Ford CA, Bearman PS, Moody J. Foregone Health Care Among Adolescents. JAMA. 1999;282(23):2227–2234. doi:10.1001/jama.282.23.2227
Author Affiliations: Departments of Medicine and Pediatrics (Dr Ford) and Sociology (Drs Bearman and Moody), University of North Carolina, Chapel Hill. Dr Bearman is currently at Columbia University, New York, NY, and Dr Moody is currently at The Ohio State University, Columbus.
Context No annual national population estimates exist of the numbers of adolescents
who think they need but do not receive health care or their risk of health
Objective To describe the proportion of adolescents who report foregone health
care each year and the influence of sociodemographic factors, insurance status,
past health care, and health risks/behaviors on foregone care.
Design Cross-sectional analyses of data from wave 1 of the National Longitudinal
Study of Adolescent Health, conducted during 1995.
Setting In-home interviews conducted throughout the United States.
Participants Of 27,000 adolescents in grades 7 through 12 who were invited to participate,
20,746 (76.8%) completed the in-home interview.
Main Outcome Measure Reported foregone health care in the preceding year by individual and
family characteristics, insurance status, past health care, health/behavior
risk factors, and symptoms.
Results On average, 2268 (18.7%) of 12,079 adolescents reported foregone health
care within the past year. Factors associated with decreased risk of foregone
care included continuous private or public insurance (adjusted relative risk
[95% confidence interval], 0.64 [0.50-0.82] to 0.82 [0.70-0.96]), and having
a physical examination within the past year (0.87 [0.78-0.97] for male and
0.79 [0.70-0.88] for female adolescents). Factors associated with increased
risk of foregone care included older age (1.12 [1.06-1.15] for male), minority
race/ethnicity (1.25 [1.06-1.46] to 1.50 [1.30-1.73]), single-parent home
(1.31 [1.18-1.46] for female), and disability (2.03 [1.61-2.52] for male and
1.66 [1.20-2.10] for female). Adolescents participating in the following behaviors
were more likely to report foregone care than those who did not: daily cigarette
use (26.0% vs 16.8%; 1.34 [1.16-1.55]), frequent alcohol use (30.3% vs 18.1%;
1.34 [1.11-1.62] for male), and sexual intercourse (25.1% vs 15.1%; 1.23 [1.09-1.39]
for male and 1.39 [1.23-1.56] for female). From 32.4% to 38.2% of adolescents
with symptoms suggesting health problems reported foregone care (1.61 [1.13-2.26]
to 2.03 [1.81-2.28]).
Conclusions Our study suggests that adolescents who forego care are at increased
risk of physical and mental health problems. Efforts to improve adolescent
health through health care should address factors influencing foregone care.
Physicians' opportunities to address adolescents' health needs depend
on adolescents at risk of health problems seeking care. Health care for emotional
problems, substance use, and risk-associated sexual behaviors is particularly
important since major causes of adolescent morbidity and mortality are related
to motor vehicle injuries, interpersonal violence, suicide, alcohol and tobacco
use, and sexual behaviors.1,2
Physicians have opportunities to screen adolescents for behaviors associated
with potential negative health outcomes and for health problems during routine
health care3 and visits precipitated by adolescents'
concerns. Both depend on adolescents having access to and using health care
Previous research shows that a variety of social and economic factors
influence access to and use of health services by children and adults.4- 6 Although frameworks
for studying health care utilization differ, most recognize the importance
of an individual's predisposition to use services (eg, related to demographic
characteristics, family size, and beliefs), ability to secure services (eg,
health insurance, availability of health care, and transportation issues),
and level of need or illness (eg, perceived symptoms, general state, and diagnoses).5- 8 Previous
research specific to or including adolescents has shown that insurance status,
income, usual source of care, race/ethnicity, and perceptions of health care
settings and professionals influence health care utilization among youth.9- 15
Within the context of research on health care utilization, there has
been relatively little focus on foregone health care. This may be an important
omission when trying to understand health care utilization during adolescence.
