Content uploaded by Regan L Bailey
Author content
All content in this area was uploaded by Regan L Bailey on Mar 18, 2014
Content may be subject to copyright.
Public Health Nutrition: page 1 of 5 doi:10.1017/S1368980013001729
Diet-related practices and BMI are associated with diet quality
in older adults
Dara W Ford
1,
*, Terryl J Hartman
2
, Christopher Still
3
, Craig Wood
3
, Diane Mitchell
1
,
Pao Ying Hsiao
1
, Regan Bailey
4
, Helen Smiciklas-Wright
1
, Donna L Coffman
5
and
Gordon L Jensen
1
1
Department of Nutritional Sciences, The Pennsylvania State University, 110 Chandlee Laboratory, University
Park, PA 16802, USA:
2
Department of Epidemiology, Emory University, Atlanta, GA, USA:
3
Center for Health
Research & Obesity Institute, Geisinger Health System, Danville, PA, USA:
4
Office of Dietary Supplements,
National Institutes of Health, Rockville, MD, USA:
5
The Methodology Center, The Pennsylvania State University,
State College, PA, USA
Submitted 25 February 2013: Final revision received 20 May 2013: Accepted 29 May 2013
Abstract
Objective: To assess the association of diet-related practices and BMI with diet
quality in rural adults aged $74 years.
Design: Cross-sectional. Dietary quality was assessed by the twenty-five-item
Dietary Screening Tool (DST). Diet-related practices were self-reported. Multi-
variate linear regression models were used to analyse associations of DST scores
with BMI and diet-related practices after controlling for gender, age, education,
smoking and self- v. proxy reporting.
Setting: Geisinger Rural Aging Study (GRAS) in Pennsylvania, USA.
Subjects: A total of 4009 (1722 males, 2287 females; mean age 81?5 years)
participants aged $74 years.
Results: Individuals with BMI , 18?5 kg/m
2
had a significantly lower DST score
(mean 55?8, 95 % CI 52?9, 58?7) than those individuals with BMI 5 18?5–24?9 kg/m
2
(mean 60?7, 95 % CI 60?1, 61?5; P 5 0?001). Older adults with higher, more
favourable DST scores were significantly more likely to be food sufficient, report
eating breakfast, have no chewing difficulties and report no decline in intake in
the previous 6 months.
Conclusions: The DST may identify potential targets for improving diet quality in
older adults including promotion of healthy BMI, breakfast consumption, improving
dentition and identifying strategies to decrease concern about food sufficiency.
Keywords
Diet quality
Ageing
Dietary-related practices
Diets consistent with dietary guidelines that are rich
in fruits, vegetables, whole grains, low-fat dairy and
lean meats are associated with decreased morbidity and
mortality
(1)
. Quality of diet becomes increasingly important
in old age due to declining physiological function, changes
in body composition and decreased energy require-
ments
(2,3)
. Risk of undernutrition is also increased in older
adults for some potentially modifiable reasons including
financial constraints, appetite decline, poor dentition
and functional and cognitive limitations
(4)
. The Dietary
Screening Tool (DST) is a validated tool that utilizes food-
based questions developed for use in assessing diet quality
in older, rural adults
(5)
. It is a simple, self-administered
questionnaire containing food- and behaviour-related
questions that assess overall dietary quality of older
adults
(2)
. The objective of the present study was to deter-
mine the relationship of the DST with diet-related practices
and characteristics known to contribute to nutritional risk
among a cohort of adults aged $74 years.
Materials and methods
Study participants
The Geisinger Rural Aging Study (GRAS) began in 1994
with adults aged 65 years or older enrolled in a Medicare-
managed health maintenance organization. Study details
have been published previously
(6)
. The participants have
been followed as a longitudinal cohort over time with
repeated measures of height, weight, medication use,
diet-related practices, living environment, self-rated
health and functional status. In-depth dietary assessment
to estimate usual intakes has been conducted only on
small subsets of the cohort in a cross-sectional manner
and such data are not available for the entire cohort
(5,7)
.
All surviving GRAS participants (n 5993) were mailed
demographic and health questionnaires and the DST for
the current study in the autumn of 2009. After follow-up,
4009 (67 %) participants (1722 males, 2287 females; mean
age 81?5 years) returned completed surveys, providing
Public Health Nutrition
*Corresponding author: Email djw5083@psu.edu r The Authors 2013
information on age, height, weight, smoking status,
diet-related practices and dietary information, among
other characteristics. Additionally, self-reporting or proxy
reporting by someone other than the participant was
noted. The study was conducted according to the
guidelines laid down in the Declaration of Helsinki and
all procedures involving human subjects/patients were
approved by the Office of Research Protections at The
Pennsylvania State University and the Human Research
Protection Program of the Geisinger Health Systems
Institutional Review Board. Consent was implied by survey
completion.
