Child Health
Call for Papers
Determinants of Access & Quality of Care
Chair: Elizabeth Ozer, Ph.D., M.A.
Implications for Policy, Delivery or Practice: These findings
indicate the importance of children having an individual PCP,
especially in Medicaid and managed care programs providing
care to large numbers of children.
Primary Funding Source: Ambulatory Pediatric Association
• Ethnic Disparities in SCHIP: The Role of Acculturation
Susan Haber, Sc.D., Susan Haber, Sc.D., Janet Mitchell, Ph.D.,
Sonja Hoover, M.P.P.
Sunday, June 6 • 11:30 a.m.-1:00 p.m.
• Pediatric Quality of Care is Associated with Primary Care
Provider Type
David Brousseau, M.D., M.S., Raymond Hoffmann, Ph.D.,
Jennifer Yauck, M.S., Glenn Flores, M.D., Ann Nattinger, M.D.,
M.P.H.
Presented by: David Brousseau, M.D., M.S., Assistant
Professor, Pediatrics, Medical College of Wisconsin, 9000 W
Wisconsin Avenue, Milwaukee, WI 53226; Tel: 414.266.2625;
Fax: 414.266.2635; E-mail: dbrousse@mail.mcw.edu
Research Objective: It is not known whether parents who
identify their child's primary care provider (PCP) as an
individual receive higher quality care than those who view the
child s PCP as a location/facility. Our objective was to
compare the quality of care received by children when the
identified PCP is an individual vs. a facility.
Study Design: Cross-sectional analysis of 2000 Medical
Expenditure Panel Survey
Population Studied: We analyzed data from the 2000
Medical Expenditure Panel Survey, weighted to give national
population based estimates. Only children with an identified
PCP were included. The identified PCP was classified as an
individual or facility; those identifying the PCP as the
emergency department were excluded. Quality of care was
assessed using parent reports for: 1) getting care without long
waits [routine care, sick care, phone help and brief wait times
with appointments], 2) getting needed care [referrals and
receiving care doctor believed necessary], and 3) doctor
communication [quality of doctor listening, respect, time and
understandability of explanations]. Multivariate logistic
regression was done to assess the ability to always receive
high quality care, adjusted for income, insurance coverage,
race/ethnicity, interview language, health status and age.
Principal Findings: 4,497 children, representing 52 million
children nationwide, had an identified PCP. Mean age was 8.0
years, 51% were male, and 24% were covered by Medicaid.
48% of parents identified their child s PCP as an individual.
Parents who identified the child s PCP as an individual rated
the ability to always get high quality care significantly higher
than those identifying a site of care in 2 of 3 areas: getting care
without long waits (22.4% vs. 15.8%, p<0.01) and doctor
communication (54.2% vs. 44.5%, p< 0.001), but not for
getting needed care (88.2% vs. 85.9%, p=0.06). In
multivariate analyses, children whose parents identified an
individual PCP had significantly higher odds of always getting
care without long waits (OR 1.52; 95% CI 1.18-2.02), and
highest quality doctor communication (OR 1.48; 95% CI 1.241.77), but only trended towards an increased ability for always
getting needed care (OR 1.22; 95% CI 0.99-1.54).
Conclusions: Parents who identify an individual rather than a
facility as their child's PCP report receiving higher quality care.
Presented by: Susan Haber, Sc.D., Senior Research
Economist, , RTI International, 411 Waverley Oaks Road, Suite
330, Waltham, MA 02452; Tel: 781.788.8100; Fax: 781.788.8101;
E-mail: shaber@rti.org
Research Objective: Ethnic disparities in access to care for
both insured and uninsured populations have been
documented in numerous studies. This study uses survey
data on children in Oregon’s Medicaid look-alike SCHIP
program, collected under the Child Health Insurance Research
Initiative (CHIRI), to compare the experience of Hispanic
children (the primary ethnic minority in Oregon) with nonHispanic children and to explore the extent to which
differences are explained by parental acculturation (as
measured by language).
Study Design: We conducted a telephone survey (in both
English and Spanish) of parents of children in Oregon’s
SCHIP program (N=1,206). Parents of Hispanic children were
oversampled. Analyses compared non-Hispanic children,
Hispanic children with English-speaking parents, and
Hispanic children with Spanish-speaking parents.
Population Studied: Children currently enrolled in or recently
disenrolled from Oregon’s SCHIP program as of January
2002.
Principal Findings: Over 70% of Hispanic parents and more
than half of Hispanic children in our survey do not speak
English at home. While there are substantial disparities
between Hispanic and non-Hispanic children in Oregon’s
SCHIP program, most of these are explained by parent’s
language and there are generally few significant differences
between English-speaking Hispanic and non-Hispanic
children. Children of Spanish-speaking Hispanic parents were
less likely to access primary care services than children of
English-speaking Hispanic and non-Hispanic parents, despite
having poorer parental rating of their health status.
Nonetheless, Spanish-speaking parents were less likely to say
their children had an unmet need for care. Although nearly all
children have a usual source of care, most Spanish-speaking
Hispanic children go to a hospital clinic or community health
center, while other children mainly receive their care through a
doctor’s office or HMO. Children from Spanish-speaking
families were more likely to be insured during the year prior to
enrolling in SCHIP. However, this was mostly through
Medicaid and Spanish-speaking Hispanic families were only
half as likely as English-speaking Hispanic and non-Hispanic
families to have had private insurance or paid a premium for
health insurance in the past. Among disenrollees, Spanishspeaking Hispanic children were the least likely to be currently
insured and the least likely to have a usual source of care (but
also the least likely to have an unmet need for care).
Conclusions: Acculturation is an important driver of ethnic
disparities in Oregon’s SCHIP program. Evidence of poorer
access to care for Hispanic children with Spanish-speaking
parents, despite their apparently poorer health status, is cause
for concern. These utilization gaps may be explained by
differences in parental perceptions of the need for care, as well
as cultural differences in rating health status.
Implications for Policy, Delivery or Practice: Policies that
address ethnic disparities should be sensitive to the role of
acculturation and variation even among members of the same
ethnic group. Covering Spanish-speaking Hispanic children
through policies that promote employer-sponsored insurance
may be challenging because of their lack of experience with
private insurance and their reliance on traditional safety net
providers.
Primary Funding Source: AHRQ, The David and Lucile
Packard Foundation, HRSA
• The Covering Kids and Families Evaluation: Findings
from Parent Focus Groups on Access to Care
Ian Hill, M.P.A., M.S.W, Holly Stockdale, M.P.H., M.P.P.,
Marilynn Evert, Katherine Gifford
Presented by: Ian Hill, M.P.A., M.S.W, Senior Research
Associate, Health Policy Center, The Urban Institute, 2100 M
Street, N.W., Washington, DC 20037; Tel: 202.261.5374; Fax:
202.223.1149; E-mail: ihill@ui.urban.org
Research Objective: The Covering Kids and Families (CKF)
program, funded by the Robert Wood Johnson Foundation,
seeks to expand health insurance coverage for children and
families by supporting the outreach and enrollment efforts of
community-based organizations across the country. As part
of an evaluation of CKF, researchers conducted a series of
focus groups to explore parents’ experiences accessing health
care for their children, and to assess whether and how these
experiences affect parents’ decisions to enroll in or renew
coverage for their children under Medicaid or the State
Children’s Health Insurance Program (SCHIP).
Study Design: In the spring and summer of 2003, 13 focus
groups were conducted in 5 cities—Boston, Massachusetts;
Denver, Colorado; Los Angeles, California; Mena, Arkansas;
and San Antonio, Texas. In each community, groups were
conducted with parents of children insured in either Medicaid
or SCHIP and parents of uninsured children. Three groups
were conducted with Spanish-speaking parents in two
communities—Denver and Los Angeles. The cities chosen,
identified from the 10 states participating in the case study
component of the evaluation, represented a range of “access
scenarios” and provided geographic diversity.
Population Studied: Parents of children insured in Medicaid
or SCHIP and parents of uninsured children.
Principal Findings: Generally speaking, access to primary
care was considered good by most parents with children
enrolled in Medicaid and SCHIP. This was true across all
communities. For the uninsured, there was considerable
variation in perceptions of access to care across communities.
For example in Boston, where large numbers of private
physicians participate in public programs, parents of
uninsured children reported few problems obtaining care.
Conversely, in Denver, where fewer providers participate in
Medicaid or SCHIP, or provide charity care, uninsurance
negatively affected parents’ ability to access care for their
children. For parents of both uninsured and insured children,
finding dentists and specialists was more difficult than finding
primary care providers. Spanish-speaking families confront
numerous additional barriers to obtaining care due to
language barriers and perceived discrimination. Among all
focus group participants, great value is placed on health
insurance. Even when parents encounter problems accessing
care, very few indicated a reluctance to enroll their children
into Medicaid or SCHIP, or to renew their children’s public
coverage.
Conclusions: Medicaid and SCHIP programs appear to
provide good access to primary health care services in both
rural and urban communities. Dentists and specialists that
accept Medicaid and/or SCHIP are more difficult to find.
Efforts to expand outreach and further simplify application
procedures to enroll families in Medicaid and SCHIP are
essential as families with uninsured children continue to
report difficulties accessing health care.
Implications for Policy, Delivery or Practice: Policymakers
should focus on strategies for expanding access to dental and
specialty care and consider investing additional resources
towards helping families navigate the health care system.
Communities with high numbers of Spanish-speaking
populations should focus on reducing barriers to care for
these families, such as hiring more Spanish-speaking case
workers, reducing discrimination, and helping families locate
Spanish-speaking providers.
Primary Funding Source: RWJF
• Mental Health Need and Access to Services for CSHCN
Moira Inkelas, Ph.D., M.P.H., Kandyce Larson, M.S.W.,
Kathryn Smith, M.S.N., M.A., Ramesh Raghavan, M.D., Ph.D.,
Susan Igdaloff, M.D.
Presented by: Moira Inkelas, Ph.D., M.P.H., Assistant
Professor, Health Services, UCLA School of Public Health,
1100 Glendon Avenue, Suite 850, Los Angeles, CA 90024; Tel:
310.794.0966; Fax: 310.794.2728; E-mail: minkelas@ucla.edu
Research Objective: To examine self-reported need and
unmet need for mental health care among CSHCN and their
families, and explore access disparities by demographics,
health care factors, and region.
Study Design: Analysis of a new, nationally representative
telephone survey of parents conducted by the National Center
for Health Statistics in 2001 - the National Survey of Children
with Special Health Care Needs - providing national and state
estimates of access and quality for children with chronic
conditions. Parents report if the child needed counseling or
mental health services during the past year, if the family
needed such care due to the childs condition, and if they
obtained needed care.
Population Studied: Parents of 38,866 children age 0-17 with
chronic health conditions lasting at least one year.
Principal Findings: About 25 percent of CSHCN are reported
as needing mental health care while 13 percent have a family
need. In logistic regression, low household income, high
impairment of the child, changing rather than stable health
care needs, and burden of the chronic condition on the family
- measured by reduced work hours and more than one hour of
weekly care coordination by the parent - are associated with
higher odds of child and of family mental health care need,
controlling for child age, gender, and race/ethnicity. About 16
percent of children and 20 percent of family members with a
mental health need do not receive care. Of those needing care,
odds of unmet need are higher for children who are African-
American OR 1.43, CI 1.05-1.96, uninsured OR 3.95, CI 2.536.15, have high impairment OR 1.76, CI 1.41-2.20, and in
households below 200 percent FPL. Similar access disparities
are found for family members with the exception of no
difference with race/ethnicity. With health care factors added
to the model, odds of unmet need for the child, and a family
member, are higher when the child lacks a personal physician
or usual source of care, the parent reports that the childs care
is not well organized, or family-centered care ratings are low.
Unmet needs are greatest in Western states even adjusting for
these demographic and health care factors.
Conclusions: CSHCN and their family members have mental
health needs that are not being adequately addressed. Rates of
mental health needs among family members are
independently associated with care-giving burden. Unmet
mental health needs are more frequent among children who
lack continuity of provider and receive lower interpersonal
quality.
Implications for Policy, Delivery or Practice: Children with
special health care needs and their families face coping
challenges that may increase their need for mental health
services. Increasing coordination in and family-centeredness
of primary and specialty care may improve identification of
needs, improve access, and reduce disparities. Patterns of
unmet need may be influenced by supply, reimbursement, and
provider awareness and screening practices. The need for
state-level supply and access-promoting policies are implied
by regional disparities that persist even adjusting for family
and health care factors.
Primary Funding Source: California Healthcare Foundation
• Access and Satisfaction with Care for CSHCN in
Medicaid Managed Care and Other Types of Health Plans:
An Analysis of the 2000 MEPS
Paul Newacheck, Dr.PH, Sue Kim, Ph.D., M.P.H.
Presented by: Sue Kim, Ph.D., M.P.H., Postdoctoral Scholar,
Department of Medicine, University of California, San
Francisco, 3333 California Street, Suite 335, San Francisco, CA
94143-0856; Tel: 415.502.4078; Fax: 415.502.8291; E-mail:
sekim@medicine.ucsf.edu
Research Objective: Children with special health care needs
(CSHCN) -- defined as those children who have or are at
increased risk for 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 -- use more health care services
than other children and consequently have higher health care
expenditures. Currently, there are no studies on this
population using a nationally representative sample to
examine access and satisfaction in managed care. The goal of
this study is to examine whether access and satisfaction
differed for Medicaid enrolled CSHCN in managed care
organizations versus those not in managed care.
Study Design: Medical Expenditure Panel Survey (MEPS) is
designed to produce national estimates of the health care use,
expenditures, and insurance coverage of the U.S. civilian noninstitutionalized population. We used the CSHCN definition
adopted by the federal and state Maternal and Child Health
programs, as operationalized by the CSHCN Screener, to
identify the children with special health care needs. We used
responses to a series of CAHPS questions included in the
MEPS, which are designed to measure quality of care from
consumer’s perspective, to examine access and satisfaction.
We also distinguished CSHCN enrolled in Medicaid and
whether or not they were in a managed care plan in the year
2000. Medicaid enrollment and health plan characteristics
were determined using parental responses to survey questions
in the MEPS.
Population Studied: The 2000 MEPS data set contains 6581
children of which 973 (14.3%) were identified with special
health care needs. There were 202 CSHCN in Medicaid
managed care and 101 CSHCN in Medicaid but not in
managed care. (Additional data from the 2001 MEPS will be
added prior to the AcademyHealth meeting).
Principal Findings: The preliminary results show that there is
no significant difference between children with special health
care need enrolled in Medicaid FFS and Medicaid
HMO/managed care organizations in receiving a routine
medical care visit in the past year or being able to get an
appointment when needed. Both groups were equally likely to
have had an illness or injury needing care in the past 12
months and receive care when wanted. The majority of
respondents (69% for the managed care group and 80% for
the non-managed care group) indicated that they did not have
a problem in getting necessary care in the past 12 months.
About 76% of those in non-managed care plans responded
that the child’s physician always listened carefully to the
parents, while only about 60% of those in managed care plans
felt that the child’s doctor always listened to them (P<.01).
Also, 74% of those in non-managed care plans responded that
the child’s physician always showed respect for what the
parents had to say, compared to 58% for those enrolled in
managed care plans (P<.01). In addition, a higher percentage
of the non-managed care group (69% vs. 49%) responded
that the child’s physician spent enough time with the child
and parent (P<.05).
Conclusions: It appears that there is no significant difference
in health care access for CSHCN enrolled in managed and
non-managed Medicaid plans. The majority of the
respondents were able to receive routine care, had no problem
receiving necessary care, or getting referrals to a specialist.
However, from the parents’ perspective, CSHCN enrolled in
managed care Medicaid plan seems worse off in terms of
family centered care. Those enrolled in managed care showed
a lower satisfaction with their physicians, regarding time spent
with child and parent, showing respect, and listening carefully
to the parents.
Implications for Policy, Delivery or Practice: Policy makers
may use this information to examine state Medicaid
enforcement since federal Medicaid law requires access to
care in HMO/managed care to be at least as good as it is in
Medicaid FFS. Before making any policy changes, it would be
important to explore whether lower satisfaction with managed
care plans is associated with lower clinical quality of care
received.