We know that adolescents' decisions about seeking health care are complex,
particularly for the types of sensitive health issues linked to adolescent
morbidity and mortality, and that factors influencing care-seeking are in
part related to unique developmental issues. For example, adolescents report
that concerns about privacy influence their willingness to seek health care
and, in contrast to adults' concerns about privacy, adolescents worry about
disclosure of sensitive health information to their parents.16- 19
A recent survey found that among adolescents who report a history of not getting
needed health care, the leading reason was that they did not want their parents
to know, followed by financial or insurance reasons.20,21
In addition, despite emerging biological maturity, adolescents may lack the
cognitive or psychosocial maturity to effectively use health care systems
during a time when they are developing increased autonomy.22
To determine the potential importance of foregone health care among
adolescents, several gaps in our knowledge need to be addressed. First, the
number of adolescents who forego health care on an annual basis is not known.
Most national population estimates of unmet health needs among adolescents
have been limited by reliance on data from parents or other adult family members9- 14
who may not be fully aware of adolescents' health needs or decisions to forego
health care for sensitive health issues.23
Previous studies that have directly asked adolescents about unmet health needs
have been regional studies and/or limited by use of measures for foregone
care that do not specify a time interval.18- 21,24
Second, we do not know if adolescents who forego care are the "worried well"
or at risk of real health problems. Third, comparisons between factors that
are associated with foregone health care and general health care utilization
among adolescents have not been performed. If factors predicting foregone
health care are substantially different than factors predicting general health
care utilization, then policies aimed solely at improving general utilization
among adolescents may not effectively reduce foregone care.
The National Longitudinal Study of Adolescent Health (Add Health) provides
opportunities to begin to address these gaps in our existing knowledge.25 The main objectives of this study are to define a
national population estimate of annual foregone health care among adolescents,
their risk of health problems, and reasons for foregone care. In addition,
this study describes the influence of sociodemographic factors, family characteristics,
insurance status, previous health care, behaviors with risk of negative health
outcomes, and symptoms suggestive of serious physical or mental health problems
on adolescents' report of foregone health care within the past year and makes
comparisons between factors influencing foregone care vs having a physical
examination in the past year.
All procedures for Add Health were approved by the University of North
Carolina Institutional Review Board on Research Involving Human Subjects.
A stratified nationally representative sample of all public and private
high schools in the United States was drawn from the Quality Education Database
in April 1994. Schools were selected with probability proportional to size
across broad strata based on region, urbanicity, school type (public, private,
and parochial), and racial composition. For each of the 80 primary sampling
strata, a school or school pair spanning grades 7 through 12 was recruited
to participate in the study. Seventy percent of the originally selected schools
agreed to participate. Replacement schools were recruited within each community.
In total, 79% of all schools contacted participated in the study.
Throughout the fall of 1994, school questionnaires were administered
to all students attending participating schools; 77% of students enrolled
at the start of the school year participated in the study, including approximately
4000 students who were not on earlier school rosters but who were attending
school in the fall. In total, 90,118 in-school questionnaires were completed.
From the students who completed an in-school questionnaire, 27,000 individuals
were selected for a 90-minute in-home interview. Approximately 220 students
were selected from each school pair, irrespective of size, to form a main
core sample of approximately 17,600 (80 × 220). In addition to individuals
selected for the main sample, special oversamples of race/ethnic groups, disabled
youth, and individuals eligible for participation in a behavior-genetic sample
were selected for in-home interviews.
Wave 1 in-home interviews were conducted during 1995. Of the 27,000
adolescents invited to participate in the in-home interview, 20,746 (76.8%)
completed the interview. Of those completing the interview, 12,102 represented
the main core sample, and the remainder represented adolescents from oversampled
subgroups (Table 1). The sample
for this study consists of all adolescents who completed the wave 1 in-home
Add Health interview.
