International Journal of Behavioral

Nutrition and Physical Activity BioMed Central

Methodology Open Access

The validation of a home food inventory
Jayne A Fulkerson*1, Melissa C Nelson2, Leslie Lytle2, Stacey Moe2,
Carrie Heitzler2 and Keryn E Pasch2

Address: 1School of Nursing, University of Minnesota, Minnesota, USA and 2Division of Epidemiology & Community Health, University of
Minnesota, Minnesota, USA
Email: Jayne A Fulkerson* - fulke001@umn.edu; Melissa C Nelson - nels5024@umn.edu; Leslie Lytle - lalytle@umn.edu;
Stacey Moe - gerl0056@umn.edu; Carrie Heitzler - heitz022@umn.edu; Keryn E Pasch - pasc0074@umn.edu
* Corresponding author

Published: 4 November 2008 Received: 2 June 2008

Accepted: 4 November 2008
International Journal of Behavioral Nutrition and Physical Activity 2008, 5:55 doi:10.1186/1479-5868-5-55
This article is available from: http://www.ijbnpa.org/content/5/1/55
© 2008 Fulkerson et al; licensee BioMed Central Ltd.
This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0),
which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Abstract
Background: Home food inventories provide an efficient method for assessing home food
availability; however, few are validated. The present study's aim was to develop and validate a home
food inventory that is easily completed by research participants in their homes and includes a
comprehensive range of both healthful and less healthful foods that are associated with obesity.
Methods: A home food inventory (HFI) was developed and tested with two samples. Sample 1
included 51 adult participants and six trained research staff who independently completed the HFI
in participants' homes. Sample 2 included 342 families in which parents completed the HFI and the
Diet History Questionnaire (DHQ) and students completed three 24-hour dietary recall
interviews. HFI items assessed 13 major food categories as well as two categories assessing ready-
access to foods in the kitchen and the refrigerator. An obesogenic household food availability score
was also created. To assess criterion validity, participants' and research staffs' assessment of home
food availability were compared (staff = gold standard). Criterion validity was evaluated with kappa,
sensitivity, and specificity. Construct validity was assessed with correlations of five HFI major food
category scores with servings of the same foods and associated nutrients from the DHQ and
dietary recalls.
Results: Kappa statistics for all 13 major food categories and the two ready-access categories
ranged from 0.61 to 0.83, indicating substantial agreement. Sensitivity ranged from 0.69 to 0.89, and
specificity ranged from 0.86 to 0.95. Spearman correlations between staff and participant major
food category scores ranged from 0.71 to 0.97. Correlations between the HFI scores and food
group servings and nutrients on the DHQ (parents) were all significant (p < .05) while about half
of associations between the HFI and dietary recall interviews (adolescents) were significant (p <
.05). The obesogenic home food availability score was significantly associated (p < .05) with energy
intake of both parents and adolescents.
Conclusion: This new home food inventory is valid, participant-friendly, and may be useful for
community-based behavioral nutrition and obesity prevention research. The inventory builds on
previous measures by including a wide range of healthful and less healthful foods rather than foods
targeted for a specific intervention.

Page 1 of 10
(page number not for citation purposes)
International Journal of Behavioral Nutrition and Physical Activity 2008, 5:55 http://www.ijbnpa.org/content/5/1/55

