FOOD INSECURITY AND CHILDREN WITH SPECIAL HEALTH

CARE NEEDS

By

Mary Kristine Travis

A THESIS

Presented to the Department of Dietetics and Nutrition

and the Oregon Health & Science University

School of Medicine

in partial fulfillment of

the requirements for the degree of

Master of Science

October 2008
 School of Medicine
Oregon Health and Science University

___________________________________

CERTIFICATE OF APPROVAL

___________________________________

This is to certify that the Master’s thesis of

Mary Kristine Travis

has been approved

____________________________________
Mentor/Advisor

____________________________________
Member

____________________________________
Member

ii
 TABLE OF CONTENTS

Page
List of Tables and Figures…………………………………………………………….. v

Acknowledgements…………………………………………………………………… vi

Abstract……………………………………………………………………………….. vii

CHAPTER

I. Introduction……………………………………………………………….. . 1

Problem and Significance 1

Food Insecurity…..…...………………………………………………. 2
Food Insecurity Definitions...………………………………… 2
Trends in Food Insecurity…………………………………….. 3
Predictors of Food Insecurity…………………………………. 4
Adverse Effects of Food Insecurity on Child Health and
Development………………………………………….. 5
Food Insecurity and Obesity………………………………...... 10
BMI Z Score………………………………………………….. 12
Identifying Food Insecurity...……….……………………………….. 13
Assessment of Household Food Security…………………….. 13
USDA 6-Item Subset…………………………………………. 14
Clinical Screening Tests………………………………………. 15
Screening at Primary Care Level……………………………... 16
Receiver Operating Characteristic Curves for Assessment
of Screening Tools……………………………………. 18
Children with Special Health Care Needs……………………………. 19
Definition and Description……………………………………. 19
Financial Impact of Children with Special Health Care
Needs…………………………………………………. 20
Impact on the Child 22
Identifying Children with Special Health Care Needs..……………… 22
Relationship between Food Insecurity and Children
with Special Health Care Needs……………………… 23
Tillamook County Preschool Screening Description and
Demographics……………………………………………….... 24
Research Questions and Specific Aims……………………………… 25

II. Methods…………………………………………………………………… 27
Study Design………………………………………………….. 27
iii
 Study Population……………………………………………… 28
Inclusion/Exclusion Criteria…………………………………… 28
Data Management and Definition of Dependent Variable…….. 28
Recoding of Variables…………………………………………. 29
Descriptive Characteristics…………………………………….. 31
Receiver Operating Characteristic…………………………….. 31
Chi Square and Multivariate Logistic Regression...................... 31

III. Results….………………………………….................................................. 33
Study Demographics…………………………………………… 33
Income and Education…………………………………………. 33
Food Insecurity and Children with Special Health Care
Needs………………………………………………….. 36
Receiver Operating Characteristic Analysis…………………... 39
Chi Square Analysis…………………………………………… 40
Univariate Logistic Regression Analysis……………………… 44
Multivariate Regression Model………………………………... 47

IV. Discussion..……………………………………………………………….. 49
Prevalence of Food Insecurity and Children with
Special Health Care Needs……………………………. 49
Single Question Screening……………………………………. 51
Predictors of Food Insecurity…………………………………. 55
Multivariate Predictors of Food Insecurity…………………… 56
Limitations……………………………………………………. 57
Contributions…………………………………………………. 58
Future Directions……………………………………………... 58

V. Summary and Conclusions………………………………………………... 60

References…………………………………………………………………………….. 61

Appendices

A. Multi Modular Preschool Screening: Survey of Household Food Needs… 71

B. USDA 6-Item Subset (Short Form) of the 12-Month Food Security Scale
-Questionnaire………………………………………………………… 79
C. Children with Special Health Care Needs (CSHCN) Screener©....………. 81

iv
 LIST OF TABLES AND FIGURES

TABLE PAGE

1. Trends in Food Insecurity in the United States and Oregon, 1998-2006………….. 4

2. Demographic Profile of Study Sample..... ………………………………………… 34

3. Food and Income Behaviors, Food Security and Children with Special Health

Care Needs Status……………………………………………………………. 37

4. Frequency of Responses to Food Security Questions……………………………... 38

5. Receiver Operating Characteristic Results………………………………………… 39

6. Chi Square of Household Demographic Characteristics by Household Food

Security Status……………………………………………………………….. 42

7. Chi Square of Selected Household Demographic Characteristics by Having

a Child with Special Health Care Needs……………………………………... 43

8. Univariate Logistic Regression Associations between Household Food

Insecurity and Household Characteristics ………...…………………………. 45

9. Associations between Food Insecurity and Family Characteristics,

Multivariate Logistic Regression Model……………………………………… 48

FIGURE

1. Receiver Operating Characteristic Plot……………………………………………. 41

v
 ACKNOWLEDGEMENTS

I would like to thank my advisor, Dr. Liz Adams, for her time, guidance and support.

Thank you for your flexibility with project changes and long distance mentoring, and for

always being supportive. I would also like to thank my committee members, Dr. Dana

Hargunani and Dr. Victoria Warren-Mears, for their time and encouragement.

Thanks to my parents, for supporting me in my cross-country move and in following my

dreams. I love you both so much.

I have many many thanks for my fiancé Deacon. Thank you for allowing me to voice my

irrational fears in the thesis writing process, for reminding me that thousands of people

do this every year and for always supporting me. I truly could not have done this without

you.

vi
 Abstract

Food Insecurity and Children with Special Health Care Needs

by

Mary Kristine Travis

Master of Science in Clinical Nutrition

Oregon Health and Science University

Fall 2008

Background: Food insecurity (FI) is the “limited or uncertain availability of nutritionally

adequate and safe foods or limited ability to acquire acceptable foods in socially

acceptable ways.” In 2006, 15.6% of households with children in the United States were

classified as being food insecure, including 12.6 million children. Many adverse effects

are seen in food insecure children; spanning physical, developmental, cognitive, and

social realms. Screening for childhood food insecurity is crucial to identify and provide

aid for those in need but is not often done in primary care settings.

Children with Special Health Care Needs (CSHCN) are “those who have a chronic

physical, developmental, behavioral, or emotional condition and who also require health

and related services of a type or amount beyond that required by children generally.”

Having CSHCN puts an additional financial burden on the household, increasing the

likelihood for food insecurity. CSHCN are already at a higher risk for nutrition related

problems, and coupled with food insecurity, the risk for adverse effects may be much

greater for this group.

vii
Research Questions: 1. Is a single question from the USDA 6-Item Subset (Short Form)

of the 12-month Food Security Scale adequate for use as a screening tool for food

insecurity? and, 2. What are the household and family characteristics that predict

household food insecurity in the study population?

Methodology: A survey was administered to parents of pre-school children attending the

annual Multi-Modular Pre-School Health Screening in Tillamook, OR, with questions

pertaining to food security, CSHCN, and household/family characteristics. Statistical

analysis was performed using receiver operating characteristic (ROC) plots, chi square

analysis and multivariate logistic regression. SPSS 15.0 software was used.

Results: Question 1 from the USDA 6-Item Subset (Short Form) of the 12-month Food

Security Scale resulted in the highest area under the ROC curve, the highest sensitivity,

and an acceptable measure of specificity. Question 1 would be a valid single question

screening tool for identifying food insecure individuals. The predictors of household

food insecurity adjusted for covariates were having a CSHCN in the family, having a

family income below $25,000, and using no-cost food (such as from a food bank).

Significance: The results will be shared with Tillamook County to help with

identification of the predictors of household FI and with planning strategies to address the

prevalence of food insecurity. Determining a valid screening tool is the first step towards

implementing screening in primary care, in order to identify and provide support to food

insecure families at a community level.

viii
 Chapter I

Introduction

Problem and Significance. Childhood is a critical time for growth and development,

and nutritional demands for a healthy child are high. Food insecurity, or not having

enough safe, nutritious food, is a persistent problem that affects children as well as adults

in the United States (US) every year. Lack of food or lack of variety of food poses a risk

for children in their physical, mental and psychological growth.1-11 The US has long been

aware of the pervasiveness of food insecurity in this country, and the goal of Healthy

People 2010 Objective 19-18 is to reduce the incidence of food insecurity to six percent

of the population.12 Household food insecurity currently affects 10.9% of households in

the US and 11.9% of households in Oregon, indicating that there is much progress to be

made.13 Additionally, as households face economic challenges such as rising costs or

unemployment, the prevalence of food insecurity is likely to rise.

Assessment of household and child food insecurity is carried out regularly in large

national surveys at the population level to determine the prevalence of food insecurity,

but is rarely done in primary practice settings.14,15 Reasons for not asking about food

insecurity in primary care settings are numerous, including time constraints and lack of

knowledge of food insecurity.16,17 However, identifying children in primary care settings

who are living in food insecure households may be an optimal avenue for follow up with

the family to obtain appropriate services and programs to combat the food

insecurity.5,15,16 Identifying food insecurity in households with children in this setting

1
may also lead to a more in-depth assessment by the physician or other health care

provider of food insecurity-related issues, such as iron-deficiency anemia.7 Developing a

reliable one-question screening tool is the first step toward reducing the prevalence of

childhood food insecurity in the primary care setting.

Children with special health care needs are those with any physical,

developmental, behavioral, or emotional condition and are already at an increased

nutritional risk.18,19 The families of these children may be at greater risk for food

insecurity due to high out-of-pocket costs for treatment and management of the

condition(s) of the children.20-23 Children with special health care needs may be

especially susceptible to the adverse effects of food insecurity. More information about

this specific population and their experiences with food insecurity is needed to make a

crucial step towards resolving food insecurity in the US.

Food Insecurity.

Food Insecurity Definitions. Food insecurity (FI) is defined as the “limited or uncertain

availability of nutritionally adequate and safe foods or limited or uncertain ability to

acquire acceptable foods in socially acceptable ways.”13 In short, it describes people and

families who do not have enough food from month to month, or at times throughout a

given year. Conversely, food security (FS) is “access by all people at all times to enough

food for an active, healthy life.”13 Beginning in 2006, the United States Department of

Agriculture (USDA) introduced new labels to define varying levels of food insecurity.13

2
The old terms “food insecurity without hunger” and “food insecurity with hunger” are

now referred to as “low food security” and “very low food security,” respectively. The

classifications are the same, such that previous and future reports are comparable. Low

food security designates those who experience decreases in diet quality and variety rather

than quantity, while very low food security delineates those who experience “disrupted

eating patterns” and decreases in food quantity.13 For the purposes of this paper, the term

food insecurity will encompass both “low food security” and “very low food security”

unless otherwise stated. Child food insecurity is a term used throughout this thesis. It is

assessed by the USDA 18 question Household Food Security Survey Module (HFSS).13

The last seven questions in the HFSS refer to the children of the household specifically,

and by assessing responses to these questions, national percentages of child food

insecurity can be determined.

Trends in Food Insecurity. Since 1998, the national prevalence of household FI has

ranged between 10.5% and 11.9%; currently it is 10.9% and very low food security, the

more severe form, is at 4.0%.24 The percentage of children who are food insecure

nationally is much higher, ranging from 17.2% to 19.7% from 1998 to 2006.24 Currently

the percentage of children who are food insecure is 17.2%, the lowest since 1998.24 Food

insecurity in Oregon has for many years been above the national average though recently

the prevalence has decreased to near the national percentage (see Table 1). The reasons

for this are not well understood, as strictly income or poverty level information does not

3
Table 1

Trends in Food Insecurity in the United States and Oregon, 1998-200624,25

Household Food Insecurity Child Food Insecurity

Year
US Rate (%) OR Rate (%) US Rate (%)

1998 11.8 12.3 19.7

2000 10.5 13.7 18.0

2002 11.1 12.9 18.1

2004 11.9 11.9 19.0

2006 10.9 11.9 17.2

explain the discrepancy, although action by anti-hunger advocacy groups may have

helped to reduce FI.25 Though the level of food insecurity in Oregon has neared the

national level, the state of Oregon still has the need for better services and programs to

continue to decrease the prevalence to be equal to or below the national average.

Predictors of Food Insecurity. Characteristics that predict household food insecurity

have been described in analyses of data from national surveys including the Current

Population Survey, the National Health and Nutritional Examination Surveys (NHANES)

and various others. Food insecurity disproportionately affects single mother-headed

households (30.8%), African American households (22.4%), Hispanic households

(17.9%), and those living below the poverty line (36%).13 Additional household

4
demographic and socioeconomic predictors of FI include: presence of children,13,26-28

adults with less than 12th grade education,26-28 having a disabled person in the

household,28 renting rather than owning a home25 and living in principal cities of

metropolitan areas or living in rural areas.13 Understanding the predictors associated with

FI is important in being able to identify and assist families likely experiencing food

insecurity, gaps in public services, and may suggest methods of preventing long and short

term food shortages in families.

