The Journals of Gerontology Series A: Biological Sciences and Medical Sciences 63:67-75 (2008)
© 2008 The Gerontological Society of America
Unintentional Weight Loss Predicts Decline in Activities of Daily Living Function and Life-Space Mobility Over 4 Years Among Community-Dwelling Older Adults
Christine S. Ritchie,
Julie L. Locher,
David L. Roth,
Theresa McVie,
Patricia Sawyer and
Richard Allman
Departments of 1 Medicine (Division of Gerontology, Geriatrics, and Palliative Care),2
Health Care Organization and Policy, and 3 Sociology, 4 Center for Aging, and 5 Department of Biostatistics (School of Public Health), University of Alabama at Birmingham.
6 Geriatric Research Education and Clinical Center at the Birmingham Veterans Affairs Medical Center, Alabama.
Address correspondence to Christine S. Ritchie, MD, MSPH, UAB School of Medicine, 1530 3rd Ave. South, Ch-19 Rm. 219, Birmingham, AL 35294. E-mail: critchie{at}uab.edu
 |
Abstract
|
|---|
Background. The relationship between body mass index (BMI), weight loss, and changes in activities of daily living (ADL) function and mobility in older adults is not clear. We sought to study the relationship between BMI and weight loss on the rate of decline in ADL function and life-space mobility over a 4-year period among older African Americans and whites.
Methods. The participants were 983 enrollees in the University of Alabama at Birmingham (UAB) Study of Aging, a longitudinal study of mobility among community-dwelling older adults stratified to achieve a balanced sample in terms of sex, race, and residence. Primary outcome measures were changes in ADL function and mobility assessed by the UAB Study of Aging Life-Space Assessment (LSA) which were measured every 6 months.
Results. Relative to normal weight participants, those with BMI levels in the obese range did not show more rapid ADL functional decline, but a history of unintentional weight loss predicted more rapid decline. Relative to normal-weight participants, other BMI categories were not associated with more rapid decline in LSA scores. However, unintentional weight loss predicted more rapid declines in LSA. Intentional weight loss had no relation to ADL function or LSA decline.
Conclusions. In this population of community-dwelling older African Americans and whites, neither BMI nor intentional weight loss had an association with rate of functional decline. Unintentional weight loss had a negative relation with rate of functional decline, regardless of baseline BMI. Whether this is causal remains to be determined.
Key Words: Weight loss • Body composition • Function • Mobility
THE relationship between body mass index (BMI), weight loss, and functional change in older adults is not clear. Low BMI and obesity have been shown to be associated with lower mobility and function (1–3). Many studies suggest that weight loss may be predictive of limitations in activities of daily living (ADL) function (1,3) whereas others suggest that weight loss protects against functional decline (4). Reynolds and colleagues (5) suggested that, although obesity has little effect on life expectancy in adults 70 years old and older, obese older adults live a higher proportion of their remaining lives disabled. Launer and colleagues (6) evaluated older women in the Epidemiologic Follow-Up Study of the National Health and Nutrition Examination Survey NHANES I (1971–1987) and found that women with high past or baseline BMI were at increased risk for reporting disability.
What is not clear from these studies is whether older adults who are obese and show restricted function or ability to navigate their environment continue at the same rate of decline with aging on these variables as do less obese older adults, or whether rates of decline are more pronounced in obese older adults relative to normal weight or underweight older adults. An additional issue that remains unresolved is how weight loss, intentional or unintentional, affects functional change. We sought to identify the impact of BMI and weight loss on the pattern of change in ADLs and life-space mobility over a 4-year period among older African Americans and whites.
|
METHODS
|
|---|
Sample
The participants were 983 enrollees for whom BMI could be assessed in the University of Alabama at Birmingham (UAB) Study of Aging, a population-based, longitudinal study of mobility among community-dwelling older adults. Participants were recruited from a stratified random sample of Medicare beneficiaries aged 65 years or older living in five counties of central Alabama. Two counties were classified as urban and three counties as rural. This study oversampled African Americans, men, and rural residents to provide a balanced sample in terms of race, gender, and urban–rural residence. Persons in nursing homes, persons who were unable to set their own appointments, and persons from whom height and weight could not be obtained (n = 17) were excluded. For interested participants, an in-home interview was scheduled. Prior to the interview, written informed consent was obtained. The study protocol was reviewed and approved by the UAB Institutional Review Board.