Background Crockett and colleagues [8] developed a shelf inventory of

Availability of foods in the home has been shown to be 80 foods in 12 categories based on foods targeted in a can-
significantly associated with dietary practices, intake, and cer risk reduction program. Only perishable foods tar-
eating patterns [1-5]. Although this area of research has geted in the intervention program were included on the
been growing and several instruments have been devel- inventory. They conducted two criterion-related valida-
oped to assess the home food environment (see [6] for a tion studies that compared participant and staff-reported
comprehensive review), often the instruments include a inventories. In the two studies, sensitivity was reported as
limited number of items selected to address a specific aim 0.86 and 0.87, and specificity was reported as 0.92 and
such as fruit and vegetable availability for cancer preven- 0.90, respectively. In addition, Cohen's kappa showed sig-
tion or high-fat food availability for cardiovascular health. nificant overall agreement (p < .0001) between partici-
Instruments that provide a more comprehensive assess- pant and staff reports in both studies. Furthermore,
ment of food and energy availability in the home are cur- construct validation was assessed by comparing inventory
rently unavailable. Given current scientific interest in the responses with food frequency questionnaires, with over-
contextual and environmental influences on energy bal- all agreement of 73.6%
ance and obesity, there is a great need for such inventories.
Miller and Edwards [9] assessed the face, content and cri-
When developing such assessment inventories, thorough terion validity for a 166-item shelf inventory that was
evaluation is needed, including several dimensions of based on previous work [8,11] with modifications for fat-
validity testing. Criterion validity tests the performance of and sugar-modified foods that would be relevant to the
an instrument by comparing it to a gold standard [7]. A purchases of individuals with diabetes. Similar to the pre-
rigorous but logistically-challenging method for assessing vious inventory, mostly perishable foods were included.
criterion validity for a home food inventory is to require Thirty-one older adults diagnosed with Type 2 diabetes
participants and study staff to independently complete a completed the inventory and within 48 hours study staff
home food availability instrument, and then compare directly observed foods in the home. Cohen's kappa sta-
their responses using the staff report as the "gold stand- tistic was 0.87 and sensitivity and specificity were 0.90
ard." Comparisons are evaluated for consistency of results and 0.97, respectively. Construct validity was not
between the two methods using kappa or correlation sta- assessed.
tistics. In addition, comparisons may be made by examin-
ing sensitivity (i.e., proportion of foods in the home Marsh and colleagues [10] conducted a study with 48 par-
assessed as present by the staff that were accurately identi- ents that demonstrated criterion validity for a 34-item
fied as present in the home by the participant) and specif- fruit, juice, and vegetable availability questionnaire. Par-
icity (i.e., proportion of foods in the home assessed as ents reported the presence of these foods in the home
absent by the staff that were accurately identified as absent within the last seven days, and staff conducted inventories
in the home by the participant). Construct validity is sug- on the same visit. In addition to perishable foods, frozen,
gested when expected relationships are shown between canned and dried fruit/vegetable products were included.
the measure and other variables in a conceptual frame- Cohen's kappa values ranged from 0.24–0.53 for juices,
work [7]. For example, if an HFI shows high availability of 0.12–0.76 for fruits, and 0.22–0.66 for vegetables. Overall
high fat foods in the home, one might expect family mem- sensitivity and specificity values for total fruit, juice and
bers to report a high calorie intake in their diet. vegetables were 36.8 and 39.1, respectively. Construct
validity was not assessed.
The development of many of the home food availability
measures described in the literature include a very limited Lastly, Raynor and colleagues [2] completed a type of cri-
number of foods, and have not included comprehensive terion validity and construct validity of a household food
validity testing, particularly criterion and construct valid- availability instrument (22 high-fat and 22 low-fat items)
ity. Criterion validity is very important if participants in with 165 adults. Criterion validity was accomplished by
research studies are expected to complete self-report home correlating the number of high-fat and low-fat food items
food inventories on their own while in their homes. Con- reported by two adults in the same household (r = 0.69, p
struct validity is also important because an instrument < .001 and r = 0.59, p < .001, respectively). However, gold
should be associated with expected health outcomes. standard criterion validity comparing participant report to
Only four studies have demonstrated criterion validity of trained staff report was not completed. Construct validity
home food inventories [2,8-10], and only two have dem- considered correlations between availability of high-fat
onstrated construct validity [2,8], and each has its limita- foods and low-fat foods with fat consumption via a food
tions (as described below). frequency questionnaire, and showed significant (p <
.001) correlations (r = 0.25 and r = -0.33, respectively).

Page 2 of 10
(page number not for citation purposes)
International Journal of Behavioral Nutrition and Physical Activity 2008, 5:55 http://www.ijbnpa.org/content/5/1/55

In comparison to other published home food inventories pleted inventories at the same time, they began their
[6], these four studies represent the most comprehensive assessments in different parts of the home and were
methods that have been used to test validity to date. instructed not to communicate with each other as they
Despite their relatively rigorous validity testing, however, completed the inventory. Participants were provided with
only two studies demonstrated construct validity, and a $30 gift card for their participation.
each of these measures is somewhat limited in scope, with
none providing a comprehensive assessment of overall Sample 2
home food and energy availability. Thus, given the lack of For the construct validation phase, 349 families (one stu-
comprehensive home food inventories that have been rig- dent between the ages of 10 and 17 years and one parent/
orously validated, there is a need for additional instru- guardian or other adult caregiver) were recruited from the
ment development in this area. An inventory assessing a following sources: 1) an existing cohort of youth partici-
wide range of food exposures in the home, including both pating in the Minnesota Adolescent Community Cohort
healthful and less healthful foods, may be important in (MACC) Tobacco Study [13], 2) a Minnesota Department
understanding contextual influences on obesity, weight of Motor Vehicle (DMV) list restricted to the seven-county
gain, and nutritional intake. In addition, such an inven- metro area, or 3) a convenience sample drawn from local
tory could be useful in determining appropriate interven- communities. Of the 349 youth/adult pairs measured,
tion strategies that may fit the needs of individual 26% were recruited from the MACC cohort, 49% were
households or might identify targets for public health recruited from the DMV sample and 25% were recruited
messages. The purpose of the present study was to develop from the convenience sample [12].
and validate a home food inventory that is easily com-
pleted by adults in their homes and includes both health- In the larger IDEA study, youth and adults pairs scheduled
ful and less healthful foods as well as reduced-fat and a visit to an IDEA clinic where anthropometric measures
regular-fat varieties of foods potentially related to the were taken and psychosocial surveys (that included demo-
obesity epidemic. graphic characteristics) were administered to both stu-
dents and adults. Instructions for additional measures
Methods were given at this time and parents received a packet of
Procedures, Participants, and Measures instruments to take home to complete and return by mail.
The University of Minnesota's Institutional Review Board Included in this packet were the HFI and the DHQ. The
approved this study, and all participants signed the appro- DHQ is a food frequency that has been widely used with
priate parental consent and student assent forms. The adults (NCI). Students were told to expect that three die-
study was conducted in three phases. First, brief field test- tary recall interviews would be conducted with them by
ing of the newly developed HFI was conducted with a telephone within the next month, and were provided with
small sample of adults (n = 5) to assess ease of completion a two-dimensional food model packet to help them esti-
and comprehension as the participants indicated which mate portion size. The final sample for construct valida-
foods were difficult to inventory. The inventory food list tion includes data from the 342 families who completed
was re-evaluated at this stage and foods were added if par- the HFI, DHQ and dietary recalls (98% of sample).
ticipants frequently indicated that there were foods to
code without a place to do so on the inventory. Second, Measures
adults in the community were recruited to complete the Home Food Inventory
home HFI in their homes while allowing trained staff To develop the initial set of food items, the investigators
members in their homes to complete the inventory inde- examined existing instruments [8,14] and reviewed the
pendently (criterion validity testing with Sample 1). literature that identified major contributors to overall
Third, parents and students were recruited for participa- energy intake (e.g., [15]). This process allowed us to eval-
tion in the IDEA (Identifying Determinants of Eating and uate the foods listed in inventories that were developed
Activity) study [12] in which parents completed the HFI for a specific limited purpose (e.g., a diabetic population,
and Diet History Questionnaire (DHQ) and students par- nonperishable foods, etc) as well as expand the items in
ticipated in 24-hour recall dietary interviews (construct our inventory to include foods known to be associated
validity with Sample 2). with energy intake in the population. In addition, based
on literature that demonstrates a high correlation between
Sample 1 readily accessible foods (i.e., foods in plain view) and
For the criterion validation phase, 51 adults were recruited their intake [3], two items were added to assess the acces-
from 19 area Minneapolis Park and Recreational Centers sibility of healthful foods within the main kitchen area
via posted flyers. Trained research staff traveled to the par- and the refrigerator. Thus, we evaluated the literature and
ticipants' homes to complete consent procedures and instruments to date, and added foods that provided a
complete to the HFI. Although participants and staff com-