After obtaining written parental consent and adolescent assent, adolescent
in-home interviews were conducted using Computer-Assisted Personal Interview
and Audio Computer-Assisted Self-Interviewing technology.26
Adolescents listened to all potentially sensitive questions through earphones
and directly entered their responses into a laptop computer, thereby minimizing
potential for interviewer or parental effects on their responses. In addition,
1 parent in each household completed a written questionnaire that included
questions about the adolescents' insurance status.
Age was computed from date of birth. Race/ethnicity was measured by
self-identification and was coded "Hispanic" if participants identified themselves
at least in part to be Hispanic, white, African American, and other. Verbal
ability was assessed using a modified Peabody Picture Vocabulary Test.27
A socioeconomic status (SES) score for each parent was computed based
on the sum of 5-point ordinal rankings of education and occupation (low to
high) as reported by the adolescent participant; family SES was coded as the
maximum score of either parent. Families with only 1 parent living in the
home were designated "single-parent" families.
Based on parent/guardian questionnaire responses, insurance status was
coded as "private" if the adolescent had individual or group private coverage
(such as Blue Cross or Cigna) or used a prepaid health plan (such as a health
maintenance organization or the Civilian Health and Medical Program of the
Uniformed Services [CHAMPUS]). Insurance status was coded as "public" if the
adolescent had Medicare (from Social Security) or Medicaid coverage. Interrupted
vs continuous coverage was determined by response to the following question:
"In the past 12 months, has there been a time when (name) had no health insurance?"
Responses to "When did you last have a physical exam by a doctor or
nurse?" were dichotomized to less than a year ago vs 1 or more years ago.
Height and weight were measured at the time of interview, from which
body mass index (BMI) was calculated (weight in kilograms divided by height
in meters squared). Participants were considered overweight if their BMI was
greater than or equal to 2 SDs above the sample mean (31.4 kg/m2
for male and 30.9 kg/m2 for female participants) and underweight
if their BMI was less than or equal to 2 SDs below the sample mean (13.6 kg/m2 for male and 13.5 kg/m2 for female participants). Level
of sports activity was measured by response to "During the past week, how
many times did you play an active sport, such as baseball, softball, basketball,
soccer, swimming or football?" with responses ranging from 0 to 5 or more
times per week. Presence of disability was measured by response to "Do you
have difficulty using your hands, arms, legs, or feet because of a permanent
physical condition?" Participants who reported ever smoking at least 1 cigarette
every day for 30 days were considered to have a history of regular smoking.
Participants who reported "getting drunk or very, very high on alcohol" at
least 1 day a week over the past 12 months were considered frequent alcohol
users. Participants who responded affirmatively to "Have you ever had sexual
intercourse? When we say sexual intercourse, we mean when a male inserts his
penis into a female's vagina" were considered sexually active.
Participants were asked about symptoms over the preceding 12 months.
Those who reported 1 or more of the following symptoms almost every day or
every day were considered to be at high risk of serious mental or physical
health problems: headache, a stomachache or upset stomach, cold sweats, feel
physically weak for no reason, sore throat or cough, feeling really sick,
dizziness, or chest pain. Participants who reported crying almost every day
or every day were considered to be at high risk of mental health problems.
Participants who reported painful or very frequent urination (dysuria) almost
every day or every day were considered to be at high risk of genitourinary
tract problems, including sexually transmitted infections.
Foregone health care was measured by response to "Has there been any
time over the past year when you thought you should get medical care, but
Data were prepared using SAS Version 6.12 software (SAS Institute Inc,
Cary, NC) and statistical analyses were performed using STATA Release 5.0
software (Stata Corp, College Station, Tex). Descriptive statistics were computed
for the core sample weighted to correct for the stratified sampling design,
yielding estimates that are representative of the national adolescent population.
Multivariate analyses were performed using a theoretically derived model constructed
from variables hypothesized to influence foregone health care because of their
association with health care utilization in existing literature. Risk factors
for health problems were selected to represent common health-protective and
health-compromising domains in this age group. Because of interactions between
sex and several independent variables (eg, level of sports activities, frequency
of crying, frequency of painful urination), multivariate analyses were performed
separately for males and females.