Dietary screening tool
Detailed information on the development and validation
of the DST has been described elsewhere
(5,7)
. The DST
consists of twenty-five questions originally derived from
extensive secondary analysis of the dietary intakes of
rural older adults in the GRAS (see online supplementary
material). The possible score range is from 0 to 100 points
with 5 ‘bonus’ points for multivitamin/mineral supple-
ment use (score could not exceed 100). Responses to
questions were then scored according to the previously
validated scoring algorithm with a score ,60 considered
‘unhealthy’, 60–75 considered ‘borderline’ and .75 con-
sidered ‘healthy’
(5)
. An example of a DST question is
‘How often do you usually eat whole grain breads?’
Participants then chose from ‘never’, ‘less than once a
week’, ‘1 or 2 times a week’ and ‘3 or more times a week’
to classify their intake. Cognitive interviewing was used to
ensure understandability of questions for the population
of interest
(7)
. Points were allotted for each question based
upon breakdown of major dietary components of the
Healthy Eating Index-2005
(8)
. Dietary quality was estab-
lished by comparison with nutrient intakes
(5,7)
and food
group intakes
(5)
derived from multiple 24 h recalls.
Eating behaviour measures
Nine total questions identified the presence of problems
associated with diet-related practices through yes-or-no
responses. All questions were self- or proxy reported.
These questions addressed inadequate food or concerns
about sufficient food, not eating on one or more days per
month, having a decline in intake, eating alone, skipping
breakfast, having more than one alcoholic drink per day
for women or more than two per day for men, reporting
chewing difficulty and mouth pain. Associations between
all diet-related practices and DST score were analysed.
Statistical analyses
All data were analysed using the Statistical Analysis Software
Package 9?3. Descriptive data were generated using PROC
MEANS and PROC FREQ for all adults and by gender.
Multivariate linear regression models were used to analyse
associations of continuous DST score as the dependent
variable with BMI and each of the nine diet-related practices
after controlling for age (continuous), gender, education
(,high school v. $high school), smoking (ever/never)
and self- v. proxy reporting. BMI was calculated from self-
reported height and weight collected in the demographic
and health questionnaires, and was assessed both as a
continuous variable and categorically according to
National Institutes of Health guidelines (,18?5 kg/m
2
,
18?5–24?9kg/m
2
,25?0–29?9 kg/m
2
and $30?0 kg/m
2
). All
dietary behaviours that were related significantly to DST
score at P , 0?05 were retained as potential candidates for
the multivariate model. Results are presented as mean DST
scores with 95 % confidence intervals adjusted for age,
gender, self- or proxy reporting, and BMI when BMI was
not the independent variable of interest. P values are for
the tests of between-group differences from the multi-
variate models. Interactions between the predictors of
interest (diet-related practices and BMI) and each covariate
(gender, BMI, age, education, smoking, self- v. proxy
reporting) were assessed by including each individual
factor (e.g. gender) and its cross-product term in separate
models. Significance was considered at P , 0?05.
Results
Descriptive characteristics of the sample are shown
in Table 1. Compared with those who completed the
DST, non-responders were older (83?2 v. 81?4 years;
P , 0?0001) and more likely to be female (OR 5 1?3, 95 %
CI 1?2, 1?5; P , 0?0001). Less than 9 % (n 333) of partici-
pants used proxy reporters and those who did were more
likely to be male (OR 5 1?5, 95 % CI 1?2, 1?9; P 5 0?0002),
less likely to report education beyond high school
(OR 5 0?5, 95 % CI 0?3, 0?7; P 5 0?0002), older (mean 83?7
(
SD 5?5) years v. 81?2(SD 4?1) years; P , 0?0001) and had
lower DST scores (mean 57?6(
SD 12?3) v. 60?6(SD 12?7);
P , 0?0001). The cohort was comprised almost exclu-
sively of non-Hispanic whites (98?7 %) with at least a high
school degree. Less than half the sample was male (43 %).
BMI did not differ by gender. Although over half of the
respondents lived with a spouse (n 2095), 46 % of female
respondents lived alone compared with only 20 % of male
respondents. The mean unadjusted DST score for the
sample was 60?3(
SD 12?7), with females (mean 61?9
(
SD 12?6)) reporting a significantly higher score than males
(mean 58?2(
SD 12?4); P , 0?0001).