Page 3 of 10
(page number not for citation purposes)
International Journal of Behavioral Nutrition and Physical Activity 2008, 5:55 http://www.ijbnpa.org/content/5/1/55

more comprehensive inventory of foods associated with 86%). Dietary recalls were conducted for two weekdays
dietary intake of adults in the US. and one weekend day, with the aim of having each of the
three recalls completed within a 2-week period. In gen-
During the course of initial validity testing (Sample 1), eral, multiple dietary recalls are widely accepted as a valid
changes in the number of items and instructions occurred. and reliable method for dietary assessment, and have
For example, participants were allowed to "write in" foods yielded acceptable validity in children as young as 10
within given categories (e.g., dairy, fruits, vegetables), and years [16]. Trained and certified staff from the Nutrition
if "write in" responses were found to be frequent, they Coordination Center (NCC) at the University of Minne-
were added as items on the inventory. During the validity sota administered the recalls, using the Nutrition Data
testing with Sample 1, the HFI had 186 items while the System Research (NDS-R) software [2006, Nutrition
final inventory administered to Sample 2 included 190 Coordinating Center, University of Minnesota, Minneap-
items. olis, MN] with an interactive, interview format with direct
data entry linked to a nutrient database [17]. The NDS-R
HFI items are listed in a checklist type format with yes/no data set allows for the examination of both nutrient
(1/0) response options. Higher scores represent greater intakes and food group information (e.g., servings of
availability. In addition, participants were instructed to fruits and vegetables consumed).
check whether the vegetable, fruit, and bread items were
fresh, frozen, dried or canned, as appropriate. The cate- Diet History Questionnaire
gory order was set up to facilitate ease of completion, Dietary assessment for parents was conducted using the
beginning with the refrigerated items, followed by frozen Diet History Questionnaire (DHQ) food frequency instru-
items, and non-perishable items. Participants are ment developed by the National Cancer Institute (NCI).
instructed to look for these foods in all areas of the home Parents received the DHQ at the clinic visit and were
where food is stored, including the refrigerator, freezer, asked to mail it in when completed. This instrument con-
pantry, cupboard, and other areas (e.g., basement). Partic- sists of 144 food items and includes both portion size and
ipants were informed that lower fat products may be dietary supplement questions. Requiring approximately
labeled as "reduced-fat," "low-fat," "light," "nonfat," or one hour to complete, it has been widely used to charac-
"skim." Foods in the dairy, added fats, frozen desserts, terize usual food and nutrient intakes in numerous adult
prepared desserts, and savory snacks were categorized into populations.
regular-fat or reduced-fat groupings; beverages were cate-
gorized into regular sugar and low sugar categories; and Several studies have been conducted to assess the validity
foods in the two ready-access categories were further sub- and calibration of the DHQ. Findings indicate that the
grouped into healthful and less healthful categories. DHQ provides reasonably valid estimates for usual intake
Although the categorization of foods into healthful and of most nutrients and that it performs as well or better
less healthful categories may not be entirely straightfor- than other well-known food frequency instruments avail-
ward, we assessed each food by its typical fat and sugar able in the field [18-20]. The food list and nutrient data-
content when determining its category. To assess the over- base used for standardized analysis of the DHQ are
all obesogenic home food availability, a summative score derived using national dietary data from the US Depart-
was created that includes regular-fat versions of cheese, ment of Agriculture's Continuing Survey of Food Intakes
milk, yogurt, other dairy, frozen desserts, prepared des- by Individuals (1994–96) [20].
serts, savory snacks, added fats; regular-sugar beverages;
processed meat; high-fat quick, microwavable foods; Data Analysis
candy; access to unhealthy foods in refrigerator and To assess criterion validity using data from Sample 1, par-
kitchen. The obesogenic home food availability score ticipants' and research staffs' assessment of home food
potential range was from 0–71 (present sample: range = availability were compared. Consistent in research of cri-
9–53, M = 29.4, SD = 7.6). The HFI can be requested from terion validity, the staff report was considered the gold
the primary author. A table reflecting which foods are standard as they were trained on how to use the inventory.
included in each food group/subgroup is provided in Validity was evaluated by calculating kappa, sensitivity,
Additional file 1. The inventory took approximately and specificity between participant and staff reports on
30–45 minutes to complete depending upon the amount the presence of individual foods. To summarize these
of food stored in the home. results, we calculated the average of these individual kap-
pas across both major and minor food groupings. In addi-
24-Hour Recall Interviews tion, to test the performance of the instrument's
Students in Sample 2 completed three telephone-admin- assessment of broad food categories (rather than individ-
istered 24-hour dietary recalls following their clinic visit ual foods), we assessed the extent of agreement between
(response rate for students completing three recalls was food group summary scores between participant and staff

