosocial, physical, and environmental factors (Glasgow et al., 1997; Glasgow, Toobert, & Gillette, 2001; Jack et al., 2004; Murata et al., 2004). Prior research on overcoming barriers to dietary adherence has primarily focused on motivation, knowledge, and psychosocial characteristics (Hunt, Pugh, & Valenzuela, 1998; Murata et al.).
In addition, though, economic factors—such as adequacy of economic resources—likely influence food selection and the degree of adherence to dietary self-care behaviors (Hunt et al., 1998; Jack, Liburd, Vinicor, Brody, & Murry, 1999; Nelson et al., 2001; Sharkey, 2003; Wen et al., 2004). Prior studies have shown that diminished food sufficiency, which results in an inadequate amount of food intake (primarily due to insufficient economic resources), is associated with low food intakes of individual and multiple nutrients (Briefel & Woteki, 1992; Dixon, Winkleby, & Radimer, 2001; Lee & Frongillo, 2001; Nelson et al.; Rose, 1999; Sharkey; Sharkey et al., 2002). In this study, I used a conceptual model (Figure 1)—influence of individuals' economic context on healthful eating, diabetes self-management, and diabetes-related clinical outcomes—to categorize fundamental, intermediate, and proximate resource-related factors that may influence healthful eating and served as a guide for data analysis. This model suggests that out-of-pocket expenses likely contribute (either directly or indirectly) to the level of food sufficiency.
|
Although it is hypothesized that homebound older adults with diabetes are at greater risk for lower levels of food sufficiency than those without diabetes, our understanding of whether the extent and dynamics of food sufficiency differs over time between these two groups is limited. Information is especially lacking for older persons who rely on the traditional model of home-delivered meals for food assistance, namely, five nutritionally balanced meals each week (i.e., one per weekday), which each provide at least 33% of daily nutrient requirements.
My purpose in the present study is to determine whether diabetes status increases the likelihood for heightened food insufficiency. That is, I seek to determine whether such status increases the likelihood of diminished food sufficiency or persistent food insufficiency over 1 year, independent of baseline fundamental factors and intervening events (e.g., hospitalization or major illness) or change in economic or health-related factors (e.g., loss of drug coverage or food stamps or increased medication use).
|
Methods |
|---|
|
TOPAbstractMethodsStatistical AnalysisResultsDiscussionReferences |
|---|
Data Collection: Food-Sufficiency Status—A Measure of Resources
Interviewers asked homebound participants three questions that have been used previously to assess level of food sufficiency in the past 6 months: (a) Were there days when there was no food in the house and no money or food stamps for food?; (b) Were there days when you had to choose between buying food and buying medication?; and (c) Were there days when you had to choose between buying food and paying bills? I placed participants into one of three groups of diminished food sufficiency: (a) being food sufficient (FS), that is, responded no to all three questions; (b) at risk of becoming food insufficient (RFI), that is, responded yes to either Question 2 (food or medication) or Question 3 (food or bills) and no to Question 1 (no food); or (c) food insufficient (FI), that is, responded yes to Question 1 (no food; see Sharkey, 2003).
Baseline Fundamental Factors
Population Characteristics
Participants reported their gender, race (Black or non-Hispanic White), age, marital status, living arrangement (lived alone or lived with others), and education completed.
Economic Factors
This category included current monthly income, drug coverage (no drug insurance, government coverage, or supplemental drug coverage), and participation in food assistance programs (participation in the Food Stamp Program or other local food assistance programs such as a food pantry). Government drug coverage included Medicaid, indigent medication assistance, or Veterans Affairs pharmacy benefits.
Health-Related Factors
Interviewers asked participants if a physician had ever told them that they had specific health conditions, including arthritis, congestive heart failure, diabetes, heart disease or angina, high blood pressure, kidney disease, lung disease (e.g., asthma, emphysema, or chronic bronchitis), osteoporosis, or stroke. I constructed a summary score for coexisting disease conditions from a total of these nine health conditions. I defined a dichotomous comorbidity variable as having reported three or more of the nine disease conditions (compared with fewer than three disease conditions). I visually inspected and listed prescription-medication containers, and I divided the total number of unique medications into tertiles (i.e., zero to four, five to seven, and eight or more medications).