Logistic regression analyses using unweighted data from the entire sample,
with all independent and stratifying variables simultaneously entered into
the regression model, produced adjusted odds ratios (ORs).28,29
We elected to simultaneously enter all variables into the regression model
because stepwise models that incrementally enter only a portion of the known-to-be
important variables are misspecified; this may produce biased equations and
parameter estimates.30 Because adjusted ORs
may overestimate risk ratios greater than 1 and underestimate risk ratios
less than 1 when main outcomes are common (>10%), adjusted relative risk (aRR)
and 95% confidence intervals (CIs) were calculated to better represent true
risk ratios.31 Thus, the aRR represents the
association of each independent variable with the dependent variable after
accounting for the influence of all other independent variables.
The mean (SD) age of the participants was 15.9 (1.8) years with approximately
equal representation from grades 7 through 12. Participants were predominantly
white (white, 65.2%; African American, 15.0%; Hispanic, 12.3%; and other,
7.6%) with approximately equal representation from all SES groups. Nearly
one fourth (23.5%) lived in a single-parent home. During the 12 months preceding
the interview, 71.1% of participants had continuous private insurance, 9.4%
had continuous public insurance, 1.6% had interrupted private insurance, 4.9%
had interrupted public insurance, and 13.0% had no insurance (Table 1).
The mean (SD) BMI was 22.5 (4.47) for male and 22.2 (4.37) kg/m2 for female participants. On average, participants reported playing
an active sport 2 to 3 times a week. Approximately 3% of participants reported
a significant physical disability. Behaviors associated with potential negative
health outcomes were reported by many participants: 38.3% reported a history
of vaginal intercourse, 21.6% reported a history of daily cigarette use, and
5.7% reported alcohol intoxication at least weekly during the preceding year.
Approximately 13% of participants reported physical symptoms, 1.8% reported
crying, and 1.4% reported dysuria at least several times a week over the preceding
year, suggesting high risk of physical or mental health problems (Table 2).
On average, 2268 (18.7%) of 12,079 adolescents reported at least 1 time
over the past year when they thought they should get medical care, but did
not (Table 1). Uninsured adolescents
were more likely to report foregone care than adolescents with continuous
private or public insurance (23.9% vs 17.0% vs 15.7%, respectively) and interruption
of insurance attenuated this relationship. Adolescents with lower-SES families
were more likely to report foregone care than adolescents from higher-SES
families (20.8% vs 16.5%). Adolescents who did not have an examination in
the past year were more likely to report foregone care compared with adolescents
who did (21.5% vs 17.4%).
Older adolescents were more likely to report foregone care compared
with younger adolescents (21.2% for adolescents >15 years vs 14.0% for adolescents
<15 years) and minority adolescents were more likely to report foregone
care than white adolescents (African American [22.7%] vs Hispanic [20.6%]
vs white [17.2%]). Adolescents who lived in a single-parent home were more
likely to report foregone care than adolescents living in a 2-parent home
(21.9% vs 17.8%).
Adolescents participating in behaviors that place them at risk of negative
health outcomes were more likely to report foregone care than adolescents
not participating in risk-associated behaviors (Table 2). Adolescents who had a history of daily smoking were more
likely to forego care than adolescents who did not (26.0% vs 16.8%), and adolescents
who frequently used alcohol were more likely to forego care than those who
did not (30.3% vs 18.1%). One fourth of sexually active adolescents reported
foregone care in comparison to only 15.1% of those who were not sexually active.
In addition, disabled adolescents and adolescents experiencing symptoms suggestive
of serious physical or mental health problems were approximately twice as
likely to report foregone care as adolescents in the comparison groups.