Page 4 of 10
(page number not for citation purposes)
International Journal of Behavioral Nutrition and Physical Activity 2008, 5:55 http://www.ijbnpa.org/content/5/1/55

reports by creating additive summary scores for major and Criterion Validity
minor food groupings (e.g., overall dairy score, cheese Table 2 provides information regarding criterion validity
score, respectively). Using data from Sample 2, construct with comparisons of reports of the presence of individual
validity was assessed by examining Spearman correlations food items between the trained staff data (gold standard)
of five major food category scores on the HFI (i.e., dairy, and participants' data, including the number of items, the
vegetables with and without potatoes, fruit, and meats & average kappa statistic, the average sensitivity value, and
other nondairy protein) with number of servings of the the average specificity value for each food category. In
same foods as well as nutrients that should be correlated addition, Spearman correlations of major food categories
with these foods (e.g., calcium with dairy, Vitamin C with comparing data from staff and participants are provided.
fruit) with foods and nutrients from the DHQ and 24- The number of items in each major food category ranged
hour recall interviews. These five categories were chosen from five to 26.
since it was possible to create similar categories across the
three measures. In addition, we assessed construct validity For all 13 major food categories and the two accessibility
of the obesogenic home food availability score by com- categories, Cohen's kappa ranged from 0.61 (prepared
paring it to both parental and adolescent reports of energy desserts) to 0.83 (Fruits), sensitivity values ranged from
intake. All analyses were conducted in SAS (v9.1, SAS 0.69 (prepared desserts) to 0.89 (vegetables), and specifi-
Institute, Inc., Cary, NC, 2003). city ranged from 0.86 (meat & other nondairy protein,
bread, and beverages) to 0.95 (fruits, microwavable/
Results quick-cook foods, and kitchen accessibility). All kappas
Sample Characteristics assessing agreement between staff and participant reports
As shown in Table 1, Sample 1 consisted of 51 adults (pre- of major food categories were greater than 0.60, indicating
dominantly female). About two-thirds of the sample was substantial agreement. Spearman correlations between
white, followed by African American, American Indian, staff and participant major food category scores ranged
Mixed race/ethnicity, Latino, and Asian. More than half of from .71 (frozen desserts) to .97 (candy). All correlations
the sample had a college degree. Student gender for Sam- between staff and participant reports of major food group
ple 2 was equally distributed and most identified them- scores were greater than 0.70. All validity statistics for the
selves as white. The majority of students reported living obesogenic home availability score were equally accepta-
with their mother and father together. Students were pri- ble.
marily attending public schools, and more than two-
thirds of the sample was in 9th–12th grade. Parent partici- Construct Validity
pants in Sample 2 were predominantly white females with Results for construct validity (correlations between five
some college education. HFI major food categories and the same food categories
and expected associated nutrients from the DHQ and 24-
hour recalls) are presented on Table 3. All of the HFI

Table 1: Characteristics of criterion validity sample (Sample 1) and construct validity sample (Sample 2)

Sample 1 Sample 2

Characteristic Adults (n = 51) Parents (n = 342) Students (n = 342)

Gender, %
Females 94.1 75.6 51.0
Education, %
Less than HS 4.0 0.6 100.0
HS or GED 8.0 8.6 --
Some college or vocational school 26.0 26.1 --
College degree 28.0 35.2 --
Training beyond college degree 34.0 27.8 --
Race/ethnicity, %
White 68.6 98.8 93.4
African American 13.7 0.3 1.4
American Indian 5.9 0.0 0.0
Asian 2.0 0.0 0.3
Hispanic/Latino 3.9 0.0 0.0
Mixed 5.9 0.9 4.6
Age (years; M, SD) 39.4 46.7 (0.3) 15.4 (0.1)