Participants who had BMI , 18?5 kg/m
2
had sig-
nificantly lower DST scores (OR 5 55?8, 95 % CI 52?9,
58?7) than those participants with BMI 5 18?5–24?9 kg/m
2
(OR 5 60?8, 95 % CI 59?5, 60?9; P 5 0 ?001) after adjust-
ment for age, sex, education, smoking status and self- v.
proxy reporting. The adjusted DST score for those parti-
cipants with BMI , 18?5 kg/m
2
remained significantly
lower (OR 5 55?8, 95 % CI 52?9, 58 ?7) compared with the
DST score for all other BMI classes combined (OR 5 60?5,
95 % CI 60?1, 60?9; P 5 0?002). In contrast, compared with
Public Health Nutrition
2 DW Ford et al.
participants with BMI 5 18?5–24?9 kg/m
2
, there were no
statistically significant differences in DST score for either
overweight or obese individuals (see Table 2). There
were also no significant associations between BMI and
any of the diet-related practices.
Four of the nine diet-related practices were significantly
associated with DST score after adjustment for BMI, age,
sex, education, smoking status and self- v. proxy reporting
(Table 2). Significantly lower DST scores were found in
participants who reported a decline in intake over the
previous 3 months, skipping breakfast, concern about
having enough food and difficulty with chewing or
swallowing. The remaining five diet-related practices were
not significantly associated with DST score. No meaningful
and significant effect modifications were observed between
any variables tested (data not presented).
Discussion
It was our goal to investigate the associations between
BMI, diet-related practices and diet quality in a population
Public Health Nutrition
Table 1 Characteristics of study participants: rural adults aged $74 years, Geisinger Rural Aging Study (GRAS),
Pennsylvania, USA, autumn 2009
Men (n 1722; 43?0 %) Women (n 2287; 57?0%)
Characteristic Mean or n
SE or % Mean or n SE or %
Age (years)* 81?34?281?54?4
Race
White 1654 98?2 2234 99?1
Non-Hispanic black 29 1?715 0?7
Other 1 0?14 0?2
Education
,High school 1327 77?1 1942 84?9
$High school 395 22?9 345 15?1
BMI (kg/m
2
)
,18?5140?859 2?6
18?5–24?9 460 26?7 696 30?4
25?0–29?9 814 47?3 839 36?7
$30?0 434 25?2 693 30?3
Ever smoke
Yes 61 3?682 3?7
No 1629 96?4 2159 96?3
Eat breakfast
Yes 1660 96?4 2190 95?8
No 62 3?697 4?2
Eat alone
Yes 305 17?7 851 37?2
No 1417 82?3 1436 62?8
Intake decline
Yes 111 6?5 161 7?0
No 1611 93?5 2126 93?0
Excess alcohol
Yes 94 5?537 1?6
No 1628 94?5 2250 98?4
Food insufficient
Yes 8 0?59 0?4
No 1714 99?5 2278 99?6
Enough food each day
Yes 1681 97?6 2254 98?6
No 41 2?433 1?4
No food some days
Yes 4 0?27 0?3
No 1718 99?8 2280 99?7
Chewing difficulty
Yes 69 4?087 3?8
No 1653 96?0 2200 96?2
Mouth pain
Yes 41 2?453 2?3
No 1681 97?6 2234 97?7
DST score* 58?212?461?912?6
DST category-
,60 917 53?3 925 40?4
60–75 629 36?5 976 42?7
.75 176 10?2 386 16?9
DST, Dietary Screening Tool.
*These data are presented as mean and standard error; all other data are presented as number and percentage.
-Categories utilized from previously published data
(1)
.
Diet practices and quality in older adults 3
of adults aged $74 years. There are limited data on dietary
quality for large cohorts of older adults, particularly those
living in rural areas. Our results indicate that a low DST
score is associated with low BMI and poor diet-related
practices including chewing difficulties, skipping breakfast,
concerns of food sufficiency and decline in intake.
Older adults with low BMI had a much poorer diet
quality than all other older adults, including those who
were obese. Population studies suggest that risk of
mortality is doubled in older adults who have a BMI
,18?5 kg/m
2
compared with 18?5–24?9 kg/m
2
indepen-
dent of recent weight change
(9,10)
. The association
between obesity and mortality in older adults is complex,
with overweight and mild obesity being associated with
reduced mortality in cohort studies of adults $65 years old
with follow-up periods ranging from 3 to 18 years
(9–11)
.