Primary Funding Source: Maternal and Child Health Bureau,
DHHS
Call for Papers
Child Health Challenges: New Research Approaches
Chair: Patrick Vivier, M.D., Ph.D.
Tuesday, June 8 • 11:15 a.m.-12:45 p.m.
• Racial Disparities in Hospital Admissions and Surgical
Management of Children with Appendicitis
T.M. Bird, M.S., Evan Kokoska, M.D., J. Alex Reading, M.A.,
John Tilford, Ph.D., James Robbins, Ph.D.
Presented by: T.M. Bird, M.S., Research Associate, Pediatrics,
University of Arkansas for Medical Sciences, 800 Marshal
Street, Little Rock, AR 72202; Tel: 501.364.4936; Fax:
501.364.1552; E-mail: birdtommym@uams.edu
Research Objective: To assess racial and insurance
disparities and trends in hospital admissions and surgical
management of children with appendicitis.
Study Design: We analyzed data from the Healthcare Cost
and Utilizations Project (HCUP) Kids’ Inpatient Database
(KID) for the years 1997 and 2000. The KID consists of an
80% sample of pediatric discharges from all hospitals in
participating states; 22 in 1997 and 27 in 2000. Data are
weighted to provide national estimates. We reviewed all
records for children = 18 years of age at time of admission
with an ICD-9 discharge diagnosis code for acute appendicitis.
Outcomes included type of surgery (open or laparoscopic),
length of stay (LOS), operative delay, incidence of surgical
complications, and hospital charge.
Population Studied: The 1997 KID has a total of 1.9 million
discharge records from 2,521 hospitals of patients 18 years of
age and younger. The 2000 KID includes 2.5 million discharge
records from 2,784 hospitals of patients 20 years of age and
younger. We restricted our analysis to those patients age 18
years or less at the time of admission with an ICD-9 diagnosis
code of simple (540.9) or complicated (540.0 and 540.1)
appendicitis.
Principal Findings: Weighted to represent national estimates,
73,330 cases were diagnosed with appendicitis in 1997 and
91,100 cases in 2000. Of these, only 6.6% in 1997 and 6.8%
in 2000 were African American. In the year 2000, African
Americans accounted for 15% of children 0 to 18 years of age
in the U.S. population and 18% of hospitalizations in the KID.
Even though the hospital incidence rate is lower than
expected, African American children were more likely than
Caucasians to be admitted with complicated appendicitis
(odds ratio (OR) 1.16, p = 0.002), more likely to have a
complication (OR = 1.34, p = 0.038), and less likely to undergo
laparoscopic surgery (OR = 0.75, p < 0.001). When compared
to Caucasians, African Americans also had a greater surgical
delay, a longer length of stay, and a greater hospital charge.
Conclusions: This national study demonstrates substantially
lower rates of hospitalization for appendicitis and worse
outcomes among African American children. The finding of a
lower hospitalization rate for African Americans has not been
previously reported in the epidemiology of appendicitis.
Implications for Policy, Delivery or Practice: Higher
perforation and complication rates, longer wait times until
first procedure, and lower rates of laparoscopic techniques
among African American children illustrate disparities in
access to care for minority youth.
• Changes in Newborn Delivery during a Period of Rapid
Expansion of Medicaid Managed Care in Los Angeles and
Orange County, California
Ruey-Kang Chang, M.D., M.P.H., Sandra Rodriguez, B.S.,
Thomas Klitzner, M.D., Ph.D.
Presented by: Ruey-Kang Chang, M.D., M.P.H., Assistant
Professor, Pediatrics, David Geffen School of Medicine at
UCLA, B2-427 MDCC, 10833 Le Conte Avenue, Los Angeles,
CA 90095; Tel: 310.825.6197; Fax: 310.825.9524; E-mail:
rkchang@ucla.edu
Research Objective: We sought to determine changes in
newborn deliveries during a period of rapid Medi-Cal
managed care (MMC) expansion in Los Angeles and Orange
County, two of the most populated counties in California. The
objectives of this study were to determine if: 1) The Medi-Cal
newborn delivery market became decentralized over time with
MMC expansion; 2) Decentralization increased the rate of
inter-hospital transfer; 3) Decentralization changed the
pattern of care (rate of cesarean section); 4) Decentralization
changed the outcomes of care.
Study Design: The California statewide abstracted hospital
discharge data from the January 1990 to December 1999 were
used to conduct the analysis. California Vital Statistics on
birth record were used to compare the numbers of births each
year. American Hospital Association database on hospitals
was used to determine the ownership of the hospitals.
Herfindahl-Hirschman Index (HHI) was calculated to measure
the distribution of newborn deliveries among hospitals.
Percentages of deliveries by cesarean section were calculated.
Outcomes of birth and deliveries were measured by
calculating percentages of in-hospital deaths, birth
weight<1500 gm, and inter-hospital transfers.
Population Studied: All newborn deliveries in Los Angeles
County and Orange County, CA hospitals between 1990 and
1999.
Principal Findings: There were 2,351,199 newborn deliveries
(LAC 1,832,287, OC 518,922) in 124 hospitals 1990-1999. The
number of newborn dekiveries per year decreased by 21% in
that period. Dramatic decline in the numbers of deliveries (up
to 80%) was seen in all four LAC county hospitals. The overall
market share of public hospitals decreased from 22% in 1990
to 8% in 1999 and the loss of market share was mostly picked
up by not-for-profit hospitals, which increased market share
from 54% to 68%. The newborn delivery market for Medi-Cal
patients in both LAC and OC became less concentrated as
evidenced by decreasing HHI. During the same period, the
HHI for private patients increased slightly (more
concentrated) for both counties. The rate of cesarean section
for private patients decreased between 1990-1996, followed by
a rise 1996-1999. In contrast, cesarean section rate for MediCal patients did not decline but continued to rise over the
study period. Patterns of convergence in the CS rate over the
study period between private and Medi-Cal patients were seen
in both LAC and OC. The rate of newborn transfer for patients
with Medi-Cal in LAC remained unchanged (around 0.8%).
The newborn in-hospital mortality rates in both LAC and OC
decreased and there was no significant changes in the rate of
low birth weight newborns in LAC and OC over the 10-year
study period.
Conclusions: The expansion of MMC in LAC and OC was
associated with a dramatic shift of newborn deliveries from
public hospitals to not-for-profit hospitals. The increase in
cesarean section rate among Medi-Cal patients may be a
result of this market shift. At the same time, outcomes of
newborn deliveries, including in-hospital mortality, interhospital transfer and low birth weight rate, did not change
significantly.
Implications for Policy, Delivery or Practice: Medicaid
managed care programs appear to have resulted in
decentralization of newborn deliveries among hospitals in LAC
and OC. Such decentralization was associated with changes
in the patterns of care, but did not have adverse effect on the
outcomes of care.
Primary Funding Source: AHRQ
• Use of Atypical Antipsychotic Drugs by Children and
Adolescents in the United States: A Retrospective Cohort
Study
Lesley Curtis, Ph.D., Leah Masselink, B.A., Truls Ostbye, M.D.,
Ph.D., Steve Hutchison, Ph.D., Ranga Krishnan, MB, ChB,
Kevin Schulman, M.D.
Presented by: Lesley Curtis, Ph.D., Assistant Research
Professor, Center for Clinical and Genetic Economics, Duke
Clinical Research Institute, Duke University Medical Center,
P.O. Box 17969, Durham, NC 27715; Tel: 919.668.8101; Fax:
919.668.7124; E-mail: lesley.curtis@duke.edu
Research Objective: Previous research has suggested
increased use of atypical antipsychotic drugs in pediatric
populations. We examined the prevalence of atypical
antipsychotic use by age and sex among children and
adolescents in the United States.
Study Design: We conducted a retrospective cohort study
using the administrative claims database of a large
pharmaceutical benefit manager. The main outcome measure
was prevalence of prescriptions, which we calculated for the 5
atypical antipsychotic drugs and for concurrent use of
antidepressant drugs. We stratified the analysis to explore age
and sex differences in use of the drugs, and we used chisquare or Fisher exact tests to test for differences in
proportions and prevalence.
Population Studied: Subjects were included if their insurance
carriers required the benefit manager to track claims at the
subject level, excluding those whose plans used a single
identifier for multiple family members. The data set consisted
of 6,213,824 outpatients aged 19 years and younger who were
enrolled continuously from January through December 2001
and who filed at least 1 claim for a prescription drug during
that period. Each subject's claims were linked using a unique
identifier encrypted to ensure confidentiality. A total of 1171
insurers were represented, covering all 50 states, the District
of Columbia, and US territories.
Principal Findings: Prevalence of atypical antipsychotic use
was 267.1 per 100,000 subjects (16,599/6,213,824). Compared
to the overall study population, subjects with at least 1 claim
for an atypical antipsychotic were more likely to be male
(70.7% vs 51.2%; P<.001) and were older (76.9% vs 51.6% of
subjects aged 10 through 19 years; P<.001). Prevalence peaked
at 594.3 per 100,000 among male subjects aged 10 through 14
years, and at 291.0 per 100,000 among female subjects aged
15 through 19 years. Nearly one quarter (3830/16,599) of
subjects with a claim for an atypical antipsychotic were aged 9
years or younger, and nearly 80% of those (3021/3830) were
male. In contrast, the prevalence of concurrent antidepressant
use was significantly higher among female subjects than
among male subjects (192.6 vs 106.8 per 100,000; P<.001).
Conclusions: Although evidence regarding the safety and
efficacy of atypical antipsychotics in pediatric populations is
limited, nearly one quarter of children and adolescents with
claims for those drugs were younger than 9 years.
Understanding the long-term effects of early and prolonged
exposure to atypical antipsychotics is crucial given their
extensive use in pediatric populations.
Implications for Policy, Delivery or Practice: Off-label use of
prescription drugs in children and adolescents is common,
and some evidence suggests that primary care physicians are
more likely than experts in pediatric psychopharmacology to
prescribe psychotropic drugs. For patients covered by
Medicaid, the incentive to treat behavioral problems using
drug therapy may be especially strong, because Medicaid
provides almost complete coverage for drug therapy and only
limited reimbursement for psychiatric evaluation. Researchers
must develop a better understanding of the circumstances in
which physicians prescribe atypical antipsychotics to children
and adolescents, especially given the scarcity of data regarding
long-term effects of these drugs.
Primary Funding Source: AHRQ
• Newborn Hospitalizations for Birth Defects In the Pre
and Post Folic Acid Fortification Periods
James Robbins, Ph.D., John Tilford, Ph.D., T. M. Bird, M.S.,
Mario Cleves, Ph.D., Joseph Thompson, M.D., M.P.H.,
Charlotte Hobbs, M.D. Ph.D.
Presented by: James Robbins, Ph.D., Professor, Pediatrics
and Psychiatry, University of Arkansas for Medical Sciences,
800 Marshall Street, Little Rock, AR 72202; Tel: 501.364.8370;
Fax: 501.364.1552; E-mail: robbinsjamesm@uams.edu
Research Objective: To compare newborn hospitalization
rates for birth defects during 1997, prior to mandatory
fortification of grains with folic acid, and during 2000, after
mandatory fortification.
Study Design: Analyses use the Kid s Inpatient Database
(KID) of hospital discharges for 1997 and 2000. The KID is the
only all-payer, nationally representative, inpatient care
database for children in the United States. Data are weighted
to represent all births in the U.S. and are comparable across
years. Population estimates and standard errors were
calculated from weights provided by the Agency for Healthcare
Research and Quality. Analyses compared newborn
hospitalization rates for select birth defects per 10,000 live
births across years for defects expected to be influenced by
folic acid and defects not known to be influenced by folic acid.
Population Studied: The 1997 and 2000 KID were accessed
to identify hospital discharges for congenital anomalies easily
recognized at birth or likely to require inpatient care within the
first few days of life. ICD-9 discharge codes for newborns less
than 10 days of age were used to define anomalies. To avoid
duplicate counting of cases, infants discharged to an acute
care hospital were excluded. These patients would likely be
captured in the sample of the referral hospital.
Principal Findings: Among birth defects thought to be
affected by folic acid fortification, newborn hospitalization
rates decreased significantly for four defects: spina bifida (3.30
to 2.75 per 10,000 live births), upper limb defects (2.13 to
1.88), lower limb defects (1.37 to 1.11), and common truncus
arteriosis (0.80 to 0.57). Rates of orofacial clefts did not
change. Newborn hospitalization rates of defects not known
to be associated with folic acid (hypospadias, diaphragmatic
hernia, obstetric genitourinary defects) did not change.
Conclusions: Decreases in newborn hospitalization rates for
spina bifida and limb defects are consistent with the known
protective effect of folic acid. Results do not support a
reduction in rates of orofacial clefts, despite literature
suggesting that folic acid should reduce the risk of these
defects. Differences over time may have been influenced by
differences in diagnostic precision, maternal age, or other
maternal characteristics.
Implications for Policy, Delivery or Practice: In response to
strong evidence that folic acid can reduce the prevalence of
neural tube defects, the FDA mandated fortification of grain
products with folic acid to begin in 1998. This study of
national hospitalization rates for birth defects supports a more
pervasive impact of the fortification policy.
Primary Funding Source: CDC
• Age-Specific Pathophysiologic Mortality Models for
General Pediatric Inpatients
Ying Tabak, Ph.D., Richard Johannes, M.D., M.S.
Presented by: Ying Tabak, Ph.D., Senior Biostatistician,
Research, CardinalHealth - Data and Clinical Information, 500
Nickerson Road, Marlborough, MA 01752; Tel: 508.571.5120;
Fax: 508.571.6101; E-mail: ying.tabak@cardinal.com
Research Objective: Studies of risk-adjusted mortality in
pediatric patients have been limited to relatively small
samples of ICU patients in teaching hospitals. The validity for
community pediatric patients remains unknown. Moreover,
age-specific weights have been applied only to vital signs (VS).
No age-specific effect of laboratory values has been studied in
spite of rapid developmental changes in laboratory values
across childhood, especially early childhood. The goal of this
study is to develop and validate age-specific pediatric mortality
risk models for general pediatric inpatients using commonly
measured objective pathophysiologic variables, including
laboratory assessment, vital signs, and status of neurologic
function.
Study Design: Retrospective lab data were imported from
hospitals’ electronic systems or abstracted from chart review,
if automated data were not available. Vital signs and the
Glasgow Coma Scores were abstracted from chart review.
Age-specific normal ranges of laboratory and VS variables
were crafted by referring to the medical literature and
empirical examination of the data for three age groups,
Neonates (0-30 days), Young Children (1-23 months), and
Older Children (2-17 years). Logistic regressions were used to
derive risk adjustment models, which were validated with
Bootstrapping techniques.
Population Studied: A total of 243,173 pediatric inpatients,
excluding healthy newborns, admitted in 2000-2001 from 215
acute care hospitals were included. There were 93,011
neonates (507 deaths, .55% mortality), 46,152 young children
(170 deaths, .37% mortality), and 104,010 older children (421
deaths, .40% mortality).
Principal Findings: Variables that entered into all three
models with generic thresholds included low platelets (<=150
10^9/L) and hypothermia (<=95F). Variables that entered into
all three models with age-specific thresholds included elevated
creatinine, hyperglycemia, lower heart rates, and systolic
hypotension. The thresholds of elevated creatinine decreased
with age (from .8mg/dL to .5mg/dL). The thresholds for
hyperglycemia increased with age (>110 mg/dL, 200 mg/dL,
and 300 mg/dL respectively). The risk of lower heart rates
posed greater risk for neonates and younger children and the
threshold was much higher for neonates (<=120) than older
children (<=60). The threshold of systolic hypotension
increased significantly across age groups (<=50, <=70, and
<=80 respectively). All of these variables had odds ratios
greater than 1.5 and p<.05. The three models contain between
11-15 variables, all of which were significant at p<.05 or better.
