Child Health Call for Papers Medicaid, SCHIP & Access to Care: National, State & Local Perspectives Chair: Paul Wise, Brigham and Women’s Hospital Sunday, June 26 • 8:30 am – 10:00 am ●Changes in Children’s Mental Health Care, 1997-2002. Tatiana Andreyeva, MA, MPHil, Roland Sturm, Ph.D. Presented By: Tatiana Andreyeva, MA, MPHil, Doctoral fellow, RAND Graduate School, 1776 Main Street, Santa Monica, CA 90401; Tel: (310) 393-0411 x6047; Fax: (310) 2608155; Email: tatiana@rand.org Research Objective: Multiple forces have changed the practice of mental health care for children during the past decade. Limited empirical data is available on the consequences of these reforms for children's mental health care. The objective of this study is to describe changes in mental health care in a large representative sample of schoolage children over 1997-2002, and to examine differential rates of change across socioeconomic populations and geographic regions. Study Design: The study uses data from the 1997, 1999 and 2002 cross-sectional rounds of the National Survey of America’s Families, a nationally representative survey of children, adults under the age of 65, and their families fielded in 13 states and the balance of the nation for the Urban Institute. The main outcome measures are: 1) utilization of any mental health services and number of visits among users, 2) unmet need for mental health services (no service use among children with identified need according to the Child Behavior Checklist), and 3) need among users of mental health services. Population Studied: Nationally representative children ages 6-17 surveyed in 1997 (N = 21,824), 1999 (N = 23,423), and 2002 (N = 22,064). Principal Findings: Utilization of mental health services among school-age children increased significantly between 1997 and 2002, with most of the growth occurring between 1999 and 2002. While the average utilization of any mental health services grew nationwide from 7.1% in 1997 to 8.9% in 2002, there was substantial variation in growth rates across socioeconomic populations and geographic regions. Among all socioeconomic groups examined, only uninsured children had a decrease in utilization of mental health services over 1997-2002, from as low levels as about 2/3 of the national average in 1997 to 1/2 of the national levels in 1999 and 2002. Utilization rates for black children and children with public insurance were the slowest to grow (1.3-1.4% annually), whereas white children and children in high-income families had the opposite extreme (5.6-6.0% annually). Changes in overall means mask significant differences in trends across states. Although all 13 states in the survey increased their use of children’s mental health services between 1997 and 2002, the slowest to grow (2.1-2.2% per year) were the states with initial utilization levels well below the national average, e.g. Alabama and Mississippi. Similar patterns occurred in unmet need for mental health services and need among users. Conclusions: Utilization of mental health services among school-age children is increasing at substantially different rates among socioeconomic populations and geographic regions. As children with the highest need, predominantly black children and children in low-income families, are increasing utilization at a much lower rate than children overall, the well-known and disturbing socioeconomic disparities in health care may threaten to increase in the future. We show multi-fold differences in the growth of children’s mental health utilization across states (up to 5-fold) that exceed the disparities in trends across race/ethnicity and income groups. Implications for Policy, Delivery, or Practice: Changes in the practice of mental health care are associated with disproportionate effects on children in already marginal groups and lagging states. Providers can make a difference by delivering mental care focused on children with the highest needs rather than trying to increase the overall rates of service use. Primary Funding Source: National Institutes of Mental Health ●Insurance Trends: Rural, Minority Children Lag Over 20 Years Keith Elder, Ph.D., MPH, MPA, Janice Probst, Ph.D., Daniel Patterson, MPH, Ph.D., Charity Moore, MSPH, Ph.D. Presented By: Keith Elder, Ph.D., MPH, MPA, Assistant Professor, Health Services Policy and Managment, University of South Carolina Arnold School of Public Health, 800 Sumter Street Health Sciences Building, Room 120, Columbia, SC 29208; Tel: (803) 777-5041; Fax: (803) 777-1836; Email: kelder@gwm.sc.edu Research Objective: Over the past decade, Medicaid coverage for poor children has significantly expanded, principally through SCHIP. We examined the degree to which this expansion may have increased overall levels of insurance coverage for children, with a focus on rural and minority children (African American and Hispanic), across the period 1980-2000. We also assessed health care utilization patterns over time among rural children and whether the patterns were different for white and minority children. Study Design: Data were drawn from the National Health Interview Survey, a nationally representative household survey conducted annually by the National Center for Health Statistics. Public use data sets for the years from 1980 through 2001 were accessed. The dependent variables were health insurance status, family poverty index, family structure (one or two parents in household), highest education level in the household, and doctor visits in past 12 months. All analyses were stratified by race (white, African American, Hispanic) and residence (urban, rural). Population Studied: Children ages 0 –17. The number observed varied by survey year, from a high of 35,196 children in 1987 through 12,389 children in 1999. Rural was defined as residing outside an SMSA. Principal Findings: In 1979, the proportion of children who lacked health insurance ranged from 10.4% among urban whites to 36.3% among rural Hispanics. Among white children, the proportion without insurance peaked in 1994 at 20.9% (urban), declining to 5.9% in 2001. By 2001, only 8.7% of rural white children, 14.2% of rural African American children, and 26.7% of rural Hispanic children remained without coverage. Throughout the period, urban white children were least likely to be uninsured. Within each race/ethnicity, rural children were less likely to have insurance coverage than urban children, a disparity that remained unchanged across the period. The proportion of children with no physician visit in the preceding year declined over the period. Among urban whites, the proportion of children with no visit declined from 22.4% in 1979 to 9.4% in 2001. Among rural Hispanics, the most at-risk group, the proportion of children with no visit declined from 40.2% in 1979 to 24.4% in 2001. Conclusions: The proportion of children without insurance declined between 1979 and 2001, with parallel declines in the proportion of children with no access to health services. However, racial disparities remain, with rural minority children most likely to be uninsured and least likely to have seen a physician within the past year. Implications for Policy, Delivery, or Practice: The proportion of rural children without insurance declined sharply with the introduction of SCHIP in 1996. In the current budgetary climate, however, states are unlikely to be able to support Medicaid expansions without federal matching funds, slated to expire in 2007. Given the erosion of private insurance during the late 1990s and through the present, the proportion of uninsured children is likely to increase across the next decade, with adverse effects on children and the institutions that serve them. Primary Funding Source: HRSA, Federal Office of Rural Health Policy ●Out-of-pocket Financial Burden for Low-income Families with Children: Socioeconomic Disparities and Effects of Insurance Alison Galbraith, M.D., MPH, Sabrina T. Wong, RN, Ph.D., Sue E. Kim, Ph.D., MPH, Paul W. Newacheck, DrPH Presented By: Alison Galbraith, M.D., MPH, Instructor, Department of Ambulatory Care and Prevention, Harvard Pilgrim Health Care and Harvard Medical School, 133 Brookline Avenue, 6th Floor, Boston, MA 02215; Tel: (617)5099893; Fax: (617)859-8112; Email: alison_galbraith@hms.harvard.edu Research Objective: To determine whether socioeconomic disparities exist in the financial burden of out-of-pocket health care expenditures for families with