The problem of food insecurity affects adults and children alike, and is not

necessarily restricted to the homeless or to those below the poverty line.13 In 2006, 11%

of households and 15.6% of households with children were classified as having low to

very low food security,13 which means a substantial number of children are exposed to

household food insecurity, or are food insecure themselves. As the severity of poverty

increases, the presence and severity of low food security increases as well, but there is

also a small percentage of households living above the poverty line that report the

presence of food insecurity, and not all impoverished households experience low food

security.13 Thus food security or lack of security cannot be predicted solely by income;

multiple factors are associated with food insecurity.

Adverse Effects of Food Insecurity on Child Health and Development. In 2006, 12.6

million children were classified as having low food security, and of those, 3.4 million had

very low food security.13 A variety of adverse effects of food insecurity have been

described among children, including physical, developmental, cognitive, and social

effects. Through mainly cross sectional studies with a sample size of several thousand,
5
the main categories of adverse health effects described in food insecure children are:

chronic health conditions,3-7 ratings of fair/poor health by their parents or caregivers,4-6,8,

34
poor diet,29-33 psychosocial functioning,1,2,5,9,10 behavior problems2,5,9-11,35 and academic

issues.1,2,5

Poor Health

Adverse health effects experienced by food insecure children are often illustrated

by using health ratings by a caregiver, by number of hospitalizations, and by frequency of

ailments. Alaimo and colleagues have done extensive research on food insecurity and its

effects on children, using data from the third National Health and Nutritional

Examination Survey (NHANES III). Common to large surveys is the health rating

question, where a caregiver is asked to rate the child’s health as “excellent,” “very good,”

“good,” “fair,” and “poor.” Ratings of fair and poor are often grouped together to show a

low level of health. Among children who live in food insecure households, this study and

others4,6 found that preschool aged and school-aged children were rated by their parents

to be in fair or poor health more often than their food secure peers.3 Furthermore, the

greater the severity of poverty combined with FI, the poorer the children were rated in

overall health by their caregiver. The food insecure children were found to have more

stomachaches, more headaches, and more colds, after adjusting for sociodemographic

and family characteristics and health risks. Cook and colleagues supported the higher

incidence of ailments in food insecure children in their research, and showed the trend

was also the same for lifetime hospitalizations, though to a lesser degree.4 Food insecure

and hungry children have been found to be absent from school more often,34 which may
6
contribute to the link between FI and poor school performance. These findings depict just

a few of the many adverse health effects that FI has on children.

Behavior and Psychological Effects

Behavior problems and lower psychosocial functioning are adverse effects found

to be associated with food insecurity in children. The common method for determining

these outcomes is by survey either of the caregivers or of the teacher. Surveys such as

the Pediatric Quality of Life Initiative, Child Behavior Checklist (CBCL), and Social

Skills Rating System are examples of tools that might be used to assess behavior or

psychosocial functioning. Studies using these and other surveys found that food insecure

children were more likely to have at least one of the following behavior problems:

aggressiveness, feeling anxious/depressed, or displaying attention deficit,11 and they had

impaired social skills1 and lower psychosocial function.35 Food insecure children were

also more likely to have externalizing and internalizing behavior problems, 9 and be

classified as dysfunctional.10 Lower scores on the CBCL and the Children’s Global

Assessment Scale (measures overall functioning) were also found to be more likely in

this population when compared to their food secure peers.34

Alaimo, et al, (2001) analyzed NHANES III data, with sample sizes of 3286

children aged 6-11 and 2063 adolescents aged 12-16, looking at cognitive, academic, and

psychosocial performances of the surveyed children. They found that food insecure

children aged 6-11 were more likely to have seen a psychologist (Odds Ratio [OR] 1.89,

P ≤ 0.05), food insecure children aged 12-16 were also more likely to have seen a

7
psychologist (OR 1.82, P ≤ 0.05), ever have been suspended (OR 1.95, P ≤ 0.05) and had

difficulty getting along with others (OR 1.74, P ≤ 0.05).2 The same researchers analyzed

the NHANES III data for associations between depression and food insecurity, and found

that adolescents who were food insecure were four times more likely to have had

dysthymia, two times more likely to have had thoughts of death, three and a half times

more likely to have had the desire to die, and five times more likely to have attempted

suicide compared to their food secure peers when adjusted for various household

characteristics; though these odds ratios were not significant at the 0.05 level.36 It is

clear, however, from the studies noted here, that food insecurity has varied and broad

negative effects on the psychosocial, psychological and behavioral aspect of children’s

health.

Academic Performance

Food insecurity and its association with school performance is a key query of

many studies involving children and FI. A study by Jyoti et al, (2005) assessed children

in kindergarten and then again when they reached the third grade. The researchers found

that children who were food insecure throughout had smaller gains in math and reading

scores than their food secure peers, and those transitioning from food secure to food

insecure had much smaller gains in reading scores.1 Other studies reported lower math

scores in food insecure children,2 and increased math scores when breakfast was given to

at risk children.37 Studies often do not reveal a significant relationship between FI and

school performance. This may be due, in part, to inadequate sample size and multiple

8
confounding factors, but evidence suggests a trend of decreased school performance with

increased food insecurity, and improvements in performance with adequate nutrition.

Diet Quality

Analysis of the diets of children is a helpful, yet complicated method for showing

the potential diet inadequacies of children experiencing food insecurity. Deficiencies in

nutrients over extended periods of time can have adverse effects on a child’s health and

development. Furthermore, excesses in nutrients like saturated fat and cholesterol from

cheaper foods may adversely affect a food insecure child into adulthood by causing

conditions such as high cholesterol and atherosclerosis. Most studies investigating FI use

data from large-sample surveys that have limited data on actual or reported dietary

intakes. Using the Continuing Survey of Food Intakes by Individuals (CSFII), Casey, et

al, (2001) were able to obtain a sample size over five thousand children, as well as two

24-hour diet recalls for each, reported by the children six years and older, with parental

help, and by the parents of children five years and younger.8 With this data, nutrient

intakes were available for analysis in addition to health and development factors. The

children were categorized according to household income level as well as food security

status. The low income, food insecure children had significantly lower total energy and

carbohydrate intakes, and higher cholesterol intakes than their higher income, food secure

peers. In addition, food insecure children across all age groups reportedly ate fewer dark

green vegetables, nuts, fruits, and yogurt, and ate more eggs, dry beans, and peas than the

food secure group.8

9
 A study of adult diets in lower income counties in Mississippi, Louisiana, and

Arkansas found similar results. Prior to Hurricane Katrina, one 24 hour food recall was

taken over the phone, as well as the 18-Item HFSS, and results demonstrated that food

insecure adults scored lower on the Healthy Eating Index (HEI) (P<0.0001) and

consistently achieved intakes that fell further below the Dietary Reference Intakes than

food secure individuals.38 Particularly low were intakes of vitamin A, copper, and zinc

with significant differences between FI and FS adults (P< 0.01). The findings of poorer

diets when looking at food insecure individuals may be a partial explanation for the array

of symptoms that are related to food insecurity found in children, such as poor health and

behavior.

Food Insecurity and Obesity. There has been much debate recently on the topic of

weight and its associations with childhood FI. Results from large studies range from

showing evidence that food insecure children are less likely to be overweight or obese31,39

to evidence that the children are more likely to be overweight.30,40,41 Other studies show

no significant associations at all, or significant associations within some categories of

age, gender or race/ethnicity but not others.1,3,7,33 The analysis of weight associations

with FI is complicated by many variables, including income, age, race/ethnicity, the

presence or absence of hunger, and comparisons between mild to severe FI, all of which

can contribute to confounding and decreased statistical significance.

Casey, et al, (2006), analyzed body mass index (BMI) data from NHANES

surveys, with a sample size of 6995 children aged 3-17. The 18 question HFSS was used

to identify FI, and the researchers evaluated both overweight status (BMI ≥ 95%) and at
10
risk for overweight (BMI ≥ 85%).30 The researchers found that children in food insecure

households and children that were child food insecure were more likely to be both

overweight and at risk for overweight at P< 0.01. Broken down into groups by gender

and age, the analyses again exhibited that children in food insecure households and those

that were child-food insecure were more likely to be overweight and at risk for

overweight, or in some cases, trends in the data displayed findings in these directions

despite not reaching statistical significance. In multivariate regression, however,

household FI was not statistically significantly associated with overweight or at risk for

overweight, although child FI was significantly associated with at risk for overweight.30

A separate study of 1514 preschoolers found that children in the study population were

3.4 times more likely to be overweight (P < 0.05) if they lived in food insecure houses,

even after adjusting for variables such as birth weight, parents’ weight status, income and

education.41 This study used one question to assess FI, and followed the children from

birth to age 4.5 years, allowing for plentiful data collection.

Rose and Bodor (2006) found nearly the opposite associations between food

insecurity and child overweight status. The odds ratio (OR) for overweight in food

insecure children was 0.80, indicating that “after controlling for other possible

confounders, children from food-insecure households were 20% less likely to be

overweight.”39 The total sample size was 12890 children assessed twice, once in

kindergarten and again in first grade in the Early Childhood Longitudinal Study,

Kindergarten Cohort. This survey used the 18 question HFSS to assess FI status.

Interestingly, in addition to the above result, the researchers found a significant inverse

11
relationship between income and overweight status.39 Though there was no mention in

the article of the effect of the age of the children, since all the subjects were aged five to

six at their second assessment, it may be that this cohort was simply too young to be

displaying the BMI/weight consequences of FI.

Other studies have found a mix of results. For example, Alaimo, et al, (2001)

used the NHANES III data and found no associations in the two through seven year age

group between food insecurity and overweight. Results suggested higher levels of

overweight in non-Hispanic white girls aged 8-16 years, although the P value was not

significant.42 Jyoti, et al, (2005) used data from the Early Childhood Longitudinal Study,

Kindergarten Cohort and found significant associations between FI and increased BMI in

females only.1 The trend in research suggests a positive association between food

insecurity and childhood overweight, but currently, no definitive statement can be made

due to the multitude of non-significant findings.

BMI Z Score. Measurements of anthropometrics for children, such as height and weight,

change in meaning with age and gender, in that one measurement will be interpreted

differently depending on whether the child is a nine year old female or a four year old

male. Anthropometric data are most meaningful when compared to national standards

such as the National Center for Health Statistics (NCHS). Use of anthropometric Z-

scores for age and sex allows the comparison across age and sex categories. Z scores

also have a normal distribution, which is better for statistical analysis. Z scores of -2.0 to

2.0 encompass 95% of the population in the normal distribution, and are the cuff-off

points for the range of “normal.”43 In regards to child BMI, above a Z score of 2.0

12
indicates overweight and a score below -2.0 indicates underweight. Z scores are a

method for accurately analyzing and reporting pediatric BMI data, and are therefore often

used in studies.

Identifying Food Insecurity.

Assessment of Household Food Security. Since the 1980’s the American public has

been aware of and concerned about hunger in the US; the concept of food insecurity has

been defined since 1990.13 The awareness of hunger led to the development of numerous

methods for identifying hunger and FI in the general population. To assess FI the most

recognized instrument is the 18 question Household Food Security Survey (HFSS). This

has been developed and utilized by the USDA and has been contained in the Current

Population Survey (CPS) since 1995.13 Testing of the HFSS began with a cognitive

assessment and field test by the US Census Bureau. It has also been tested extensively

since its conception by the US Census Bureau, Mathematica Policy Research, Inc, IQ

Solutions and numerous other independent agencies to assess its validity, reliability and

applicability across various household types.13 The questions range from asking about

the uncertainty of having enough food for the members of the household, to the children

of the household missing meals because the parents are unable to provide food. This

questionnaire can be used to assess the severity of food insecurity from fully food secure

to the most severe food insecure level, low food security among the children.

13
 NHANES is another USDA survey and includes a variety of questions pertaining

to health and nutrition. Since 1999, NHANES has contained the 18 question HFSS.