The models showed excellent predictive power with c-statistics
of .85, .86, and .89 respectively.
Conclusions: The thresholds for pathophysiologic
determinants vary by age. Risk adjustment using
pathophysiologic variables should consider the age specificity
in order to capture and quantify risks more precisely. Agespecific weights are more clinically plausible. The current
models also demonstrated that commonly measured
pathophysiologic variables are adequate in predicting
mortality for general pediatric populations.
Implications for Policy, Delivery or Practice: This study
provides age-specific risk adjustment tools that are clinically
valid and cost-effective. The pathophysiologic data are either
automated or have potential for automation. The
parsimonious models have practical values to providers,
purchasers, government agencies, and public health groups
for quality improvement, cost containment, and performance
reporting in hospitalized general pediatric patients.
Related Posters
Poster Session A
Sunday, June 6 • 6:45 p.m.-8:00 p.m.
• Parental Factors Associated with Childhood
Immunization Status: Results from the National
Immunization Survey
Norma Allred, M.S.N., Ph.D., Kate Shaw, M.S., Tammy
Santibanez, Ph.D., Donna Rickert, Dr.PH
Presented by: Norma Allred, M.S.N., Ph.D., Epidemiologist,
National Immunization Program, Centers for Disease Control
and Prevention, 1600 Clifton Road, N.E., Mailstop E-52,
Atlanta, GA 30333; Tel: 404.639.8722; Fax: (404) 639-8614; Email: NAllred@cdc.gov
Research Objective: To measure the association of parental
knowledge, attitudes and beliefs about immunization and the
immunization status of their 19-35 month old children.
Study Design: A supplemental module was administered
from July 2001 to December 2002 to randomly selected
parents in the National Immunization Survey (NIS), a
national random digit telephone survey of households with
children aged 19-35 months conducted annually to determine
immunization status. The module included selected
questions regarding vaccine safety and side effects,
simultaneous vaccine administration, new vaccines and
vaccinating an ill child. Chi square tests and logistic
regression analyses were performed to examine associations
of attitudes and beliefs with demographic characteristics and
4:3:1:3:3 up-to-date status (4+ doses of diphtheria and tetanus
toxoids and pertussis vaccine, 3+ doses of polio vaccine, 1+
dose of measles-containing vaccine, 3+ doses of Haemophilus
influenzae type b vaccine and 3+ doses of hepatitis B vaccine).
Population Studied: Nationally representative, populationbased sample of households with children between the ages
of 19-35 months
Principal Findings: Between July 2001 and December 2002,
48,531 NIS telephone interviews were completed. For 32,707
of those interviews, verbal consent to contact identified
providers was obtained and adequate provider data were
obtained, thus yielding validated vaccination coverage data for
67.4% of interviews. Of the 9908 parents who were selected
and completed the additional module, 7810 (79%) had
provider-verified data and are included in this analysis. There
was a significant association between parental concern about
vaccine safety and their child’s immunization coverage; for
those with the highest concern, coverage was 75%, for
intermediate concern 65%, and for the lowest concern 53%,
(p<0.01, n=7789). Children of parents who disagreed that
children who stayed at home should receive the same
vaccines as children in day care or pre-school had lower
coverage levels (65% vs. 75%, p<0.05, n=7727). These
differences remained significant after controlling for
demographic characteristics. No other significant
associations were found.
Conclusions: The finding from this study that parental
concern over vaccine safety is associated with immunization
status has been reported in previous research. The significant
association found between parental disagreement over which
children should be immunized based on attendance in a day
care or pre-school is a new finding that should be further
explored.
Implications for Policy, Delivery or Practice: Strategies that
health care providers can use to reassure parents about
vaccine safety and educate them about the importance of
immunizations for all children may have a positive impact on
vaccine coverage.
Primary Funding Source: CDC
• Prolonged Bottle-Feeding Associated with Iron Deficiency
Jane Brotanek, M.D., Jill Halterman, M.D., M.P.H., Michael
Weitzman, M.D., Karen Nead, M.D., M.P.H., Peggy Auinger,
M.S.
Presented by: Jane Brotanek, M.D., Fellow, General
Pediatrics, Pediatrics, HRSA, 601 Elmwood Avenue, Box
278881, Rochester, NY 14642; Tel: 585.275.1544; Fax: (585) 5060150; E-mail: jane_brotanek@urmc.rochester.edu
Research Objective: Iron deficiency in infancy and early
childhood is associated with behavioral and cognitive delays.
Few studies have explored the contribution of infant feeding
practices, particularly prolonged bottle-feeding, to iron
deficiency among toddlers.
The objective of this study is to examine the association
between prolonged bottle-feeding and iron deficiency among a
nationally representative sample of children ages 1-3 years.
Study Design: The National Health and Nutrition
Examination Survey III 1988-1994 provides cross-sectional
data on the feeding practices of children 1-3 years of age and
contains measures of iron status including transferrin
saturation, free erythrocyte protoporphyrin, and serum ferritin.
Children were considered iron-deficient if any 2 of these values
were abnormal for age and gender. The prevalence of iron
deficiency and the duration of bottle-feeding were determined
for toddlers from three major ethnic groups (Black, White,
Mexican-American). Bivariate comparisons of the prevalence
of iron deficiency with duration of bottle-feeding were
performed. Logistic regression was used to examine the
relationship between iron deficiency and duration of bottlefeeding, controlling for the effects of age, gender, lead level,
obesity, maternal education and poverty status, and duration
of infant breastfeeding and/or formula feeding.
Population Studied: A nationally representative sample of
children ages 1-3 years living in the United States between
1988-1994.
Principal Findings: The prevalence of iron deficiency was 3%
among White children, 4.7% among Black children, and 10.6%
among Mexican-American children ages 1-3 years (p<0.001).
With increasing duration of bottle-feeding, the prevalence of
iron deficiency among all children increased (2.2%, bottle-fed
<= 12 months; 6.6%, stopped bottle-feeding between 13 and
23 months; and 7.4%, stopped bottle-feeding between 24 and
48 months, p<0.001). At 24-48 months, 35.5% MexicanAmerican children were still being bottle-fed, as compared
with 17.3% White children and 15.7% Black children (p<0.001).
In a multivariate model, bottle-feeding into the 24-48 month
age range and Mexican ethnicity both independently predicted
iron deficiency (OR 2.7, 95% CI 1.3, 5.5 and OR 3.1, 95% CI 1.7,
5.7, respectively).
Conclusions: Children with prolonged bottle-feeding are at
risk for iron deficiency.
Implications for Policy, Delivery or Practice: Screening
practices may need to be modified to reflect this increased
risk, especially the substantial risk among Mexican-American
toddlers.
Primary Funding Source: HRSA
• The Effect of Primary Care Practice Characteristics on
Pediatric Emergency Department Utilization
David Brousseau, M.D., MS, John Meurer, M.D. M.B.A.,
Evelyn Kuhn, Ph.D., Mayme Isenberg, M.S., Marc Gorelick,
M.D., MSCE
Presented by: David Brousseau, M.D., MS, Assistant
Professor, Pediatrics, Medical College of Wisconsin, 9000 W
Wisconsin Avenue, Milwaukee, WI 53226; Tel: 414.266.2625;
Fax: (414) 266-2635; E-mail: dbrousse@mail.mcw.edu
Research Objective: To determine the relationship between
the characteristics of a pediatric practice and ED utilization
through analysis of practice and provider characteristics.
Study Design: Retrospective cohort
Population Studied: Retrospective cohort of children who
received their primary care from a single network of twelve
pediatric practices (28 pediatricians) and who were insured by
United Healthcare of Wisconsin. From an insurance
administrative dataset, demographic data and utilization data
from 12/1/1999 through 7/31/2001 were collected for all
children who had at least one visit to a practice physician.
Primary care practices were contacted and characteristics of
the physicians/practices were obtained. A backward stepwise
hierarchical multiple regression model, controlling for person
level demographics, was used to evaluate the association
between these characteristics and ED reliance (the percentage
of all visits for a patient that occur in the ED).
Principal Findings: A total of 9,208 children were included in
the analysis. Their average age was 7.7 years, 51% were male,
and 21% were insured through Medicaid. The children made
52,717 visits to healthcare providers, of which 4,512 were to the
emergency department. Children whose primary care
providers had fewer weekends of call (p<0.05), fewer weekend
office hours (p<0.001) and a higher percentage of Medicaid
patients (p<0.002) showed higher ED reliance. Fewer patients
per hour worked (p=0.08) also trended towards higher ED
reliance, but this may be explained by the correlation between
a lower number of patients and fewer years of service.
Conclusions: Children cared for by more accessible primary
care providers, as measured by more frequent weekend call
and working in practices with more weekend hours, rely less
on the ED for care.
Implications for Policy, Delivery or Practice: Changes in
practice characteristics, which make an individual provider
more available, may decrease emergency department
utilization.
Primary Funding Source: Children's Hospital Foundation
• When the Bough Breaks: Provider-Initiated
Comprehensive Care is More Effective and Less Expensive
for Sole-Support Parents on Social Assistance
Gina Browne, Ph.D., Reg.N., Carolyn Byrne, Ph.D., Reg.N.,
Jacqueline Roberts, M.Sc., Reg.N., Amiram Gafni, Ph.D.,
Susan Whittaker, M.Sc., B.Sc., B.A.
Presented by: Gina Browne, Ph.D., Reg.N., Founder &
Director, System-Linked Research Unit; Professor, Nursing &
C.E.&B., School of Nursing, McMaster University, Faculty of
Health Sciences at Frid Street, 75 Frid Street, Building T30,
Hamilton, Ontario, L8P 4M3; Tel: 905.525.9140 Ext. 22293;
Fax: (905) 528-5099; E-mail: browneg@mcmaster.ca
Research Objective: This five-year study conducted in
Ontario, Canada is designed to assess the effects and expense
of adding a mix of provider-initiated interventions to the
health and social services typically used in a self-directed
manner by sole-support parents and their children receiving
social assistance in a national system of health and social
insurance.
Study Design: Families were randomized to receive proactive
in-home visits by nurses, employment retraining, and age
appropriate childcare, recreation, or skills development for all
children in the household alone or combined and further
compared to their own self-directed use of services.
Population Studied: This was a 5-arm prospective
randomized trial of 765 single parents on social assistance
and their 1330 children (aged 0 - 24 years).
Principal Findings: Results from a 2-year interim analysis
show that providing social assistance families with proactive
comprehensive care (health promotion, employment
retraining, and recreation activities for children) compared to
allowing families to fend for themselves in a self-directed
manner, results in 15% more exists from social assistance
within one year and substantial savings to society in terms of
social assistance payouts.
Conclusions: It is more effective and creates more savings,
yet no more expensive to provide health and social services in
a comprehensive fashion to parents and all children in a
household on social assistance.
Implications for Policy, Delivery or Practice: This study
funded by Health Canada presents clear evidence that
providing comprehensive care to social assistance recipients
produces tremendous short and long term financial gains and
societal benefits.
Primary Funding Source: Health Canada
• Predictors of High Cost Admissions for Congenital Heart
Surgery
Jean Connor, DNSc., R.N., CPNP, Kimberlee Gauvreau, ScD,
Kathy Jenkins, M.D., M.P.H.
Presented by: Jean Connor, DNSc., R.N., CPNP, Research
Fellow, Cardiology, Children's Hospital Boston, 300
Longwood Avenue, Boston, MA 02115; Tel: 617.355.8890; Fax:
617-739-5802; E-mail: Jean.Connor@cardio.chboston.org
Research Objective: To identify patient, institutional, and
regional factors associated with high cost admissions for
congenital heart surgery.
Study Design: We used hospital discharge data from the
Healthcare Cost and Utilization Project (HCUP) Kid’s
Inpatient Database (KID) year 2000(data from 27 states).
High cost admissions were defined as those in the highest
decile for total hospital charges. Univariate and multivariate
analyses with and without patients who died were used to
determine demographic and hospital predictors for an
increased frequency of high cost cases. Case mix severity was
approximated using RACHS-1 risk groups. Regional and state
differences were also examined.
Population Studied: Cases of congenital heart surgery < 18
years of age.
Principal Findings: Among 10,569 cases of congenital heart
surgery identified, median hospital charges were $53,828.
Statewide differences in the number of high cost admissions
were present; CA, CO, FL, HI, PA and TX had more high cost
cases and ME and SC had fewer. Subsequent analyses were
performed adjusting for baseline state effects. Multivariate
analyses with and without patients who died using generalized
estimating equations revealed RACHS-1 risk category (OR
1.69-14.7), age (OR 3.9), prematurity (OR 4.7), the presence of
other major non-cardiac structural anomalies (OR 2.5),
Medicaid insurance (OR 1.47) and admission during a
weekend (OR 1.64) to be independent predictors of a higher
frequency of high cost cases (p<0.05). Although some
institutional differences were noted in univariate analyses,
bedsize, teaching and children’s hospital status, hospital
ownership, and hospital volume of cardiac cases were not
independently associated with greater numbers of high cost
admissions.
Conclusions: States varied in the frequency of high cost
admissions for congenital heart surgery. Patients with greater
disease complexity, younger age, prematurity, other
anomalies, Medicaid and admitted during a weekend were
more likely to result in high cost. Institutions of various types
did not differ in frequency of high cost admissions, regardless
of children’s hospital or teaching status.
Implications for Policy, Delivery or Practice: Using hospital
charge data as a surrogate for cost could be considered useful
to health care providers, institutions, payors, and policy
makers when formulating interventions and policies to
allocate resources for complex congenital heart surgery
• Children's Healthcare in the National Healthcare Quality
and Disparities Reports: Findings, Challenges, and
Opportunities
Denise Dougherty, Ph.D., Susan Meikle, M.D., Pamela Owens,
Ph.D., Edward Kelley, Ph.D., Ernest Moy, Ph.D.
Presented by: Denise Dougherty, Ph.D., Senior Advisor, Child
Health, U.S. Department of Health and Human Services,
Agency for Healthcare Research and Quality, 540 Gaither
Road, Rockville, MD 20850; Tel: 301.4271868; Fax: 301 4271347; E-mail: ddougher@ahrq.gov
Research Objective: In 2003, the UDSDHHS, led by the
Agency for Healthcare Research and Quality, released two
landmark reports, the National Healthcare Quality and
National Healthcare Disparities Reports. We will present
selected findings from these reports on children's healthcare
quality and disparities, and provide an analysis of why the
reports did not pay more attention to children's issues, and
the challenges and opportunities for the reports in the future.
Study Design: The NHQR and NHDR were developed by
AHRQ and 2 federal interagency advisory committees who
relied in turn on non-federal advice. NHQR reports on the six
IOM quality domains. The NHDR reports on disparities in
quality, access, and use for AHRQ's 8 congressionallymandated priority populations (children, women, elderly, lowincome, inner-city, rural, people with special health care
needs, racial and ethnic minorities) by race/ethnicity and
income. Data reported in the reports were driven by criteria
of national representativeness, collection over time, and a
focus on high-cost, high-prevalence conditions at a national
level. In addition, the first reports were intended to provide
broad national snapshots of healthcare quality and do not
provide in-depth analyses of quality for particular populations.
Our analysis of the reports from a child healthcare perspective
focuses on the implications of these criteria for reporting on
children's healthcare quality.
Population Studied: This is more a policy analysis than a
study of a population per se, although findings from the
reports on children will be presented. Available data on
children's healthcare quality, using the reports' criteria, were
studied by the report development team. For this analysis,
we contrasted the reports' criteria for inclusion with the reality
of children's healthcare and availability of children's healthcare
quality and disparities information.
Principal Findings: Quality and disparities issues for children
were identified in the NHQR and NHDR. For example, the
reports found that the poorest children are nearly twice as
likely to be admitted to the hospital for gastroenteritis (an
indicator of possible poor ambulatory care), and that black
children are nearly twice as likely to be admitted to the hospitl
for asthma (another indicator of possibly poor ambulatory
care). A wealth of data are available in the appendix tables
accompanying the report, but may not be highlighted in the
main reports. Quality measures for children are relatively new,
and few measures are being used to collect data in routine
nationally representative data collection efforts. Children are
less likely than adults to suffer from the high-prevalence, highcost chronic conditions that the report focused on.