Page 5 of 10
(page number not for citation purposes)
International Journal of Behavioral Nutrition and Physical Activity 2008, 5:55 http://www.ijbnpa.org/content/5/1/55

Table 2: Inventory major and subgroup food category validity indices (Sample 1, n = 51)

Food category # Items Average kappa for category Sensitivity Specificity Staff/participant correlationa

Dairy 21 0.72 0.81 0.91 0.92

Cheese 11 0.64 0.74 0.90 0.85
Regular fat 5 0.62 0.76 0.84 0.70
Reduced fat 6 0.65 0.72 0.95 0.80
Milk/other dairy beverages 6 0.89 0.90 0.97 0.90
Regular fat 1 0.94 1.00 0.98 0.88
Reduced fat 5 0.87 0.88 0.97 0.89
Yogurt 2 0.71 0.85 0.86 0.89
Regular fat 1 0.70 0.78 0.91 0.70
Reduced fat 1 0.72 0.91 0.81 0.73
Other Dairy 3 0.78 0.87 0.90 0.87
Regular fat 2 0.73 0.90 0.86 0.78
Reduced fat 1 0.88 0.82 1.00 0.88
All vegetables, including potatoes 20 0.80 0.89 0.90 0.88
All vegetables, no potatoes 19 0.80 0.89 0.90 0.88
Fruits 26 0.83 0.87 0.95 0.95
Meats & other nondairy protein 16 0.74 0.88 0.86 0.85
Processed meat 4 0.74 0.84 0.91 0.78
All other protein 11 0.74 0.89 0.85 0.83
Added Fat 13 0.76 0.84 0.92 0.79
Regular fat 8 0.78 0.88 0.91 0.76
Reduced fat 5 0.72 0.78 0.94 0.77
Frozen Desserts 7 0.64 0.70 0.94 0.71
Regular fat 3 0.83 0.86 0.95 0.82
Reduced fat 4 0.50 0.58 0.93 0.70
Prepared Desserts 8 0.61 0.69 0.93 0.73
Regular fat 6 0.58 0.68 0.92 0.65
Reduced fat 2 0.81 0.80 0.98 0.81
Savory Snacks 18 0.73 0.84 0.91 0.95
Regular fat 10 0.71 0.88 0.89 0.93
Reduced fat 8 0.76 0.78 0.95 0.91
Microwavable/quick-cook foods 8 0.71 0.78 0.95 0.81
Breadb 12 0.71 0.80 0.91 0.86
Wheat 5 0.77 0.76 0.95 0.74
White 7 0.66 0.83 0.87 0.79
Dry breakfast cerealc
Whole grain 1 -- -- -- 0.75
High sugar 1 -- -- -- 0.87
Low sugar 1 -- -- -- 0.77
Candy 5 0.79 0.87 0.94 0.97
Beverages 9 0.76 0.86 0.88 0.84
Regular sugar 6 0.74 0.86 0.86 0.78
Low sugar 3 0.82 0.85 0.93 0.89
Kitchen accessibility 12 0.74 0.74 0.95 0.75
Access to healthy foods 6 0.66 0.63 0.95 0.67
Access to unhealthy foods 6 0.83 0.84 0.96 0.86
Refrigerator accessibility 15 0.63 0.75 0.89 0.73
Access to healthy foods 9 0.59 0.70 0.91 0.72
Access to unhealthy foods 6 0.68 0.81 0.86 0.73
Obesogenic food availability score 71 0.73 0.83 0.91 0.94

aSpearman correlation between staff and participant report of foods present in home.
bPilot
form did not include croissant item.
cKappa statistic not available owing to multiple category response options for item.

major food category scores among the parents were signif- A, Vitamin C, fiber, and calcium were significantly and
icantly and positively correlated with the corresponding positively associated with their respective HFI major food
food group servings on the DHQ. Furthermore, Vitamin categories (e.g., calcium with the dairy score, Vitamin C

Page 6 of 10
(page number not for citation purposes)
International Journal of Behavioral Nutrition and Physical Activity 2008, 5:55 http://www.ijbnpa.org/content/5/1/55

Table 3: Correlations between HFI food categories and DHQ food serving and nutrient data (Sample 2, n = 342)

HFI major food category DHQ servings and nutrients Spearman Correlation p-value

Total number of dairy products Total number of dairy servings 0.15 <.01
Average calcium (mg) 0.16 <.01
Total number of vegetables (including potatoes) Total number of vegetable servings (including potatoes) 0.34 <.0001
Vitamin A (IU) 0.27 <.0001
Vitamin C (mg) 0.27 <.0001
Fiber (g) 0.20 <.001
Total number of vegetables (excluding potatoes) Total number of vegetable servings (excluding potatoes) 0.34 <.0001
Vitamin A (IU) 0.26 <.0001
Vitamin C (mg) 0.26 <.0001
Fiber (g) 0.20 <.001
Total number of fruits Total number of fruit servings 0.37 <.0001
Vitamin A (IU) 0.26 <.0001
Vitamin C (mg) 0.30 <.0001
Fiber (g) 0.26 <.001
Total number of meats & other nondairy protein Total number of non-dairy protein servings 0.23 <.0001
Protein (g) 0.13 <.01
Obesogenic food availability score Energy (kcals) 0.16 <.01