Design
In-home interviews were conducted by trained interviewers and are described in detail elsewhere (7). They included a number of measures, including sociodemographic factors, weight and height, Life-Space Assessment (LSA), and self- or proxy-reported ADLs. Telephone interviews were subsequently conducted every 6 months for 48 months.
Measurements
Dependent variables.--
We chose two measures of function to evaluate in these analyses. First, we chose a composite measure of ADL to capture an individual's perception of difficulty in performing specific tasks necessary for independence in a community setting. Second, we chose the LSA to evaluate what participants report that they did in the 4 weeks before the assessment, using the assistive tools and supports they have available to them. The LSA provides a broad functional perspective of individuals' response to their current conditions, including how the participant's nutritional status, or the conditions associated with that status, may be interfering with engagement in the community or with mobility in the home.
Difficulty with ADLs.--
Difficulties with ADLs were measured by self-report. The six ADL items included eating, using the toilet, dressing, transferring, bathing, and walking. For each item, participants were asked: "Do you have any difficulty performing the task?" If the answer was "no," a score of zero was assigned. If the answer was "yes," the participants were asked to rate, using a Likert-type scale, the level of difficulty for the task. Responses were scored as: 1 = some, 2 = a lot, 3 = unable to do the task. A composite score for ADL difficulty was calculated using the sum of scores for the individual tasks (range: 0–18). The sum was reversed such that higher scores indicated less reported difficulty with the functional tasks. ADL difficulty was measured at baseline and at 6, 12, 24, 30, 36, 42, and 48 months after baseline over the 4 years of follow-up.
LSA.--
The UAB LSA was used to identify the areas through which a person reported moving during the 4 weeks prior to the assessment. It reflects the distance that persons travel from the room where one sleeps, the frequency of such movement, and whether equipment or assistance from another person is used. The LSA was administered at baseline and every 6 months over the 4 years of follow-up. Specific questions were: "During the past 4 weeks, have you (i) been to other rooms of your home besides the room where you sleep (level 1); (ii) been to an area immediately outside your home such as your porch, deck or patio, hallway of an apartment building, or garage (level 2); (iii) been to places in your immediate neighborhood, but beyond your own property or apartment building (level 3); (iv) been to places outside your immediate neighborhood but within your town (level 4); and (v) been to places outside your immediate town (level 5)?" For each life-space level, participants were asked how often they traveled to that area (less than once a week; 1–3 times each week; 4–6 times each week; daily) and if they had assistance from another person or from an assistive device (yes vs no). The LSA was scored by assigning a value to each of the five levels, and then summing the five scores. The level scores were obtained by multiplying the level number (1–5) by a value for independence (2 = no assistance, 1.5 = use of equipment only, 1 = use of another person) times a value for frequency of movement (1 = less than once a week, 2 = 1–3 times each week, 3 = 4–6 times each week, and 4 = daily). LSA scores ranged from 0 (mobility confined to one's bedroom) to 120 (traveled out of town every day without assistance from another person or an assistive device).
The test–retest reliability of LSA scores between the in-home, face-to-face interview and a telephone interview within 2 weeks of the baseline assessment is 0.96 (8). LSA is also highly correlated with observed physical performance (8).
Independent Variables
BMI.--
Height without shoes was measured to the nearest half inch using a folding ruler. Participants were measured standing on an uncarpeted surface against a wall if possible. Weight to the nearest pound (without shoes) was obtained using a calibrated floor scale for participants able to stand on scales. For participants unable to stand, height and weight were calculated from the knee-height measure and arm circumference (N = 89). If the knee-height measure was unavailable, self-reported height and weight were used for the calculation (N = 37). Among 180 participants who both provided self-reported weight and were weighed, the correlation between self-reported weight and measured weight was 0.98.
BMI was calculated from weight in kilograms divided by meters squared and was categorized according to the National Heart, Lung, and Blood Institute (NHLBI) Clinical Guidelines thresholds into underweight (BMI < 18.5), normal weight (BMI 18.5–24.9), overweight (25–29.9), obese (BMI 
30 and < 35), and very obese (BMI 35) categories (9).
Weight loss.--
Weight loss was assessed at the baseline interview by asking, "In the past year, have you lost weight (> 10 pounds)?" If the participants answered "yes" to this question, they were then asked "Did you try to lose weight?" These questions were then coded into the following categories: no weight loss, intentional weight loss, and unintentional weight loss.