Conclusions: The reports provide guidance on quality and
disparities gaps to close in children's health care. AHRQ
seeks input on how to improve reporting on children's
healthcare in future reports. In the longer term, attention to
children may improve 1) with development and
implementation of additional quality measures for children; 2)
if there are changes in the ways that priorities are set in health;
or 3) both. Meeting participants will be asked for input on
these issues.
Implications for Policy, Delivery or Practice: Findings on
children in the first NHQR and NHDR can be used to set
priorities for local quality improvement and disparities
reduction efforts. The reports, in combination with other
initiatives on children's healthcare quality, also provide a
roadmap for where measurement of quality and disparities in
children's health care needs to go.
Primary Funding Source: AHRQ
• Recommendations for the Development of Child Health
Information Systems to Meet Medical Care and Public
Health Needs
Kristin Saarlas, M.P.H., Nicole Fehrenbach, M.P.P., Alan
Hinman, M.D., M.P.H., David Ross, ScD, Nicole Fehrenbach,
M.P.P.
Presented by: Nicole Fehrenbach, M.P.P., Senior Research
and Evaluation Associate, Public Health Informatics Institute,
Emory University, 750 Commerce Drive, Suite 400, Decatur,
GA 30030; Tel: 404.687.5622; Fax: (800) 765-7520; E-mail:
nfehrenbach@taskforce.org
Research Objective: The objectives of the December 2003
conference were to:
1. Review national initiatives and other factors influencing the
development of child health information systems
infrastructure.
2. Develop concrete recommendations, reflecting the input of
stakeholders, for the development of immediate and 3-5 year
actions to integrate the information about the health care that
a child receives, from the moment he or she is born, in a
simple yet comprehensive format so that the end user
(parents, families, health professionals, insurers and
policymakers) can take all appropriate actions.
3. Enlist stakeholders in communicating, supporting, and
implementing the recommendations.
Study Design: All Kids Count, a national program funded
since 1992 by the Robert Wood Johnson Foundation, hosted
an invitational conference in Atlanta, December 3-4, 2003. The
conference was co-sponsored by 17 organizations and
government agencies, including AHRQ, HRSA, CDC, CMS,
AAP, ASTHO, AHCP, Commonwealth Fund, Family Voices,
March of Dimes, and NICHQ. Over 100 representatives from
55 organizations and agencies attended the conference and
worked to forge recommendations in the areas of governance
and stakeholder involvement; information infrastructure,
economic and funding issues, and use of information.
Population Studied: The participants addressed the needs of
all children, especially those with special health care needs
and at-risk populations, in the course of developing the
recommendations.
Principal Findings: Recommendations for each of the four
areas were made and will be available in the upcoming
conference proceedings (www.allkidscount.org) . Common
threads in the recommendations were:
1. The need to develop a national coalition of stakeholders to
promote integration of separate child health information
systems within the context of ongoing national initiatives such
as NHII and PHIN;
2. The need to develop the business and policy cases for
integrated child health information systems;
3. The need to develop agreement on standards for collecting
and transferring information; and
4. The need to get the word out about the importance of
integrating separate child health information systems to
improve health and health services.
Conclusions: There are many efforts underway to develop
clinical or public health child health information systems but
little attention is being paid to integrating the information
from those systems. Such integration would improve our
ability to identify the multiple health care needs of children
and ensure services were provided in a timely and effective
manner. Broad dissemination and endorsement of the
recommendations from the conference are essential next
steps.
Implications for Policy, Delivery or Practice: The quality and
delivery of health care services for children can be greatly
improved through information technology. However, public
health and medical care agencies must work closely together
to develop information systems that integrate and link
information in a complete and comprehensive way. The
recommendations and action plan from this conference are
significant in that they were developed with broad stakeholder
input, including consumer and family advocates. Policy
makers and program officers from each of the co-sponsoring
agencies can use these recommendations when addressing
issues related to health information systems for children.
Primary Funding Source: RWJF
• Detection of Health-Related Social Problems in Pediatric
Primary Care Settings Using a Computer-Based
Assessment Tool
Eric Fleegler, M.D., Tracy Lieu, M.D., M.P.H., Paul Wise, M.D.,
Sharon Muret-Wagstaff, Ph.D., M.P.A.
Presented by: Eric Fleegler, M.D., Fellow, Division of
Emergency Medicine, Children's Hospital Boston, 300
Longwood Avenue, Boston, MA 02115; Tel: 617.355.2565; Fax:
(617) 975-3691; E-mail: eric.fleegler@childrens.harvard.edu
Research Objective: To assess the acceptability of and
interest in a computer-based survey for identifying major
health-related social problems and referral of families to
appropriate social services. To use this tool (The Online
Advocate, www.onlineadvocate.org) to determine the
percentage of families in pediatric primary care clinics with
one or more health-related social problems in the following
domains: 1. Access to health care; 2. Housing; 3. Food
insecurity; 4. Income insecurity; 5. domestic violence
Study Design: Cross-sectional, self-administered, computerbased survey. Research assistants approached participants to
take the survey on a laptop computer in the clinic waiting
room and/or examination rooms. The survey was
administered to the next eligible participant when a computer
became available.
Population Studied: The survey was administered in two
primary care pediatric clinics in Boston, MA, to primary
caregivers of children up to 6 years old attending well-child
visits. Enrolled subjects were literate in either English or
Spanish. Of 450 subjects approached 260 were eligible and
205 (79%) consented.
Principal Findings: Acceptability: Of 260 eligible subjects, 12
(4.6%) were illiterate. Of the 205 consented, 194 (95%)
completed the survey in an average of 16 minutes. Ninety two
percent of families expressed acceptance of The Online
Advocate as a tool for assessment and referral within their
pediatrician’s office (welcome it, 62%; not mind at all, 30%;
be mildly annoyed, 6%; be very annoyed, 2%).
During informal discussion with providers at both clinics most
stated that the survey did not interfere with clinic flow and
allowed patients to use the wait time constructively. Providers
felt these were important issues for patients to receive help.
Social Problems: Eighty percent of families experience one or
more health-related social problems; 34% one domain, 26%
two domains, 20% three or more domains.
Twenty seven percent of families experienced problems with
access to health care; including 23% of parents without health
insurance and 11% of children without health insurance. 56%
of families identified housing problems; 43% reported one or
more major household problems as defined by the American
Housing Survey, 15% were either homeless or doubled up,
and 8% had their utilities shut off in the past year. 39% of
families experienced problems with food security as assessed
by the Community Childhood Hunger Identification Project 8
question scale; 28% of families were defined as food insecure
and 11% as hungry. 17% of families had income insecurity as
defined by the primary provider out of work and looking for a
job. 15% of mothers (n=178, 92% of participants) experienced
domestic violence; 13% experienced verbal domestic violence
and 8% physical domestic violence.
Conclusions: Families of young children are significantly
burdened by health-related social problems. The Online
Advocate allows families to self-assess and identify these
problems within the context of a primary care visit. The
Online Advocate model demonstrated a high level of usability
and acceptability among participants.
Implications for Policy, Delivery or Practice: The pediatric
office has the potential to assess, refer, and help families with
social problems that impede the proper health and
development of their children. A computer-based tool takes
full advantage of emerging technologies to make the
identification of social problems efficient and realistic given
the cadence of clinical practice.
Primary Funding Source: AHRQ
• Keeping Asthmatic Children Out of Hospitals: Parents'
and Physicians' Perspectives on How Pediatric Asthma
Hospitalizations Can Be Prevented
Glenn Flores, M.D., Milagros Abreu, M.D., Sandra Tomany,
M.S., Johm Meurer, M.D., M.B.A.
Presented by: Glenn Flores, M.D., Director, Community
Outcomes; Associate Director, Center for the Advancement of
Urban Children, Pediatrics, Medical College of
Wisconsin/Children's Hospital of Wisconsin, 8701 Watertown
Plank Road, Milwaukee, WI 53226; Tel: 414.456.4454; Fax:
(414) 456-6385; E-mail: gflores@mail.mcw.edu
Research Objective: 214,000 hospitalizations occur each year
among the 6.3 million US children who have asthma. Not
enough is known about how to prevent pediatric asthma
hospitalizations. The goal of this study was to identify the
proportion of preventable pediatric asthma hospitalizations
and how such hospitalizations might be prevented, according
to parents and physicians of hospitalized asthmatic children.
Study Design: Cross-sectional survey of parents, primary care
pediatricians (PCPs), and inpatient attending physicians
(IAPs) of a consecutive series of all children admitted for
asthma to an urban hospital in a 14-month period.
Population Studied: Consecutive series of all children
admitted for asthma to an urban hospital over a 14-month
interval. The comparison group for bivariate analyses
consisted of a consecutive series of all children admitted for
other ambulatory-sensitive conditions to the same urban
hospital over a 14-month period (N = 325).
Principal Findings: The 235 hospitalized children had a mean
age of 4 years; most were poor (mean annual family income =
$14,795), non-white (91%), and had public (74%) or no (14%)
health insurance. Compared with children hospitalized for
other ambulatory-sensitive conditions, hospitalized asthmatic
children were significantly more likely to be African-American
(70% vs. 57%; P = 0.02), older (mean age = 4.4 vs. 3.3 years
old; P < 0.001), and to not have made a physician visit or
phone contact prior to admission (52% vs. 41%; P = 0.008).
Only 26% of parents said that their child's admission was
avoidable, compared with 36% of PCPs and 43% of IAPs. The
proportion of asthma hospitalizations assessed as preventable
varied according to the source or combination of sources,
from 17% for agreement among all three sources to 52% as
identified by any one of the 3 sources. PCPs (81%) and IAPs
(71%) significantly more often (P < 0.001) than parents (39%)
cited parent/patient-related reasons for how hospitalizations
could have been avoided, including adhering to and refilling
medications, better outpatient follow-up, and avoiding known
disease triggers. Parents (28%) and IAPs (28%) significantly
more often (P < 0.008) than PCPs (10%) cited physicianrelated reasons for how hospitalizations could have been
avoided, including better education by physicians about the
child's condition, and better quality of care. Multivariate
analyses revealed that an age > 10 years, finding medical care
to be too expensive, no physician contact prior to the
hospitalization, working poor family income, and having no
health insurance were associated with approximately 2-9 times
the odds of a preventable asthma hospitalization.
Conclusions: The proportion of asthma hospitalizations
assessed as preventable varies from 17-52%, depending on the
source making the assessment. Adolescents, families who find
care to be too expensive, families who fail to contact
physicians prior to hospitalization, children from working-poor
families, and uninsured children are at greatest risk for
preventable hospitalizations.
Implications for Policy, Delivery or Practice: Many pediatric
asthma hospitalizations might be prevented if parents and
children were better educated about the child’s condition,
medications, the need for follow-up care, and the importance
of avoiding known disease triggers.
Primary Funding Source: RWJF, Agency for Healthcare
Research and Quality
• Engaging Parents in Decision Making
Shirley Girouard, Ph.D., R.N., FAAN
Presented by: Shirley Girouard, Ph.D., R.N., FAAN, Associate
Professor, Nursing, Southern Connecticut State University,
P.O. Box 2458, Branford, CT 06405; Tel: 203.488.5135; Fax:
(203) 488-5165; E-mail: sgirouard@aol.com
Research Objective: To evaluate the impact of a civics
training program on parental engagement in the community
and to assess knowledge and skills related to democracy
benchmarks. As the nation and states struggle with budget
deficits and resource allocations and with the increasing
demand for public accountability, society’s ability to meet the
needs of children is compromised. Given the vulnerability of
children, their dependence on parents and other adults to
speak for them and the necessity of providing children’s
services that will assure healthy and productive adults of the
future, programs and policies are needed to assure positive
outcomes. Parents, especially in the public arena, are an
underutilized resource for developing, implementing and
assessing programs and policies to positively affect the health,
education and overall well-being of children. This paper
describes an innovative program of the Connecticut
Commission on Children aimed at enhancing parental
involvement through democracy and civic skills training and
provides the results of an evaluation of one program
component.
Study Design: A descriptive study design assessed the foci
and outcomes associated with the community projects of the
civics training program participants. All projects conducted
over a three-year period (N=333) in 13 communities in one
state were analyzed using content analysis and descriptive
statistics.
Population Studied: Parents who developed and
implemented the 333 community projects represented diverse
social, economic, ethnic and racial groups.
Principal Findings: Of the 333 projects, 40% addressed early
childcare or educational issues; 19%, health and health care
issues; and 13%, safety concerns. Remaining projects
reflected skill development and personal growth in civic skills.
The projects met one or more of the democracy benchmarks:
all demonstrated knowledge about and skill in democratic
participation. A number of projects focused on specific
democracy benchmarks: 17% addressed social or economic
growth issues; 19%, tolerance of diversity; and 14%,
promotion of shared values and the public good.
Conclusions: The civic training program provided parents
with knowledge and skills in civics and democracy that were
translated into community activities and programs to enhance
child health, education and safety. In addition, parents used
their knowledge and skills to influence decision making by
others, including policy makers at the community and state
levels, to benefit children.
Implications for Policy, Delivery or Practice: When parents
have knowledge and skills to affect the public domain of
decision-making about children’s issues, they contribute
significantly to better programs and policy. By providing
opportunities for parents to enhance their civic training, this
important constituency can bring about needed and
appropriate change. Providers, educators, public officials,
researchers and others should engage parents to improve
children’s present and future health and well being.
Primary Funding Source: State funding
• Impact of State Children’s Health Insurance Program
(SCHIP) on Children’s Access to Primary Care in South
Carolina: A Study of Hospitalizations with Ambulatory
Care Sensitive Conditions (ACSC)
Whiejong Han, Ph.D.
Presented by: Whiejong Han, Ph.D., Research Assistant
Professor, Health Services Policy and Management, University
of South Carolina, Arnold School of Public Health, Columbia,
SC 29208; Tel: 803.777.3460; Fax: (803) 777-1836; E-mail:
hanwj@gwm.sc.edu
Research Objective: The purpose of this study was to
evaluate the effect of State Children’s Health Insurance
Program (SCHIP) on children’s health in South Carolina as
measured by the incidence of hospitalizations with
Ambulatory Care Sensitive Conditions (ACSC).
Study Design: To compare rates of ACSC admissions before
and after SCHIP, longitudinal and cross-sectional hospital
discharge data were used. This encounter-level data was linked
with Medicaid data to identify the children enrolled in SCHIP.
For multivariate analysis logistic regression was applied to
find out likelihood of being hospitalized for ACSC. Study also
used Generalized Estimating Equations (GEE) with a negative
binomial model to investigate differences in rates of ACSC
admissions across time.
Population Studied: Population of this study included
children (ages 0 to 18) in South Carolina who experienced a
hospital admission during the period of 1995 ~ 2001.
Principal Findings: During the study period 269,403 hospital
admissions were detected from children aged 0 to 18 years
old. Of those, 77,339 (28.7%) were with ACSC. After
controlling other variables children under Medicaid had a
lower probability to be hospitalized with ACSC, while younger
children, male, African-American children, or children living in
rural and very-rural areas had higher probability for ACSC
admissions. Having community health centers in county was
negatively related to the proportion of ACSC admissions.
Number of primary care physicians per 10,000 children had
significant association with ACSC admission but the effect
was a minimal. Although the number of uninsured children
declined after SCHIP introduced, no program effect was found
on rates of ACSC over time. During study period, bacterial
pneumonia, asthma and dehydration were most prevalent
pediatric ACSC in South Carolina.
Conclusions: The results detected no significant changes in
ACSC admissions after SCHIP introduced. This study,
however, confirmed that race, insurance status, and children’s
residence are strong predictors for ACSC hospitalizations.
Followings can be mentioned as possible explanations for the
absence of positive findings in rates of ACSC over time. First,
this study examined only four years of data after program
implementation. Some new initiatives, however, take a while
to be effective. Secondly, those who first enrolled in the
program could have relatively poor health status because they
were mostly uninsured previously. Study with more years of
data could show different results on preventable
hospitalizations.