* Summative score that includes regular-fat versions of cheese, milk, yogurt, other dairy, frozen desserts, prepared desserts, savory snacks, added
fats; regular-sugar beverages; processed meat; high-fat quick, microwavable foods; candy; access to unhealthy foods in refrigerator and kitchen.

with the fruit score). Similarly, the obesogenic home the 24-hour recalls was not found to be associated with
availability was significantly and positively associated availability of related products in the home. The
with parental energy intake. As shown in Table 4, signifi- obesogenic home food availability score was significantly
cant, but attenuated, correlations were found between sev- and positively associated with adolescent energy intake.
eral of the HFI major food category scores and 24-hour
recall nutrient intake values among the adolescents (HFI Discussion
dairy score with recall dairy servings and recall calcium, The present study describes the validation of a home food
HFI vegetable scores with recall vegetable servings and inventory designed to include a wide range of foods that
recall Vitamin A, and HFI fruit score and recall Vitamins contribute to energy intake, including more and less
A, C, and fiber). However, for nearly half of the food healthful foods. The wide range of foods included on the
group categories and nutrients, dietary intake based on inventory speak to its content validity and study findings

Table 4: Correlations between HFI food categories and dietary recall food serving and nutrient data (Sample 2, n = 342)

HFI major food category Dietary recall servings and nutrients Spearman Correlation p-value

Total number of dairy products Total number of dairy servings 0.15 <.01
Average calcium (mg) 0.13 <.05
Total number of vegetables (including potatoes) Total number of vegetable servings (including potatoes) 0.16 <.01
Vitamin A (IU) 0.13 <.05
Vitamin C (mg) 0.10 .08
Fiber (g) 0.06 .28
Total number of vegetables (excluding potatoes) Total number of vegetable servings (excluding potatoes) 0.17 <.01
Vitamin A (IU) 0.13 <.05
Vitamin C (mg) 0.10 .08
Fiber (g) 0.05 .34
Total number of fruits Total number of fruit servings 0.07 .17
Vitamin A (IU) 0.11 <.05
Vitamin C (mg) 0.13 <.05
Fiber (g) 0.17 <.01
Total number of meats & other nondairy protein Total number of non-dairy protein servings 0.03 .56
Protein (g) 0.02 .68
Obesogenic food availability score* Energy (kcals) 0.13 <.05

* Summative score that includes regular-fat versions of cheese, milk, yogurt, other dairy, frozen desserts, prepared desserts, savory snacks, added
fats; regular-sugar beverages; processed meat; high-fat quick, microwavable foods; candy; access to unhealthy foods in refrigerator and kitchen.

Page 7 of 10
(page number not for citation purposes)
International Journal of Behavioral Nutrition and Physical Activity 2008, 5:55 http://www.ijbnpa.org/content/5/1/55

indicate substantial criterion, and construct validity for present study, all of the correlations between the HFI
the new inventory, particularly for adults. In addition, the major food categories and DHQ servings and nutrients
checklist type format is easily completed by research par- were statistically significant in the expected direction. Sev-
ticipants in their homes without undue response burden. eral correlations between the HFI major food categories
and child reported servings and nutrients were significant
The demonstrated criterion validity of the new home food but attenuated in comparison to those of their parents.
inventory as shown by high kappa, sensitivity, specificity Perhaps it should be expected that food availability and
values and high correlations between participants' and intake would be more similar from the same informant
staffs' reports of foods in the home suggests that the (in this case parents) either because he/she purchases
instrument could be used effectively for data collection by foods he/she prefers and eats. Another potential explana-
participants, thus, alleviating the need for staff home visits tion for the poorer construct validity for the youth is that
which are expensive, time-consuming and potentially the youth's dietary intake was assessed using 24-hour
intrusive. In comparing our criterion validity indices to recalls while the parents completed a food frequency that
previous research, kappa, sensitivity, and specificity values assessed usual food intake over a longer period of time.
appear to be similar to two previous studies reporting Further, our findings that an obesogenic food availability
these values [8,9], and substantially higher than those score for the household is significantly and positively
reported by Marsh and colleagues [10]. Raynor and col- associated with energy intake of both parents and adoles-
leagues [2] are the only previous investigators to demon- cents indicate that high fat foods available in the home
strate significant criterion validity of groupings of high-fat and captured on the inventory are potentially good start-
and low-fat foods by showing significant correlations ing points for public health messages for healthful eating.
between reports of two adults living in the household.
Our study examined several validity indices regarding staff The use of previous instruments and literature associated
and participant reports of regular fat and reduced fat ver- with energy consumption to determine the selection of
sions of dairy, added fats, frozen and prepared desserts, foods for the instrument makes the home food inventory
and savory snacks and demonstrated substantial validity useful for many purposes. Previous home food availabil-
as well, and extended the previous work by examining ity measures were developed for specific study objectives
subgroups of foods within major categories (e.g., cheese such as fruit/vegetable consumption or foods associated
within the dairy category) rather than grouping all low-fat with cancer. The broader selection of foods in our inven-
foods together. However, it should be noted that, tory increases its utility in nutrition- and obesity-based
although the comparison of staff and participant intervention programs.
responses is an accepted practice for measuring criterion
validity, this testing does not attest to the inventory's cap- There are several limitations that should be noted when
ture of all relevant foods, nor does it eliminate the possi- interpreting the results of the present study. The present
bility that participants may have altered their responses study did not assess test-retest reliability and therefore
since the research staff were in their homes. cannot address consistency of foods available in homes
over time. However, consistency of foods may be less of
Although all of the HFI major food groups and many of an issue given that Raynor and colleagues [2] conducted
the food subgroups showed substantial criterion validity, two-week test-retest reliability of the absolute number of
several of the food subgroups did not perform as well and high-fat and low-fat foods and showed substantial stabil-
deserve mention. In particular, the reduced-fat frozen des- ity. We also did not assess time since last shopping trip
sert, regular fat prepared dessert, and white bread catego- which could have influenced the home availability of per-
ries had lower than desired criterion validity. Our findings ishable items or preferred foods which may be consumed
regarding lower validity for prepared desserts is similar to more readily [6]. However, our significant correlations
that found in previous research [8]. Several anecdotal between the HFI scores and the DHQ food servings and
observations early in the study indicate that the wide vari- nutrients indicate that perhaps this potential confounder
ety of dairy, soy and other frozen desserts available in the was not influential. Another potential limitation is that
marketplace may make it difficult to assess nutrient and the new home food inventory does not assess quantity;
fat content. Similar observations were made for prepared participants either check "yes" if the food is present in the
desserts. In addition, the proliferation of whole grain home or "no" to indicate that the food is not present in
white breads and light wheat breads available today con- the home. Accordingly, a household may "score" high on
fuse participants and staff alike when categorizing bread the number of fruits and vegetables even when quantity is
types. limited or low on the number of fruits and vegetables
when quantity is high for only a few foods. In addition,
Our findings regarding construct validity are similar to the list of foods is not an exhaustive list of all possible
those reported by Raynor and colleagues [2]. In the foods contributing to the obesity epidemic; however, in