Numerous food security studies have used NHANES data for analysis.1,4,6,11,29,30 The

HFSS is used in a variety of surveys, for both national and smaller populations. Other

surveys that use the HFSS include the Continuing Survey of Food Intakes by Individuals

(CSFII), which is a USDA survey that includes 24 hour recalls of food intake;13 and the

Food Research and Action Center’s Community Childhood Hunger Identification Project

(CCHIP), which was the first survey designed to provide data on food insecure families

with children.44

USDA 6-Item Subset. The USDA also utilizes a 6-Item Subset (Short Form) of the 18

question HFSS, which is a shorter version of the form used in the Current Population

Survey (See Appendix B). The short form was evaluated for effectiveness, and was

found to be accurate and reliable for classification of food security status.45,46 The six

questions come directly from the full 18-question model, and include questions such as,

“the food we bought just didn’t last and we didn’t have money to get more, was that

often, sometimes or never true?” and “in the last 12 months, did you ever eat less than

you felt you should because there wasn’t enough money to buy food?” with the responses

being yes, no, or I don’t know.47 The severity of the questions increases as they are read,

in that an affirmative answer to the last question would indicate a more dire food insecure

situation than an affirmative answer to the first. There are three main advantages of this

shortened form: (1), a shorter time response burden for families can be useful for

inclusion in longer surveys or for screening criteria, (2), it is able to provide prevalence

14
estimates of low and very low food security with minimal bias relative to those on the 18-

item module, and (3), because it was developed from the full version, the six-item subset

is standard and can be compared and related to the longer survey.13

Due to its condensed form, there are some disadvantages to use of this form over

the 18-item questionnaire. This survey is less precise and somewhat less reliable than the

full module. It cannot measure the most severe levels of food insecurity, and does not

contain questions pertaining to child food insecurity and hunger, rather all questions ask

about the entire household.13 Despite these drawbacks, this survey instrument has been

employed by numerous studies and has been well tested.13,45,46,48

Clinical Screening Tests. Clinical screening may be used to identify food insecurity on

an individual level. At this level, assessment of FI is not for descriptive or census

purposes. Instead it is used to identify individuals and families experiencing FI in order

to provide information about programs providing income and food assistance. A second

purpose of clinical screening for FI is to address problems related to FI. Use of a one

question screening tool has been assessed to detect the prevalence of hunger in

households by Kleinman et al.49 Hunger is different from low or very low food

insecurity; it is “an individual-level physiological condition that may result from food

insecurity.”13 A one question written screening form was given to all parents in primary

care pediatric clinic, and responses were compared to the 18 question HFSS that a sub

group of participants completed during in-depth interviews. The one question to detect

hunger had an 83% sensitivity and an 80% specificity, which the authors said represents

an accurate and reliable screening tool.49 Though the focus of this study was on hunger,
15
the results demonstrated that it is possible to implement a one question screening tool in a

primary care setting. This study showed that it is also possible to ask briefly about food

insecurity as part of the routine for health clinics serving low-income populations.

Screening at Primary Care Level. It is unclear how much screening for FI occurs at the

level of hospitals and clinics for general pediatric care, although anecdotal reports

indicate that access to food or food insecurity is rarely addressed in primary care.14,15

Screening for FI at this level may be crucial to reducing the prevalence of FI, in that

clinicians can make resources and services known to those experiencing FI. Granger and

Holben (2004) state, “Physicians are well positioned to be aware of food insecurity

among their patients before health and nutritional deterioration occurs and play a critical

role in improving food security in the United States,” but also concede that this is rarely

done.15 Not all food insecure families make use of governmental and nongovernmental

food assistance programs,49 and screening at the primary care level may serve to increase

use of the programs and decrease prevalence of FI. Many articles investigating FI

concluded with phrases such as, “this study supports the need to educate [clinicians]

about food insecurity and incorporation of food security practices into the healthcare

process.”16 Another study concluded with saying, “healthcare providers, along with their

office staff, have an opportunity to reach eligible families.”5 However no research was

found addressing the actual implementation of food insecurity screening practices in

clinical settings.

A study assessing the knowledge and practices of nurse practitioners regarding

food insecurity found that 47% of nurses surveyed were “not knowledgeable enough to
16
identify a source [of their knowledge]” about FI, and only 33% strongly agreed or agreed

that they were familiar with the topic of FI.16 The authors reported that the survey was

optional (sent by mail), and only those concerned with FI may have responded;

consequently an overestimation of knowledge of FI may have been obtained.16 A similar

study assessing physicians’ knowledge and practices towards FI found that practices and

referrals related to FI were not regularly performed, and that the majority were not

familiar with the concept of FI.15 These authors also noted only those concerned with FI

may have responded. The development of a standardized, validated question to address

the concept of food insecurity may be a key to increasing both the knowledge and

practices of clinicians regarding FI. The hunger screening study by Kleinman et al,

(2007) discussed above documents the feasibility of applying a written screening

question in a primary care setting. Follow-up of any personal patient-clinician interaction

and discussion of available services was not assessed.

The Oregon Childhood Hunger Initiative (CHI) recently reported on a survey

mailed to randomly selected healthcare providers in the Portland, Oregon area that

addressed hunger and the primary care setting.17 The goal of the CHI project was to

determine what providers surveyed know about FI and its health implications, as well as

to find out the tools needed to ask about FI and if providers would take action if given the

tools. When asked what factors prevent providers from asking about hunger during

clinical appointments, 81% marked “time constraints.”17 In addition, “I don’t know

enough about the issue” was marked by 45.5% and “I don’t know how to ask this type of

question” was cited by 22.3% of healthcare providers.17 Despite these barriers, 89%

17
reported that they would be willing to use a standardized screening question to identify

FI. Assessing FI in national surveys and in various other populations is necessary to

determine the population prevalence of FI and impact or need for government food

assistance programs; interventions incorporating screening for FI in clinics and hospitals

may be an additional avenue for reduction of the incidence of FI, particularly in children.

Receiver Operating Characteristic Curves for Assessment of Screening Tools.

Analysis of Receiver Operating Characteristic (ROC) curves provides a method to select

optimal screening tools of desired sensitivity and specificity and discard others, and can

be applied to a screening tool for food insecurity.50 In essence, the ROC curve is a plot of

the sensitivity versus 1-specificity of a classifier system. Sensitivity is defined as the

conditional probability that the indicator (for instance, one screening question) will

correctly give a food insecure result, provided that the person screened is food insecure.

Specificity is defined as the conditional probability that the indicator will correctly give a

negative result, such as food secure, provided that the person screened is food secure (or

NOT food insecure). When looking at the results in ROC analyses, the closer to 1.0 the

measure of sensitivity, the better the performance of the question, and the closer to 1.0

the measure of specificity (meaning the value of 1-specificity closer to zero, as results are

given in this format), the better the performance of the question as well. A properly

discerning screening tool will have both a high sensitivity and a low 1-specificity, in

order to identify those who are food insecure as such (or “True Positives”), but not also

falsely identify those who are not food insecure as food insecure (known as a “False

18
Positive”). However, it is not possible for a tool to have both perfect sensitivity and

perfect specificity; there is a trade-off or balance in the two measures.

In the ROC curve, the data are plotted as sensitivity versus 1-specificity. A

diagonal line runs through the ROC plot from bottom left to top right, and is the line of

indifference.51 A result falling on this line represents an indicator (screening question)

that does not discriminate between food security and food insecurity better than chance

alone. The closer a point is to the upper left corner of the graph, the better the sensitivity

and 1-specificity are for the question being tested.51 Refer to Figure 1 in the Results

section for an illustration of a ROC plot. The area under the curve (AUC) is an additional

calculation used to assess ROC plots. The AUC specifies the probability that, when you

pick one positive and one negative example at random (for instance, a food insecure

person and a food secure person), the decision function assigns a higher value to the

positive than to the negative example.50 In other words, AUC is the best balance between

sensitivity and specificity for a certain screening question. Indicators that discriminate

perfectly between food secure and food insecure have an AUC of 1.0, whereas a question

that discriminates no better than chance has an AUC of 0.5.

Children with Special Health Care Needs.

Definition and Description. Children with Special Health Care Needs (CSHCN) are

“those who have or are at increased risk for a chronic physical, developmental,

19
behavioral, or emotional condition and who also require health and related services of a

type or amount beyond that required by children generally.”52 As examples, this

definition includes children with type 1 diabetes, cerebral palsy, autism, and cancer. In

2006, 13.9% of children nationwide and 13.6% of Oregon children had one or more

special health care need,52 and 20% of households nationwide had CSHCN.53,54 Services

needed by CSHCN beyond those utilized by children without chronic conditions are

varied, and include case management, special therapies, special equipment and supplies,

home health, respite care, and special nutrition.

Financial Impact of Children with Special Health Care Needs. The prevalence of

having a child with special health care needs (SHCN) increases as poverty status

increases, and creates an additional financial burden on the family.20,21 One study from

1998 found that children in families with income levels at or below the federal poverty

line are one third more likely than those in families above the poverty line to have a child

with SHCN.21 A nation-wide analysis found that over 40% of families with CSHCN

experience finance-related family problems.22 Employment problems, either reduced

hours worked or the need to stop work altogether, affect 30% of this population

nationally,52,53 and a smaller percentage of parents of CSHCN work full time than those

of children without chronic conditions. 55 Working less than full time not only reduces

income; this may also prohibit eligibility for insurance through the employer,

compounding the financial load. A study by Chung, et al, (2007) found that 41% of

parents were not able to miss work on days they believed they needed to in order to care

for their ill child.56 The reasons were mainly because the income was needed or they

20
feared being fired.56 This statistic illustrates that parents may have to choose between

providing for their family financially and caring for their child. The entire household or

the child with SHCN may suffer the consequences of either decision. Beyond simply the

financial burden, but perhaps intrinsically linked to it, numerous studies have found

adverse effects of having a child with SHCN on the family, including reduced parental

employment, 22,53 increased stress,57 and poor mental health.

Insurance does play a protective financial role in the costs associated with caring

for a child with SHCN, and there are expanded insurance opportunities for parents with

SHCN children, such as Medicaid, and assistance programs like the Supplemental

Security Income (SSI). Despite these resources, mean out-of-pocket expenses for a

SHCN child are almost double that of a child without chronic health conditions. 22,23

Viner-Brown (2005) reported that only half the families of CSHCN that are impacted

financially have adequate insurance to cover needed services,53 and families of CSHCN

face greater burdens even finding coverage to meet the child’s needs.55 In a focus group

of parents with CSHCN, parents frequently reported that insurance companies would

deny coverage of the child for an extended time, or would impose lengthy waiting

periods for needed services, both resulting in large out of pocket expenses.58 It is clear

that insurance does not fully compensate for the abundant needs of CSHCN, and the

resulting gap may contribute to an increased risk of FI or a greater severity of FI for the

family.

21
Impact on the Child. Due to the higher costs to the families, children with special

health care needs are more likely to go without various types of care, posing a real threat

to their well-being.55,59 Porterfield, et al, (2007) found that children from poor families

were less likely to use specialized physician services and prescription medications than

children of families with incomes above 200% of the federal poverty level due to cost

and health plan problems.59 Despite that, CSHCN average about four times as many

hospital stays23 and three times the medical encounters and absences from school when

compared with other children.21 Because the definition of children with special health

care needs is so broad, and because each child’s condition or conditions are so

personalized, it is difficult to assess or describe the health ramifications of a SHCN on a

child beyond showing issues such as increased medical needs, mental health care needs,

missed school days due to hospitalizations, unmet prescription needs, etc. However,

CSHCN from poorer families may not receive the spectrum of care that is consistent with

recommended practices.

Identifying Children with Special Health Care Needs.

The definition of children with special health care needs is broad and varied. The

one nationally accepted tool to identify CSHCN is the Children with Special Health Care

Needs Screener© developed by The Child and Adolescent Health Measurement Initiative

in accordance with the definition specifications of the Federal Maternal and Child Health

Bureau (MCHB).60 The survey is a five item, parent-completed module that takes only a

22
minute to fill out, yet provides identification of children across the range of conditions

denoted in the definition of CSHCN.60 The five items included in the survey are: 1. the

need or use of prescription medications, 2. an above-routine use of services, 3. the need

or use of specialized services or therapies, 4. the need or use of mental health services,

and 5. a functional limitation. The full survey can be found in Appendix C. The screener

is applied by the NCHS in the National Survey of Children with Special Health Care

Needs (NS-CSHCN) and in numerous other comprehensive surveys.60

Relationship between Food Insecurity and Children with Special Health Care

Needs. It has been shown above that FI negatively impacts the general health, school

performance, behavior and physical growth of children. It was also stated above that the

prevalence of households with CSHCN increases as poverty increases, a factor

intrinsically tied to FI. Regardless of food security status, CSHCN are at an increased

nutritional risk18,19 and with food security factored in, the nutritional risk of CSHCN is

likely to increase. A position paper from the American Dietetic Association states that,

“persons with . . . special health care needs frequently have nutrition problems including

growth alterations (such as failure to thrive, obesity, and growth retardation) metabolic

disorders, poor feeding skills, medication-nutrient interactions, and partial or total

dependence on enteral or parenteral nutrition.”18 In families that are struggling to pay

bills and potentially facing problems obtaining adequate health insurance coverage to

meet the needs of CSHCN, feeding the child anything may be the main priority, making

an appropriate diet for the specific needs of the child a lower priority. CSHCN in food

insecure households may be at increased risk for one or more of the adverse physical,

23
developmental, or psychological health issues associated with food insecurity in addition

to the already present issues brought on by the SHCN. For the health of the nation’s

children, food insecurity among families with CSHCN is a problem that demands further

attention in the form of services and support that will help families address their needs.