Covariates.--
In addition to gender, race, and age, covariates included in the model included smoking status (whether the participant smoked in the past year), physical activity, education, and comorbidity score. Physical activity was measured using the activity assessment of the Cardiovascular Health Study. This activity assessment was based on a modified Minnesota Leisure Time Activity Questionnaire and included questions about the frequency and duration of 15 different types of activities over the previous 2 weeks. Activities were assigned metabolic equivalents according to intensity, and leisure-time energy expenditure (kcal/wk) was calculated for each person (10). Education, a marker of socioeconomic status, was divided into year categories: 1 
6; 2 = 7–11; 3 = 12; 4 13. Comorbidity score was created by summing the total number of verified comorbidities that make up the Charlson Comorbidity Index (11). Diseases were considered verified if the participants were taking a medication for the disease, if their primary physicians reported that they had the condition, or if the condition was documented on a hospital discharge within the 3 years before entry into the study.
Statistical Analysis
Descriptive statistics were performed first to characterize the sample (Table 1). Dummy-coded vectors were created for the BMI categories at baseline, and these vectors were entered as a set along with the covariates in all statistical models. Next, analysis of covariance was used to estimate ADL difficulty scores and life-space composite scores as a function of BMI group for cross-sectional analyses at baseline. Adjustments were made for the effects of age, gender, race, urban versus rural residence, smoking status, recent weight loss, and comorbidity score. If a significant overall effect was found for BMI group, pairwise comparisons were conducted between BMI groups on the covariate-adjusted means of the dependent variable. Dummy-coded vectors were created for the BMI categories at baseline, and these vectors were entered as a set along with the covariates in all statistical models.
Longitudinal changes in ADL difficulty scores and life-space composite scores were examined using multilevel growth curve models as estimated in SAS Proc Mixed (SAS, Cary, NC) (12,13). These models were used to fit trajectories of change beginning 6 months after baseline, with the baseline observation of each dependent variable serving as a covariate. BMI group and weight loss served as primary predictor variables, with age, gender, and race also serving as covariates. Smoking status was not included in these models because it had no unique predictive effect in the baseline cross-sectional models. All participants with any follow-up data (N = 958) were included in these longitudinal models because the models are able to estimate intercepts and slopes for each participant based on the amount of data collected, even for participants who died or dropped out prior to the end of the 4-year follow-up period. For all covariates, mean-centered predictors were used. These were derived by subtracting the mean at baseline from each participant's observed score. This method improves the interpretability of the findings and the calculation of model-predicted scores because a zero score on any predictor now represents the mean on that predictor.
Both linear and curvilinear (i.e., quadratic) growth trajectories were estimated, and the fit of these trajectories to the observed data was examined using deviance statistics and the Akaike Information Criterion. In general, the quadratic models did not fit significantly better than the more parsimonious linear models, so only the results of the linear models are reported here.
The time variable in the longitudinal models was scaled to represent the time prior to the 4-year follow-up assessment. That is, time was scaled as –3.5 for the 6-month assessment, –3.0 for the 12-month assessment, and so on, up to 0 for the 4-year assessment. This allowed the main effect tests to consist of group comparisons at the 4-year point, and the Group x Time interaction effects to represent tests of the differences in linear slopes across time between the groups.
|
RESULTS
|
|---|
Baseline sample characteristics are displayed in Tables 1 and 2 for the 983 participants who had valid BMI data. Mean age for the cohort was 75.30 (± 6.72) years. The average ADL difficulty score was 16.49 (± 2.47). The average LSA score was 64.5 (± 24.5), which is representative of a score for someone going out of the house daily and into the neighborhood 4–6 times per week. The majority of participants were overweight (38%), with 31% obese or very obese and only 2% underweight. Participants in the overweight or obese categories were more likely to be younger (mean age = 73 years in the highest BMI group compared to mean age = 76 years in the lowest BMI group). Women and African American participants were more likely to be in the very obese (BMI 35) group. Smokers, those who had experienced recent weight loss, and those with higher comorbidity scores were more likely to be in the lowest BMI category.