children, and whether health insurance coverage decreases financial burden for lowincome families. Study Design: Cross-sectional family-level analysis. The main outcome was financial burden, defined as the proportion of family income spent on out-of-pocket health care expenditures. We aggregated annual out-of-pocket expenditures for all family members. Family insurance coverage was categorized as: 1) all members publicly insured all year, 2) all members privately insured all year, 3) all members uninsured all year, or 4) a mixture of private, public, and/or no insurance. We used bivariate statistics to examine if financial burden varied by poverty level. Multivariate linear regression models were used to assess whether family insurance coverage was associated with level of financial burden for low-income families. Population Studied: Families with a child <19 years old in the Household Component of the 2001 Medical Expenditure Panel Survey. Principal Findings: There was a regressive income gradient in financial burden such that families with incomes <100% of the Federal Poverty Level (FPL) had disproportionately greater financial burden than families with higher incomes. For lowincome families (< 200% FPL), financial burden was significantly less for families with full-year public insurance compared to those without insurance all year and compared to those with full-year private insurance. Conclusions: Socioeconomic disparities exist in the financial burden of out-of-pocket health care expenditures for families with children. However, financial burden is reduced for lowincome families with full-year public coverage. Full-year public coverage provides substantially greater protection for lowincome families than full-year private coverage. Implications for Policy, Delivery, or Practice: Public insurance programs play an important role in protecting lowincome families with children from burdensome out-of-pocket expenses. Private insurance plans may place low-income families at greater risk for financial burden from health care expenditures. Policies to expand health insurance coverage to the uninsured should consider the effects of cost sharing on financial burden for families. Primary Funding Source: AHRQ ●The Effect of County Characteristics on Children’s Participation in California’s Medicaid and SCHIP Programs Jennifer Kincheloe, Ph.D., MPH Presented By: Jennifer Kincheloe, Ph.D., MPH, Research Scientist, Health Services, UCLA Center for Health Policy Research, UCLA School of Public Health, 845 Coronado Drive, Glendale, CA 91206; Tel: (310)909-8725; Email: jkinchel@ucla.edu Research Objective: This study examines the impact of county-level factors on children’s enrollment in California’s Medicaid (Medi-Cal or MC) and SCHIP (Healthy Families or HF) programs, while controlling for individual and family-level factors. Among other factors, it examines the impact of outreach funding, media expenditures, outstationed eligibility workers, and county-sponsored health insurance expansion initiatives (expansion initiatives) for undocumented children and children whose families make slightly too much money to qualify for HF. Study Design: Cross-sectional; simple and hierarchical multivariate logistic regression analysis. Independent variables include family-level sociodemographic characteristics and county characteristics such as money spent on outreach, the number of outstationed eligibility workers, and whether there was an expansion initiative. Population Studied: Individual-level and family-level data from the 2001 California Health Interview Survey (CHIS 2001) were merged with county-level data from the California Department of Health Services and Managed Risk Medical Insurance Board, the Medi-Cal Policy Institute, and the 100% Campaign. These included data on whether a county had an expansion initiative, the number of outstationed eligibility workers, and money dispersed to each county from the state for various types of outreach in FY 1999-2000 and FY 20002001. The CHIS 2001 random-digit dial telephone survey collected information from a representative sample of about 55,000 California households. The study focuses on children ages 0-17 who were eligible for either the MC or the HF program and had no private or employer-sponsored insurance coverage. Principal Findings: 1) Controlling for other factors, eligible children who lived in counties with an expansion initiative had three times the odds of enrolling in MC compared with those who lived in counties without expansion initiatives (p=0.06). There was no significant effect on HF enrollment, perhaps because eligibility is administered at the state level while MC eligibility is administered locally. 2) Among children eligible for MC, the odds of being enrolled increase by 6% for each additional dollar spent on outreach per eligible child in the county, excluding media (p=0.08). Among HF eligible children the odds increased by 7% for each additional dollar spent on outreach per eligible child. 3) For each additional outstationed eligibility worker per 1,000 eligible children in the county, the odds of MC enrollment decreased by two thirds. 4) Media dollars spent per eligible child had a negative impact on enrollment in HF (OR .89). Conclusions: This study demonstrates the positive “spillover effect” that local expansion initiatives have on children’s enrollment in the MC program and the importance of outreach money in reducing uninsurance among program eligible children. The spillover effect was not significant for HF— perhaps because HF enrollment is administered at the state level while MC enrollment is administered locally. The study highlights the need to evaluate outreach strategies and to dedicate scarce resources where they will have the greatest impact. California’s “one size fits all” media campaign actually had a negative effect on HF enrollment, as did outstationed eligibility workers. The latter surprising result may be explained by the fact that outstationing is expensive, diverting resources away from other outreach programs, and that eligibility workers in hospitals target high cost patients, not children, in order to reduce bad debt. Implications for Policy, Delivery, or Practice: States may increase children’s enrollment in the Medicaid and SCHIP programs by funding proven program outreach and through expansion initiatives. Local administration of program enrollment may have a positive effect on enrollment. Primary Funding Source: California Program on Access to Care ●Uninsurance Among Children Who Are Eligible for SCHIP: Results from A Nationally Representative Survey Hao Yu, Ph.D. Presented By: Hao Yu, Ph.D., Health Policy Analyst, RAND, 201 North Craig Street, Suite 202, Pittsburgh, PA 15213; Tel: (412)683-2300 x4460; Fax: (412)683-2800; Email: hao_yu@rand.org Research Objective: While studies in specific states have examined children’s enrollment pattern in the State Children’s Health Insurance Program (SCHIP), little is known at the national level about those who are eligible for SCHIP but uninsured. This study aims to assess the: (1). Proportion of children eligible for SCHIP across the U.S., (2). Proportion of SCHIP-eligible children who are uninsured, (3). Risk factors for these children to be uninsured, and (4). Their parents’ knowledge of and experience with SCHIP. Study Design: Data were from the first National Survey of Children with Special Health Care Needs, 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. The uninsurance referred to being uninsured for the past 12 months Population Studied: 215,162 children sampled by the survey, including 38,866 children with special health care needs. Principal Findings: Nationally, 6% of children, or 4.3 million, were eligible for SCHIP in 2000, and among these eligible children, 36%, or 1.5 million, were uninsured. Substantial variation existed across states in terms of SCHIP eligibility and the proportion of eligible but uninsured children. After controlling for other socio-demographic characteristics, older children were more likely to be uninsured. Black children had lower probability of being uninsured than white children. Children from families, for which the interview language was not English, were two times more likely to be uninsured than those from English-speaking families. Children with special health care needs were 50% less likely to be uninsured than those without special needs. The main