Page 8 of 10
(page number not for citation purposes)
International Journal of Behavioral Nutrition and Physical Activity 2008, 5:55 http://www.ijbnpa.org/content/5/1/55

selecting foods to be included on the inventory, we bal- and edited the manuscript; CH and KP assisted in scale
anced the number and types of foods with response bur- score development and conducted data analyses. All
den. Our goal was to create an inventory that was simple authors have given final approval of this manuscript.
and quick to complete. Attempting to collect data on
more foods, quantities of foods, or more specifics about Additional material
foods such as brand, would have impacted the time and
complexity of completing the inventory and added to
response burden. Moreover, our construct validity testing Additional file 1
A table reflecting which foods are included in each food group/subgroup.
indicated that the foods measured on the inventory were Click here for file
significantly associated with energy intake from other [http://www.biomedcentral.com/content/supplementary/1479-
measures, suggesting that we have captured a significant 5868-5-55-S1.doc]
amount of the variation in the adult diet. Furthermore,
the sample used for construct validity (Sample 2) over
represented educated, Caucasian adults and findings may
not generalize to less educated or minority populations. Acknowledgements
We would like to thank Anne Samuelson who coordinated and collected
The present study had several strengths. It is one of the few data; and Jeanna Rex, Rachel Cope, and Dawn Nelson who collected data.
measures of the home food environment that has under- We would also like to thank the parents and students in the study for their
time and willingness to share their personal lifestyles. The study was funded
gone criterion and construct validity testing, and it also
as part of the IDEA study (PI: Leslie Lytle, PhD) funded by NCI's Transdis-
has content validity for a broad range of foods that may be ciplinary Research in Energetics and Cancer Initiative (NCI Grant 1 U54
useful for assessing the obesogeneity of the home food CA116849-01, Examining the Obesity Epidemic Through Youth, Family,
environment. In particular, the criterion validity in which and Young Adults, PI: Robert Jeffery, PhD).
staff visited participants' homes to assess food availability
was strong as was the construct validity between the HFI We thank Data Collection and Support Services (DCSS) in the School of
and the adult dietary intakes. Public Health at the University of Minnesota for their forms design, data
entry, and quality control work for the IDEA study.