Implications for Policy, Delivery or Practice: Since certain
population (eg., younger children, children in rural area) were
more likely to have an ACSC admission, population-specific
insurance eligibility might help to reduce ACSC admission
rates. Allocation of SCHIP funding to counties with more
target population, community-based initiatives for areas with
higher ACSC rates, extend after-hour care for those areas
could be other ways to improve children’s health. Finally,
hospital discharge data provides an efficient way to analyze
trends of preventable hospitalizations, and analyzing this
trend can be a strong predictor of impact of SCHIP. Therefore,
continuous study with extended data is suggested to keep
policy makers informed on this issue.
• The Role of Medicaid, SCHIP, and Other Public Sectors
in the Financing of Children's Mental Health Services
Embry Howell, M.S.P.H., Ph.D.
Presented by: Embry Howell, M.S.P.H., Ph.D., Principal
Research Associate, Health Policy Center, The Urban Institute,
2100 M Street, N.W., Washington, DC 20037; Tel:
202.261.5714; Fax: 202-223-1149; E-mail: ehowell@ui.urban.org
Research Objective: To identify the financing streams for
children's mental health services and the roles of each sector.
To describe how the various sectors interact to fund services,
and how those roles are changing over time. To provide new
data on the prevalence of mental health conditions and use of
mental health services by income and insurance status.
Study Design: Three types of information were used in this
study. We performed a comprehensive review of the literature
on the topic; we interviewed 38 knowledgeable key informants
around the country; and we analyzed new data from the
National Survey of America's Families.
Population Studied: Low income children with mental health
problems.
Principal Findings: Medicaid is a very prominent funder of
children's mental health services and its role is growing.
SCHIP is not yet a very prominent funder, but it provides
relatively generous coverage compared to the private sector,
and its role is also growing. The role of other sectors is also
very important (public mental health agencies, education,
child welfare, and juvenile justice), but there is a need for
coordination across sectors. There are new collaborations
evolving to blend funding and coordinate services that serve
as models. Low income children have higher rates of mental
health problems than non-poor children, and they have similar
rates of use of mental health services. Uninsured children
with mental health problems have much lower rates of use
than Medicaid/SCHIP children.
Conclusions: Medicaid and SCHIP funding of mental health
services provide important access to mental health services
for low income children, and their roles are increasing. There
is a need for greater coordination across the various sectors
involved in providing and funding such services.
Implications for Policy, Delivery or Practice: For the benefit
of low-income children, Medicaid/SCHIP agencies should play
increased roles in monitoring the quality and cost of the
mental health services that are provided under their programs.
Primary Funding Source: RWJF
• The Direct Medical Cost of Diabetes in Children
Kumiko Imai, Ph.D., Ping Zhang, Ph.D., Giuseppina
Imperatore, M.D.
Presented by: Kumiko Imai, Ph.D., Centers for Disease
Control and Prevention, 4770 Buford Highway, MS K-10,
Atlanta, GA 30341; Tel: 770.488.4876; Fax: (770)488-5966; Email: kimai@cdc.gov
Research Objective: To estimate the excess direct medical
cost of diabetes in children and to demonstrate a use of an
administrative claims database in analyzing children's health
care utilization.
Study Design: We used a large commercial administrative
claims database for 1999, 2000 and 2001. Total direct
medical costs were calculated using inpatient, outpatient, and
pharmaceutical claims for more than 740,000 children, and
were fitted by a modified two-part model that controlled for
age, sex, and other factors specific to each family. The first
part of the two-part model used a logit model that estimated
the likelihood of incurring any medical costs, while the second
part of the model was based on a generalized linear model
that predicted the amount of medical costs for children who
had positive costs. Inclusion of family random effects allowed
us to control for unobservable factors shared by siblings.
Population Studied: Claims and enrollment data came from
737,561-914,372 children aged 0-19 who were enrolled in a
variety of employer-sponsored insurance plans in 1999, 2000
and 2001. For those, 2,266-3,755 children were identified as
diabetic.
Principal Findings: Compared to children without diabetes,
children with diabetes on average spent an excess of $2,500 if
they were not on insulin and $4,000 if they were on insulin.
Outpatient services accounted for about half of total
expenditures for non-insulin treated diabetic children, while
drugs accounted for about a quarter for insulin-treated
diabetic children. Also, estimates from the three years
showed that the medical cost of diabetes in children grew
considerably during the period.
Conclusions: The cost of diabetes during childhood is
substantial.
Implications for Policy, Delivery or Practice: Considering
the large economic burden of diabetes during childhood,
more research is needed to find ways to prevent the disease in
younger populations. Also, large administrative claims
databases might be a useful source of information for
researchers analyzing children's health care utilization and its
trends over time.
• Association of Having a Personal Physician with Quality
of Care for CSHCN
Moira Inkelas, Ph.D., M.P.H., Kandyce Larson, M.S.W.,
Kathryn Smith, M.A., M.S.N., Paul Newacheck, Dr.PH
Presented by: Moira Inkelas, Ph.D., M.P.H., Assistant
Professor, Health Services, UCLA School of Public Health,
1100 Glendon Avenue, Suite 850, Los Angeles, CA 90024; Tel:
310.794.0966; Fax: (310) 794-2728; E-mail: minkelas@ucla.edu
Research Objective: Determine what percentage of children
with special health care needs have a usual place for care
and/or personal health care provider, and if report of these
relationships is associated with sociodemographic factors,
health insurance, and experiences with care when adjusting
for condition severity.
Study Design: Data are from National Survey of Children with
Special Health Care Needs, a 2001 nationally representative
telephone survey conducted by the National Center for Health
Statistics providing national and state-based estimates. In
addition to prevalence of usual place and of personal
physician, a provider continuity measure is constructed with
categories of: 1-having both a usual place for health care, and
a provider (doctor or nurse) who knows the child best, 2-usual
place but no personal provider, 3-personal provider but no
usual place; and 4-no usual place or personal provider. 95
percent of parents reporting a personal provider say it is a
physician. Family-centered care - FCC - is the 5 item FACCT
measure on a 0-100 scale. Bivariate and multivariate analyses
are used.
Population Studied: Parents of 38,866 children with a chronic
health condition that has lasted or is expected to last at least
one year.
Principal Findings: About 10 percent of CSHCN have no
usual place for care and another 9 percent have a usual place
but receive sick care and routine preventive care from different
places. Parents of 11 percent of CSHCN report that no single
provider knows the child best. The provider continuity
measure shows that more uninsured and publicly-insured
than privately insured children have a usual place but lack a
personal physician. In logistic regression controlling for child
age, condition severity, and total visits in the past year, odds
of having a personal physician are lower for children who are
uninsured OR 0.38, CI 0.31-0.48, publicly insured OR 0.73, CI
0.61-0.86, Latino with Spanish-speaking parent OR 0.54, CI
0.36-0.82, and whose mother has lower educational
attainment. In multivariate OLS using these same control
variables, compared to children with both a usual place and
personal provider, FCC is similar for those with no usual place
who have a personal physician b= -0.55, p=0.48, but lower for
those with a usual place but no personal provider b= -11.9,
p<0.001 and those with no usual place or a personal provider
b= -16.2, p<0.001.
Conclusions: Although causality in the association and all
reasons for variability in physician relationship among CSHCN
are not known, having a single physician whom the family
counts on for information and coordination appears to
improve access and parent-provider communication for
CSHCN. Having only a usual source confers little benefit.
Implications for Policy, Delivery or Practice: CSHCN may
receive care from one or more specialists in addition to a
primary care physician. Even with multiple providers, having a
physician who knows the child best - whether a specialist or
PCP - may lead to better coordination of care and quality of
physician-parent communication. Ascertaining if parents feel
they lack this relationship may enable the specialist or primary
care physician to help the family determine where the medical
home should be.
Primary Funding Source: California Healthcare Foundation
• Medicaid Enrollment and Utilization In Cuyahoga
County: Evaluating the Early Childhood Initiative Amid
Other Health Systems Changes
Siran Koroukian, Ph.D., M.S.N., M.H.A., Engel Polousky, M.S.,
Rob Fischer, Ph.D., Claudia Coulton, Ph.D.
Presented by: Siran Koroukian, Ph.D., M.S.N., M.H.A.,
Assistant Professor, Epidemiology and Biostatistics, Case
Western Reserve University, 10900 Euclid Avenue, Cleveland,
OH 44106-4945; Tel: 216.368.3197; Fax: 216-368-3970; E-mail:
sxk15@case.edu
Research Objective: A primary goal of the Cuyahoga County
Early Childhood Initiative (ECI) is to facilitate early and
sustained enrollment in Medicaid, which in turn is likely to
promote receipt of timely and adequate amount of preventive
services, through increased access to health services. Due to
a coincidence in timing of the expansion of Medicaid
managed care programs (MCPs) and the State Child Health
Insurance Program (SCHIP) with that of the ECI, it was
necessary to evaluate the performance of the ECI
independently from that of MCPs or the SCHIP – programs
with goals consistent with that of the ECI. The objective of
this study was to compare outcome measures across counties
with similar expansions in MCPs.
Study Design: Ohio Medicaid enrollment and claims history
files for 1998-2001, encompassing fee-for-service claims and
managed care encounter data, were used to obtain the
relevant measures. These measures included age at first
enrollment in Medicaid; length of enrollment in Medicaid; as
well as the timing and number of comprehensive preventive
visits (CPVs) in the first 15 months of life. Multivariable
survival analysis and logistic regression analyses were
employed to assess the effects of residence in Cuyahoga
County, as an indicator for having been exposed to ECI,
independently from the effects of the expansion of MCPs, or
the SCHIP, relative to the measures of interest, and to control
for differences in demographic profile across the counties.
Adjusted hazard ratios (HRs) and odds ratios (ORs) were
derived respectively from multivariable survival and logistic
regression analyses.
Population Studied: Children enrolled in Ohio Medicaid
during 1998-2001, and residing in seven large urban counties
with mandatory managed care programs.
Principal Findings: The study population included
approximately 140,000 children 6 years of age or younger.
The measures were comparable across the counties, and
followed the same trends of improvement over time. A small,
but statistically significant advantage was observed in these
measures relative to Cuyahoga County. In particular,
Cuyahoga children were significantly more likely than others to
enroll early in Medicaid and have sustained enrollment in the
program (HR of 1.02 and 0.88, respectively, p < 0.01 for each
comparison). Relative to the receipt of CPVs, Cuyahoga
children were significantly less likely not to have received any
CPV, or less than 4 visits during the first 15 months of life (OR
of 0.76 and 0.88 respectively, p < 0.01 for each comparison).
They were more likely than others to receive at least 5 CPVs in
the first year of life – the recommended number of preventive
visits by the American Academy of Pediatrics (OR=1.11, p <
0.01).
Conclusions: The findings reflect favorably upon the ECI, and
support the notion that the ECI may have yielded modest
benefits above and beyond the effect of MCPs or the SCHIP
program. Additional analyses on programmatic details and
the logistics entailed in accessing and using Medicaid and
health services may help identify factors that could further
improve the relevant outcomes.
Implications for Policy, Delivery or Practice: The provision
of additional services at the community level seems to
enhance the existing synergy yielded by the expansion of
MCPs and SCHIP, two programs with goals similar to that of
the ECI.
Primary Funding Source: Cuyahoga County
• Feasibility of Influenza Immunization for Children 6-23
Months in the Inner-City
Chyongchiou Lin, Ph.D., Richard Zimmerman, M.D., M.P.H.,
Mary Patricia Nowalk, Ph.D., RD, Alejandro Hoberman, M.D.,
David Greenberg, M.D., Stuart Weinberg, M.D.
Presented by: Chyongchiou Lin, Ph.D., Assistant Professor,
Health Policy and Management, University of Pittsburgh, A613
Crabtree Hall, 130 DeSoto Street, Pittsburgh, PA 15261; Tel:
412.624.3625; Fax: 412-624-3146; E-mail: cjlin@pitt.edu
Research Objective: Determine the feasibility of vaccinating
typically hard-to-reach, low-income children against influenza,
and assess its effect on timely receipt of other vaccines.
Study Design: Before/after trial with the health care center as
the unit of intervention in three types of settings: pediatric
residency, family practice residencies and faith-based
practices, at ten offices in low-income urban locations.
Population Studied: Children aged 6-23 months, during the
2002-2003 influenza season, were selected for intervention.
The comparison group for analysis was children who were in
the same age range in the previous influenza (2001-2002)
season and were active patients of the practice, that is, they
had been seen in the office in the last 6 months. Active
patients born between October 1, 2000 and April 1, 2002 were
the 2002-2003 intervention group (n = 1,534) and children
born between October 1, 1999 and April 1, 2001 comprised the
pre-intervention comparison group (n = 1,210).
Principal Findings: Immunization rates of the two-dose
influenza vaccine improved significantly at all three types of
health care centers. Compared with the pre-intervention
season, vaccination rates for the first dose during the
intervention season significantly increased from 7.6% to
42.3% at pediatric residency offices, from 4.3% to 15.6% at
family practice residencies; and from 4.3% to 47% at faithbased practices (P<0.001). Rates for the second dose were
significantly improved over pre-intervention, but lower than
first dose rates (P<0.001). Average ages at vaccination for
other recommended childhood vaccines did not differ or were
significantly younger (MMR and varicella) for children who
received influenza vaccine vs. those who did not. Moreover, a
higher percentage of influenza-vaccinated than unvaccinated
children received their MMR, DTaP 3, IPV 2 and Hib 2
vaccines within a two-month grace period of the
recommended age (P<0.05), with no change in the timing of
receipt of varicella, DTaP 1, IPV 1 and Hib 1.
Conclusions: With directed effort, it is possible to increase
influenza vaccination at health care centers serving lowincome children. The addition of a two-dose vaccine does not
delay receipt of other vaccines.
Implications for Policy, Delivery or Practice: Our findings,
presented to the Advisory Committee on Immunization
Practices (ACIP), provided a basis for full recommendation of
the influenza vaccine for all 6-23 month old children,
beginning in 2004.
Primary Funding Source: CDC, the Association of Teachers
of Preventive Medicine (ATPM)
• Effects of Maternal Characteristics on Cesarean Delivery
Rates among U.S. Department of Defense Healthcare
Beneficiaries, 1996 - 2002
Andrea Linton, Ph.D., Michael Peterson, Dr.P.H., Thomas
Williams, Ph.D.
Presented by: Andrea Linton, Ph.D., Senior Health Program
Analyst, TRICARE Management Activity, Dept. of Defense, 5111
Leesburg Pike, Falls Church, VA 22041; Tel: 703.681.3636; Email: andrea.linton@tma.osd.mil
Research Objective: To identify rates of cesarean birth for the
Military Health System (MHS) healthcare beneficiary
population from 1996 to 2002 and compare those rates to the
national experience. We also sought to identify non-clinical
factors associated with cesarean delivery including maternal
age, race, socioeconomic status, and insurance coverage.
Study Design: Hospital discharge and claims records for
babies born in the military and civilian hospitals that comprise
the Military Health System (MHS) healthcare network were
used to calculate total and primary cesarean rates and vaginal
birth after cesarean (VBAC) rates from 1996 to 2002. Annual
cesarean rates for subgroups defined by maternal age, race,
and socioeconomic status were calculated to examine rate
variations and rate trends within the study population. Pooled
data from 1999 to 2002 were used to compare rates across
socioeconomic status, stratified by age and race.
Population Studied: Beneficiaries who had a childbirth
experience at any one of the military and civilian hospitals
within the MHS between 1996 and 2002.
Principal Findings: Total and primary cesarean rates among
the MHS population were lower than those reported nationally
for every year examined. The cesarean and VBAC rate trends
in the national and MHS populations were similar. Within the
MHS population, total cesarean delivery increased with
increasing maternal age and was more highly associated with
racial minorities relative to white women. The higher
socioeconomic subgroup (defined as active duty, retired, and
warrant officers, and their families in this study) was generally
associated with reduced cesarean section rates.