reasons for not having insurance were “costs too much” (40%) and “can’t get insurance through employer” (14%). Among the SCHIPeligible but uninsured children, 63% of their parents had heard about SCHIP, compared with 88% of them hearing about Medicaid. The top two sources that these parents first heard about Medicaid/SCHIP were “friends and family members” (28%) and “health care providers” (21%). Although 82% of these parents indicated that they would enroll their child if told eligible, only 22% of them had ever applied for SCHIP, and 19% of them perceived difficulty in completing SCHIP application. Conclusions: A substantial number of children nationally were eligible for SCHIP but were uninsured, and their parents had limited knowledge of SCHIP and had difficulty in applying for SCHIP. Implications for Policy, Delivery, or Practice: It may be helpful for states with high uninsurance rates to review their eligibility criteria, enrollment practices, and outreach strategies. In particular, specific enrollment efforts should be taken to target older children, white children, children from non-English-speaking families, and children without special health care needs. An implication for health care providers is that many uninsured children may in fact be eligible for SCHIP. Health care providers can play a role in educating about SCHIP, and in helping families enroll. Primary Funding Source: No Funding Call for Papers Organizational & Community Factors in Quality & Safety of Care for Children Chair: David Bergman, Stanford University Sunday, June 26 • 10:30 am – 12:00 pm ●Outcomes at 6 Months Post-Admission to Pediatric Intensive Care: Report of a National Study of Pediatric Intensive Care Units in the United Kingdom. Samantha Jones, BSc Pharmacology, GJ Parry, K Rantell, K Stevens, C McCabe, K Rowan Presented By: Samantha Jones, BSc Pharmacology, Research Fellow, Health Services Research, ScHARR, University of Sheffield, Regent Court, 30 Regent Street, Sheffield, England, S1 4DA; Tel: 44 (0)114 222 0772; Fax: 44 (0)114 272 4095; Email: sam.jones@sheffield.ac.uk Research Objective: To measure the health status of children using the Health Utilities Index (HUI2) 6 months after admission to a pediatric intensive care unit in the United Kingdom and to assess the relationship between long term outcomes and measures of illness severity at admission to pediatric intensive care. Study Design: Prospective observational cohort study. Population Studied: A total of 3551 patients admitted to 23 paediatric intensive care units in the United Kingdom between March 2001 and February 2002. Principal Findings: Consent forms agreeing to subject participation were received from 3843 parents/guardians of children admitted to pediatric intensive care. Following exclusions due to eligibility checks, 2895 HUI2 questionnaires were sent out to parents/guardians, and 2044 (71%) were returned. No difference was found in the initial illness severity of those children who consented to participate (versus those who did not) and those for whom a HUI2 questionnaire was returned (versus those for whom a questionnaire was not). The primary outcomes measured by the HUI2 consist of six dimensions: sensation, cognition, emotion, mobility, self-care and pain. At 6 months post admission to pediatric intensive care, the percentage of children in the highest (best possible health status) category of the six HUI2 dimensions was 61% for sensation, 70% for cognition, 66% for emotion, 71% for mobility, 69% for self-care and 65% for pain. Overall, 28% of children were in full health 6 month post admission to pediatric intensive care, whilst 9% had some level of impairment on all the outcome measures at 6 months post admission. To assess the relationship between longer-term outcomes and measures of illness severity at admission, ordinal regression models were estimated for each of the primary outcomes, using the illness severity measures included in the Pediatric Index of Mortality (PIM) and the Pediatric Risk of Mortality (PRISM, PRISM III-12 and PRISM III-24) as explanatory variables. Only the illness severity measures included in PIM were significantly related to outcome at 6 months. The remaining measures were related to some of the outcomes, but not all. None of the existing illness severity measures provided satisfactory discriminatory power for any of the health outcomes at 6 months. Conclusions: At 6 months post admission to pediatric intensive care, there is a significant sustained morbidity which must be taken into account when examining strategies to improve or optimize pediatric intensive care unit configurations. Whilst the degree of morbidity appears to be related to initial illness severity, this information alone is insufficient to predict long term outcomes. Work is ongoing to examine the predictive potential of other explanatory variables. Implications for Policy, Delivery, or Practice: Mortality following pediatric intensive care is currently only around 6% in the United Kingdom. In determining the quality and performance of pediatric intensive care units, it is important to take into account variations in the health status of survivors post discharge. However, these results suggest that illness severity at admission has little impact on the longer-term health status of children. In attempting to develop policies to improve longer-term outcomes of children post pediatric intensive care, it may be important to also take into account other factors, such as pre-existing co-morbidities, socioeconomic status and the quality of care received by children following discharge from intensive care. Primary Funding Source: Medical Research Council (MRC) ●The Impact of Chronic Care Coordination on Young Children (Age 0 to 5) with Asthma: A Statewide Evaluation of the California Community Asthma Intervention Mina Lai, MPH, David Núñez, M.D., MPH, Toshi Hayashi, Ph.D., Pradeep Gidwani, M.D., MPH Presented By: Mina Lai, MPH, Research Analyst, Epidemiology and Program Evaluation Unit, California Department of Health Services, 1616 Capitol Avenue, Suite 74.317, P.O. Box 997413, MS 7212, Sacramento, CA 958997413; Tel: (916) 552-9870; Fax: (916) 552-9911; Email: mlai@dhs.ca.gov Research Objective: During the past two decades, the prevalence of asthma increased dramatically, with the greatest increase reported for children under age five. California’s statewide Community Asthma Intervention (CAI) aims to decrease asthma morbidity and improve the quality of life for these children and their families. The objective of CAI is to demonstrate that asthma care coordination is an effective approach to alleviating the burden of asthma on young children and their families. Study Design: Community health workers and other health professionals were established as Asthma Coordinators (ACs) at eight project sites in California. The ACs were based in both clinics and community organizations and worked with the children, their parents, clinical practitioners, and childcare providers to affect change in asthma management skills, including elimination of home environmental asthma triggers. To measure this change, the ACs conducted interviews at enrollment with the parents of enrolled children. Follow-up interviews were conducted at 6-month intervals and at time of discharge. Outcome measures addressed in the interview included: possession of an asthma management plan; hospitalizations and emergency room visits; medication usage; childcare or preschool days missed; and parent work days missed. The Cochran-Mantel-Haenszel chi-square test was performed, using results from the enrollment and latest follow-up interviews. Population Studied: Children age 0 to 5 with asthma, residing in the nine counties served by CAI, were eligible for enrollment in the program. 34% of enrollees resided in rural or semi-rural communities. 92% had health insurance coverage, with 68% enrolled in the state Medicaid program. 