Intervening Events
I collected data during the 1-year in-home assessment related to exposure to intervening events that might have occurred since the initial in-home assessment: events related to consequences of prior health risk or change in economic or health-related factors. Consequences of prior health risk included occurrence of a major illness or overnight hospitalization (i.e., occurrence and number of times and days). Change in economic or health-related factors included loss of food stamps (i.e., receiving food stamps at baseline and not at 1 year), loss of drug coverage (i.e., having government or supplemental drug coverage at baseline and having neither at 1 year), or increased drug use (i.e., taking a greater number of unique prescription medications at 1 year than at baseline).
Proximate Factors at 1-Year Assessment
In addition to being asked the same food-sufficiency questions as at the baseline, participants were asked three questions that measured the perceived adequacy of economic resources (Mitchell, Mathews, Hunt, Cobb, & Watson, 2001): (a) Do you worry about having enough money in the future?; (b) Do you have trouble making ends meet?; and (c) Do you not have enough money for little extras? From the summary score (range = 0–3), I constructed a dichotomous inadequate economic resource variable: 0 = fewer than two inadequate economic resource items and 1 = two or three items.
|
Statistical Analysis |
|---|
|
TOPAbstractMethodsStatistical AnalysisResultsDiscussionReferences |
|---|
I estimated multivariate logistic regression models, using a three-category nominal dependent variable, to evaluate the relationship between baseline fundamental factors and increasingly severe levels of food insufficiency at baseline and 1 year. In the first model (food-sufficiency status at baseline), I simultaneously entered all baseline characteristics to adjust for sample characteristics. I used backward elimination strategy, which sequentially removes statistically nonsignificant variables, to obtain the "best" set of independent variables (Kleinbaum et al., 1998). In the second regression model (food-sufficiency status at 1 year), I added baseline food-sufficiency status to the independent variables in the first model to adjust for level of food sufficiency at baseline. I present regression results in terms of odds ratios for being RFI or being FI relative to being FS, along with their 95% confidence intervals.
Finally, I used two multivariate logistic regression models (backward elimination strategy) to examine the relationship of diabetes status, intervening events, and heightened food insufficiency status at 1 year. In Model 1, the dependent variable was diminished food sufficiency (DFS), which I defined by the negative change in level of food sufficiency between baseline and 1 year: from FS to RFI or FI, and from RFI to FI. In Model 2, the dependent variable was DFS/PFI—DFS at 1 year or PFI (persistent food insufficiency), which was food insufficiency at both baseline and 1 year.
|
Results |
|---|
|
TOPAbstractMethodsStatistical AnalysisResultsDiscussionReferences |
|---|
|
|
|
|
|
|
Discussion |
|---|
|
TOPAbstractMethodsStatistical AnalysisResultsDiscussionReferences |
|---|
In this study, not only did food-sufficiency status diminish over time for the entire sample, but this status became or remained worse for older adults with diabetes; this observation remained even after I took into consideration intervening events that may influence the adequacy of economic resources. More specifically, proportionately more homebound elders with diabetes (than without) experienced diminished food insufficiency (1-year negative change in level of food sufficiency) or persistent food insufficiency (food insufficiency at baseline and 1 year later). This observation is of special concern for older persons with diabetes, given that prior studies found that the risk or presence of food insufficiency was associated with diet and health; that is, lowest levels of nutrient intake from food, increased burden of multiple chronic conditions, increased risk for restricting prescription-medication use, or higher physician utilization were seen in these individuals (Nelson et al., 2001; Sharkey, 2003; Sharkey et al., 2005). Because increased out-of-pocket costs and poor dietary adherence may both lead to poorer clinical outcomes in individuals with diabetes (Karter et al., 2003; Murata et al., 2004), the economic-context model and the study results suggest that food-sufficiency status is likely linked to out-of-pocket costs (and economic conditions) and poor dietary adherence.