Conclusions: Cesarean deliveries are performed less
frequently for the MHS healthcare beneficiary population,
relative to the national population. Associations between
socioeconomic factors and cesarean rates reported for the
national population were not apparent in the study
population. The consistent pattern of rate variation across
racial subgroups in the MHS population suggests that factors
beyond those examined in this study are needed to explain the
elevated cesarean rates for racial minorities.
Implications for Policy, Delivery or Practice: The historically
lower cesarean rates and higher VBAC rates in the DoD
suggest that further studies of DoD healthcare practice
patterns, institutional culture, and outcomes may offer
insights for developing interventions to slow the increase in
cesarean use and ultimately reverse a three decade-long trend
without compromising maternal or fetal health.
Primary Funding Source: Dept. of Defense
• The Economic Impact of Persistent Pulmonary
Hypertension of the Newborn (PPHN): Use of the KID
Database for a Common Neonatal Disease
Scott Lorch, M.D., MSCE, Joel Portnoy, M.D.
Presented by: Scott Lorch, M.D., MSCE, Neonatologist,
Neonatology and Center for Outcomes Research, The
Children's Hospital of Philadelphia, 3535 Market Street, Suite
1029, Philadelphia, PA 19104; Tel: 215.590.1714; Fax: (215) 5902378; E-mail: lorch@email.chop.edu
Research Objective: Much of the epidemiologic data on
PPHN has come from randomized controlled trials, which
may not demonstrate the societal impact of the disease.
Administrative databases would allow for such populationbased studies, but only include codes for comorbidities and
thus do not capture clinical data, such as the use of inhaled
nitric oxide. Because of this fact, neonatal diseases have not
been studied using asuch dministrative databases. The
objective of this study is to (1) determine national estimates of
total charges and length of stay for infants with PPHN; and (2)
to develop prediction models for death, need for extracorporeal membrane oxygenation rescue treatment (ECMO),
and length of stay (LOS) in infants with PPHN, and to validate
these results with models using clinical data.
Study Design: Infants <14 d of age with an ICD-9 code for
PPHN (416.0 or 747.89) were identified using the 2000 KIDS
database, an 80% stratified random-sample of pediatric
hospitalizations in 27 states. 1 cohort required an ICD-9 code
for meconium aspiration, pulmonary hypoplasia, asphyxia,
pneumonia, or sepsis, and a second included all infants
except those with an ICD-9 code for complex congenital heart
disease, diaphragmatic hernia, or birth weight <1750g. A third
cohort included all infants transferred to CHOP with PPHN
between 1996 and 2001 (N=91). Patient-weighted, stratumspecific methods estimated total hospital charges, ECMO use,
LOS, and mortality. Logistic and linear regression models
determined the effect of demographic and comorbid
conditions on LOS and total charge for all three cohorts and
were validated using bootstrap methods.
Population Studied: Infants <14 d of age with an ICD-9 code
for PPHN (416.0 or 747.89) in the 2000 KIDS database.
Principal Findings: Using the narrow cohort, there were an
estimated 7149 infants treated with PPHN in the US in the
year 2000 (95% CI 6525-7773), with an average LOS of 12 days
(95% CI 11.4-12.6). 42% were female, 5.1% required ECMO
(95% CI 3.4%-6.7%), and 4.9% died (95% CI 4.0%-5.9%).
69.7% were treated in teaching hospitals. Their estimated
hospital charge was $378.2 million (95% CI $325 million-$432
million). Similar results were seen using the broad cohort.
Predictive models using either administrative cohort had
excellent predictions for risk of death (c-statistic 0.86, 95% CI
.83-.89) and need for ECMO (c-statistic 0.87, 95% CI .85-.9),
but weaker validity for LOS (r2=.27, 95% CI .25-.3). Clinical
data improved the predictive validity only for the LOS model
(r2=.6) by including a major determinant of LOS, total days on
oxygen, which cannot be ascertained in administrative
datasets. Higher number of days on oxygen predicted fewer
residual hospital days prior to discharge.
Conclusions: PPHN has substantial economic costs. ECMO
use and mortality rates are lower than previously published
rates. Administrative databases allow for the development of
strong predictive models for death or need for ECMO, but
detailed clinical data are needed for adequate LOS predictions,
especially length of oxygen administration.
Implications for Policy, Delivery or Practice: Administrative
databases allow for population-based studies of neonatal
diseases. Understanding the factors leading to more days on
oxygen could aid in reducing the economic burden of PPHN.
Primary Funding Source: AHRQ
• Planning for Medical Transitions for Adolescents with
Special Health Care Needs
Debra Lotstein, M.D., Moira Inkelas, Ph.D., Kandyce Larson,
M.S.W., Neal Halfon, M.D., M.P.H.
Presented by: Debra Lotstein, M.D., Research Fellow, Robert
Wood Johnson Clinical Scholars Program, UCLA, 911 Broxton,
3rd Floor, Los Angeles, CA 90024; Tel: 310.794.0612; E-mail:
lotstein@ucla.edu
Research Objective: There is growing consensus that
adolescents with special health care needs should eventually
transition to adult-oriented health systems. Transition
planning has been suggested as a way to ease the process of
transition for chronically ill adolescents and to ensure that
teens are assisted with age-related changes in medical care,
insurance coverage and independent living. The objectives of
this study were: 1) to estimate the proportion of teens with
special health care needs receiving transition assistance from
health care providers, and 2) to assess the factors associated
with receiving transition assistance.
Study Design: Analysis of a nationally-representative, random
digit dial telephone survey of parents of youth with special
health care needs. Dependent variables measuring transition
assistance were (1) general discussion of changing needs in
adulthood (“transition discussion”), and (2) transition
discussion plus more specific planning for future needs
(“transition planning”). We carried out bivariate and
multivariate analyses to identify predictors of transition
discussion and planning.
Population Studied: Parents of 5,555 adolescents aged 13-17
with at least one chronic condition included in the National
Survey of Children with Special Health Care Needs (conducted
by the National Center for Health Statistics in 2000-2002).
Principal Findings: 50% of parents reported having transition
discussions with their child’s physicians and fewer (34%)
reported receiving transition planning. Logistic regression
models with social, disease severity and medical care process
variables were performed, first for transition discussion and
secondly for transition planning. Hispanic teens had lower
odds than whites of discussing transition (OR=0.47, 95% CI
0.29-0.76). Adolescents with more than 10 physician visits in
the last year were more likely to discuss transition (OR=1.69,
CI 1.31-2.17) as were those with more family-centered
communication (OR=2.10, CI 1.71-2.57). Age of the adolescent
was not associated with transition discussion. Odds of
transition planning were higher for teens with more visits
(OR=1.91, CI 1.45-2.51) and for those with better ratings of
communication (OR=2.21, CI 1.77-2.74). Older teens were
more likely than younger teens to have received transition
planning from their pediatric provider (17 year olds vs. 13-16
year olds, OR 1.61, CI 1.24-2.08).
Conclusions: Most adolescents with chronic health
conditions are not receiving the transition assistance that
guidelines suggest is needed, and some racial/ethnic
disparities are evident. Even among 17 year olds, planning
rates are low. Those with a large number of provider contacts
have greater odds of receiving transition assistance. Transition
discussion and planning are associated with the quality of
provider-patient communication
Implications for Policy, Delivery or Practice: Efforts to
increase the number of youth receiving transition planning
might effectively be linked with strategies to improve familyprovider communication. Further research is needed to
understand 1) the ways that patient characteristics shape
transition needs and 2) which forms of transition assistance
are most helpful in improving adult outcomes for young
adults with chronic conditions.
Primary Funding Source: RWJF
• An Evaluation of the Accountability and Quality of
Managed Care for Medicaid and SCHIP Children: State
Approaches to Monitoring and Enforcing Oral Health and
Lead Screening Contractual Standards
Sara Rosenbaum, J.D., Anne Markus, J.D., Ph.D., Colleen
Sonosky, J.D., Lee Repasch, M.A., D. Richard Mauery, M.P.H.
Presented by: Anne Markus, J.D., Ph.D., Assistant Research
Professor, Health Policy, The George Washington University
Medical Center, 2021 K Street, N.W., Suite 800, Washington,
DC 20006; Tel: 202.530.2339; Fax: (202) 296-0025; E-mail:
armarkus@gwu.edu
Research Objective: To examine state experiences with
monitoring and enforcing contractual standards of pediatric
care imposed on managed care organizations (MCOs).
Study Design: Review of Medicaid/SCHIP managed care
contracts and semi-structured telephone interviews with
Medicaid/SCHIP officials involved in contract oversight.
Questions focused on the logic behind the selection of one
high and one low prevalence condition (oral disease and lead
poisoning) to emphasize in the contract, and the development
of expectations regarding MCO performance in these two
areas of child health and the monitoring of actual
performance against those expectations. If a state was
sufficiently concerned with a particular child health condition
to address it with some specificity in the contract, it was
expected to focus on it as part of its oversight efforts. A
conceptual framework on key components of an effective state
monitoring plan was used to guide the analysis and organize
findings.
Population Studied: 16 states with 27 Medicaid and
separately-administered SCHIP full-risk managed care
programs, representing more than half of the nation’s
population, a broad range of fiscal capacity, child well-being,
managed care markets, and approaches to government
programs; 9 states—15 programs—participated (RR=56%).
Principal Findings: Medicaid and SCHIP contractual
provisions varied in specificity across/within states but
generally addressed the service components and periodicity
schedules of oral health and lead screening. Contractual
standards differed between Medicaid and SCHIP because of
different federal benefit requirements and state health needs.
All states had a formal monitoring plan. All agencies collected
data regularly to measure MCO compliance with quantifiable
standards but few used quality benchmarks specific to oral
health and lead screening. Many programs designed a
graduated incentive and penalty system, believed to favor
compliance. Most states, even those with high managed care
penetration, approached enforcement more as collaboration
with MCOs because of underlying systemic issues beyond
state and MCO control.
Conclusions: States play a significant role in ensuring
pediatric quality and they adapt their quality improvement
strategies to the conditions of local markets to achieve the
best possible outcomes for children under the circumstances.
States pay attention to the contract and generally perform
some contract monitoring as one avenue to ensuring quality.
They do not consider the contract as all encompassing on
quality improvement because its reach falls short of
addressing underlying systemic issues.
Implications for Policy, Delivery or Practice: Findings have
implications for Medicaid/SCHIP quality improvement.
Legislation is broad in that regard and should remain broad so
that states can tailor their approaches to what is happening at
the local level. Because underlying systemic issues appear to
lessen the influence of contract language and enforcement,
states should be encouraged to turn to additional, noncontractual strategies, such as community awareness, to
address these problems. Findings also have implications for
children’s right to coverage under Medicaid and SCHIP, and
for children with special health care needs whose higher
prevalence in Medicaid and SCHIP warrants particular efforts
on the part of states to monitor the quality of care they
receive.
Primary Funding Source: The David and Lucile Packard
Foundation.
• Translating Evaluation Research Into Practice: EvidenceBased Strategies to Enhance Service Delivery and Quality
of Care to Vulnerable Children
Maryam Navaie-Waliser, Dr.P.H., Laura Ensler, MS.E.d.,
Aubrey Spriggs, M.A., David Jones, C.S.W.
Presented by: Maryam Navaie-Waliser, Dr.P.H., Senior
Research Associate, Center for Home Care Policy and
Research, Visiting Nurse Service of New York, 5 Penn Plaza,
11th Floor, New York, NY 10001; Tel: 212.609.5762; Fax: (212)
290-3756; E-mail: maryam.navaie@vnsny.org
Research Objective: Although evaluation research in health
settings aims to improve understanding of care processes and
outcomes, successfully translating results into practice to
enhance service delivery and quality of care remains a
challenge to many professionals. This study will present
evidence-based strategies that demonstrate multiple pathways
of using research data to impact practice to enhance service
delivery and quality of care to vulnerable children within homeand-community-based health settings.
Study Design: A prospective cohort study was begun in 20022003 with participants in an Early Head Start (EHS) program
of a large, urban home health agency targeting
adolescent/young parents and their children. Trained staff
performed interviews and observational assessments around
the time of admission and every 4 to 6 months thereafter for
follow-up using standardized, reliable, and validated
instruments to capture multiple health and social outcomes.
Outcomes data were linked with service utilization and
administrative data for comparative assessments and analytic
purposes.
Population Studied: Children between ages 0 to 3 years (n =
101) and their parents (n = 99 mothers and 70 fathers) served
by an EHS program of a large, urban home health agency in
2002-2003. Over 90% of children were African-Americans or
Hispanics living in poverty.
Principal Findings: Process and outcome measures revealed:
(a) a need for more timely charting and comprehensive
documentation, (b) high levels of need among families of
infants/toddlers in understanding child development, building
self esteem, managing depression and stress, developing
parenting competence and skills, reducing anxieties
surrounding attachment issues, reducing social isolation, and
engaging in community building and civic activities; and (c)
an increased need among children for preventive health care
and use of standardized child developmental assessments.
Translating these findings into constructive and meaningful
practice changes was achieved by administrators and
practitioners using three key strategies: (a) conducting
rigorous staff development to better prepare practitioners
through knowledge and skill enhancements by training,
reflective supervision, and enhanced in-service education; (b)
tailoring and prioritizing home-based and center-based health
and social services to reflect needs through implementation of
multidimensional individualized care plans; (c) performing
thorough quality assurance monitoring through various
enhancements of administrative processes; and (d) adoption
of age-appropriate, standardized child developmental
evaluation tools. To date, follow-up data have been
promising, suggesting that evaluation research results have
been successfully incorporated into practice improvement
initiatives as evidenced by program expansion, increased
advocacy efforts, improved parent-child interactions, better
child and parent outcomes, improvement in staff knowledge
and readiness skills, and earlier identification and referral of
children with special needs.
Conclusions: Although there is no simple formula for
translating research into practice, through collaborative
efforts, creative strategies can be devised and employed
incrementally to enhance service delivery and quality of care to
vulnerable children and their families.
Implications for Policy, Delivery or Practice: Within homeand-community-based health settings, researchers,
administrators, and front-line practitioners can maximize
research applications to practice. Short-term and long-term
implications of the study will be discussed, emphasizing use
of multiple data sources to establish priorities, identifying and
targeting service delivery, and monitoring improvements in
care processes and outcomes.
• Utilization of Intracranial Pressure Monitors in Critically
Ill Children with Meningitis
Fola Odetola, M.D., John Tilford, Ph.D., Matthew Davis, M.D.,
MAPP
Presented by: Fola Odetola, M.D., Clinical Instructor,
Department of Pediatrics and Communicable Diseases,
University of Michigan, F6884 mott/0243, 1500 East Medical
Center Drive, Ann Arbor, MI 48103; Tel: 734.615.8418; Fax:
734.615.5153; E-mail: fodetola@med.umich.edu
Research Objective: To describe patient- and hospital-level
factors associated with the use of Intracranial pressure
monitors (ICPM) in critically ill children with meningitis.
Study Design: A retrospective cohort study of hospitalizations
for meningitis requiring mechanical ventilation was conducted
using the Kids’ Inpatient Database (KID). The KID includes
over 2 million pediatric discharge records obtained from 22
states in 1997 and 27 states in 2000, drawn from the 4 major
US census regions. Records were weighted to obtain national
estimates of meningitis hospitalizations and ICPM use. ICPM
cases were identified using ICD-9 procedure codes. Meningitis
was defined using ICD-9 codes for a primary or secondary
diagnosis of meningitis. Hospitalizations with ICD-9 codes for
traumatic brain injury or ventriculo-peritoneal shunts were
excluded, as were pre-transfer hospitalizations. Using data
pooled from 1997 and 2000, bivariate comparisons of factors
related to ICPM use were performed, and multivariate logistic
regression models were fit to determine which factors were
independently associated with ICPM use.
Population Studied: Hospitalizations for meningitis requiring
mechanical ventilation,in children 0-18 years old in 22 states in
1997 and 27 states in 2000.