70% of enrollees were Latino, and 15% were African-American. The data used in evaluation were based on 1,920 individuals, out of 2,497 total enrollees, whose parents completed at least two interviews. The number of individuals included in the analyses varied for each outcome measure. Principal Findings: Among other improvements in outcome measures, there were significant increases in the proportion of children who received asthma management plans from their health care providers (39% to 81%); a decrease in children who experienced daytime asthma symptoms more than twice a week (54% to 13%); a decrease in asthma-related hospitalizations (17% had one or more hospitalization in a sixmonth period, decreasing to 6%); and a decrease in missed school days (57% of children attending childcare/preschool missed at least one day or more due to asthma at enrollment, decreasing to 27% at follow-up). Conclusions: Chronic care coordination for children age 0 to 5 with asthma improves communication between healthcare providers, childcare providers, and families; addresses issues of cultural competency in asthma education; and enables parents to take an active role in their child’s asthma care by helping them to better understand the condition. The resulting improvements in health and other important outcomes demonstrate the success of the intervention. Implications for Policy, Delivery, or Practice: Coordinating asthma-related care for young children effectively addresses the critical health issues that affect this age group. Efforts should be made to integrate this model into current health care delivery systems for young children with asthma. Primary Funding Source: First 5 California ●Patient Safety for Infants and Children in Academic Medical Center Hospitals: Organizational and Human Factors Related to Harmful Medical Event Outcomes– Evidence using Electronic Medical Error-Event Reporting Systems Sandra Magnetti, MS, DrPH, Raj Behal, M.D., MPH Presented By: Sandra Magnetti, MS, DrPH, Post doctoral fellow, Clinical Practice Advancement Center, University HealthSystem Consortium, 2001 Spring Road, Oak Brook, IL 60523; Tel: (630) 954-1719; Fax: (630) 954-5879; Email: magnetti@uhc.edu Research Objective: To identify human and organizational factors, which are significantly related to harmful vs. nonharmful medication events in infants and children. Study Design: We analyzed medication error-event data reported over a three-year period submitted by 23 academic medical centers, which participated in a collaborative, on-line, web based event reporting system. Front-line clinical staff entered information on the events and scored the level of harm using a progressive 10-point scale ranging from near miss to patient death due to the event. Unit managers had the option to review the report and comment on factors that contributed to the event. The input screen format allowed them to select one or more of 36 human and organizational factors, depending on their perception of the event. We used the nurse unit manager’s report to correlate these human and organizational factors with a harm/no harm score. The score was developed by dichotomizing the progressive 10-point scale into those events that caused harm to the patient ranging from temporary harm requiring extra intervention to patient death versus those events that caused no harm to the patient. The Fisher’s exact test was used to evaluate these relationships using a p value <= 0.05. We analyzed the children’s and infant’s data separately. Population Studied: Medication error-event data from “children” (inpatients between one month and 18 years) and “infants” (inpatients from birth to less than one month of age) were utilized. The data came from 23 academic medical center hospitals’ web based reporting systems for years 2001-2004. Principal Findings: We found that of the medication events in “children” between one month and 18 years (N=4208), 6.5% were related to harm; in “infants” from birth to less than 30 days of age (N=757), 6.9% were related to harmful events. The remaining reported events did not reach the patients or require additional treatment. Of the 36 human and organizational factors available for selection, the following were significantly related to the harmful events for “children”: “insufficient staffing”, “inadequate resident supervision”, “bed availability”, and “language barrier”. For “infants”, “shift change, equipment malfunction, “equipment availability” and “no 24-hour pharmacy were listed as related to some form of harm. Conclusions: This study reveals that there are specific human and organizational factors related to medication error-events, which, either necessitated additional treatment measures, were life threatening, or resulted in death. It also indicates that these factors may differ in treatment of infants versus children. Implications for Policy, Delivery, or Practice: This project provides information as to which events are more closely associated with harm to the infants and children being treated in hospitals. Endeavors to target error-event prevention programs should take note of possible differences in organizational factors related to children versus infants. Primary Funding Source: No Funding ●Nurse Staffing and Pediatric Quality of Care Barbara Mark, Ph.D., Wallace F. Berman, M.D., David W. Harless, Ph.D. Presented By: Barbara Mark, Ph.D., Professor, School of Nursing, University of North Carolina at Chapel Hill, Carrington Hall CB#7460, Chapel Hill, NC 27599; Tel: (919) 843-6209; Fax: (919)843-3168; Email: bmark@email.unc.edu Research Objective: Recent research has demonstrated the relationship between hospital nurse staffing and various measures of quality and patient safety, but this research has focused almost exclusively on adult, rather than pediatric patients, despite the estimated $1 billion in excess charges due to adverse pediatric safety events. Two studies of inpatient pediatric acute care incidentally included a measure of nurse staffing and came to different conclusions, demonstrating the need for targeted investigation of nurse staffing and pediatric quality of care. The purpose of our study is specifically to examine the relationship between nurse staffing and pediatric patient safety outcomes (mortality, complications, and failure to rescue). We control for community level characteristics that can affect demand for hospital care (per capita income, hospital use, population, race, infant mortality rate), market characteristics (managed care penetration), and hospital characteristics (hospital size, technological complexity, ownership, teaching status, pediatric patient volume, and net income). Study Design: A longitudinal panel study of 200 general acute care hospitals in California from 1996 – 2001. Data are from: the Area Resource Files, American Hospital Association’s annual survey of hospitals, InterStudy, and the California Office of Statewide Health Planning and Development’s financial, utilization, and patient discharge files. Risk adjustment was conducted by MEDSTAT. We apply a dynamic panel data regression model for count data. Differential levels of patient acuity are accounted for by using the expected number of mortalities, complications, and failures to rescue as the measures of exposure for a hospital in a particular time period. The fixed effects specification controls for differences in initial conditions by including the initial count of outcomes as a regressor in addition to the value of the lagged count of outcomes. Staffing levels are measured using RN FTEs per 1,000 expected- mortality-indexadjusted inpatient days. Population Studied: Approximately 3 million pediatric patient discharges from 200 general acute care hospitals. Principal Findings: In our initial models, although the point estimates were not significant (i.e., p < .05), for a hospital with the mean level of expected mortalities per one thousand inpatient days, parameter estimates suggest that: 1) the number of mortalities decreases 0.1% per 1% increase in RN FTEs per one-thousand inpatient days; 2) the number of incidents of failure to rescue decreases by 0.09% per 1% increase in RN FTEs per one-thousand inpatient days; 3) the number of complications decreases by 0.03 per 1% increase in RN FTEs per one-thousand inpatient days; and 4) for-profit hospitals and teaching hospitals had significantly lower failure to rescue rates than not for profit and non-teaching hospitals. Conclusions: The relationship between nurse staffing and pediatric quality of care was not statistically significant in the models tested. However, other variables related to nurse staffing (i.e., tenure, experience, turnover) are not captured in the data set. Amplified models that include race and infant mortality, as well as additional nursing variables (extent of agency personnel) are being developed. Implications for Policy, Delivery, or Practice: These results do not support