Because greater proportions of homebound elders with diabetes experienced a reduction in their level of food sufficiency between the first and second in-home assessment, this study utilized two strategies to evaluate whether this relationship was independent of other factors. First, I examined the association of baseline fundamental factors with food-sufficiency status 1 year later, after adjustment for baseline food-sufficiency status. With respect to 1-year level of food sufficiency, having diabetes progressively increased the odds for RFI or FI. This progression occurred despite continued receipt of home-delivered meals and may be a consequence of worsening personal financial conditions (e.g., greater out-of-pocket expenses) that are due to the shifting of health costs to individuals (e.g., copays), increased costs for basic needs (e.g., food, utilities, and transportation), or intervening and perhaps unexpected events (e.g., hospitalizations or loss of drug or food-stamp coverage). In the second approach, I considered intervening events and perceived inadequacy of economic resources—indicators of increasing demands on economic resources or worsening financial conditions—as determinants of heightened food insufficiency. I used two definitions of heightened food insufficiency: (a) DFS between baseline and 1 year, and (b) a composite (DFS/PFI) of either DFS or PSI (FI at both assessments). Regardless of definition, participants with diabetes were almost two times as likely to report heightened food insufficiency as persons without diabetes, independent of intervening events or inadequacy of economic resources. In addition, the perception of having inadequate economic resources increased the odds for both definitions of heightened food insufficiency, with a greater influence on a composite DFS/PFI than DFS alone. Because perceived inadequacy of economic resources was associated with having a low income, being married (e.g., health-related expenses of the spouse were not measured), and suffering from comorbidity (but not medication use or drug coverage), the results suggest that, in light of the conceptual model, this proximate factor may describe overall financial condition and help health care providers and others understand how fundamental factors interrelate and contribute to diminished food sufficiency or persistent food insufficiency over time.
These findings are especially important in light of certain contextual factors. First, all participants received the traditional home-delivered-meals service—one meal a day for each of the five weekdays. Thus, it is possible that these results underestimate the relationship between diabetes and food sufficiency among characteristically similar older adults who do not regularly receive home-delivered meals as a result of meals service unavailability (e.g., waiting lists, service limited to fewer meals, geographic location, or inadequate community resources, including financial or lack of volunteers; see Ponza et al., 1996). Second, the traditional "school-lunch" model of home-delivered-meals service may no longer be sufficient to prevent inadequate nutrient intakes, especially among individuals with diabetes. Homebound older adults, especially individuals who receive home-delivered meals, report a proportionally greater prevalence of diabetes and other nutrition-related chronic conditions than does the overall older adult population. Furthermore, these individuals rely on home-delivered meals for a greater proportion of daily intake to meet long-term nutritional needs (Ponza et al.; Sharkey et al., 2002; U.S. Department of Health and Human Services, 2004). Third, meal programs that assess nutritional risk and function seldom collect detailed information on diabetes status, medication use, or food-sufficiency status, because such data collection is not part of OAANP reporting requirements (Administration on Aging, 2000). Finally, a paradigm shift in diabetes care has occurred from a traditional approach that is reactive, fragmented, and focused on the individual to a new chronic care model, which incorporates community resources and policies to support self-management (Norris & Olson, 2004). Concomitantly, a paradigm shift within OAANP also has occurred that holds service providers to greater accountability to demonstrate effectiveness and document program outcomes that recognize that dietary modification is part of the disease prevention-and-management process—especially for diseases that are highly prevalent among older persons, such as diabetes.
Strengths and Limitations of the Study
The present study has several particular strengths. First, using a conceptual model of the influence of individuals' economic context, in this evaluation I examined the influence of intervening events, perceived adequacy of income, and diabetes status on declining food sufficiency over time among OAANP homebound participants; this is a methodology not typically utilized in this context. Second, using two waves of data from the North Carolina NAFS provided a breadth of clinical, nutritional, and social information often not found in food-insufficiency studies of older persons, such as detailed information on the burden of diseases, medication use from visual inspection, a three-level indicator of food-insufficiency severity that included forced-resource decisions, and intervening events. Finally, two multivariate regression approaches allowed for an examination of the association between diabetes status and food-sufficiency status at two time points and provided possible explanations for heightened food insufficiency over 1 year.
Several limitations to this study are acknowledged. Food-sufficiency data, a measure of resources, did not include information on duration and frequency; as a result, this study was unable to differentiate between acute and chronic risk for, or presence of, food insufficiency. This issue is a recurring concern with studies of food insufficiency in elderly persons (Nord, Andrews, & Winicki, 2002). Additional limitations included measurement of most study variables through self-report only, and absence of information on spouse, actual out-of-pocket costs for medications, and intervening events. Furthermore, lack of data on duration of diabetes, history of glycemic control, or diabetes-related complications must be acknowledged as study limitations.