Tillamook County Preschool Screening Description and Demographics.

The data used for this study was obtained from a free health screening for

preschool children called the Tillamook Multi Modular Preschool Screening. This

screening has taken place in the city of Tillamook, Oregon for the past 20 years. The

Northwest Regional Education Service District (NWRESD) and the Child Development

and Rehabilitation Center (CDRC) of the Oregon Health and Science University sponsor

this event, which provides free health screening to children in Tillamook County prior to

entering kindergarten. The children are assessed for medical, developmental, physical or

educational problems. Results are used to identify children in need of health-related

services and make appropriate referrals prior to school entry. The screening is well

known and well advertised, and served as the ideal medium through which to obtain the

data used for this analysis.

Tillamook County, Oregon is found in the North West coastal region of Oregon,

and encompasses 1,333 square miles.61 The city of Tillamook had 4,300 residents in

24
2000. Tillamook County had 24,000 residents, as reported by the most recent Current

Population Survey.61 The distribution of county residents by race /ethnicity is as follows:

the majority (91%) are non-Hispanic White, 5.1% are Hispanic or Latino and 0.2% are

Black. Female headed households make up 4.8% of the households in the county.

Eighty four percent of people over the age of 25 are high school graduates or higher. A

full 21.5% of children are in poverty in the county compared to 18.8% in the state.62

Research Questions and Specific Aims.

This thesis utilized data obtained from the Tillamook Multi-Modular Preschool

Screening using the Survey of Household Food Needs to address two main questions.

(See Appendix A for survey) The questions are as follows:

1. Is a single question from the USDA 6-Item Subset (Short Form) of the 12-month

Food Security Survey Module valid for use as a screening tool for food

insecurity?

2. What are the household and family characteristics that predict household food

insecurity in the study population?

The objectives of this study are:

1. To test the hypothesis that a single question from the 6-Item Subset (Short Form)

of the 12-Month Food Security Scale has appropriate sensitivity and specificity to

be an accurate screening tool in primary care.

25
 2. To test the hypothesis that having a child with special health care needs in the

family predicts household food insecurity.

Testing these hypotheses will be achieved by using data obtained from the Tillamook

Multi-Modular Preschool Screening using the Survey of Household Food Needs to

accomplish the following specific aims:

1. Determine the prevalence of low food security (food insecurity) and very low

food security in the study population.

2. Determine the prevalence of children with special health care needs in the study

population.

3. Identify the most appropriate single question for use as a screening tool from the

6-Item Subset (Short Form) of the 12-Month Food Security Survey Scale using

Receiver Operating Characteristic curve analysis to describe the sensitivity,

specificity and area under the curve of each question of the 6-Item Subset.

4. Identify associations between household food insecurity status and status of

households with children with special health care needs using chi square and

logistic regression analysis.

5. Identify the household characteristics that predict food insecurity in the study

population using the variables obtained from the Tillamook Multi-Modular

Preschool Screening: Survey of Household Food Needs.

26
 Chapter II

Methods

Study Design. This retrospective analysis of existing data was determined to be exempt

from need for review from the Oregon Health and Science University Institutional

Review Board. Data for this cross-sectional descriptive study were obtained in 2006

from Tillamook, Oregon by the Northwest Regional Education Services District

(NWRESD). Data were collected from parents of children who participated in the

Annual Tillamook Multi-Modular Preschool Screening (TMMS) in 2006. The

instrument used for data collection was the Multi-Modular Preschool Screening: Survey

of Household Food Needs (See Appendix A). The annual screening is optional, and

available to all families in the area who wish to have their pre-kindergarten child

screened. The Survey of Household Food Needs was mailed to parents signed up for the

screening, along with other screening paperwork. The parents who had not completed

the survey before arriving at the screening event were invited to complete the survey on

site as well as ask questions of staff.

The survey is a 17 question tool. It contains the USDA 6-Item Subset (Short

Form) of the HFSS (See Appendix B), as well as three questions modified from questions

on the Children with Special Health Care Needs Screener© (See Appendix C). The

remaining questions ask about demographic characteristics and other household

characteristics and behaviors. Questionnaires were available in Spanish and English

versions.

27
Study Population. The participants in the study were families with at least one child

aged three to six years who attended the TMMS. Families living in Tillamook, Oregon,

and Tillamook County, Oregon, were recruited for the screening, although participation

was not restricted to those living within the area. Addresses were not recorded, and

children living outside of Tillamook County may have been included. A total of 221

families representing 228 children completed the survey, out of 239 children attending

the screening.

Inclusion/Exclusion Criteria. The main inclusion criterion was having a child of pre-

kindergarten age, although the age range in the study sample was three to six years. All

families who wanted to participate could do so. The survey was available in English and

Spanish. If the family did not have a mailing address (for example, had moved recently

or were homeless) they were not excluded from the survey; families were able to fill out

paperwork and surveys during the screening. Families bringing in more than one child

for the screening filled out one survey of demographic and household characteristics, and

then answered child-specific health and nutrition questions for each child screened.

Respondents not providing answers to the food security questions or who provided partial

responses only were excluded from analysis.

Data Management and Definition of Dependent Variable. Data from the surveys

were entered into Microsoft Access, and transferred to an SPSS data file. All analyses

were performed using SPSS Version 15.0.63 Missing data were excluded from all

analyses. Households were classified as either food secure or food insecure, depending

on the answers to the 6-Item Subset of the HFSS. Families responding affirmatively to
28
two or more food security survey questions were categorized as food insecure according

to the scoring protocol.47 The 6-Item Subset asks questions only of the household, not of

the children, so assumptions cannot be made about the food security status of the

children. Food insecurity in this sample refers to household FI only. Respondents

answering affirmatively to five or more questions were categorized to have “very low

food security” in the household.47 Ten families, or 4.3% of the study population, had

very low food security, which was too small a number to analyze with reportable results.

Therefore, low FS and very low FS were collapsed into two categories, either food secure

or food insecure; household food insecurity was used as the dependent variable of

interest.

Recoding of Variables. Classifying a child as one with a special health care need was

based on responses to three questions in the survey (questions 16, 17 and 18). All three

questions needed to be answered in the affirmative to identify the child as one with a

SHCN. To increase cell size of the CSHCN variable for analysis, affirmative answers for

the first two questions were used to represent the CSHCN variable, and the requirement

of the condition lasting 12 months or longer (the content of the third question) was not

included. Though this resulted in an alteration of the standard definition of CSHCN, this

allowed analysis of having a child with SHCN in this population, regardless of the

duration, as this question was of particular interest to the study.

Measured heights and weights were recorded for each child and used to calculate

the body mass index (BMI), which is weight in kilograms divided by height in meters,

squared. The BMIs were converted to Z scores using Epi Info64 by applying the NCHS
29
growth chart data. The BMI Z score was tested in models as both a continuous variable

and as a trichotomous categorical variable of low BMI (< -2.0 Z score), normal BMI (-2.0

– 2.0 Z score) and high BMI (> 2.0 Z score).

All variables except child age were analyzed as categorical variables reflecting

how questions were asked, and required recoding to provide adequate cell sizes. When

necessary and where feasible, categories were collapsed to provide adequate cell sizes for

analysis. Need for recoding was determined by assessing frequency tables and

histograms. For example, question number 7 asked about family type and provided seven

answer choices including “other.” The frequency table for this variable showed that the

option “single father living with children” had only 1 respondent, and therefore this

category was collapsed with “single mother living with children.” Similarly,

“grandparents living with children” had only 2 respondents, and was grouped with

“couple living with children.”

Health status of the child was addressed as, “In general, would you say this

child’s health status is A. Excellent, B. Very good, C. Good, D. Fair or E. Poor.” There

were zero responses for the option of “poor” and four responses of “fair.” Because of

this, the responses were collapsed and recoded into a dichotomous variable with excellent

and very good as one category and good and fair as the second. The latter represented the

lower end of health reported in this survey. Collapsing the health status variable is

common practice in other analyses.3,4,6

30
Descriptive Characteristics. Demographic characteristics of the study sample were

obtained with descriptive statistics such as mean and frequency and are reported in detail

in the Results section. Prevalence of household FI and families with CSHCN were

described using descriptive statistics.

Receiver Operating Characteristic. Receiver operating characteristic (ROC) curve

analysis and corresponding areas under the ROC curve were carried out to address Aim

3: identify the most appropriate single question for use as a screening tool from the 6-

Item Subset (Short Form) of the 12-Month Food Security Survey Scale using Receiver

Operating Characteristic curve analysis to describe the sensitivity, specificity and area

under the curve of each question of the 6-Item Subset. Each question in the 6-Item

Subset was compared to the entirety of the 6-Item Subset. Analyses included measures

of sensitivity, specificity and area under the curve (AUC).

Chi Square and Multivariate Logistic Regression. Chi square and univariate and

multivariate logistic regression analyses were used to test Aims 1, 2, 4 and 5. For all

results, P values < 0.05 were considered statistically significant. Frequency tables and

chi square were used to determine the prevalence of FI and CSHCN, while univariate and

multivariate logistic regression were utilized to identify predictors of FI. Correlation

between pairs of variables were evaluated, and those correlated above a 0.35 cut-off point

were run in simple logistic regression to evaluate the strength of their relationship with

food insecurity; the variable with the larger P value was left out of subsequent model

building. A 0.35 correlation was chosen as this is a common rule-of-thumb cut-off for

this type of analysis. For instance, having a home garden and child gardening correlated
31
highly at 0.818 (P = 0.01), and so child gardening was not included in the model

building, as having a home garden explained so much of the variable and child gardening

was less significant in simple logistic regression. Correlations were also checked

between the dependent variable, household FI, and the independent variables to observe

the relationships between variables.

Simple logistic regression was run first for each variable and variables with a P

value of > 0.2 were excluded from the model building. All remaining variables were put

into a model together for backwards elimination, unless they had been eliminated because

of a high correlation with another variable. Variables were eliminated one-by-one

according to largest P value, unless the variable was of interest to the end interpretation,

in which case it was left in regardless of P value.

Confounding was assessed by calculating changes in P values and betas (β) after

each variable was eliminated in the multiple logistic regression model. If the change was

10% or more, the variable was considered to be a confounder and left into the model to

control for the effect.

32
 Chapter III

Results

Study Demographics. Characteristics of the study sample are presented in Table 2 and

Table 3. The 221 families that participated in the survey represented 228 children. The

majority of the children were preschool age, and the mean age was 53.5 months, or 4.5

years. A majority of participants were non-Hispanic White (n = 165, 77.1%) and almost

one fifth were Hispanic (n = 39, 18.2%). Ten participants were of another race or

ethnicity (4.7%). Almost three fourths of families were parent or grandparent couples

with children (n = 156, 72.6%), one fifth were single parent families (n = 42, 19.5%), and

the remaining family types made up 7.2% (n = 17). The mean family size in this sample

was 4.36, and the average number of children per household participating was 1.25. The

majority of families brought one child to the screening; seven families brought two

children and zero families brought three or more children.

Income and Education. Over a third of the families participating reported an income of

$35,000 or above (35.1%, n = 72), whereas the second largest income category was $0-

14,999 reported by 26.3% (n = 54). Fewer reported incomes of $15,000-$24,999 at

21.0% (n = 43) and $25,000-$34,000 at 17.6% (n = 36). Almost half (42.7%, n = 90) of

the sample population had finished high school or the General Educational Development

tests (GED), while 31.8% (n = 67) had completed some college, 11.8% (n = 25) were

college graduates, and 13.7% (n = 29) had finished some high school or less. A full

56.4% of persons completing the study had a high school diploma or less.

33
Table 2

Demographic Profile of Study Sample

Study Population Response

Characteristic
Percent % n

Number of Families in Survey n = 221

Number of Children in Survey n = 228

Age of Child, Mean in Months (SD)1 53.5 (10.69)

Weight of Child in Pounds, Mean (SD)1 42.41 lb (8.512)

Race/Ethnicity: White 77.10% n = 165

Hispanic 18.20% n = 39

All Other 4.70% n = 10

Family Type: Single Parent 19.50% n = 42

Grandparent/Parent Couple w/
Children 72.60% n = 156

Other 7.20% n = 17

Number of Children Participating Per

Family, Mean (SD)1 1.25 (0.496)

Family Size, Mean (SD)1 4.36 (1.39)

Income Category: $0-14,999 26.30% n = 54

$15,000-$24,999 21.00% n = 43
$25,000-$34,999 17.60% n = 36

$35,000 or Above 35.10% n = 72

Table 2 continued next page . . .