Principal Findings: There were an estimated 1067 and 1170
hospitalizations for childhood meningitis requiring
mechanical ventilation in 1997 and 2000, respectively. Most
(79%) of the hospitalizations involved infants (<1 year old).
There was no difference in the frequency of ICPM use by
calendar year. In bivariate analyses, ICPM use differed by age
(4% in infants, 14% in children 1-4 years old, 21% in children
5-18 years old; p<0.01) and by census region. There was no
difference in ICPM use by gender, race, etiologic organism,
median household income, primary payer, or hospital
characteristics (rural/urban, teaching status, bed size). In
multivariate analyses, increasing age in 1-year increments (OR
1.14, 95%CI 1.09-1.20) and living in the West (OR 2.59, 1.384.87) compared to other census regions were independently
associated with ICPM use, while controlling for other factors.
Conclusions: The findings indicate heretofore unrecognized
variation in the use of ICPM for childhood meningitis by child
age and by census region. Further study is warranted to
examine the reasons for this variation and effects on clinical
outcomes related to ICPM use.
Implications for Policy, Delivery or Practice: No clear
guidelines exist for the use of ICPM in children hospitalized
for meningitis, neither have outcomes in these children been
well described. Additional research incorporating these study
findings, could improve the care of critically ill children with
meningitis for whom ICPM is implemented.
• Regionalization of Non-Cardiac Neonatal Surgery in
California: Is it Geographically Feasible?
Ciaran Phibbs, Ph.D., Roderic Phibbs, M.D., Susan Schmitt,
M.S., Diana Farmer, M.D., Laurence Baker, Ph.D., Beate
Danielsen, Ph.D.
Presented by: Ciaran Phibbs, Ph.D., Research Economist,
Health Economics Resource Center (152), VA Medical Center,
795 Willow Road, Menlo Park, CA 94025; Tel: 650.493.5000
Ext. 22813; Fax: (650) 617-2639; E-mail: cphibbs@stanford.edu
Research Objective: While the association between surgical
volume and mortality risk has been demonstrated for neonatal
cardiac surgery, there have been no studies that look at this
association for non-cardiac neonatal surgery, or at the
feasibility of regionalizing these cases. The objectives of this
study were to:
- Look at where neonatal surgery was being performed.
- See if it was geographically feasible to concentrate neonatal
surgery in hospitals that perform high-volumes of these
procedures.
- See if the number of non-cardiac neonatal surgeries a
hospital performs was associated with mortality for neonates
who require major surgery.
Study Design: California birth certificates for 1991-2000 were
linked to death certificates, and hospital discharge data,
including neonatal transfers. ICD-9 procedure codes were
used to identify major non-cardiac surgical procedures.
Descriptive data were summarized by NICU level of care and
hospital surgical volume. Distance was measured using
straight-line distance between hospitals. Data from the birth
certificate and discharge abstracts were used to control for
clinical risk factors in logistic regression models to examine
the association between surgery volume and mortality risk.
Population Studied: All births in California that 1991-2000
who underwent major non-cardiac surgery before they were
discharge home.
Principal Findings: A majority of the non-cardiac neonatal
surgery was performed at tertiary centers that had high
surgical volumes. In 1991 the 12 hospitals that performed at
least 50 surgeries did a total of 948 (58%) of the 1638
surgeries performed in California. But, in 1991 there were 37
hospitals that performed less than 15 major non-cardiac
surgical procedures on neonates, and an additional 21
hospitals that performed between 15 and 49 of these
surgeries. These numbers were similar over the 1990s, except
that there was a modest decline (21 to 16) in the number of
hospitals with moderate surgery volumes. Most of the
moderate surgery volume hospitals were State-certified to
perform neonatal surgery (16 of 21 in 1991 and 16 of 16 in
2000). Only 5 of the 37 low-volume providers were certified to
perform neonatal surgery in 1991. This increased to 17 of 38 in
2000. All of these moderate volume hospitals and virtually all
of the low volume hospitals were located in major urban areas
within reasonable geographic access to hospitals that had
high surgical volumes. There was no significant volumeoutcome effect on mortality when all non-cardiac surgical
cases were considered, but many of these procedures were for
conditions not associated with high levels of mortality. The
volume effect may be significant for higher-risk procedures, or
for other measures of outcome.
Conclusions: While 60% of non-cardiac neonatal surgery was
performed in hospitals that were high-volume providers of
pediatric surgery, there were relatively large numbers of
hospitals that performed low volumes of neonatal surgery. All
but 6 of the hospitals with low surgery volume were close to a
high-volume provider.
Implications for Policy, Delivery or Practice: Regionalizing
non-cardiac neonatal surgery in California to high-volume
providers is geographically feasible; virtually all of the surgery
is already being performed at hospitals near a high-volume
provider.
Primary Funding Source: NICHD
• Nationwide Estimates of the Use of Implantable Medical
Devices in Children
Joel Portnoy, M.D., Scott Lorch, M.D., MSCE, Chris Feudtner,
M.D., Ph.D., M.P.H.
Presented by: Joel Portnoy, M.D., Assistant Professor, Center
for Outcomes Research, Department of Anesthesiology and
Critical Care Medicine, Children's Hospital of Philadelphia,
3535 Market Street, Suite 1029, Philadelphia, PA 19104; Tel:
215.590.5758; Fax: 215-590-2378; E-mail:
portnoy@email.chop.edu
Research Objective: Implantable medical devices (IMDs) can
have a substantial impact on the lifespan and quality of life of
children with chronic diseases, yet little is known about the
epidemiology of these devices in children. The aim of this
study is to calculate national estimates of the frequency of
placement of IMDs in children, describe characteristics of the
associated hospital admissions, and test the hypothesis that
the number of IMDs placed is increasing.
Study Design: We conducted a descriptive study used the
Kids Inpatient Database, a stratified random sample of
pediatric discharges from non-federal community hospitals in
1997 and 2000. We identified cases by procedure codes
corresponding to the placement of an IMD. National
estimates of the number of devices implanted in patients less
than 18 yrs old were obtained using survey estimation
techniques, and the increase over time was tested using a ttest.
Population Studied: Pediatric patients admitted to all
nonfederal, short-term, general and other specialty hospitals
in the years 1997 and 2000
Principal Findings: There were an estimated 503,653
(standard error 14,013) IMDs placed in pediatric inpatients
during 1997 and 2000. The procedures were done mostly in
large, urban, teaching hospitals; half were performed in
children's hospitals. For an admission where a device was
placed the average length of stay was 15.9 days, the average
charge was $55,281, 4.5% of the patients died, 4.0% were
transferred to long-term care, and 8.0% were discharged with
home health care. Between 1997 and 2000 respectively there
were significant increases in the total number of devices
(222,224 vs. 281,429, p<.001), internal fixations (24,925 vs.
44,419, p<.001), gastrostomies (10,462 vs. 13,087, p=.011),
spinal fusions (7,157 vs 11,418, p<.001), and skin grafts (221 vs.
476, p=.021).
Conclusions: The number of IMDs placed in children in 1997
and 2000 was large and increasing. Prospective trials are
needed to study the safety, efficacy, costs, and benefits of
these devices in pediatric patients.
Implications for Policy, Delivery or Practice: Because the
population of children using implantable medical devices is
large and growing, policymakers may need to reevaluate the
steps required to bring an IMD to market for pediatric use.
• Health Insurance and Children’s Reliance on Emergency
Departments for Non-Urgent Care
Mary Rimsza, M.D., William Johnson, Ph.D., Tricia Johnson,
Ph.D.
Presented by: Mary Rimsza, M.D., Director of Health, School
of Health Administration & Policy, Arizona State University,
Box 2104, Tempe, AZ 85287-2104; Tel: 480.965.1145; Fax:
(480) 965-4605; E-mail: Mary.Rimsza@asu.edu
Research Objective: This study uses micro-data from a
unique community health data system to estimate the effect of
insurance on emergency department use for non-urgent care,
controlling for demographics, health care needs and
accessibility.
Study Design: Data from a unique community health
information system are used to examine factors that affect
access, including (a) predisposing characteristics, (b)
characteristics that enable access, and (c) characteristics that
represent a need for care. We estimate the probability of
reliance on an emergency department using multivariate logit
equations.
Population Studied: Participants include children 0 to 19
years of age, living in a semi-rural border county in Arizona,
who either used the emergency department or relied on
primary care providers for routine, non-traumatic care in 1999.
Principal Findings: Uninsured children were 7.7 times more
likely to rely on the emergency department than insured
children. The children most likely to rely on the emergency
department were adolescents, Native-American, uninsured
and those living in the city of Yuma. The majority of children
whose usual source of care is the emergency department, in
terms of numbers, were privately insured, ages 0 to 4, Latino,
and living in the city of Yuma. Children insured by Medicaid
were 90% less likely than privately insured children to rely on
the emergency department as their usual source of non-urgent
care.
Conclusions: Providing health insurance for children living in
the border counties will reduce, but not eliminate, reliance on
the emergency department for care, since the majority of
children who rely on the emergency department for care
already have health insurance. The highest utilization rates are
observed for Native American Children. Controlling for
insurance status, Latino children are not more likely than
White, non-Latinos to rely on emergency department care.
Implications for Policy, Delivery or Practice: Reliance on the
emergency department for routine care is more costly, less
effective and less coordinated than care provided by a primary
care provider. There is a widespread belief that expanding
health insurance coverage for children will effectively end
children’s reliance on emergency departments as the usual
source of care for non-urgent conditions. This research
suggests that health insurance coverage alone will not
substantially reduce the numbers of children who
inappropriately rely on emergency departments for care. The
search for alternatives should include targeting children with
characteristics that are associated with a high probability of
reliance on the emergency department, weighted by the
numbers of children with those characteristics. The ranking of
interventions could then be evaluated by analyzing the
probability of success for the interventions associated with
each of the weighted targets.
Primary Funding Source: The Flinn Foundation & Arizona
State University
• Children with Special Health Care Needs with Emotional,
Developmental and/or Behavioral Needs: Barriers to
Utilization
Roberta Scheinmann, M.P.H., Gerry Fairbrother, Ph.D.,
Melinda Dutton, JD, Beth Osthimer, JD, Rachel Cooper, M.A.
Presented by: Roberta Scheinmann, M.P.H., Research
Associate, Division of Health and Science Policy, The New
York Academy of Medicine, 1216 Fifth Avenue, New York, NY
10029; Tel: 718.788.3216; E-mail: rs536@hotmail.com
Research Objective: To examine how Children with Special
Health Care Needs (CSHCN) with emotional developmental
and/or behavioral problems differ from other CSHCN and
from children without special needs (non-CSHCN), in
measures of their barriers to utilization of health services.
Study Design: A cross-sectional random-digit dial survey of
parents of publicly insured or uninsured children was
conducted in New York City in 2002/2003. Questions about
access to and use of medical care were asked about a focal
child.
Population Studied: Publicly insured and uninsured children
in New York City
Principal Findings: Of the 1172 respondents, 24% (n=266)
identified their child as a CSHCN using a standard tool
developed by FACCT. Of the 266 CSHCN children, 33% had
an emotional, developmental or behavioral problem that was
expected to last more than one year. Children with emotional,
developmental or behavioral problems were more likely than
other CSHCN and non-CSHCN to have seen a mental health
professional in the past year [54% vs. 5% vs. 4%; p<.001] and
were more likely to have delayed or gone without emotional
developmental and behavioral care [18% vs. 2% vs. 2; p<.001].
Among the problems these children and their families face in
getting care compared to the other CSHCN and non-CSHCN
were cost of care [23% vs. 9 % vs. 8%; p=.01]; not knowing
where to go for care [22% vs. 10% vs. 8%; p=.01]; not getting a
referral or approval [25% vs. 13% vs. 7%; p<.01]; and
transportation to care problems [25% vs. 14% vs. 7%].
Conclusions: Even after controlling for insurance status,
CSHCN with emotional, developmental or behavioral
problems, experience greater health access difficulties than
other CSHCN and than children without special needs.
Implications for Policy, Delivery or Practice: Outpatient
mental health services need to be expanded or made more
accessible in order to effectively link all CSHCN with care.
Primary Funding Source: WT Grant and Altman Foundations
• Measuring Health-Related Quality of Life (HRQOL) in
Pediatric Cerebral Palsy: Hearing the Voices of Children
Sandra Sherman, B.A., Tasha Burwinkle, Ph.D., Psy.D., James
Varni, Ph.D., Kanela Hanna, B.A., Susan Berrin, B.A., Jason
Miller, B.A.
Presented by: Sandra Sherman, B.A., doctoral student,
Motion Analysis Laboratory, SDSU/UCSD Joint Doctoral
Program in Clinical Psychology, Children's Hospital of San
Diego, 6363 Alvarado Court, Suite 103, San Diego, CA 92120;
Tel: 858.455.1887; Fax: 619-594-6780; E-mail:
ssherman@ucsd.edu
Research Objective: Outcomes research has typically focused
on physical measures, proclaiming them quantifiable and
scientifically valid. Recently, a paradigm shift resultant in a
holistic approach to healthcare emphasizes measurement of
HRQOL, encompassing both physical and psychosocial
domains. Due to cognitive deficits associated with pediatric
CP, however, outcomes research using children’s selfperceptions of their HRQOL is limited. Furthermore, research
has focused on physical functioning, as studies suggest only
modest correlations between parent and child measures of
internal states. We hypothesize that: (1) children with CP and
their parents can reliably self and proxy-report HRQOL, (2)
higher correlations between parent-proxy and child self-report
on physical than psychosocial functioning; (3) CP patients will
report lower HRQOL than a healthy sample; (4) greater
disease severity will be associated with lower HRQOL scores.
Study Design: The Pediatric Quality of Life Inventory™
(PedsQL™) 4.0 Generic Core Scales, which generates a total
score and two main subscales: physical and psychosocial
(comprised of emotional, social, and school functioning), was
administered to 60 CP patients (5-18) and 183 parents of
patients (2-18). Internal consistency reliability coefficient
alphas for child and parent reports were determined for group
comparisons, and correlations between parent-proxy and child
reports were calculated. ANOVAs to determine the
relationship between disease severity and HRQOL, and t-tests
for independent samples to compare CP patients with a
healthy sample (401 children, 718 parents) were performed.
Population Studied: CP patients and their parents
Principal Findings: Internal consistency reliability alpha
coefficients for total, physical, and psychosocial functioning
were: child = .84, .76, .77; parent = .90, .89, .85. Correlations
between parent and child reports were higher for physical (.61,
p<.01) than psychosocial functioning (.22, n.s.). Within the
psychosocial domain, these correlations were moderate and
significant (p<.01) for social (.46) and school (.32), indicating
emotional functioning (.12, n.s.) was the source of crossinformant variance. For both parent-proxy and self-report,
HRQOL was significantly (p<.001) lower for CP patients than
healthy children, with large effect-sizes for differences between
these groups for both physical (self-report = .94, parent-proxy
= 1.70) and psychosocial functioning (self-report = .89, parentproxy = 1.76). For child and parent-proxy, greater CP disease
severity was associated with lower total HRQOL (hemiplegia >
diplegia, quadriplegia, p<.05).
Conclusions: Our most striking finding is that children with
CP can reliably report their HRQOL. Although parent-child
correlations for physical, social, and school functioning are
moderate to strong, the small, non-significant correlation for
emotional functioning (the most internalized psychosocial
domain) indicates parents are not accurate reporters of their
children’s emotional state, and emphasizes the importance of
child self-report. Additionally, large effect-sizes between CP
patients and healthy children in both physical and
psychosocial functioning suggest the latter as an important
domain that should be measured and addressed, in addition
to traditional physical measures.
Implications for Policy, Delivery or Practice: Children with
CP suffer distress in both physical and psychosocial
functioning. The feasibility of self-report in CP patients
suggests a unique opportunity to evaluate HRQOL in this
population, allowing for targeted interventions that address
areas of concern for the child, including physical and
emotional difficulties, to begin to hear the voices of these
children.
Primary Funding Source: March of Dimes
• Obesity and Medicaid: A Life-Cycle Model of Pediatric
Obesity
Judith Shinogle, Ph.D., Melayne Morgan McInnes, Ph.D., Eric
Johnson, Ph.D.