changing nurse staffing ratios to improve quality of care for pediatric patients in California – the only US state that currently has mandated minimum nurse staffing ratios. Primary Funding Source: AHRQ ●Follow-Up Care for Children Prescribed Attention Deficit Hyperactivity (ADHD) Medication Sally Turbyville, MA, Phil Renner, MBA, Russell Mardon, Ph.D., Shaheen Halim, MA, Richard Herman, M.D., Christy L. Beaudin, Ph.D. Presented By: Sally Turbyville, MA, Senior Health Care Analyst, Quality Measurement, NCQA, 2000 L Street, Suite 500, Washington, DC 20036; Tel: (202)955-1756; Fax: (202) 955-3599; Email: turbyville@ncqa.org Research Objective: To assess follow-up care for children with a new prescription for Attention Deficit Hyperactivity Disorder,ADHD, and to determine the validity of using administrative pharmacy data to identify the study population. ADHD is one of the more common chronic conditions of childhood. Study Design: Observational study conducted in six health plans. All submitted administrative claims and pharmacy data for commercial plan members and three submitted Medicaid data. For this measure, children were identified as having ADHD if they had one new prescription claim for an ADHD specifc drug. These drugs were specified on a NDC code list based on expert panel’s recommendations. Diagnosis of ADHD in administrative data was not required to qualify for the study population. Analyses were run to calculate follow-up rates, and to determine the validity of administrative data as compared to the medical record and the ability of pharmacy data to accurately capture ADHD diagnosis. Population Studied: The study population included 5,288 children ages 6 to 12 years. Medical records were examined for 867 patients. Principal Findings: The validation study showed that of children confirmed with a new medication treatment episode a diagnosis of ADHD was present for nearly all patients identified by pharmacy data (average: 96.8 percent). Approximately 43 percent of the children in the study population enrolled in a commercial plan and 45 percent of children enrolled in Medicaid had an initiation visit (i.e., a follow-up visit within 30 days after the ADHD medication was dispensed). Approximately 40 percent of the children in the study population enrolled in a commercial plan and 42 percent of children enrolled in Medicaid had at least two additional follow-up visits with a provider within 11 months of their initiation visit. There is moderate variation among health plans, and none had follow-up rates greater than 50 percent, indicating significant room for improvement. Requiring an ADHD diagnosis to be captured in the follow-up claim to count toward numerator compliance, the initiation visit rates drop to 19 percent and 23 percent for commercial and Medicaid, respectively. The continuation and maintenance follow-up visits (at least two additional visits) rates drop as well, to an average of 14 percent (commercial) and 16 percent (Medicaid). Children enrolled in Medicaid and commercial plans had similar rates. Conclusions: Low rates of follow-up for children with a new ADHD medication prescription were observed in commercial and Medicaid populations. Pharmacy data in commercial and Medicaid populations accurately identified children with ADHD on medication. Implications for Policy, Delivery, or Practice: Despite the guidelines recommending timely follow-up care for children dispensed medication for ADHD this study shows variation and significant room for improvement. Performance measurement for follow-up through NCQA HEDIS® measurement for children on ADHD medication can be used to demonstrate variable follow-up rates and improve data capture for population management. Primary Funding Source: Unrestricted Educational grants and SAMHSA grant Call for Papers Childhood Obesity & New Data & Findings for Children with Special Health Care Needs Chair: Jonathan Klein, University of Rochester Monday, June 27 • 9:00 am – 10:30 am ●Comparing and Interpreting Findings on the Prevalence and Health and Health Care Service Need Characteristics of Children and Youth with Special Health Care Needs (CYSHCN) Across Three new National Data sets Christina Bethell, Ph.D., MBA, MPH, Debra Read, MPH, Paul Newacheck, DrPH, Stephen Blumberg, Ph.D. Presented By: Christina Bethell, Ph.D., MPH, MBA, Associate Professor, Department of Pediatrics, Oregon Health & Science University, Mailcode CDRC-P, 707 SW Gaines Street, Portland, OR 97239; Tel: (503)494-1892; Fax: (503)494-2475; Email: bethellc@ohsu.edu Research Objective: To compare findings on the prevalence and health and health care service need characteristics of children and youth with special health care needs (CYSHCN)identified using the same screening method (CSHCN Screener) in the 2001 National Survey of Children with Special Health Care Needs (NS-CSHCN), the 2000 Medical Expenditures Panel Survey (MEPS) and the 20032004 National Survey of Children's Health (NSCH). Study Design: Data from the NS-CSHCN, MEPS and NSCH are used to calculate the proportion of children meeting criteria for having one or more type of special health care need using the standardized CSHCN Screener. Child health and health care service needs as well as sociodemographic characteristics of CYSHCN are compared to further understand similarities and differences in the prevalence and profile of CYSHCN provided by each of these datasets. Several analyses consider whether differences observed may be due to variations across surveys in (1) sampling (random vs. oversampling; household vs. target child interview; English vs. Non-English language), (2) administration (survey introduction, order of items, incentives, survey length) and (3) weighting. Differences in shifts in practice patterns between survey years (2000 to 2004), such as increased use of prescription medication among children and higher rates of provider identification of children with mental or behavioral health problems (e.g. ADHD), are also considered. In addition, differences in characteristics of CYSHCN identified using alternative methods (CSHCN Screener vs. condition checklists) are compared to further understand the overall performance of the non-condition specific, consequencesbased CSHCN Screener across these three new national data sets. Population Studied: Children age 0-17 who met standardized parent-reported criteria for having a special health care need in the NS-CSHCN, MEPS and the NSCH. (N= 68,217) Principal Findings: The rate of CYSCHN in the 2001 NSCSHCN is 13.6 (13.4-13.9) vs. 16.2 (14.7-17.7) for 2000 MEPS for English language cases (MEPS English only) and 17.6% (17.2-18.0) for the 2004 NSCH. Once identified, CYSHCN identified in each survey do not vary dramatically in the health and health care service need consequences they experience. We hypothesize that higher rates of identification for the MEPS and NSCH vs. NS-CSHCN are due to variations in sampling and administration of this surveys and real shifts in practice patterns in recent years. We also hypothesize that significantly fewer children will be identified using the CSHCN Screener vs. a chronic condition check-list in both MEPS and the NSCH and that these children experience more functioning and health service need consequences compared to children only identified using a condition check-list. Conclusions: Important differences in the profile of CYSHCN provided across the new national data sets exist and involve differences in methodology as well as potentially real differences due to shifts in practice patterns and epidemiology of CYSHCN between survey years. Implications for Policy, Delivery, or Practice: Users of the new national data sets need to understand differences in methods used across surveys in order to validly interpret findings on CYSHCN emerging from the NS-CSHCN, MEPS and the soon to be released NSCH. Primary Funding Source: HRSA ●Access to Care for Adolescents with Special Health Care Needs Debra Lotstein, M.D., Nicole Garro, MA, Moira Inkelas, Ph.D. Presented By: Debra Lotstein, M.D., Assistant Professor, Pediatrics, UCLA, 10833 Le Conte Avenue 12-358 MDCC, Los Angeles, CA 90095; Tel: (310)825-9346; Email: dlotstein@mednet.ucla.edu Research Objective: Multiple studies of the general population have found that adolescents have poorer access to care than younger children. Studies of youth using various definitions of disability also suggest poorer access to care among teenagers with disabilities. It is unknown if this agerelated pattern applies to the new broadly defined population of children with special health care needs (CSHCN). The objective of this study is to determine if the prevalence and predictors of unmet health care needs differ for adolescents compared to younger children with special health care needs. Study Design: Data are from the 2001 National Survey of Children with Special Health Care Needs, a nationally representative telephone survey conducted by the CDC. Our main dependent variable is unmet need, defined as having at least one unmet need for a health service (from a list of 18 types) in the last 12 months. We compute the prevalence of unmet need by age group (0-1, 2-5, 6-11, and 12-17 years old), and compare the odds of having unmet need controlling for insurance coverage, type of health care need and condition severity. Next, we compare predictors of unmet need for health services using age-stratified logistic regression analyses, adjusting for predisposing, enabling and need factors associated with access to care in prior studies. 