34
Table 2 continued:

Study Population Response

Characteristic
Percent % n
Parents Education: Some High School or
Less 13.70% n = 29

High School or GED 42.70% n = 90

Some College 31.80% n = 67

College Graduate 11.80% n = 25

1
Results reported as mean in the first column and standard deviation (SD) in the second
GED = General Educational Development tests

Table 3 represents questions on the survey pertaining to food and income

behaviors in the household, and also shows the prevalence of food insecure families and

families who have CSHCN. About a quarter of families have gardens, and children

typically aid in gardening (family has home garden, 24.9%, n = 54, and child gardens

27.2%, n = 58). In general, there was substantial use of many assistance programs: over a

third (38.30%) reported food stamp usage, just under a third (31.90%) make use of the

National School Lunch Program (which provides low- or no-cost lunch in schools to

eligible children), and nearly half use the Special Supplemental Nutrition Program for

Women, Infants and Children (WIC) (46.50%). A lesser number of families reported

35
using Head Start or Early Head Start and obtained food at no cost from churches or food

pantries, but these percentages were considerable as well (17.3%, n = 35 and 18.2%, n =

38 respectively).

“Program use” was a variable made to describe families that used at least one of

the programs, to show the proportion of families who used programs versus those who

did not. Almost three quarters (73.6%, n = 173) of the survey population used at least

one of the programs listed in the questionnaire, showing that a large number of the

families make use of federal and public food assistance.

It is not common that a child will skip breakfast for any reason in the study

population, just 15% (n = 32) skip breakfast a few days a week or more. The question

about skipping breakfast was followed by a query of why the child would skip breakfast,

but since the majority of respondents did not answer this question (71.5%, n = 168) or

marked “none of the above” (21.3%, n = 50), this question was not reported or analyzed.

No other information was obtained about mealtime habits.

Food Insecurity and Children with Special Health Care Needs. Household food

insecurity in this sample is 23.2% (n = 53). Food insecurity with hunger, or very low

food security, occurred in 4.3% (n = 10) of the households. The proportion of affirmative

responses to the food security questions are shown in Table 4. Children with a special

health care need made up 10.8% (n = 25) of the study population. Children in this

sample who were both from food insecure households and have a special health care need

represent 3.4% (n = 8) of the total children included in this study.

36
Table 3

Food and Income Behaviors, Food Security and Children with Special Health Care Needs

Status

Survey Response
Behavior or Characteristic
Percent (%) N

Family has Home Garden1 24.90% n = 54

Child Gardens1 27.20% n = 58

Food Stamp Use1 38.30% n = 80

School Lunch Program Use1 31.90% n = 66

School Breakfast Program Use1 25.40% n = 52

Summer Food Program Use1 3.40% n=7

Head Start or Early Head Start Use1 17.30% n = 35

WIC Use1, 2 46.50% n = 100

Obtain No-Cost Food1 18.20% n = 38

Families who use at least one program 73.60% n = 173

Child Skips Breakfast: Almost Daily 5.60% n = 12

Few Days a Week 9.40% n = 20

Almost Never 85.00% n = 181

Table 3 continued next page . . .

37
Table 3 continued:

Survey Response
Behavior or Characteristic
Percent (%) N
Household Food Insecurity (Low Food
Security) 23.20% n = 53

Very Low Food Security 4.30% n = 10

Child has Special Health Care Need 10.80% n = 25

Children who are both in FI3 households
and have a SHCN4 3.40% n=8
1
Indicates a yes or no answer
2
WIC = The Special Supplemental Nutrition Program for Women, Infants and Children
3
FI = Food Insecure
4
SHCN = Special Health Care Need

Table 4

Frequency of Responses to Food Security Questions

Question from 6 Item Subset1 Frequency Percent (%)

Question 1 65 27.7

Question 2 53 22.6

Question 3 20 8.5

Question 4 18 7.7

Question 5 18 7.7

Question 6 15 6.4
1
See 6-Item Subset in Appendix B for list of questions.

38
Table 5

Receiver Operating Characteristic Results

Area Under Standard 95 % CI for

Test Question1 Sensitivity 1-Specificity
the Curve Error Sensitivity

Question 1 0.958 0.015 0.981 0.066 0.928, 0.987

Question 2 0.919 0.028 0.870 0.033 0.864, 0.973

Question 3 0.661 0.048 0.333 0.011 0.568, 0.754

Question 4 0.667 0.048 0.333 0.000 0.573, 0.760

Question 5 0.655 0.048 0.315 0.006 0.561, 0.748

Question 6 0.639 0.048 0.278 0.000 0.545, 0.733

1
See 6-Item Subset in Appendix B for list of questions.

Receiver Operating Characteristic Analysis. An overall question of this thesis was to

determine whether a single question from the USDA 6-Item Subset of the Food Security

Survey would be valid and appropriate as a screening tool for delineating food insecurity

in the survey population. ROC analysis was performed to test this aim, and the results

are depicted in Table 5. Question 1 performed at the highest level of area under the curve

(AUC) and sensitivity. AUC for question 1 was 0.958, where 1.00 represents a prefect

39
test. The sensitivity of question 1 was 0.981 and 1-specificity was 0.066. This question

corresponds to, “The food we bought just didn’t last, and we didn’t have money to get

more,” with possible responses being “Often true,” “Sometimes true,” or “Never true.”

Question 2 had a slightly lower AUC of 0.919 with a substantial decrease in the

sensitivity at 0.870. 1-specificity was 0.033 which indicates better specificity than

question 1 (0.066). AUC was lower for questions 3-6 than for questions 1 and 2; the

AUC for these three questions were all in the range of 0.636 – 0.661. Figure 1 shows the

ROC plot for identification of FI using question 1 of the 6-Item Subset.

Chi Square Analysis. Results of chi square analyses testing statistically significant

differences between food secure families and food insecure families and statistically

significant differences between families with and without CSHCN are presented in

Tables 6 and 7. The majority of the household demographic characteristics and program

use were significantly associated with food insecurity (P ≤ 0.05). In contrast, family type

(P = 0.258), age of child (P = 0.178) and gender (P = 0.875) of child were not

significantly related to FI. The association between food insecurity and having CSHCN

was significant in chi square at P = 0.051. Chi square was also used to test associations

between household demographic characteristics and program use and families with

CSHCN (See Table 7). There were three significant associations. The relationship

between food insecurity and having a child with SHCN was significant at P = 0.051. The

two additional significant associations with having a child with SHCN were

excellent/very good health status versus good/fair health status (P = 0.009) and Head

Start use (P = 0.023).

40
Figure 1

Receiver Operating Characteristic Plot for Identification of Food Insecurity using

Question 1 of the 6-Item Subset of the 12-Month Food Security Scale

1.0

Sensitivity
0.8

0.6

0.4

0.2

0.0
0.0 0.2 0.4 0.6 0.8 1.0

1 - Specificity

*Diagonal segment depicts line of indifference; points falling on this line discriminate

no better than chance alone

41
Table 6

Chi Square of Household Demographic Characteristics by Household Food Security

Status

Variable Chi Square Value

Food Stamp Use1 10.998 b

School Lunch Program Use1 5.486 a

School Breakfast Program Use1 5.422 a

Summer Food Program Use1 16.059 c

Head Start Use1 6.359 a

WIC Use1, 2 6.830 b

No-Cost Food Use1 10.479 b

Program Use1 15.659 c

Home Garden Use1 6.440 a

Child Garden Use1 5.837 a

Family Type3 2.713

Family Income4 31.651 c

Child Skips Breakfast 6.681 a

Having a CSHCN5 3.818

Age of Child 7.625

Health Status, 2 Categories6 4.390 a

Table 6 continued next page . . .

42
Table 6 continued:

Variable Chi Square Value

Race/Ethnicity7 25.969 c

Gender of Child 0.025

Parent's Education 24.576 c

a b c
indicates P < 0.05, indicates P < 0.01, indicates P < 0.001
1
Indicates a yes or no answer
2
WIC = Special Supplemental Nutrition Program for Women, Infants and Children
3
Family Type categories: Single Parent, Parent or Grandparent Couple, Other
4
Family Income categories: $0-$14,999, $15,000-$24,999, $25,000-$34,000, and ≥
$35,000
5
CSHCN = Children with Special Health Care Needs
6
Health Status categories: excellent/very good and good/fair
7
Race/Ethnicity categories: Non-Hispanic White, Hispanic, Other

Table 7
Chi Square of Selected Household Demographic Characteristics by Having a Child with

Special Health Care Needs

Variable Chi Square Value

Household Food Insecurity1 3.818

Health Status, 2 Categories6 6.755 b

Head Start Use 5.187 a

a b
indicates P < 0.05, indicates P < 0.01
43
Univariate Logistic Regression Analysis. Simple logistic regression results for survey

variables and FI are found in Table 8. Results show that having a home garden and

having children who garden in it are both associated with reduced odds of household FI.

Use of any of the programs offered in the survey, including using at least one program

versus using none increased the odds that the family would be food insecure. The only

program that did not predict household food security status was use of the Summer Food

Program use (P = 0.76). Families of Hispanic ethnicity were almost six times more likely

to be food insecure compared to families of White race (P = 0.000).

Families in which the parent reported the child as having good or fair health were

two times more likely to be food insecure (P = 0.038) compared to children reported to be

in excellent or very good health. A family with a child with SHCN was 2.5 times more

likely to be food insecure as well, although this result did not reach significance (P =

0.057). There was an inverse relationship between income and household FI; as income

decreased, the odds for FI the family increased (all P < 0.05). Households with a parent

who had completed some college or were a college graduate were 0.4 times less likely to

be food insecure (P = 0.003) compared to those who completed high school or less.

Family type and child’s gender were not predictors of FI (P = 0.875), and BMI Z scores

were also not significant predictors of FI when analyzed either as a continuous variable or

as a trichotomous categorical variable (normal BMI, low BMI, and high BMI).

44
Table 8

Univariate Logistic Regression Associations between Household Food Insecurity and

Household Characteristics

Variable Unadjusted OR 95% CI

Family has a Home Garden1 0.375a 0.172, 0.816

Child Gardens1 0.415a 0.201, 0.859

Food Stamp Use1 2.854b 1.515, 5.376

School Lunch Program Use1 2.070a 1.119, 3.829

School Breakfast Program Use1 2.073a 1.115, 3.853

Summer Food Program Use1 1.979 0.932, 4.202

Head Start or Early Head Start Use1 2.242a 1.187, 4.236

WIC Use1,4 2.310a 1.222, 4.366

No Cost Food Use1 2.820b 1.485, 5.356

Program Use: Used At Least One1 8.221b 2.463, 27.437

Child Skips Breakfast: Almost Reference Group
Never

Few Days a Week 2.206 0.756, 5.432

Almost Daily 3.763a 1.149, 12.330

Race: White Reference Group

Hispanic 5.824c 2.761, 12.284

All Other 0.500 0.061, 4.097

Table 8 continued next page . . .

45
Table 8 continued:

Variable Unadjusted OR 95% CI

Health Status: Good/Fair2 2.094a 1.040, 4.215

Family Has CSHCN1,5 2.534 0.973, 6.599

Income Category: $0-$14,900 15.786c 5.044, 49.406

$15,000-$24,900 7.367c 2.218, 24.462

$25,000-$34,000 4.103a 1.115, 15.104

≥ $35,000 Reference Group

Education: Some College or
College Grad3 0.351b 0.174, 0.708
Family Type: Single Mom or Dad Reference Group
with Children
Couple or Grandparents with
Children 0.537 0.254, 1.133
Other- 3 Generation House,
Foster, Other 0.615 0.169, 2.238

Gender of Child: Male 1.051 0.565, 1.957

BMI Z Score: Normal BMI6 Reference Group

Low BMI 2.743 0.589, 12.789

High BMI 1.590 0.697, 3.627

a b c
Indicates P < 0.05, indicates P < 0.01, indicates P < 0.001
1
Indicates a yes or no answer, all are compared to the “no” response
2
Compared to health status reported as excellent or very good
3
Compared to finishing high school or less education
4
WIC = The Special Supplemental Nutrition Program for Women, Infants and Children
5
CSHCN = Children with Special Health Care Needs
6
BMI = Body Mass Index (kg/m2), Low BMI Z score: < -2.0, High BMI Z score: > 2.0, Normal: –2.0-2.0

46
Multivariate Regression Model. The final multivariate logistic regression model to

determine predictors of household food insecurity in this study sample can be found in

Table 9. Variables retained in this multivariate model include: having a CSHCN,

race/ethnicity of respondent, household income, use of no-cost food, and use of food

stamps. Families with a child with SHCN are three times more likely to be food insecure

than those who do not have a CSHCN, after adjusting for the variables in the model (P =

0.046). Race/ethnicity was not a significant predictor of FI, but was identified as a

confounder due to beta change more than 10% after taking out the variable and was

retained in the model.