Conclusion
This new home food inventory is a valid and participant- References
1. Kratt P, Reynolds K, Shewchuk R: The role of availability as a
friendly tool to assess foods in the home, and may be use- moderator of family fruit and vegetable consumption. Health
ful for community-based behavioral nutrition and obesity Educ Behav 2000, 27(4):471-482.
2. Raynor HA, Polley BA, Wing RR, Jeffery RW: Is dietary fat intake
prevention research. The inventory builds on previous related to liking or household availability of high- and low-fat
measures by including a wide range of foods (both health- foods? Obes Res 2004, 12(5):816-823.
ful and less healthful) rather than foods targeted for a spe- 3. Cullen KW, Baranowski T, Owens E, Marsh T, Rittenberry L, de
Moor C: Availability, accessibility, and preferences for fruit,
cific intervention, as well as both perishable and 100% fruit juice, and vegetables influence children's dietary
nonperishable foods. Thus, the inventory may be useful in behavior. Health Educ Behav 2003, 30(5):615-626.
studies examining the contextual influences on obesity, 4. Campbell KJ, Crawford DA, Salmon J, Carver A, Garnett SP, Baur LA:
Associations between the home food environment and obes-
weight gain, and nutritional intake. It might also be help- ity-promoting eating behaviors in adolescence. Obesity (Silver
ful in studies determining appropriate intervention strate- Spring) 2007, 15(3):719-730.
5. Gattshall ML, Shoup JA, Marshall JA, Crane LA, Estabrooks PA: Vali-
gies for individual households or might identify targets dation of a survey instrument to assess home environments
for public health messages. Moreover, the present study for physical activity and healthy eating in overweight chil-
findings indicate that the inventory has both construct dren. Int J Behav Nutr Phys Act 2008, 5:3.
6. Bryant M, Stevens J: Measurement of food availability in the
and criterion validity, validity indices not typically home. Nutr Rev 2006, 64(2 Pt 1):67-76.
assessed for the same instrument. 7. Cronbach LJ: Test validation. In Educational Measurement 2nd edi-
tion. Edited by: Thorndike RL. Washington, DC: American Council on
Education; 1971.
Competing interests 8. Crockett SJ, Potter JD, Wright MS, Bacheller A: Validation of a self-
The authors declare that they have no competing interests. reported shelf inventory to measure food purchase behav-
ior. J Am Diet Assoc 1992, 92(6):694-697.
9. Miller C, Edwards L: Development and validation of a shelf
Authors' contributions inventory to evaluate household food purchases among
JF conducted the literature review, drafted the items for older adults with diabetes mellitus. J Nutr Educ Behav 2002,
34(5):261-267.
the Home Food Inventory (HFI), trained staff, collected 10. Marsh T, Cullen KW, Baranowski T: Validation of a fruit, juice,
data for field testing and criterion validity testing, con- and vegetable availability questionnaire. J Nutr Educ Behav
2003, 35(2):93-97.
ducted analyses of the pilot data, and drafted the manu- 11. Crockett SJ, Mullis RM, Perry CL: Parent nutrition education: a
script; MN and LL assisted in finalizing items for the HFI, conceptual model. J Sch Health 1988, 58(2):53-57.
assisted in scale score development, and edited the manu- 12. Lytle L: Testing an ecological approach to examine the etiol-
ogy of childhood obesity: The IDEA study. .
script; SM trained staff, coordinated and collected data,

Page 9 of 10
(page number not for citation purposes)
International Journal of Behavioral Nutrition and Physical Activity 2008, 5:55 http://www.ijbnpa.org/content/5/1/55

13. Widome R, Forster JL, Hannan PJ, Perry CL: Longitudinal patterns
of youth access to cigarettes and smoking progression: Min-
nesota Adolescent Community Cohort (MACC) study
(2000–2003). Prev Med 2007, 45(6):442-446.
14. Patterson RE, Kristal AR, Shannon J, Hunt JR, White E: Using a brief
household food inventory as an environmental indicator of
individual dietary practices. Am J Public Health 1997,
87(2):272-275.
15. Nielsen SJ, Siega-Riz AM, Popkin BM: Trends in food locations and
sources among adolescents and young adults. Prev Med 2002,
35(2):107-113.
16. Lytle LA, Nichaman MZ, Obarzanek E, Glovsky E, Montgomery D,
Nicklas T, Zive M, Feldman H: Validation of 24-hour recalls
assisted by food records in third-grade children. The
CATCH Collaborative Group. J Am Diet Assoc 1993,
93(12):1431-1436.
17. Schakel SF, Sievert YA, Buzzard IM: Sources of data for develop-
ing and maintaining a nutrient database. J Am Diet Assoc 1988,
88(10):1268-1271.
18. Thompson FE, Subar AF, Brown CC, Smith AF, Sharbaugh CO, Jobe
JB, Mittl B, Gibson JT, Ziegler RG: Cognitive research enhances
accuracy of food frequency questionnaire reports: results of
an experimental validation study. J Am Diet Assoc 2002,
102(2):212-225.
19. Subar AF, Thompson FE, Kipnis V, Midthune D, Hurwitz P, McNutt S,
McIntosh A, Rosenfeld S: Comparative validation of the Block,
Willett, and National Cancer Institute food frequency ques-
tionnaires : the Eating at America's Table Study. Am J Epide-
miol 2001, 154(12):1089-1099.
20. Diet History Questionnaire [http://riskfactor.cancer.gov/DHQ/]

Publish with Bio Med Central and every

scientist can read your work free of charge
"BioMed Central will be the most significant development for
disseminating the results of biomedical researc h in our lifetime."
Sir Paul Nurse, Cancer Research UK

Your research papers will be:

available free of charge to the entire biomedical community
peer reviewed and published immediately upon acceptance
cited in PubMed and archived on PubMed Central
yours — you keep the copyright

Submit your manuscript here: BioMedcentral

http://www.biomedcentral.com/info/publishing_adv.asp

Page 10 of 10
(page number not for citation purposes)