In a prior investigation within a small subset of the GRAS
cohort (n 179) we found that a low nutrient-dense diet was
associated with increased odds of obesity
(12)
and lower
waist circumference was associated with a prudent dietary
pattern
(7)
. In the current study, an association between
obesity and diet quality was not detected. Of note, no
participants in our previous study had BMI , 18?5kg/m
2(12)
.
Chewing difficulty, skipping breakfast, food insuffi-
ciency and decline in intake were associated with poor
diet quality. Chewing difficulty is linked to many adverse
clinical outcomes, including a variety of morbidities,
hospitalization and earlier mortality, and has been shown
to affect consistency and selection of food
(4)
. Skipping
breakfast is associated with decreased nutrient intake,
which may impact development and progression of
chronic disease
(13)
. In a nationally representative sample of
adults aged 60–90 years, those who were food-insufficient
consumed significantly less energy, carbohydrate, protein,
saturated fat, Fe and Zn among other micronutrients
and were more likely to report poor self-rated health than
their food-sufficient peers
(14)
. Decline in intake may lead
to unintentional weight loss which is often indicative of
underlying disease, and undernutrition in older adults is
strongly associated with increased mortality
(15,16)
.TheDST
is able to identify these diet-related practices as targetable
areas for improvement in diet quality and potentially other
health outcomes in older adults.
A relatively high response rate (67 %) in an aged
community-dwelling cohort is a major strength of this
investigation. However, there are some notable limita-
tions to address. The external validity of the DST remains
to be determined in other races and geographic regions.
The number of remaining underweight older adults was
quite low, likely due to decreased survivorship in elderly
individuals with a low BMI
(9,10)
. The screening ques-
tionnaires rely on self-report, making results subject
to recall bias. Additionally, only information regarding
age and sex was available for non-responders and so
additional comparisons could not be made.
Previously the DST was administered in an out-patient
clinic setting, requiring participants to visit their local
medical clinic in order to complete the questionnaire
(5)
.
Rural older adults experience many barriers to health care
including but not limited to social isolation, lack of
transportation and financial constraints
(17)
. By surveying
rural adults in their own homes, we were able to find
targetable areas for improvement of nutritional quality.
Overall food consumption decreases with age and it
becomes increasingly important for older adults to consume
high-quality nutrient-dense foods to meet nutrient needs
(18)
.
Public Health Nutrition
Table 2 Association between adjusted mean DST score, diet-related practices and BMI: rural adults aged $74 years,
Geisinger Rural Aging Study (GRAS), Pennsylvania, USA, autumn 2009
Eating practice* Adjusted mean DST score 95 % CI P value-
Skip breakfast 51?749?8, 53?7 ,0?0001
Eat breakfast 60?860?4, 61?2–
Eat alone 60?559?8, 61?30?71
Eat with others 60?459?9, 60?8–
Intake decline 56?855?3, 58?3 ,0?0001
No decline 60?760?3, 61?1–
Excess alcohol 58?756?5, 60?90?12
No excess alcohol 60?560?1, 60?9–
Food insufficient 53?948?0, 59?80?03
Food sufficient 60?460?0, 60?8–
Not enough food each day 58?956?1, 61?80?
32
Enough food each day 60?460?0, 60?8–
No food some days 57?449?7, 65?10?44
Always have food 60?460?0, 60?8–
Chewing difficulty 58?256?3, 60?20?03
No difficulty 60?560?1, 60?9–
Mouth pain 59?857?2, 62?30?63
No mouth pain 60?460?0, 60?8–
Underweight (BMI , 18?5 kg/m
2
)-
-
55?852?9, 58?70?001
Not underweight 60?560?1, 60?9–
*Controlling for sex, BMI, age, smoking status, education and self- v. proxy reporting.
-Represent differences between groups (appetite decline v. no decline, concern about food v. no concern, etc.) after adjustment for covariates.
-
-
Controlling for sex, age, smoking status, education and self- v. proxy reporting.
4 DW Ford et al.
The diet-related practices found to be associated with DST
score serve as potential targets for altering behaviour to
promote nutrient and energy intakes sufficient to meet
requirements. It should also be noted that the mean overall
DST score was below optimal (mean 5 60) with 86 % of
participants scoring #75 on the DST. According to pre-
vious studies, this indicates that 86 % of this sample has
either unhealthy or borderline diet quality, and so has
room for improvement
(5)
.