Multivariate analyses showed that the associations of age, race/ethnicity,
family composition, insurance status, past examination, disability, risk factors/behaviors,
and symptoms suggestive of serious mental or physical health problems with
foregone health care persisted while controlling for all other factors (Table 3). The only significant association
found in bivariate analyses that did not remain significant in multivariate
analyses was the relationship between family SES and foregone care.
Among the 2351 participants who reported foregone care, the most frequently
reported reason was "thought the problem would go away" (63.3%) (Table 4). Other reasons for foregone care
were fear of what the physician would say or do (15.5%), inability to pay
(14.0%), concerns about confidentiality (11.5%), parent/guardian would not
go with participant for care (11.7%), and difficulty making an appointment
Multivariate analyses were performed to assess the association of sociodemographic
and individual factors, family factors, insurance status, health risks/behaviors,
and symptoms with reasons for foregone health care among adolescents who reported
foregone care. Reasons for foregone care were coded as either nonaccess or
access issues. Responses of didn't want parents to know, afraid of what the
doctor would say or do, and thought the problem would go away were coded as
nonaccess issues because they imply ambivalence about seeking health care.
Responses of didn't know whom to see, no transportation, no one available
to go along, parent/guardian would not go, difficulty making an appointment,
and couldn't pay were coded as access issues because they imply a decision
to seek health care was followed by difficulties gaining access to a clinical
Adolescents with continuous private insurance or who were actively involved
in sports were less likely to report access reasons (aRR [95% CI], 0.80 [0.68-0.92])and
0.95 [0.91-0.99], respectively) and more likely to report nonaccess reasons
for foregone care (aRR [95% CI], 1.12 [1.07-1.17] and 1.02 [1.00-1.03], respectively)
than comparison groups. Adolescents with continuous public insurance were
more likely to report nonaccess reasons for foregone care compared with uninsured
adolescents (aRR [95% CI], 1.10 [1.05-1.15]).
Adolescents in single-parent families were more likely to report access
reasons (aRR [95% CI], 1.16 [1.03-1.29]) and less likely to report nonaccess
reasons for foregone care (aRR [95% CI], 0.93 [0.88-0.97]) compared with adolescents
in 2-parent homes. Female adolescents, adolescents with frequent physical
symptoms suggesting serious physical or mental health problems, and adolescents
with frequent dysuria were more likely to report access reasons for foregone
care than comparison groups (aRR [95% CI], 1.14 [1.02-1.26], 1.14 [1.00-1.30],
and 1.37 [1.00-1.77], respectively).
On average, an estimated 66.9% of adolescents reported having had a
physical examination within the previous 12 months. Adolescents who were white
or African American, with higher verbal skills, with higher SES families,
or with insurance were more likely to report having a physical examination
than comparison groups (Table 1).
Adolescents who reported high levels of sports activities were more likely
to report an examination compared with those with low sports activities (76.1%
vs 63.8%), and infrequent or nonsmokers were more likely to report an examination
than adolescents with a history of regular cigarette use (67.9% vs 63.8%)
Comparison of factors that were shown by multivariate analyses to have
independent effects on having had an examination and reported foregone care
show that insurance influences both outcome variables in the expected direction;
insured adolescents were more likely to report an examination and were at
decreased risk of foregone care (Table 3). However, insurance status was the only individual or family factor
found with this pattern. Age, race/ethnicity, and presence of 1 vs 2 parents
in the home were not associated with having an examination but were associated
with foregone care. Socioeconomic status continued to be associated with having
an examination but not foregone care. The presence of health risks/behaviors
or symptoms concerning health problems had varying associations with having
an examination and foregoing care and not always in the expected pattern.
For example, sexually active adolescents were both more likely to report having
had an examination and more likely to report foregone care than those who
were nonsexually active.
This study was not designed to assess the influence of source of usual
care on having an examination or foregone care. However, since participants
who reported an examination in the past year were less likely to report foregone
care compared with participants who did not, further analyses were performed
to explore the relationship between site of examination (ie, private physician's
office, community health center, school, hospital, or some other place) and
reported foregone care. Multivariate analyses failed to show an independent
effect of site of examination on foregone care.