When compared with the highest income group, families with incomes of $0 -

$14,999 were 10 times more likely to be food insecure (P = 0.003). Families whose

incomes were $15,000 - $24,999 were 4.5 times more likely to be food insecure (P =

0.035), and incomes of $25,000 - $34,999 was not a predictor of household FI.

Families who utilized some form of no-cost food were almost four times more

likely to be food insecure than non-users (P = 0.005). Food stamp use was not a

significant predictor of FI in this model, but was a confounder and was left in to control

for that effect. All other variables not included in the final model did not meet the

retention criteria for inclusion during the model building process.

47
Table 9

Associations between Food Insecurity and Family Characteristics, Multivariate Logistic

Regression Model

Variable N Odds Ratio, Adjusted 95% CI

Family Has CSHCN 25 3.271a 1.020, 10.492

Race/Ethnicity: White 165 Reference Group

Hispanic 39 2.844 0.913, 8.859

All Other 10 0.422 0.028, 6.453

Income: $0-$14,999 54 10.179 b 2.235, 46.366

$15,000-$24,999 43 4.510a 1.112, 18.302

$25,000-$34,999 36 2.673 0.614, 11.625

$35,000 or Above 72 Reference Group

No-Cost Food Use 38 3.711b 1.487, 9.262

Food Stamp Use 80 0.718 0.246, 2.100

a b
indicates P < 0.05, indicates P < 0.01

Model N = 183 (This N had complete data for all variables tested in model)

48
 Chapter IV

Discussion

Although this was a preliminary study of the Tillamook region, three main

findings were achieved through the analysis which will be reported to the Northwest

Regional Educational Service District (NWESD) for application in development of

services in the area. The three key results were a description of the prevalence of

household food insecurity and children with special health care needs, the identification

of a single question screening tool to identify food insecurity, and the predictors of food

insecurity in the Tillamook region.

Prevalence of Food Insecurity and Children with Special Health Care Needs. The

prevalence of household FI in this population was 23%, twice the state average of

household food insecurity and more than twice the national average. The current state

average is 11.9%; the national average is 10.9%.24 There are several possible

explanations that may account for the increased incidence of FI in this sample, the first

reflected in the results of the demographic survey questions. In general, the sample

captured by this survey was financially limited. Sixty five percent of families had an

income less than $35,000 with an average family size of 4.4 individuals, and almost three

quarters of the families utilized at least one of the child or adult food programs included

in the survey that provides support to families with limited resources. Due to the low

income status of the families and other demographic information obtained from the

49
survey, FI is likely to be experienced at a much higher rate by the study sample compared

to national and state averages.

Secondly, the nature of the community health screening evaluated may play a

large role in the prevalence of FI in this sample. The health screening was a free event at

which children were evaluated by health professionals or graduate students from a wide

range of health disciplines at no cost to families. If a family were facing financial

hardship, the free pre-kindergarten screening would appear to be a viable alternative to

costly physician examinations. Therefore the free screening and survey would likely

attract more families that were food insecure, and study estimates would be greater than

rates reported in the general population.

Another factor that may influence the high prevalence of FI in this study is the

rural location in which the health screening took place. A described predictor of food

insecurity is living in rural, non metro areas.13 Rural status could potentially explain

above average rates of food insecurity. And finally, as previously mentioned, there is

also data to suggest that households with children have higher rates of food insecurity

than households without children. As this study focused solely on families with children,

this may also contribute to a small percentage of the variance between study prevalence

and state and national averages.

A potential bias that may have occurred in this survey has to do with program use

and the reporting of food insecurity. Families who qualify for and use various food

assistance programs may be more aware of the concept of FI, and thus more likely to

50
report experiencing FI. This may increase reported rates of FI among those using

available programs.

Children with special health care needs accounted for 10.8% of the children in the

study. This is below the national average of 13.9% and the Oregon average of 13.6% .52

However, in this study, we applied a lenient modification of criteria used by MCHB to

identify CSHCN. The MCHB criteria for health conditions of CSHCN are 1) the child

currently experiences a consequence, 2) the consequence is due to a medical or other

health condition, and 3) the duration is 12 months or longer.60 In contrast in this study

we defined CSHCN as children who met the first two criteria but not necessarily the

third. The results of this study must therefore be interpreted cautiously; however they are

expected to provide useful information to the NWESD and Tillamook County and may

provide guidance for program development and for future studies related to food

insecurity and CSHCN.

One likely explanation for the low percentage of CSHCN is the age of the

children included in the study. The mean age of the sample was 4.5 years old, which may

be too young for a wide range of special health care needs to have emerged. Had the

mean age been a few years older, a pattern of greater prevalence of CSHCN may have

emerged. A second possibility is that because CSHCN average about four times as many

hospital stays23 and three times as many medical encounters as other children,21 parents

may have felt less need to bring their child to the free health screening as all of their basic

exams had already been completed. Thus, this group may have been under-represented at

the free health screening

51
 Families who were both food insecure and had a child with SHCN made up 3.4%

of the children in this study. Household FI and having a child with SHCN were

positively associated in chi square analysis; however, this relationship was just above

accepted levels of significance (P = 0.051).

Single Question Screening. Question 1 performed the best out of the six questions in

the USDA 6-Item Subset of the HFSS (See Appendix B for questionnaire and Table 4 on

page 38 for response frequency to each question). The sensitivity of this first question

was 98.1% and the specificity was 93.4%. In other words, this question is capable of

identifying True Positives 98.1% of the time and shows extremely low rates of

occurrence of False Positives (6.6%). This is a very precise and accurate identification

using only a single question. The high sensitivity was not surprising considering the first

question pertains to the least severe decrease in food quality or quantity and thus may

acquire the largest affirmative responses for any degree of FI. The first question also had

the largest area under the curve (AUC). AUC shows the balance between sensitivity and

specificity measures for a given question, with a larger AUC depicting a better balance.

Therefore question 1, despite not having the best specificity, had the best balance of the

six questions.

Questions 4 and 6 demonstrated high specificity; they would produce False

Positives very infrequently (Q4 and Q6 1-specificity < 0.001). However, these two

questions have unacceptable sensitivity for use as a screening tool (Q4 sensitivity =

0.333, Q6 sensitivity = 0.278). Question 2 would correctly identify food insecure

individuals (True Positives) 87% of the time while only misclassifying 3.3% of people as
52
food insecure. This shows question 2 to possess a better ability to screen out the False

Positives than question 1. However, the AUC for question 2 is lower than question 1,

showing question 1 to be the better overall choice. The intended use of this test would be

for screening and follow-up of families or individuals who are found to be food insecure,

and so a question with higher sensitivity and slightly lower specificity would be more

desirable to identify as many food insecure individuals possible. This would enable

health care professionals to reach and provide resources for the greatest number of food

insecure families. There appears to be only one potential drawback of using a screening

question with a high sensitivity and slightly lower specificity. This combination would

result in misclassifying a few individuals as food insecure when they are not. It is

possible that a person would be offended when handed the resource materials for coping

with FI, though this scenario seems unlikely to occur often.

There are no set cut-off points for measures of sensitivity, specificity and AUC

for acceptable or unacceptable levels. The decision to use a screening tool based on the

ROC analysis results depends on the type of screening tool needed and the judgment of

the researchers or the health care providers. As discussed previously, Kleinman and

colleagues reported that a written one-question screener for the presence of hunger that

showed an 83% sensitivity and an 80% specificity represented an “accurate and reliable”

screening tool.49 The methods of analysis and study design of Kleinman’s study were

different from those utilized in the present study, yet the results support the ability of a

one-question screener to accurately discriminate between food insecure and food secure

53
individuals. Kleinman et al’s judgment to accept their results also provides support for

the reliability and validity of the present results.

Utilization of a single question as a screening instrument would have limitations.

First, there would be no way of distinguishing levels of severity of food insecurity.

Second, there would be no method for discriminating whether children were also

experiencing food insecurity in the household. Finally, there would be no way to

determine if the circumstances were chronic or temporary. However, the primary

purpose of this screening question is to eventually be used as a tool in a primary care

setting, either in written or verbal form, not for prevalence estimates of FI or

demographic reasons. The question could be asked by a dietitian, social worker,

physician, or any other trained professional in the health care setting; alternately, the

question could be included in a pretreatment paperwork packet. In the event of an

affirmative answer, a conversation could be initiated in order to learn more about the

patient’s specific needs, or the patient could simply be provided information regarding

resources that provide aid.

As previously noted, a survey by the Oregon Childhood Hunger Initiative

reported that 22% of surveyed health providers did not know how to ask questions to

patients about food insecurity, but that 89% were willing to use a standardized screening

question to identify FI.17 Providing a standardized question may take the uncertainty out

of asking about this issue for health care providers, and construct a pathway for broader

intervention with food insecure children. Testing the appropriateness of a single question

54
screening tool is the first step toward implementing that tool in health care settings, and

the first step towards attempting to alleviate food insecurity with this method.

Predictors of Food Insecurity. A majority of the variables examined in this study were

predictors of FI in univariate logistic regression. Having a home garden and children

who help to work in the garden are two variables that were associated with decreased risk

of food insecurity in this population. On the other hand, use of any federal or community

assistance program aside from the Summer Food Program was a predictor of food

insecurity before controlling for other factors; so much so that a family who used at least

one program compared to families using none was 8 times more likely to report being

food insecure.

Children skipping breakfast almost daily and children rated as being in good or

fair health, as opposed to excellent or very good health, were both univariate predictors of

FI. This seems intuitive, as perhaps inadequate food in the house for the child’s breakfast

contributed to adverse symptoms of FI observed in the child. This in turn would prompt

the parent to rate their child’s health status as only good or fair as opposed to excellent or

very good. However, without asking any further questions, reasons or causes of the child

skipping breakfast or reported as being in good or fair health cannot be ascertained.

There was no statistical difference in BMI between food insecure and food secure

families. This finding is similar to that of Kaiser and colleagues (2002) in their study of

211 Mexican-American children aged 3-6 years which also found no statistically

significant difference in the BMI Z-scores of FI versus FS children.33 Because of the

55
young age of the children sampled in this study as well as Kaiser’s, it is possible that

weight problems, especially overweight, may have not yet emerged, and if the same

group were to be retested at older ages in the future, the results would differ as discussed

by Alaimo et al (2001). Those results showed no association between overweight and FI

in the 2-7 age group, but did find that older girls in the 8-16 age group were more likely

to be overweight than their FS peers.42

Multivariate Predictors of Food Insecurity. The largest predictor of FI in this

population as indicated by an odds ratio of over 10 was a mean income of below $15,000.

This is not surprising as previous data shows the prevalence of food insecurity increases

as income decreases. The other significant predictors in this model, when controlling for

other factors, were having a SHCN child in the family and usage of no-cost food. A

family with a SHCN child was 3.2 times more likely to be food insecure. The probable

reasons for CSHCN as a predictor of FI were discussed in the introduction, and include

more out of pocket costs for medical care for the child22,23 and fewer hours worked for

the parents,52,53 both of which contribute to financial strain. It makes sense that use of

no-cost food is a multivariate predictor of FI as food banks, soup kitchens, etc. are

designed to help those most in need of food. Use of food stamps was associated with a

decreased risk for FI, although this result did not reach statistical significance. Had the

sample size been larger, this result may have been significant and demonstrated that use

of food stamps actually helps to prevent household food insecurity, which is an objective

of food stamp utilization.

56
Limitations. As with any study, there are limitations to the current study. The main

limitation is with regard to the study sample. The sample was not randomly chosen; it

was a convenience sample taken from a free health screening. This would naturally draw

a select population with inherent ascertainment bias, so the ability to generalize beyond

this sample is limited.

A second limitation is the sample size. Participation in the survey was optional

with no incentives offered for completion. The relatively small sample size and limited

power may have prohibited the finding of significant results for some variables.

Multivariate predictors of food insecurity may have been more numerous had the sample

size been larger, and it may also have provided a greater ability to identify associations

between FI and CSHCN.