Conclusions
Older adults are at increased susceptibility for malnutri-
tion due to age-associated changes in metabolism and
physiology
(18)
, and with the number of aged persons
increasing rapidly in our population
(19)
improving nutri-
tional status is a priority. Low DST scores were associated
with low BMI, being food insecure, recent decline in food
intake, skipping breakfast and chewing difficulties. These
associations may help to identify opportunities for
anticipatory guidance and interventions for health-care
professionals to promote improvement in diet quality.
Acknowledgements
Sources of funding: This work was supported by the US
Department of Agriculture (grant #1950-51530-010-02G).
Conflicts of interest: The authors report no conflict of
interest. Authors’ contributions: T.J.H., G.L.J. and H.S.-W.
contributed to the writing and editing of this paper. C.W .
and D.L.C. assisted with statistical analysis. C.S., D.M., P.Y.H
and R.B. provided editorial assistance in writing this paper.
Supplementary material
To view supplementary material for this article, please
visit http://dx.doi.org/10.1017/S1368980013001729
References
1. Russell J, Flood V, Rochtchina E et al. (2012) Adherence to
dietary guidelines and 15-year risk of all-cause mortality. Br
J Nutr (Epublication ahead of print version).
2. Jensen GL & McGee M & Binkley J (2001) Nutrition in the
elderly. Gastroenterol Clin North Am 30, 313–334.
3. Brownie S (2006) Why are elderly individuals at risk of
nutritional deficiency? Int J Nurs Pract 12, 110–118.
4. Bailey R, Gueldner S, Ledikwe J et al. (2005) The oral health
of older adults: an interdisciplinary mandate. J Gerontol Nurs
31,11–17.
5. Bailey R, Miller PE, Mitchell DC et al. (2009) Dietary
screening tool identifies nutritional risk in older adults. Am
J Clin Nutr 90, 177–183.
6. Jensen GL, Kita K, Fish J et al. (1997) Nutrition risk
screening characteristics of rural older persons: relation to
functional limitations and health care charges. Am J Clin
Nutr 66, 819–828.
7. Bailey RL, Mitchell DC, Miller CK et al. (2007) A dietary
screening questionnaire identifies dietary patterns in older
adults. J Nutr 137, 421–426.
8. Guenther PM, Krebs-Smith SM, Reedy J et al. (2008)
Healthy Eating Index-2005. Alexandria, VA: Center for
Nutrition Policy and Promotion.
9. Locher JL, Roth DL, Ritchie CS et al. (2007) Body mass
index, weight loss, and mortality in community-dwelling
older adults. J Gerontol A Biol Sci Med Sci 62, 1389–1392.
10. Takata Y, Ansai T, Soh I et al. (2007) Association between
body mass index and mortality in an 80-year-old popula-
tion. J Am Geriatr Soc 55, 913–917.
11. Stessman J, Jacobs JM, Ein-Mor E et al. (2009) Normal body
mass index rather than obesity predicts greater mortality in
elderly people: the Jerusalem longitudinal study. JAm
Geriatr Soc 57, 2232–2238.
12. Ledikwe JH, Smiciklas-Wright H, Mitchell DC et al. (2004)
Dietary patterns of rural older adults are associated with
weight and nutritional status. J Am Geriatr Soc 52, 589–595.
13. Gollub EA & Weddle DO (2004) Improvements in nutri-
tional intake and quality of life among frail homebound
older adults receiving home-delivered breakfast and lunch.
J Am Diet Assoc 104, 1227–1235.
14. Lee JS & Frongillo EA Jr (2001) Nutritional and health
consequences are associated with food insecurity among
US elderly persons. J Nutr 131, 1503–1509.
15. Chapman IM (2007) The anorexia of aging. Clin Geriatr
Med 23, 735–756.
16. Miller SL & Wolfe RR (2008) The danger of weight loss in
the elderly. J Nutr Health Aging 12
, 487–491.
17. Goins RT, Williams KA, Carter MW et al. (2005) Perceived
barriers to health care access among rural older adults: a
qualitative study. J Rural Health 21, 206–213.
18. Drewnowski A & Evans WJ (2001) Nutrition, physical
activity, and quality of life in older adults. J Gerontol A Biol
Sci Med Sci 56, Suppl. 2, 89–94.
19. Centers for Disease Control and Prevention (2003) Public
health and aging: trends in aging – United States and
worldwide. MMWR Morb Mortal Wkly Rep 52, issue 6,
101–106.
Public Health Nutrition
Diet practices and quality in older adults 5