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: Prevalence of unmet need increases with age, reaching 18.9% for teens from about 16% for younger children. Even after controlling for structural and health need factors, adolescents have 1.2 times the odds of school-age children of having an unmet need (95% CI 1.06-1.36). Lack of insurance coverage, less stable health care needs and having a condition requiring emotional/behavioral or developmental services are common predictors of unmet need in all age groups. Yet among teenagers alone, increased odds of unmet need are found for females and non-white CSHCN. Conclusions: Adolescents with special health care needs have poorer access to health care compared with younger children, even accounting for their lower insurance rates and different types of health care need. Gender and race/ethnicity are of unique importance for teenagers’ access to care. Implications for Policy, Delivery, or Practice: Adolescents with special health care needs require enhanced services and supports to ensure access to comprehensive care. Creative clinical and policy solutions are needed to meet the unique needs of this adolescent population. Primary Funding Source: RWJF ●Physical Education and the Incidence of Overweight among Adolescents Chad Meyerhoefer, Ph.D., John Cawley, Ph.D., David Newhouse, Ph.D. Presented By: Chad Meyerhoefer, Ph.D., Economist, Center for Financing, Access and Cost Trends, Agency for Healthcare Research and Quality, 540 Gaither Road, Rockville, MD 20850; Tel: (301)427-1664; Fax: (301)427-1276; Email: cmeyerhoe@ahrq.gov Research Objective: To determine whether the time spent in physical education (PE) classes affects adolescent Body Mass Index (BMI) and clinical weight classification. Based on measurements taken during the National Health and Nutrition Examination Surveys, the incidence of overweight has grown rapidly among children and adolescents, increasing from 6.5 percent in the late 1970s to 15 percent in 1999-2000. Overweight status in childhood or adolescence has been linked to adult obesity and to subsequent adverse health impacts. There is now widespread recognition that the prevention and reduction of overweight among children and adolescents is an important public health objective. Among the changes under consideration by state policy makers is increasing primary and secondary school PE requirements. This research will indicate whether this policy holds promise for reducing overweight, as we test whether the time spent in PE classes impacts adolescent BMI and the probability of becoming overweight. Study Design: Height and weight data are used to calculate BMI for individuals in the Youth Risk Behavior Surveillance System (YRBSS) in order to investigate the relationship between PE classes and BMI. Multivariate regressions are used to determine the impact of minutes of exercise during PE classes on BMI, controlling for age, race, gender, level of urbanization, and consumption of certain food groups. Categorical regression methods are used to measure the impact of minutes of exercise during PE classes on the probability of being overweight or at risk of overweight. In addition, the impact of PE on different reported exercise levels is investigated to determine whether PE promotes more active lifestyles and a higher propensity to engage in physical activity outside of the school day. One concern is that minutes of exercise in PE may be correlated with unobserved socioeconomic characteristics of the school district, which may in turn be correlated with BMI; this would lead to omitted variable bias in our coefficients on minutes of exercise in PE. To address this possible bias, we estimate models of instrumental variables. Our instruments are the state requirements for PE. Population Studied: This study uses data from the 1999, 2001, and 2003 Youth Risk Behavior Surveillance System (YRBSS), a nationally representative survey of the U.S. high school student population. Principal Findings: Results from models of instrumental variables indicate that PE participation has a significant and negative impact on both BMI and the probability of being overweight. Our first-stage regressions in the IV models indicate that minutes of exercise in PE classes is strongly influenced by state-level PE requirements as well as the enforcement of those requirements. Conclusions: Raising the number of minutes that children are active in school PE classes would likely reduce BMI and is a promising method to reduce the prevalence of overweight among children and adolescents. Implications for Policy, Delivery, or Practice: Policymakers are searching for interventions to prevent and reduce overweight among adolescents. These results suggest that increasing the minutes of exercise during PE class time is one promising method. These results can also be used in conjunction with other information, such as the cost of increasing PE class offerings, to establish a cost-effective PE curriculum. Primary Funding Source: AHRQ ●Arkansas’s Response to Childhood Obesity: Year Two Assessment Joseph Thompson, M.D., MPH, James E. Bost, MS, Ph.D., Jennifer L. Shaw, MAP, MPH, Kevin W. Ryan, JD, MA Presented By: Joseph Thompson, M.D., MPH, Director, University of Arkansas for Medical Sciences, Arkansas Center for Health Improvement, 5800 West 10th Street, Suite 410, Little Rock, AR 72204; Tel: (501) 660-7551; Fax: (501) 660-7543; Email: thompsonjosephw@uams.edu Research Objective: The prevalence of obesity among U.S. children has reached epidemic proportions. To combat problems associated with childhood obesity, the Arkansas General Assembly passed Act 1220 in 2003 that includes a wide range of activities. Included in the Act is a requirement for all public school students to annually have their body mass index (BMI) for age assessed and reported to their parents. In the 2003-2004 school years, the Arkansas Center for Health Improvement (ACHI) undertook statewide implementation of BMI assessment and reporting for approximately 450,000 students. The continued BMI assessment for the 2004-2005 school year will yield comparable data to the first year’s statistics. Study Design: The data gathering and reporting on 421,973 students for the 2003-2004 school year was completed by ACHI in the summer of 2004. A 91% student match was found from 2003-2004 school year to the 2004-2005 school year. A comparative analysis of student and demographic data over the 2 school years will indicate improvement or regression. Population Studied: Arkansas school students for the past two years, and citizens of Arkansas are affected by Act 1220 through results and community-level changes may benefit societal efforts to combat obesity. Principal Findings: Based on BMI-for-age calculations, the 2003-2004 school year data showed that 21% of Arkansas school children were overweight, 17% were at risk for being overweight, 2% were underweight, and 60% were normal weight, using weight categories defined by the Centers for Disease Control and Prevention. The demographic distribution of results showed that both genders had similar percentages among weight classifications. African-American and Hispanic students had slightly higher percentages for overweight (24% and 26%, respectively) and at risk for overweight (17% and 20%, respectively) classifications