Despite these limiting factors, this study furthers knowledge about the greater vulnerability of homebound older adults with diabetes for worsening food sufficiency over time. With the recent enactment of the Medicare Modernization Act, only the passage of time will tell whether prescription-drug coverage—assuming no increase in other out-of-pocket expenses—will lessen diminished or persistent food insufficiency. Regardless, the OAANP home-delivered-meal service is a critical component of community resources and provides food assistance for low-income, minority, and food-insufficient homebound older adults (Millen, Ohls, Ponza, & McCool, 2002; Ponza et al., 1996). For many older adults, the home-delivered meal may prevent more severe and chronic food insufficiency. However, despite these intentions of the OAANP, this study shows that food insufficiency remains highly prevalent and increases over time, especially for homebound older adults who are expected to make necessary diet choices in order to optimally manage their diabetes.
Implications of the Study
Considering that continual risk or presence of food insufficiency is an unfavorable environmental condition that may counter diabetes self-management, the findings of this study have several implications for managing the care of chronic illness, such as diabetes, among the growing homebound older population who are living longer in the community with a greater burden of chronic disease, increased income constraints, and need for long-term food assistance.
Of particular significance, this research has implications for providers and planners of health care and nutrition services to older adults with diabetes. Limited economic resources clearly influence choices for day-to-day diabetes care (Hunt et al., 1998). Additionally, diminished food sufficiency poses a significant barrier to utilization of published care guidelines for diabetes, because of the strong association of food insufficiency to poor diet quality (including lower consumption of fruits and vegetables and low intakes of essential nutrients; see Rose & Oliveira, 1997; Sharkey et al., 2002). Given these factors, health care providers should attempt to identify high-risk older adults—those who have diabetes and are at risk of food insufficiency—and should develop community linkages and strategies that integrate nutrition with a diabetes care plan. The goal is to support a multidisciplinary chronic care model to improve diabetes management and outcomes in the context in which older persons live (ADA, 2003). For providers and planners of nutrition services to older adults, this process will involve the following: (a) viewing OAANP activities as part of a chronic care model that supports improved disease management and helps older individuals remain independent and in their own homes; (b) increased funding and program targeting to improve unfavorable environmental conditions (i.e., persistent RFI or FI) that may counter diabetes self-management; (c) incorporation of multiple food-insufficiency items in assessments and reassessments that recognize issues related to absence of food, forced scarce-resource decisions, access and availability of food (e.g., transportation, grocery shopping, and meal preparation), and food choice (e.g., "right foods for health"; Wolfe et al., 2003); (d) monitor changes in food-sufficiency status and economic and health-related circumstances in program participants; and (e) target interventions that seek to ameliorate difficulty with continually securing healthy foods. Interventions may include referrals to appropriate programs for food, utility, pharmacy, housing, and transportation assistance. Other interventions may involve increasing meals to targeted, high-risk participants through the provision of more than one meal per day, more than 5 days per week, and improvement in meal nutrient density through the use of fortified food products. Nutrition education should teach low-cost dietary alternatives and provide assistance in developing strategies for controlling food costs within a healthy diet (Hunt et al.). Finally, policy makers of health and social service programs servicing older adults should recognize the challenge to individual, family, and community resources for diabetes support among particularly vulnerable groups. Specifically, the issues related to homebound older adults who are in a minority group, economically disadvantaged, or lack food sufficiency should be addressed. Such efforts should include consideration of food-sufficiency data as decisions are made regarding funding for home- and community-based services and food and nutrition programs. Ultimately, it may be more cost effective to ensure food sufficiency—and hence reduce disease progression and disease-related complications—in this population than to provide extra medical service alone (Nelson, Brown, & Lurie, 1998).
|
Footnotes |
|---|
I acknowledge the hard work and dedication to this project by our North Carolina community partners who actively participated in the study as in-home coordinators: Amy Walls, RN and the Chatham County Council on Aging; Turquious Byrd and Senior Resources/Mobile Meals of Guilford County; Minnie McBurnett and the Johnston County Council on Aging; and Viki Baker and Meals on Wheels of Wake County, Inc. I also acknowledge the valuable suggestions of Dr. Marcia Ory and Dr. Barry Browne on earlier drafts of this article.
1 Department of Social and Behavioral Health, Texas Healthy Aging Research Network (TxHAN) Center, and Aging and Health Promotion Program, School of Rural Public Health, Texas A&M Health Science Center, College Station.
Decision Editor: Linda S. Noelker, PhD
Received for publication January 20, 2005. Accepted for publication June 6, 2005.
|
References |
|---|
|
TOPAbstractMethodsStatistical AnalysisResultsDiscussionReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  R. L. Ettinger Letter to the Editor. Gerontologist, June 1, 2006; 46(3): 317 - 317. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||
| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