In cross-sectional analyses of baseline functional measures adjusted for race, gender, smoking status, age, weight loss, physical activity, education, and comorbidity score (Table 3), participants with very low and very high BMI (35) were more likely to have low ADL and LSA scores, consistent with functional limitations. The individual differences in LSA by BMI category were not significantly different; however, for ADL status, individuals who were obese or very obese were significantly more likely to report difficulty than those with normal BMI. Unintentional weight loss was negatively associated with both LSA and ADL status (p <.0001). Intentional weight loss was not associated with either measure of functional status (Table 4). We evaluated to see whether there was an interaction between baseline BMI and weight loss; no interaction was found.
Linear longitudinal growth curve modeling was then conducted on the repeated observations of LSA and ADL scores after baseline, with the baseline observation serving as an additional covariate, to see if baseline BMI predicted change in life space or ADL. Race, gender, age, physical activity, education, intentional and unintentional weight loss, and comorbidity score were included in the analysis.
Relative to normal weight participants, those with BMI in the overweight or obese range did not experience ADL decline more rapidly (Table 5). These relationships did not change when persons in the very obese category were combined with persons in the obese category. ADLs in participants with low BMI appeared to decline the most rapidly, but this change did not reach statistical significance (Figure 1).
Relative to normal weight participants, other BMI categories did not show an accelerated decline in LSA scores (Table 6 and Figure 3). However, a history of unintentional weight loss did accelerate decline in LSA. Intentional weight loss had no impact on LSA decline. For both ADL score and LSA score, baseline history of unintentional weight loss predicted more rapid decline (Figures 2 and 4).
|
DISCUSSION
|
|---|
This study suggests that functional decrements associated with high and low BMI reflect differences already present at baseline that are not accelerated by virtue of BMI alone. Indeed, in this study of older participants, being in the overweight category (BMI 25–29.9) may have conferred some protection from functional decline relative to participants in the normal weight category. Baseline history of unintentional weight loss increased risk for ADL and mobility decline regardless of baseline BMI. Intentional weight loss did not contribute to declines in functional status, nor did it accrue any benefit over this 4-year time period. To make sure that there was not a differential effect of weight loss among participants who were heaviest at baseline, the data were evaluated for an interaction between baseline BMI and weight loss. No interaction was found.
Although the mechanisms for these relationships are not entirely clear, a number of theories have been postulated. First, unintentional weight loss may represent an underlying inflammatory process, disease, or frailty that in turn may have a negative effect on functioning (14). Indeed, a number of studies suggest that unintentional weight loss is often secondary to depression, occult malignancy, or metabolic or neurological conditions (15,16). Second, the finding that intentional weight loss was not protective against functional decline may be because of the possibility that individuals seeking to lose weight may attribute their weight loss to intentional behavioral changes, but may actually be losing weight due to subclinical disease (17). Weight loss of any kind may not be beneficial in older adults if it is accompanied by loss of lean body mass (18). Loss of lean body mass may in fact, accelerate functional decline. Finally, increased body fat may become more protective against functional decline because risks for decline change with aging. Short-term risks of falls (and possible hip fractures), which are mitigated by increased body fat, may become more important for short-term functional decline than longer term risk of cardiovascular endpoints associated with increased body fat (19). Similar to other cross-sectional studies, the cross-sectional analyses of this cohort showed an association between high BMI and functional limitations (20). Davison and colleagues (21) demonstrated this relationship among NHANES III participants 70 years old and older, particularly among women. Similarly, in the cross-sectional study by Coakley and colleagues of middle aged and older women in the Nurses Health Study, increasing levels of BMI were associated with increased functional limitations (22).
Longitudinal studies have also reported the impact of high and low BMI on functional status, although these data are less consistent. In the NHANES I Epidemiologic Follow-Up Study, using logistic regression analysis, high or low BMI contributed significantly to functional impairment (23). In Sarkisian's study of 6632 community-dwelling older women, a BMI of > 29 was found to be predictive of self-reported decline in ADLs over a 4-year period (2).
To our knowledge, our study is the first study to perform linear longitudinal growth curve modeling analyses, which allows one to evaluate the effect of BMI and weight loss on rate of functional change. Our findings suggest that baseline BMI alone does not affect the rate of functional change; rather it is the case that other factors (namely, unintentional weight loss, moderating the relationship between BMI and function) have a more influential effect on rate of functional decline.