than Caucasians (20% overweight, 17%, at risk for overweight). The results for the first statewide assessment in the nation indicate that Arkansas has a higher than estimated percentage of children and adolescents with weight problems than predicted by the NHANES sample of more than 7000 children. Conclusions: The Arkansas BMI assessment has revealed a more severe problem than anticipated, with the state having 38% of its children in the 2 highest risk weight categories. Implications for Policy, Delivery, or Practice: Analysis of the 2004-2005 school year data will be an initial indicator of trends associated with obesity and will enable longitudinal studies on factors that can influence the epidemic of obesity among our nation’s children. Primary Funding Source: RWJF ●User-Friendly Strategies to Expand the Use of new National Data Sets on Children's Health by Policymakers and Consumer and Health Care System Leaders Nora Wells, MEd, Christina Bethell, Ph.D., MBA, MPH, Debra Read, MPH Presented By: Nora Wells, M.Ed, Director of Research and Education, Federation of Children with Special Needs/Family Voices, 1135 Tremont Street, Boston, MA 21210; Tel: (617)2367210; Fax: (617)572-2094; Email: nwells@fcsn.org Research Objective: To develop and evaluate the use of an online data query tool to expand the use of the National Survey on Children with Special Health Care Needs (NSCSHCN)and the 2003-2004 National Survey on Children's Health among federal and state policymakers and program leaders and family and health care system leaders. Study Design: Federal and state policymakers and program administrators and family leaders across the US were engaged in a five phase process to assess interest, capacity and support needed to use data resulting from the NS-CSHCN and other similar data sets (n= 72 representing 36 states). Phases included (1) in-person focus groups, (2)a standardized self-reported survey, (3) in-depth structured interviews to identify preferred features and content for an online data query tool and resource center and (4) a workshop to further discuss needs and options for designing the Data Resource Center (DRC) and (5) an in-person website demonstration and evaluation session. Input obtained during ten in-person educational workshops on the DRC and the use of the DRC by over 12,000 users (65,000 hits)was evaluated to assess the usability and value of the DRC in expanding the use of national data sets by non-research audiences essential to engage in the use of data to inform and stimulate the design of health care programs and policies for children and youth. Population Studied: Federal and state Title V leaders in the area of children and youth with special health care needs (CYSHCN) and family CYSHCN leaders participating in the Family Voices consortium. (N = 72 for pretest; N = approx. 12,000 for beta test) Principal Findings: Over 85% of participants indicated that they were "extremely interested" in being able learn about and directly access and query national and state level child health and health care quality data sets. Over 83% indicated that they would be "extremely likley" to routinely use a Data Resource Center website designed to assist them in learning about and easily using this data. Over 60,000 hits to the DRC were recorded during the nine month start up period of the DRC. The vast majority of visits were to use the online interactive data query tool. Examples of uses of data findings derived from the DRC include grant applications, legislative testimony, popular press public education, stakeholder education and engagement and state block grant needs and performance assessments as well as many other applications. E-mail and telephone technical assistance are central to the success of any DRC-like effort as well as in-person and other more in-depth trainings on how to think about, query and communicate data findings in a scientifically valid and effective manner. Conclusions: State and federal policymakers and health care program leaders as well as family/consumer leaders have a strong interest in being able to directly access and validly use data findings from national and state surveys on child and adolescent health and health care. Implications for Policy, Delivery, or Practice: Without a DRC-like resource it is unlikely that essential policy, consumer and health system leaders will make use of the new national and state data on child and adolescent health and health care will. Design of any DRC-like resource must be developed with ongoing and in-depth input from target users. If US health care policy continues to emphasize consumer engagement it will be even more important to provide publicly available, nonproprietary DRC-like resources to assist consumers, policymakers and others to easily access data they would otherwise not have the skills, capacity or resources to use. Primary Funding Source: HRSA Related Posters Poster Session A Sunday, June 26 • 2:00 pm – 3:15 pm ●Parent Reports of Patient-Centered Communication with Pediatric Providers: Implications for Quality Improvement Melinda Abrams, MS, Susanne E Tanski, M.D., Peggy Auinger, MS, Michael Weitzman, M.D. Presented By: Melinda Abrams, MS, Senior Program Officer, The Commonwealth Fund, One East 75th Street, New York, NY 10021; Tel: (212) 606-3831; Fax: (212) 249-1276; Email: mka@cmwf.org Research Objective: To assess patient-centered communication between mothers of young children and their child’s primary care clinician in order to identify opportunities for improving the quality of preventive pediatric care. Study Design: Data from the 2001 and 2002 Child Preventive Health Supplement Section to the Medical Expenditures Panel Survey (MEPS) were analyzed to assess child and family characteristics associated with maternal reports of patientcentered communication, specifically how well the clinician listened, explained things, had respect for what the parent had to say and spent enough time with the child. A composite “overall communication” scale combining these four areas was created, with an answer of “always” = 3, “usually” = 2, “sometimes” = 1, and “never” = 0 for each area. A score of 06 was termed inadequate, 7-11 adequate and of 12 excellent. Maternal physical health and mental health were determined from the SF-12 within the MEPS. Bivariate and regression analyses were conducted. Population Studied: A nationally representative sample of 4450 parents of children under 5 years of age. Principal Findings: Estimates of internal consistency among the four communication questions was high (Cronbach’s a = .89). Overall, 58.7% of parents reported receiving “excellent” patient-centered communication and 6.5% “inadequate”. Poorer communication was associated with worse child health status, with 20.9% of parents of children with fair/poor health, 11.6% with good health and 6.3% with very good health and 4.7% with excellent health (p<.001) reporting inadequate communication. In multivariate regression, fair/poor child health was independently associated with having inadequate communication with the child’s practitioners, adjusted OR 5.9 (95% CI 2.7, 12.7), as was poor maternal mental health OR 2.3 (1.31, 4.05) and poor maternal physical health OR 1.8 (1.7, 2.94). There were no significant associations between communication and child’s age, race, ethnicity, health insurance, or single parent household. Conclusions: The majority of parents of young children report excellent patient-centered communication with their child’s pediatric clinician. However, mothers of children with poor health status report inadequate communication, as do mothers with poor physical or mental health. Implications for Policy, Delivery, or Practice: The association between poor health status and inadequate communication suggests opportunities for quality improvement. Screening children and parents for physical and mental health in advance of preventive care encounters could help clinicians identify families in need of additional time for anticipatory guidance and counseling. Such an approach could lead to variation in payment for preventive services based on family health status. Additional strategies to improve patient-centered communication include enhanced training of clinicians’ communications skills, development of materials to help parents communicate concerns and distribution of parent education materials about health care conditions and issues. To