Presented by: Judith Shinogle, Ph.D., Assistant
Professor/NCHS AcademyHealth Fellow, Arnold School of
Public Health, University of South Carolina, NCHS 3311 Toledo
Road, Rm 3310, Hyattsville, MD 20782; Tel: 301458.4799; Email: shinogle@cop.sc.edu
Research Objective: The prevalence of obesity in the U.S. has
increased at epidemic rates for both
children and adults. The increase in obesity potentially raises
Medicaid expenditures by two pathways. First, obesity has
been shown to increase the risk of a variety of chronic
conditions and diseases even in childhood. Second, obesity
may impact outcomes in the labor and marriage markets and
thereby affect the likelihood of qualifying for Medicaid. We
consider both pathways in calculating the obesity attributable
fraction of Medicaid expenditures for children and adults.
Study Design: Using data from the 1998 Medical Expenditure
Panel Survey (MEPS) and the 1996- 1997 National Health
Interview survey, we estimate a two part Medicaid
expenditures model for children ages 4-17 and adults ages 1864. We follow the Manning and Mullahy algorithm for
retransormation of logged dependent variable. The model of
Medicaid expenditure includes controls for characteristics
other than obesity that may cause expenditures to differ. Since
MEPS has a complex survey design, all analysis was
performed using survey weights, primary sampling units, and
strata as defined by the documentation and analyzed with
StataTM survey commands.
We also consider the links between adult obesity and
childhood obesity. Longitudinal data shows that obesity in
childhood raises the risk of adult obesity substantially even
after controlling for parental obesity. Using probability
estimates from longitudinal data, we calculate the fraction of
adult Medicaid expenditures that can be attributed to pediatric
obesity.
Population Studied: The population studied for this paper
are children ages 4 to 17.
Principal Findings: We find that children’s obesity accounts
for nearly 8% of children’s annual Medicaid expenditures
while adult obesity accounts for over 12% of adult Medicaid
expenditures. When we link adult obesity back to pediatric
obesity, we find that pediatric obesity is associated with an
increase in adult Medicaid expenditures of nearly 3%. The
discounted lifetime cost to Medicaid of adult obesity is found
to be over $900.
Conclusions: Our results for the pediatric obesity attributable
fraction of adult Medicaid expenditure
suggest that more emphasis needs to be given to the
prevention and reduction of childhood
obesity. Studies exploring the causes of the increase in adult
obesity may also want to consider
the role of pediatric obesity in adult obesity. Focusing solely
on adult obesity may lead one to
look for factors that affect current adults while missing the
root causes that occur much earlier.
Implications for Policy, Delivery or Practice: From the
perspective of policy makers, a better understanding of the full
life-cycle cost of obesity will determine whether there are
potential cost savings from allocating resources to the
prevention of pediatric obesity.
• Promoting Change for Children through Parent
Leadership and Civic Engagement: A Report on the
Connecticut Parent Leadership Training Institute
Michele Solloway, Ph.D., RPP
Presented by: Michele Solloway, Ph.D., RPP, Research
Associate Professor, Health Management and Policy,
University of New Hampshire, Hewitt Hall, #327, Durham,
NH 03824; Tel: 603.862.4150; Fax: 603-862-3461; E-mail:
micheles@cisunix.unh.edu
Research Objective: Evaluate the effects of Connecticut
Parent Leadership Training (PLTI) on civic engagement,
parent leadership in communities, state policy, systems
serving children, and outcomes for children, families and
communities
Study Design: Retrospective design included focus groups,
participant survey, and in-depth interviews with participants
and community leaders. The survey included assessments of
the curriculum; measures of civic engagement, literacy and
leadership skills; and outcome measures for individuals,
children, families, and communities. Descriptive statistics
were used for survey data. Content and thematic analyses
were used for qualitative data.
Population Studied: All PLTI participants over 10-year lifetime
of project n=836.
Principal Findings: The PLTI is a state-sponsored
community-based program located in 10 sites and consists of
a 20-week training curriculum designed to prepare parents to
become leaders in their communities. It teaches principles of
child development, how to effect change in communities,
tolerance for diversity, and civic skills, tools, and knowledge.
The PLTI attracts people from all socioeconomic and
demographic groups. The sites vary significantly by
demographic groups, but participant outcomes and
expectations showed consistently high scores. The curriculum
and instructors received high scores with some variation by
training module, site and sociodemographic groups. 80
percent of respondents said PLTI improved their ability to be
an agent of change in their community and improved their
self-confidence. 93-51 percent use various civic skills at least
once a month or more. 49 percent said PLTI helped them
make changes in their children’s schools or communities and
provided new opportunities for participation and civic
engagement. 20 percent returned to school for additional or
advanced education; 15 percent got new jobs; 11 percent
developed new careers; and 8 percent received raises at their
place of employment. 79 percent reported positive changes in
their children: Almost half said their children are now
participating in community activities; 41 percent said their
children showed improvements in self-confidence; and 31
percent said their children began to take more interest in the
news. About one-third said their children showed positive
changes in 4 or more outcomes. 61 percent said PLTI helped
their families, including improved family cohesiveness, family
problem solving; quality time spent with each other; and
feeling more connected with other families in the community.
Conclusions: The PLTI is an effective program for engaging
parents to become leaders and agents of change for their
children at the individual and policy level. When given civic
tools, parents can guide public policy and decision-making
through their commitment to children. Significant positive
externalities for children and families are found.
Implications for Policy, Delivery or Practice: Parents
represent a significant resource for policy makers and systems
serving children. Continuation and replication of PLTI in other
states is warranted. Funding remains critical. Methodologies
for tracking ripple effects require further investigation.
Primary Funding Source: State of Connecticut, Commission
on Children
• Psychiatric Comorbidity among Hospitalized Children
and Adolescents
Joseph Thompson, M.D. M.P.H., James Robbins, Ph.D., Mary
Aitken, M.D. M.P.H., Teresa Kramer, Ph.D., Terri Miller, Ph.D.
Presented by: Joseph Thompson, M.D. M.P.H., Director,
Arkansas Center for Health Improvement, University of
Arkansas for Medical Sciences, 5800 West 10th Street, Little
Rock, AR 72204; Tel: 501.660.7555; Fax: (501) 660-7543; E-mail:
thompsonjosephw@uams.edu
Research Objective: To examine patterns of comorbid
psychiatric diagnoses in patients 6-17 years old hospitalized
for selected illnesses and injuries and to determine length of
stay and hospital charges associated with comorbid
psychiatric disorders.
Study Design: Analyses are based on the 1997 Healthcare
Cost and Utilization Project (HCUP) Kid s Inpatient Database
(KID). The KID is the only national, all-payer database of
hospital discharges for children. KID data are weighted to
represent all pediatric discharges from general hospitals in the
U.S. Frequencies of discharges with principal diagnoses of
selected acute and chronic illnesses and injuries, and
secondary diagnoses of mental or substance-related disorders
were calculated. Length of stay and costs were compared for
patients with and without comorbid psychiatric diagnoses.
Population Studied: The 1997 KID includes an 80% sample
of discharges from all general and children's hospitals in the
22 participating states. A total of 1.9 million discharge records
of patients 18 years of age and younger from 2,521 hospitals
are included in the database. These analyses are limited to
children between the ages of 6 and 17.
Principal Findings: There was wide variation across medical
conditions in proportions of discharges with secondary
diagnoses of mental or substance-related disorders. Comorbid
psychiatric diagnoses were most common for epilepsy
(23.89%), burns (10.31%), traumatic brain injury (TBI; 9.67%),
diabetes (8.60%), and internal injuries (7.11%) and less
common for appendicitis (1.77%), sickle cell anemia (2.97%)
and gastroenteritis (3.39%). Most common comorbidities
overall included ADHD, substance-related disorders,
depression, and mental retardation. The most frequent
comorbidities within conditions included substance-related
disorders in hospitalizations for internal injuries and TBI,
depression in diabetes, and ADHD and developmental
disorders in epilepsy. The presence of psychiatric comorbidity
was consistently associated with longer mean length of stay,
and higher mean hospital charges.
Conclusions: Common comorbid psychiatric diagnoses in
hospitalized children include prevalent and readily identified
disorders that are associated with an increased burden of care.
Patterns of diagnoses often follow recognized etiologic
associations between mental or substance-related disorders
and physical illnesses or injuries.
Implications for Policy, Delivery or Practice: These national
data highlight the need for increased attention to the influence
of psychiatric disorders in the care of hospitalized youth.
Comorbid mental health and substance-related disorders
increase the length and costs of hospital stay.
• Pediatric Hospitalizations for Traumatic Brain Injuries:
1997 and 2000
John Tilford, Ph.D., Mary Aitken, M.D. M.P.H., Allen
Goodman, Ph.D., Jeffrey Killingsworth, BS, Jerril Green, M.D.,
Debra Fiser, M.D.
Presented by: John Tilford, Ph.D., Associate Professor,
Pediatrics/Health Policy and Management, University of
Arkansas for Medical Sciences, 800 Marshall Street, Little
Rock, AR 72202-3591; Tel: 501.364.3340; Fax: (501) 364-1552; Email: tilfordmickj@uams.edu
Research Objective: This study evaluates the incidence, use
of procedures, and outcomes of pediatric hospitalizations
(age 0-18) for traumatic brain injuries (TBI) by patient and
hospital characteristics using a unique database covering the
years 1997 and 2000.
Study Design: ICD-9 diagnosis codes from the Kids Inpatient
Database for the years 1997 and 2000 were used to identify
children with a TBI and classify the type of injury. ICDMAP-90
computer software also was used to develop abbreviated
injury scores and injury severity scores to risk-adjust
outcomes. The abbreviated injury scores were used to create
mild, moderate, and severe categorizations of TBI severity.
ICD-9 procedure codes were used to determine whether the
patient received therapies such as mechanical ventilation and
intracranial pressure monitoring. Patient payer status was
coded as public, private, self-pay (uninsured), or other
insurance.
Population Studied: National estimates of pediatric TBI
hospitalizations, use of procedures, and outcomes were
developed using weights supplied with the database.
Principal Findings: Hospitalizations involving mild TBI
decreased by 28% over the two periods. There were no
differences in the incidence of moderate or severe TBI
hospitalizations. Patient insurance status was associated with
use of therapies and in-hospital mortality. Self-pay patients
with moderate TBI were less likely to receive mechanical
ventilation, but there was no difference in in-hospital mortality.
Self-pay patients with severe TBI, had significantly higher rates
of unadjusted in-hospital mortality relative to privately insured
patients (21.4% vs. 10.1%) and adjusted in-hospital mortality
(OR = 3.45; p<0.001). Differences in injury severity explained
some, but not all of the increased mortality in self-pay
patients. Rates of intracranial pressure monitoring did not
differ by payer status.
Conclusions: A unique database for developing national
estimates of pediatric hospitalizations indicates a downward
shift in hospitalizations for mild TBI consistent with recent
management guidelines. For moderate and severe TBI
hospitalizations, the data show significant differences in inhospital mortality and therapies for uninsured children, but
the available evidence does not provide a link between quality
of care and outcomes.
Implications for Policy, Delivery or Practice: Understanding
why uninsured children with TBI die more frequently than
publicly or privately insured patients is necessary to generate a
link between health insurance expansions and reductions in
TBI-related deaths. This study suggests insurance may play a
role in reducing injury severity (prevention) rather than
improving technical quality of care, but the possibility that
better outcomes for insured patients are related to higher
quality care cannot be ruled out.
Primary Funding Source: HRSA, MCHB
• SCHIP Eligibility and Access to Care for Children with
Special Health Care Needs: Results from a Nationally
Representative Survey
Hao Yu, M.Sc., Andrew Dick, Ph.D, Peter Szilagyi, M.D.,
M.P.H.
Presented by: Hao Yu, M.Sc., Ph.D Candidate, Community
and Preventive Medicine, University of Rochester, 601
Elmwood Avenue, Box 644, Rochester, NY 14642; Tel:
585.275.3432; Fax: (585)461-4532; E-mail:
hao_yu@urmc.rochester.edu
Research Objective: While studies in specific states have
examined enrollment in the State Children’s Health Insurance
Program (SCHIP) among children in general and children with
special health care needs (CSHCN), little is know at the
national level about the role of SCHIP with respect to CSHCN.
This study aims to assess the:
(1). Proportion of CSHCN eligible for SCHIP across the U.S.,
(2). Proportion of SCHIP-eligible CSHCN who are uninsured,
(3). Characteristics of the SCHIP-eligible but uninsured
CSHCN, and
(4). Difference in unmet needs and satisfaction of care
between CSHCN enrolled in SCHIP and those eligible but
uninsured.
Study Design: Data were from the first National Survey of
CSHCN, which was conducted in 2000-2001. Bivariate and
multivariate analyses were performed.
SCHIP-eligibility was defined by age and income according to
state-specific policies as of October 2000.
Population Studied: All the 38,866 CSHCN interviewed by the
survey.
Principal Findings: Nationally, 19.0% of CSHCN, or 1.7
million, were eligible for SCHIP in 2000, and among these
eligible CSHCN, 8.2%, or 141,000, were uninsured.
Substantial variation existed across states in terms of SCHIP
eligibility and the proportion of eligible but uninsured CSHCN.
CSHCN who were SCHIP-eligible but uninsured were less
likely than CSHCN enrolled in SCHIP to have a personal
health care provider (70.5% vs. 85.6%, P<0.001), more likely
to have an unmet need (47.7% vs. 14.4%, P<0.05), and less
satisfied with care (77.7% vs. 89.8%, P>0.05). After controlling
for other patient features, the SCHIP-eligible but uninsured
CSHCN were 5 times more likely to have unmet needs than
those enrolled in SCHIP.
Between the SCHIP-enrolled CSHCN and CSHCN who were
SCHIP-eligible but privately insured, there were significant
differences in the rates at which CSHCN had a personal
provider (85.6% vs. 91.7%, P<0.01), although the two groups
had similar rates of usual source of care (92.0% vs. 92.8%,
P>0.05). The two groups were also similar in rates of unmet
needs (14.4% vs. 12.1%, P>0.05), and of satisfaction with care
(89.6% vs. 91.6%, P>0.05).
Conclusions: A substantial number of CSHCN nationally
were eligible for SCHIP but were uninsured, and these
uninsured CSHCN were more likely to have problems in
access to health care. The SCHIP-enrolled CSHCN resembled
CSHCN privately insured in terms of unmet needs and
satisfaction.
Implications for Policy, Delivery or Practice: It may be
helpful for states with high insurance rates among SCHIPeligible CSHCN to share strategies by which they successfully
enrolled these children into SCHIP. On the other hand, states
with low insurance rates may wish to review their eligibility
criteria, enrollment practices, and outreach strategies.
Another implication is that, despite the high rates of unmet
needs, satisfaction with care is relatively high according to the
standard CAHPS measures. These measures may not be
sensitive indicators of unmet needs for vulnerable populations
such as CSHCN.
An implication for clinicians is that many uninsured CSHCN
may in fact be eligible for SCHIP. Clinical offices can play a
role in educating about SCHIP, and in helping families enroll.
Primary Funding Source: The Maternal Child Health Bureau
Invited Papers
Strategies for Moderating the Long-Term Impact of
Childhood Obesity
Chair: Christopher Forrest, M.D., Ph.D.
Sunday, June 6 • 5:00 p.m.-6:30 p.m.
•
Panelists: Scott Gee, Kaiser Permanente, Roland Sturm,
RAND, Joseph Thompson, University of Arkansas for
Medical Sciences (no abstracts provided)
Invited Papers
Impact of SCHIP on Vulnerable Children: Findings from
the Child Health Insurance Research Initiative
Chair: Cindy Brach, M.P.P.
Monday, June 7 • 8:30 a.m.-10:00 a.m.
•
Panelists: Andrew Dick, University of Rochester; Elizabeth
Shenkman, University of Florida; Peter Szilagyi, University
of Rochester Medical Center (no abstracts provided)