As in other studies of older community dwelling adults, our study did not find weight loss to be protective against developing mobility limitation for any BMI group and, in fact, showed a consistently adverse effect on function (1,24,25). The lack of benefit of intentional weight loss among participants who were overweight or obese was in contrast to the study by Fine and colleagues (26), in which weight loss was associated with improved physical functioning among participants who were obese. However, the small number of participants with intentional weight loss in our cohort may limit conclusions related to voluntary weight loss and functional change. This study also suggests that, as concluded from other studies, a BMI in the overweight range does not appear to have the same negative functional consequences as it does in younger adults.
Limitations of this study include a relatively small sample size, especially among the lowest BMI category, and having only BMI available as a measure of body composition. Age-related changes in body composition (decrease in lean mass and increase in fat mass) and loss of height alter the relationship between BMI and percent body fat. At any given BMI, changes in body composition with aging underestimate body fatness, and changes in height overestimate body fatness. However, because BMI is readily available clinically, it is useful to understand that it has a relatively low effect on rates of functional decline compared to weight loss. An additional limitation was that outcome measures for function in this study were provided by self- or proxy report, and thus were more likely to be influenced by social, psychological, and health factors (27). Nevertheless, because self-reported functional measures are predictive of subsequent disablement and mortality, these relationships with nutritional factors remain important (28). This is one of the few studies to examine the role of BMI in affecting rates of functional decline and the only study that we know of that uses growth curve modeling to evaluate rate of functional change. Although baseline high and low BMI are associated with poor ADL function and life-space mobility, extremes in BMI seem to have less impact on rates of functional decline compared to unintentional weight loss. Unintentional weight loss, then, warrants close attention clinically, both as a potentially remediable risk factor and as a prognostic factor for decline when attempts to reverse it are unsuccessful. More studies are needed to evaluate which interventions are most successful in preventing or reversing unintentional weight loss in this population.
|
Acknowledgments
|
|---|
This work was funded by National Institutes of Health Grant R01-AG-15062.
|
Footnotes
|
|---|
Decision Editor: Luigi Ferrucci, MD, PhD
Received October 31, 2006
Accepted April 12, 2007
|
References
|
|---|
- Bannerman E, Miller MD, Daniels LA, et al. Anthropometric indices predict physical function and mobility in older Australians: the Australian Longitudinal Study of Ageing. Public Health Nutr. 2002;5:655-662.[Medline]
- Sarkisian CA, Liu H, Gutierrez PR, Seeley DG, Cummings SR, Mangione CM. Modifiable risk factors predict functional decline among older women: a prospectively validated clinical prediction tool. The Study of Osteoporotic Fractures Research Group. J Am Geriatr Soc. 2000;48:170-178.[Medline]
- Stuck QE, Walthert JM, Nikolaus T, Buula CJ, Hohmann C, Beck JC. Risk factors for functional decline in community-living elderly people: a systematic literature review. Soc Sci Med. 1999;48:445-469.[Medline]
- Hebert R, Brayne C, Spiegelhalter D. Factors associated with functional decline and improvement in a very elderly community-dwelling population. Am J Epidemiol. 1999;150:501-510.[Abstract/Free Full Text]
- Reynolds SL, Saito Y, Grimmins EM. The impact of obesity on active life expectancy in older American men and women. Gerontologist. 2005;45:438-444.[Abstract/Free Full Text]
- Launer LJ, Harris T, Rumpel C. Body mass index, weight change, and risk of mobility disability in middle-aged and older women. JAMA. 1994;271:1093-1098.[Abstract/Free Full Text]
- Peel C, Sawyer Baker P, Roth DL, Brown CJ, Bodner EV, Allman RM. Assessing mobility in older adults: the UAB Study of Aging Life-Space Assessment. Phys Ther. 2005;85:1008-1119.[Abstract/Free Full Text]
- Baker PS, Bodner EV, Allman RM. Measuring life-space mobility in community-dwelling older adults. J Am Geriatr Soc. 2003;51:1610-1614.[Medline]
- NHLBI. Clinical Guidelines on the Identification, Evaluation and Treatment of Overweight and Obesity in Adults-the Evidence Report. Available at: http://www.nhlbi.nih.gov/guidelines/obesity/prctgd_b.pdf. Accessed August 2007.