It was not possible to classify or identify CSHCN in the standard manner defined

by the MCHB. Therefore, comparisons cannot be made with other studies that use the

standard definition of CSHCN. Limited conclusions can be made about the children with

special needs in this study. Because the prevalence of CSHCN in this analysis was lower

than the state and national averages, a potential limitation that may result due to possible

misclassification of CSHCN is to assume that CSHCN status in this population does not

require consideration. Additionally, the low prevalence made it difficult to observe a

relationship between food security status and CSHCN. Evaluation of a lager sample

representative of the regional population is necessary to fully identify CSHCN in

Tillamook County.

57
Contributions. There are several potential significant contributions associated with this

study. One of the primary objectives of the study was to assess the reliability and validity

of a one question screening tool for identification of food insecure individuals for

potential use in primary care settings. As noted, the assessment question chosen showed

adequate sensitivity and specificity for identification of individuals at risk for household

FI.

The characteristics of the participants were ideal for analyzing the predictors of FI

despite a relatively small sample size, and were instrumental in testing the one-question

screening tool. The survey was succinct, yet covered a broad variety of household and

food behavior questions. The screening setting also allowed for collection of measured

weights and heights for calculation of BMI data, which is more accurate than utilizing

survey-reported heights and weights.

This study also has child and public health implications. Although use of a

convenience sample from the Tillamook Preschool screening may have introduced bias

and overestimation of FI in Tillamook County, results illustrate the need that regions

served by the NWESD and perhaps other rural counties have for expanded income and

food assistance programs. Results provide confirmation for various predictors of FI,

including having a child with SHCN in the family. This study also serves as a pilot study

for future work addressing FI and CSHCN in this location and others.

Future Directions. Future research testing the validity of the single question screening

tool in other populations would be a beneficial next step toward implementation of the

58
screening question in primary care settings and then assessing its impact. Future studies

in this area could also focus on larger sample sizes with older children in order to obtain

more information about CSHCN and the association with FI. Focus groups may be

especially advantageous in gaining insights and discovering the specific needs of families

who are both food insecure and have a child with SHCN. Focus groups may also provide

valuable information about families with CSHCN who are food secure versus those who

are food insecure. The food secure families with CSHCN may have techniques and

practices that make them more resilient to FI that can be shared with others. If steps can

be made to implement a screening tool for food insecurity in primary care or toward

learning more about the needs of the population that is food insecure and has CSHCN,

perhaps more progress can be made toward the prevention or alleviation of food

insecurity.

In the meantime, steps can be made now to decrease FI at the community level.

The county can encourage community gardens and school gardens, to both provide fresh

produce and to educate children about gardening. Churches, libraries, schools and others

can sponsor free or low cost lunches and community gatherings to provide food for

families in need. Public health professionals can advocate for state-wide or national

policy changes, such as increasing minimum wage, increasing food stamp provision and

many more. With some creativity, many things can be done in addition to the programs

available to reduce food insecurity in this county and others.

59
 Chapter V

Summary and Conclusions

In conclusion, there were two main objectives for this study. First, to test the

hypothesis that a single question from the 6-Item Subset of the HFSS would have

appropriate sensitivity and specificity to be an accurate screening tool in primary care;

and second, to test the hypothesis that having a child with special health care needs in the

family would predict household food insecurity.

Results of ROC analyses showed that the first question of the six displayed the

largest area under the curve, showing the best balance between sensitivity and specificity.

The first hypothesis was supported, in that a single question was identified as a candidate

for a screening tool for identification of food insecure individuals.

The second hypothesis was also supported as having a CSHCN in the house was

shown to be a multivariate predictor of household food insecurity. In addition to having a

child with SHCN in the family, having a family income below $25,000 and using no-cost

food were also shown to be multivariate predictors of household food insecurity.

The results of this study will be shared with Tillamook County and the NWRESD

to help with identification of modifiable risk factors (predictors) of household FI. The

results are also likely to be useful for planning strategies to address the needs of this

population and work to reduce the prevalence of food insecurity and associated impacts

on health and education.

60
References

1. Jyoti DF, Frongillo EA, Jones SJ. Food insecurity affects school children's academic

performance, weight gain, and social skills. J Nutr. 2005;135(12):2831-2839.

2. Alaimo K, Olson CM, Frongillo EA, Jr. Food insufficiency and american school-aged

children's cognitive, academic, and psychosocial development. erratum appears in

pediatrics 2001 sep;108(3):824b. Pediatrics. 2001;108(1):44-53.

3. Alaimo K, Olson CM, Frongillo EA, Jr, Briefel RR. Food insufficiency, family

income, and health in US preschool and school-aged children. Am J Public Health.

2001;91(5):781-786.

4. Cook JT, Frank DA, Levenson SM, et al. Child food insecurity increases risks posed

by household food insecurity to young children's health. J Nutr. 2006;136(4):1073-1076.

5. Weinreb L, Wehler C, Perloff J, et al. Hunger: Its impact on children's health and

mental health. Pediatrics. 2002;110(4):e41.

6. Cook JT, Frank DA, Berkowitz C, et al. Food insecurity is associated with adverse

health outcomes among human infants and toddlers. J Nutr. 2004;134(6):1432-1438.

7. Skalicky A, Meyers AF, Adams WG, Yang Z, Cook JT, Frank DA. Child food

insecurity and iron deficiency anemia in low-income infants and toddlers in the united

states. Maternal & Child Health Journal. 2006;10(2):177-185.

61
8. Casey PH, Szeto K, Lensing S, Bogle M, Weber J. Children in food-insufficient, low-

income families: Prevalence, health, and nutrition status. Arch Pediatr Adolesc Med.

2001;155(4):508-514.

9. Slack KS, Yoo J. Food hardships and child behavior problems among low-income

children. 2004;1290-04.

10. Kleinman RE, Murphy JM, Little M, et al. Hunger in children in the united states:

Potential behavioral and emotional correlates. Pediatrics. 1998;101(1):E3.

11. Whitaker RC, Phillips SM, Orzol SM. Food insecurity and the risks of depression and

anxiety in mothers and behavior problems in their preschool-aged children. Pediatrics.

2006;118(3):e859-68.

12. Healthy People 2010. US Department of Health and Human Services. Accessed June

19, 2008 at http://www.healthypeople.gov/document/html/objectives/19-18.htm.

13. Food Security in the United States. United States Department of Agriculture,

Economic Research Service; 2006. Accessed November 2007-September 2008 at

http://www.ers.usda.gov/Briefing/FoodSecurity/

14. Ching D. Health consequences of food insecurity and hunger. Oregon Food Bank;

2007. Accessed June 19, 2008 at

www.oregonfoodbank.org/research_and_action/HealthConsequencesofHunger.html

62
15. Granger LE, Holben DH. Self-identified food security knowledge and practices of

family physicians in ohio. Top Clin Nutr. 2004;19(4):280-285.

16. Tscholl E, Holben DH. Knowledge and practices of ohio nurse practitioners regarding

food access of patients. J Am Acad Nurse Pract. 2006;18(7):335-342.

17. Childhood hunger: A public health concern. Screening for risk in the clinic setting.

Oregon State University Extension Service: Oregon Childhood Hunger Initiative; 2008.

Accessed May 21, 2008 at

www.oregonfoodbank.org/research_and_action/documents/CHI_report_final_for_web_0

00.pdf

18. American Dietetic Association. Providing nutrition services for infants, children, and

adults with developmental disabilities and special health care needs. J Am Diet Assoc.

2004;104(1):97-107.

19. Clark MP, Oakland MJ, Brotherson MJ. Nutrition screening for children with special

health care needs. Child Health Care. 1998;27(4):231-245.

20. van Dyck PC, Kogan MD, McPherson MG, Weissman GR, Newacheck PW.

Prevalence and characteristics of children with special health care needs.see comment.

Arch Pediatr Adolesc Med. 2004;158(9):884-890.

21. Newacheck PW, Strickland B, Shonkoff JP, et al. An epidemiologic profile of

children with special health care needs. Pediatrics. 1998;102(1 Pt 1):117-123.

63
22. Kuhlthau K, Smith Hill K, Yucel R Perrin JM. Financial Burden for Families of

Children with Special Health Care Needs. Maternal & Child Health Journal. 2005; 9(2):

207-218.

23. Newacheck PW, Kim SE. A national profile of health care utilization and

expenditures for children with special health care needs. Arch Pediatr Adolesc Med.

2005;159(1):10-17.

24. Nord M, Andrews M, Carlson S. Household food security in the united states, 2006.

United States Department of Agriculture, Economic Research Center. Report No. ERR-

49; 2007. Accessed May-September, 2008 at

http://www.ers.usda.gov/Publications/ERR49/ERR49.pdf

25. Bernell S, Edwards M, Weber B. Food insecurity in Oregon. Oregon Progress Board,

Oregon State University; 2005. Accessed July 20, 2008 at

www.oregon.gov/DAS/OPB/docs/BdUp05/Jan/Hunger.ppt.

26. Quandt SA, Arcury TA, Early J, Tapia J, Davis JD. Household food security among

migrant and seasonal latino farmworkers in north carolina. Public Health Rep.

2004;119(6):568-576.

27. Kaiser L, Baumrind N, Dumbauld S. Who is food-insecure in california? findings

from the california women's health survey, 2004. Public Health Nutr. 2007;10(6):574-

581.

64
28. Bartfeld, J., Dunifon, R. Electronic report from the economic research service: State-

level predictors of food insecurity and hunger among households with children. United

States Department of Agriculture. www.ers.usda.gov. 2005.

29. Bhattacharya J, Currie J, Haider S. Poverty, food insecurity, and nutritional outcomes

in children and adults. J Health Econ. 2004;23(4):839-862.

30. Casey PH, Simpson PM, Gossett JM, et al. The association of child and household

food insecurity with childhood overweight status. Pediatrics. 2006;118(5):e1406-13.

31. Matheson DM, Varady J, Varady A, Killen JD. Household food security and

nutritional status of hispanic children in the fifth grade. Am J Clin Nutr. 2002;76(1):210-

217.

32. McIntyre L, Glanville T, Raine KD, Dayle JB, Anderson B, & Battaglia N. Do low-

income lone mothers compromise their nutrition to feed their children? CMAJ.

2003;168(6):686-691.

33. Kaiser LL, Melgar-Quinonez HR, Lamp CL, Johns MC, Sutherlin JM, Harwood JO.

Food security and nutritional outcomes of preschool-age mexican-american children. J

Am Diet Assoc. 2002;102(7):924-929.

34. Murphy JM, Wehler CA, Pagano ME, Little M, Kleinman RE, Jellinek MS.

Relationship between hunger and psychosocial functioning in low-income american

children. J Am Acad Child Adolesc Psychiatry. 1998;37(2):163-170.

65
35. Casey PH, Szeto KL, Robbins JM, et al. Child health-related quality of life and

household food security. Arch Pediatr Adolesc Med. 2005;159(1):51-56.

36. Alaimo K, Olson CM, Frongillo EA. Family food insufficiency, but not low family

income, is positively associated with dysthymia and suicide symptoms in adolescents. J

Nutr. 2002;132(4):719-725.

37. Kleinman RE, Hall S, Green H, et al. Diet, breakfast, and academic performance in

children. Ann Nutr Metab. 2002;46(Suppl 1):24-30.

38. Champagne CM, Casey PH, Connell CL, et al. Poverty and food intake in rural

america: Diet quality is lower in food insecure adults in the mississippi delta. J Am Diet

Assoc. 2007;107(11):1886-1894.

39. Rose D, Bodor JN. Household food insecurity and overweight status in young school

children: Results from the early childhood longitudinal study.see comment. Pediatrics.

2006;117(2):464-473.

40. Oh SY, Hong MJ. Food insecurity is associated with dietary intake and body size of

korean children from low-income families in urban areas. Eur J Clin Nutr.

2003;57(12):1598-1604.

41. Dubois L, Farmer A, Girard M, Porcherie M. Family food insufficiency is related to

overweight among preschoolers. Soc Sci Med. 2006;63(6):1503-1516.

66
42. Alaimo K, Olson CM, Frongillo EA,Jr. Low family income and food insufficiency in

relation to overweight in US children: Is there a paradox? Arch Pediatr Adolesc Med.

2001;155(10):1161-1167.

43. Standard Score. Wikipedia, the Free Encyclopedia; 2008. Accessed September 24,

2008 at http://en.wikipedia.org/wiki/Z_score.

44. Hunger in the US. Food Research and Action Center; 2007. Accessed April 4, 2008 at

http://www.frac.org/html/hunger_in_the_us/hunger_index.html

45. Blumberg SJ, Bialostosky K, Hamilton WL, Briefel RR. The effectiveness of a short

form of the household food security scale. Am J Public Health. 1999;89(8):1231-1234.