Our findings suggest that every year 1 of 5 adolescents feel that they
should get health care but do not. These adolescents are at increased risk
of real health problems. Many adolescents who forego health care are participating
in behaviors placing them at increased risk for short- and long-term negative
health outcomes or have symptoms suggesting they may have serious emotional
or physical health needs. If health care professionals are to address major
causes of adolescent morbidity and mortality, strategies are needed to decrease
The risk for foregone care is clearly diminished when adolescents have
continuous private or public health insurance. Independent of insurance status
and in contrast to factors influencing having an examination in the past year,
older male adolescents, minority adolescents, female adolescents who live
in single-parent homes, disabled youth, and adolescents participating in risk-associated
behaviors or experiencing symptoms suggesting serious health problems are
at increased risk for foregone care. This supports the conclusions of others
that insurance is important but not alone sufficient to ensure use of health
care services.32- 34
In Canada, where there is universal health insurance, from 40% to 60% of adolescents
who are worried about issues such as emotional problems, birth control, alcohol
use, or sexually transmitted infections have not consulted a physician or
nurse even though they would like to.35 Together
these findings confirm that factors influencing adolescents to forego health
care are complex and suggest that some are distinct from factors influencing
general use of health care services.
Reasons that adolescents forego health care need to be viewed within
a developmental context. Among adolescents who forego care, 15% report that
this is in part because a parent or guardian will not go or no one is available
to go with them, and 12% report that it is partly because they do not want
their parents to know. This reflects the developmental complexities of parent-child
relationships during adolescence. On one hand, many adolescents want parents
to be involved in health care, depend on parents for transportation, and need
their parents' consent to obtain health care. On the other hand, some adolescents
have important health issues that they want to keep private16- 21
and certain types of medical care can be provided without parents' consent.36 All adolescents have the developmental task of learning
to become responsible for their own health and health care and developing
a private relationship with their physician as they approach adulthood. Difficulties
negotiating the health care system may in part be due to adolescents' relative
inexperience as consumers of health care.
Strategies to reduce foregone health care among adolescents must consider
these developmental issues. Professional organizations support increasing
adolescents' access to health services in general and confidential health
services when needed, because of concerns about adolescents' unmet health
care needs.37- 39
Approaches that have been suggested to increase adolescents' access to health
care include increasing the availability of convenient, affordable, and developmentally
appropriate health services. This includes ensuring that private, confidential
health services are available for sensitive health issues, thereby encouraging
adolescents to seek and receive needed health care for concerns related to
substance use, sexual behaviors, and mental health. Our results suggest that
these strategies have the potential to increase clinicians' opportunities
to address issues related to major causes of adolescent morbidity and mortality
by decreasing the numbers of adolescents who forego health care.
Several potential limitations of this study should be noted. Self-report
data has inherent limitations; however, adolescent reports of their own health
care behaviors are probably at least as accurate as adult reports of adolescent
health care behaviors and probably more accurate in regard to foregone care.
Measures for symptoms associated with risk of health problems were oriented
toward chronic conditions rather than acute conditions. The reliability of
our measure for foregone care is unknown, and health concerns associated with
specific episodes of foregone care were not assessed. However, concerns associated
with a single episode of foregone care might not be critically important among
at-risk adolescents; if a depressed adolescent foregoes care for a minor illness,
an opportunity to diagnose depression is missed. Finally, this study used
a school-based design and the results may not apply to out-of-school youth.
Despite these potential limitations, our findings highlight the importance
of a comprehensive approach to increasing adolescents' use of health care
services. The newly enacted State Children's Health Insurance Programs represent
an important step toward achieving the goal of continuous insurance for all
adolescents. However, insurance alone does not ensure that adolescents will
receive needed health care. Adolescents who choose to forego health care are
of particular concern because they appear to be at disproportionately high
risk of having health problems. Factors that influence adolescents to forego
care must be considered when designing systems to address adolescents' unique