- Siscovick DS, Fried L, Mittelmark M, Rutan G, Bild D, O'Leary DH. Exercise intensity and subclinical cardiovascular disease in the elderly. The Cardiovascular Health Study. Am J Epidemiol. 1997;145:977-986.[Abstract/Free Full Text]
- Charlson ME, Pompei P, Ales KL, MacKenzie CR. A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chronic Dis. 1987;40:373-383.[Medline]
- Littell RC, Milliken GA, Stroup WW, Wolfinger RD. SAS System for Mixed Models. Cary, NC: SAS Institute; 1996.
- Singer JD, Willett JB. Applied Longitudinal Data Analysis: Modeling Change and Event Occurrence. New York: Oxford University Press; 2003.
- Martin CT, Kayser-Jones J, Stotts N, Porter C, Froelicher ES. Nutritional risk and low weight in community-living older adults: a review of the literature (1995–2005). J Gerontol A Biol Sci Med Sci. 2006;61A:927-934.[Abstract/Free Full Text]
- Thompson MP, Morris LK. Unexplained weight loss in the ambulatory elderly. J Am Geriatr Soc. 1991;39:497-500.[Medline]
- Lankisch P, Gerzmann M, Gerzmann JF, Lehnick D. Unintentional weight loss: diagnosis and prognosis. The first prospective follow-up study from a secondary referral centre. J Intern Med. 2001;249:41-46.[Medline]
- Lee JS, Kritchevsky SB, Tylavsky FA, et al. Weight-loss intention in the well-functioning, community-dwelling elderly: associations with diet quality, physical activity, and weight change. Am J Clin Nutr. 2004;80:466-474.[Abstract/Free Full Text]
- Kyle U, Zhang F, Morabia A, Pichard C. Longitudinal study of body composition changes associated with weight change and physical activity. Nutrition. 2006;22:1103-1111.[Medline]
- Margolis KL, Ensrud KE, Schreiner PJ, Tabor HK. Body size and risk for clinical fractures in older women. Study of Osteoporotic Fractures Research Group. Ann Intern Med. 2000;133:123-127.[Abstract/Free Full Text]
- Jensen GL, Friedman JM. Obesity is associated with functional decline in community-dwelling rural older persons. J Am Geriatr Soc. 2002;50:918-923.[Medline]
- Davison KK, Ford ES, Cogswell ME, Dietz WH. Percentage of body fat and body mass index are associated with mobility limitations in people aged 70 and older from NHANES III. J Am Geriatr Soc. 2002;50:1802-1809.[Medline]
- Coakley EH, Kawachi I, Manson JE, Speizer FE, Willet WE, Colditz GA. Lower levels of physical functioning are associated with higher body weight among middle-aged and older women. Int J Obes. 1998;22:958-965.[Medline]
- Galanos AN, Pieper CF, Coroni-Huntley JC. Nutrition and function: is there a relationship between body mass index and the functional capabilities of community-dwelling elderly? J Am Geriatr Soc. 1994;42:368-373.[Medline]
- Lee JS, Kritchevsky SB, Tylavsky F, et al. Weight change, weight change intention, and the incidence of mobility limitation in well-functioning community-dwelling older adults. J Gerontol A Biol Sci Med Sci. 2005;60A:1007-1012.[Abstract/Free Full Text]
- Tully CL, Snowdon DA. Weight change and physical function in older women: findings from the Nun Study. J Am Geriatr Soc. 1995;43:1394-1397.[Medline]
- Fine JT, Colditz GA, Coakley EH. A prospective study of weight change and health-related quality of life in women. JAMA. 1999;282:2136-2142.[Abstract/Free Full Text]
- Daltroy LH, Larson MG, Eaton HM, Phillips CB, Liang MH. Discrepancies between self-reported and observed physical function in the elderly: the influence of response shift and other factors. Soc Sci Med. 1999;48:1549-1561.[Medline]
- Reuben DB, Siu AL, Kimpau S. The predictive validity of self-report and performance-based measures of function and health. J Gerontol. 1992;47:M106-M110.[Abstract]
This article has been cited by other articles:
|
|
|
|
 
J. L. Locher, C. S. Ritchie, C. O. Robinson, D. L. Roth, D. Smith West, and K. L. Burgio
A Multidimensional Approach to Understanding Under-Eating in Homebound Older Adults: The Importance of Social Factors
Gerontologist,
April 1, 2008;
48(2):
223 - 234.
[Abstract]
[Full Text]
[PDF]
|
|
|
Top
Abstract
Methods