46. Gulliford MC, Mahabir D, Rocke B. Reliability and validity of a short form

household food security scale in a caribbean community. BMC Public Health. 2004;4:22.

47. Bickel G, Nord M, Price C, Hamilton W, Cook J. Measuring food security in the

united states: Guide to measuring household food security. United States Department of

Agriculture, Food and Nutrition Service; 2000. Accessed January 11, 2008 at

http://www.fns.usda.gov/fsec/FILES/FSGuide.pdf

48. Dastgiri S, Mahboob S, Ostadrahimi A, Tutunchi H. Sensitivity and specificity of a

short questionnaire for food insecurity surveillance in iran. food and Nurtition Bulletin.

2007;28(1):55.

67
49. Kleinman RE, Murphy JM, Wieneke KM, Desmond MS, Schiff A, Gapinski JA. Use

of a single-question screening tool to detect hunger in families attending a neighborhood

health center. Ambul Pediatr. 2007;7(4):278-284.

50. Receiver Operating Characteristic. Wikipedia, the Free Encyclopedia; 2008.

Accessed September 24, 2008 at

http://en.wikipedia.org/wiki/Receiver_operating_characteristic.

51. Rice AL, Stoltzfus RJ, de Francisco A, Kjolhede CL. Evaluation of serum retinol, the

modified-relative-dose-response ratio, and breast-milk vitamin A as indicators of

response to postpartum maternal vitamin A supplementation... including commentary by

Habicht J. Am J Clin Nutr. 2000;71(3):799-806, 672-3.

52. National Survey on CSHCN. American Academy of Pediatrics; 2002. Accessed

January 7, 2008 at http://www.medicalhomeinfo.org/about/cshcn.html

53. Viner-Brown SI, Kim HK. Impact of caring for children with special health care

needs on the family: Rhode island's experience. Maternal & Child Health Journal.

2005;9(2 Suppl):S59-66.

54. National Survey of Children with Special Health Care Needs. The Child and

Adolescent Health Measurement Initiative; 2006. Accessed January 7, 2008 at

http://cshcndata.org/Content/Default.aspx

68
55. Davidoff AJ. Insurance for children with special health care needs: Patterns of

coverage and burden on families to provide adequate insurance. Pediatrics.

2004;114(2):394-403.

56. Chung PJ, Garfield CF, Elliott MN, Carey C, Eriksson C, Schuster MA. Need for and

use of family leave among parents of children with special health care needs. Pediatrics.

2007;119(5):e1047-55.

57. Hartley S, Ojwang P, Baguwemu A, Ddamulira M, Chavuta A. How do carers of

disabled children cope? the ugandan perspective. Child: Care, Health & Development.

2005;31(2):167-180.

58. Lutenbacher M, Karp S, Ajero G, Howe D, Williams M. Crossing community

sectors: Challenges faced by families of children with special health care needs. J Fam

Nurs. 2005;11(2):162-182.

59. Porterfield SL, McBride TD. The effect of poverty and caregiver education on

perceived need and access to health services among children with special health care

needs. Am J Public Health. 2007;97(2):323-329.

60. The Child and Adolescent Health Measurement Initiative. Children with special

health care needs (CSHCN) screener©; 1998. Accessed November 12, 2007 at

http://www.cahmi.org/pages/Home.aspx?menuid=1

69
61. Profile of General Demographic Characteristics: 2000. Current Population Survey,

US Census Bureau; 2000. Accessed May-September, 2008 at

http://censtats.census.gov/data/OR/1604173700.pdf

62. 2005 County-Level Poverty Rates for Oregon. United States Department of

Agriculture, Economic Research Service; 2006. Accessed April 2008 at

http://www.ers.usda.gov/Data/Povertyrates/PovListpct.asp?st=OR&view=Percent&longn

ame=Oregon

63. SPSS, Inc. Statistical Package for the Social Sciences: Version 15.0. Chicago, Il:

SPSS, Inc.; 2007.

64. What is Epi InfoTM? Centers for Disease Control and Prevention. 2008. Accessed

November 2008 at http://www.cdc.gov/epiinfo/

70
 Appendix A

Multi Modular Preschool Screening: Survey of

Household Food Needs

Please help us learn about food use and food needs of Tillamook preschool children
and their families. The information you provide will be used by the NW Regional
Education Service District to help improve programs and services for children and
families in Tillamook.

Participation is voluntary and all responses are completely confidential.

We do not need your name.

Section 1: These questions ask about household food use and food needs in the past
12 months

1a. Families get food in a variety of ways. During the past 12 months, has anyone in
your household grown vegetables or fruits in a home garden?

1 2
Yes No

1b Has/have your preschool child/children helped with growing vegetables or fruits

in a home garden during the past 12 months?

1 2
Yes No

71
2. There are several different programs that families use to get food. During the past
12 months, did anyone in your household use:

Circle Yes or No

a. Food stamps? Yes No

b. School lunch program? Yes No

c. School breakfast program? Yes No

d. Summer Food Program? Yes No

e. Head Start or Early Head Start? Yes No

f. Food or food vouchers through the WIC program? Yes No

g. Food from a food pantry, church, or other place that gives food Yes No
at no cost to families in need?

These next questions are about the food eaten in your household in the last 12 months
and whether you were able to afford the food you need.

3. Were the following statements OFTEN, SOMETIMES, or NEVER true for you or
the other members of your household in the last 12 months:

(a) The food that we bought just didn't last, and we didn't have money to get
more. (Please mark one answer.)

1 2 3
Often true Sometimes true Never true

(b) We couldn't afford to eat balanced meals.

1 2 3
Often true Sometimes true Never true

72
4. In the past 12 months, did you or other adults in your household ever cut the size
of your meals or skip meals because there wasn’t enough money for food?
(Please mark one answer.)

1 2 3
Yes No I don’t know

4b. If #4 is Yes, how often did this happen

1
Almost every month
2
Some months but not every month
3
Only one or two months
4
I don’t know

5. In the last 12 months, did you ever eat less than you felt you should because there
wasn't enough money to buy food?

1 2 3
Yes No I don’t know

6. In the last 12 months, were you ever hungry but didn't eat because you couldn't
afford enough food?
1 2 3
Yes No I don’t know

73
Section 2: this section asks about the people in your household.

7. Which best describes the people in your household? (Please mark one answer.)

1
Single mother living with child(ren)
2
Single father living with child(ren)
3
Couple living with child(ren)
4
Grandparent(s) living with child(ren)
5
3-generation household--grandparent(s), parent(s),
and child(ren)
6
Foster parent(s) living with child(ren)
7
Other:

8. How many persons live in your household, counting all adults and children
including yourself? ____ persons

9. What ethnic or racial group(s) do you consider yourself?

(Please mark all that apply.)

1
African American or Black
2
American Indian or Alaskan Native
3
Asian, Native Hawaiian or other Pacific Islander
4
Latino, Hispanic
5
White
6
Some other group

74
10. How much income do you expect your household to get this year from all
sources, including wages, social security, public assistance, and all other cash
income? (Please mark one answer.)

1
$0 - $5,000
2
$5,000 - $9,999
3
$10,000 - $14,999
4
$15,000 - $24,999
5
$ 25,000 - $34,999
6
$35,000 and over

11. What is the highest grade or year of school you've completed? (Please mark one
answer.)

1
I never went to school
2
8th grade or less
3
Some high school, but I did not graduate
4
High school (or I got a GED)
5
Some college or junior college, but I did not graduate from a four-year
college
6
College graduate (from a four-year college or university) or more

12. How many children do you have participating in the Multi Modular screening this
year? (Please mark one number.)

1 2 3 4
 1  2  3  4

75
Please complete questions on the next 2 pages (questions13-18) once for each child
participating in the Multi Modular Screening.

• If you have more than one child participating, please answer the following
questions one time for each child.
If you need additional forms, please ask for them at the screening.

What is the age of the child you are answering these questions for?
_______ years and _______ months

Section 3: These questions ask about your preschool child’s (or children’s) diet and
health

13. In a typical 5-day week (Monday-Friday), about how often does this preschool
child skip breakfast? (Please mark one answer.)

1
Almost every day

2
A few days

3
Almost never

14. If #13 is “almost every day”, which of the following is true:

(Please mark one answer.)

1
My child usually does not want to eat early in the morning

2
I don’t usually have time to offer breakfast

3
All of the above

4
None of the above

76
15. In general, would you say this child’s health status is:
1
Excellent

2
Very good

3
Good

4
Fair

5
Poor

16. Does this child need or use more medical care, mental health, or education
services than is usual for most children of the same age? (Please mark one
answer.)
1
Yes

2
No Æ skip to end of last page. Thank you.

3
Don’t KnowÆ skip to the last page

17. If #16 is Yes,

Is this child’s need for medical care, mental health or educational services
because of ANY medical, behavioral, or other health condition? (Please mark one
answer.)

1
 Yes

2
No

3
Don’t Know

77
 18. If #17 is Yes:
Is this a condition that has lasted or is expected to last 12 months or longer?
(Please mark one answer.)

1
 Yes

2
No

3
Don’t Know

Thank you very much for taking the time to answer these questions. Please give
these pages to assistants at the last station of the multi modal screening.

Your responses will be used to help programs and services for the health and
nutrition of children and families in Tillamook.

Please be sure to stop at the nutrition station if you have questions about child
nutrition, or if you would like to pick up some handouts about food and nutrition.

78
 Appendix B

USDA 6-Item Subset (Short Form) of the 12-month Food Security Scale
– Questionnaire
These next questions are about the food eaten in your household in the last 12 months
and whether you were able to afford the food you need. Please circle the answer that best
describes your household.
Q1 The food that I bought just didn't last, and I didn't have money to get more. Was that
often, sometimes, or never true for your household in the last 12months?
[ 1 ] Often true
[ 2 ] Sometimes true
[ 3 ] Never true
[ 4 ] Don’t Know

Q2 I couldn't afford to eat balanced meals. Was that often, sometimes, or never true for
your household in the last 12 months?
[ 1 ] Often true
[ 2 ] Sometimes true
[ 3 ] Never true
[ 4 ] Don’t know

Q3 In the last 12 months, since (date 12 months ago), did you or other adults in your
household ever cut the size of your meals or skip meals because there wasn't enough
money for food?
[ 1 ] Yes
[ 2 ] No (GO TO 5)
[ 3 ] Don’t know (GO TO 5)

79
Q3a If you answered yes to the last question, how often did this happen --almost every
month, some months but not every month, or in only 1 or 2 months?
[ 1 ] Almost every month
[ 2 ] Some months but not every month
[ 3 ] Only 1 or 2 months
[ 4 ] Don’t know

Q4 In the last 12 months, did you ever eat less than you felt you should because there
wasn't enough money to buy food?
[ 1 ] Yes
[ 2 ] No
[ 3 ] Don’t know

Q5 In the last 12 months, were you ever hungry but didn't eat because you couldn't afford
enough food?
[ 1 ] Yes
[ 2 ] No
[ 3 ] Don’t know

80
 Appendix C

Children with Special Health Care Needs (CSHCN) Screener©

1. Does your child currently need or use medicine prescribed by a doctor (other than
vitamins)?

Yes ÆGo to Question 1a

No ÆGo to Question 2

1a. Is this because of ANY medical, behavioral or other health condition?

Yes ÆGo to Question 1b

No Æ Go to Question 2

1b. Is this a condition that has lasted or is expected to last for at least 12 months?

Yes

No

2. Does your child need or use more medical care, mental health or educational
services than is usual for most children of the same age?

Yes ÆGo to Question 2a

No ÆGo to Question 3

2a. Is this because of ANY medical, behavioral or other health condition?

Yes ÆGo to Question 2b

No ÆGo to Question 3

2b. Is this a condition that has lasted or is expected to last for at least 12 months?

Yes

No

3. Is your child limited or prevented in any way in his or her ability to do the things
most children of the same age can do?

Yes ÆGo to Question 3a

81
No ÆGo to Question 4

3a. Is this because of ANY medical, behavioral or other health condition?

Yes ÆGo to Question 3b

No ÆGo to Question 4

3b. Is this a condition that has lasted or is expected to last for at least 12 months?

Yes

No

4. Does your child need or get special therapy, such as physical, occupational or speech
therapy?

Yes ÆGo to Question 4a

No ÆGo to Question 5

4a. Is this because of ANY medical, behavioral or other health condition?

Yes ÆGo to Question 4b

No Æ Go to Question 5

4b. Is this a condition that has lasted or is expected to last for at least 12 months?

Yes

No

5. Does your child have any kind of emotional, developmental or behavioral problem for
which he or she needs or gets treatment or counseling?

Yes ÆGo to Question 5a

No

5a. Has this problem lasted or is it expected to last for at least 12 months?

Yes

No

82