implement suggested changes requires restructuring preventive health care encounters at the parent, provider, practice and system levels. Primary Funding Source: CWF ●Association Between Remitted, Current and Chronic Maternal Depression and Behavior in 3- Year-Old Children Melissa Azur, MS, Philip Leaf, Ph.D., Anne Duggan, Sc.D. Presented By: Melissa Azur, MS, NIMH Predoctoral Fellow, Mental Health, Johns Hopkins Bloomberg School of Public Health, 624 North Broadway, 8th Floor, Baltimore, MD 21205; Tel: (410)719-6172; Email: mazur@jhsph.edu Research Objective: To investigate the association between remitted, current and chronic maternal depression and internalizing behavior problems, externalizing behavior problems and adaptive behavior in three-year-old children. Second, to investigate the role of maternal social support in moderating this association. Study Design: The data come from a randomized longitudinal study of an early home visitation program in Hawaii. Maternal depression was measured using the CES-D. Internalizing behavior and externalizing behavior were measured based on maternal reports from the CBCL, and adaptive behavior was measured using the Vineland Adaptive Behavior Scale. Logistic regression was used to assess the association between maternal depression and each of the child outcomes. Interactions between maternal depression and maternal social support were explored and the models controlled for poverty, maternal education, teen parenthood, race, and child’s gender. Population Studied: The study sample consisted of 583 families participating in the larger randomized longitudinal study. These families were identified as at risk for child abuse and child maltreatment at the time of entry into the original study. The study sample was limited to families where the mother completed the interview when the target child was 3 years old. Principal Findings: Maternal depression was highly prevalent with 35% of women reporting current or chronic symptoms of depression and an additional 23% of women reporting remitted depression. There was an increasing trend in the odds of having internalizing problems among children with remitted, currently depressed and chronically depressed mothers. Children living below the poverty line and children of teen mothers were even more likely to experience internalizing problems. Remitted maternal depression and chronic maternal depression were also associated with externalizing problems. Once we controlled for maternal social support and child’s gender, there was no association between maternal depression and adaptive behavior. There were no significant interactions between social support and maternal depression. Conclusions: Children with depressed mothers are at increased risk for internalizing and externalizing behavior problems and this risk continues after the depression has remitted. Children born to adolescent mothers and children living in poverty may be at particular risk of internalizing problems. Implications for Policy, Delivery, or Practice: Pediatricians should consider a family-centered approach in treating children and consider screening mothers for a history of depression. Children living in poverty and children born to adolescent mothers may be at particular risk for internalizing problems. If depression is detected, support and treatment should be provided to both the parent and the child. Primary Funding Source: Other Government Funding ●Physical Activity Among Adolescents: The Role of Physical Education Susan Babey, Ph.D., Allison L. Diamant, M.D., MSHS, Theresa Hastert, E. Richard Brown, Ph.D. Presented By: Susan Babey, Ph.D., Research Scientist, UCLA Center for Health Policy Research, 10911 Weyburn Avenue, Suite 300, Los Angeles, CA 90024; Tel: (310)794-6961; Fax: (310)794-2686; Email: sbabey@ucla.edu Research Objective: To estimate the prevalence of regular physical activity and physical inactivity and to examine the relationship between physical education (PE) requirements and physical activity among adolescents in California. Study Design: We used data from the 2001 and 2003 California Health Interview Survey (CHIS). CHIS, a randomdigit dial (RDD) telephone survey of households drawn from every county in California, completed interviews with over 55,000 households in 2001 and 42,000 households in 2003. Bivariate and multivariate analyses were used to examine the relationship between PE and level of physical activity. Other factors examined include age, gender, race/ethnicity, and income. Population Studied: We analyzed responses from 5,858 adolescents age 12-17 interviewed for CHIS 2001 and 4,010 adolescents interviewed for CHIS 2003. Principal Findings: In 2003, 7.3% of adolescents were physically inactive, a significant increase from 2001 when 5.2% were inactive (p<0.01). Inactivity among adolescent girls nearly doubled from 2001 (9.2% vs. 5.0%, p<0.001) whereas prevalence of inactivity among boys remained about the same. Latino, African American, and Asian adolescents were less active than white adolescents. Level of physical activity also varied with PE requirements. The prevalence of regular physical activity was higher among adolescents whose schools require PE (73.0%) than among those whose schools do not require or do not offer PE (59.4%, p<0.001); and the prevalence of physical inactivity was nearly twice as high among adolescents whose schools do not offer or do not require PE (11.8% vs. 6.2%, p<0.01). Furthermore, the prevalence of regular physical activity varied dramatically by race/ethnicity and PE requirements. Among adolescents whose schools do not require PE, two-thirds (67.9%) of white adolescents engage in regular physical activity compared with less than half of Asian adolescents (43.0%, p<0.05), and less than one-third of African American adolescents (29.0%, p<0.001). Although there was significant variation in physical activity by race/ethnicity among adolescents whose schools require PE, the differences were considerably smaller, ranging from 65.4% among African Americans to 78.3% among whites. PE requirements had a differential impact on boys’ and girls’ physical activity. Boys whose schools require PE have a lower prevalence of inactivity than girls (4.4% vs. 8.1%, p<0.01). However, there were no significant differences between boys’ and girls’ prevalence of physical inactivity when their schools did not offer or did not require PE (11.1% vs. 12.4%). Conclusions: These findings suggest that physical education requirements at school are an important factor related to adolescent’s participation in physical activity and may be a promising area for intervention efforts for groups that are less active. Implications for Policy, Delivery, or Practice: Physical inactivity contributes to complications and death from chronic conditions such as type 2 diabetes, coronary heart disease, obesity, and hypertension. Despite the well-documented benefits of physical activity, many California adolescents do not get regular physical activity or get no activity at all. Identification of factors, particularly environmental factors, related to physical activity among adolescents can aid in the development of effective interventions to increase physical activity and help adolescents establish a lifetime of healthy activity behaviors. Primary Funding Source: The California Endowment ●Correct Identification of Childhood Obesity and its Association with Reported Practice Behavior Sarah E. Barlow, M.D., MPH, Michael P. Elliott, Ph.D., Sonal Bobra, MPH Presented By: Sarah E. Barlow, M.D., MPH, Assistant Professor, Pediatrics, Saint Louis University, 1465 South Grand Boulevard, Saint Louis, MO 63104; Tel: (314) 577-5647; Fax: (314) 268-2775; Email: barlowse@slu.edu Research Objective: To determine how often pediatricians correctly identify overweight/obesity during health supervision visits of children 6 to 17 years. To assess the association of overweight/obesity identification with evaluation for obesityrelated medical conditions and with reported health behavior counseling. Study Design: Pediatricians at federally-funded health centers and a random sample in private practice in Saint Louis, Missouri were asked to provide information from up to 30 visits as part of a study of care of chronic conditions during health supervision visits of children 6 to 17 years. Participants returned a copy of the visit note and the growth chart and completed a questionnaire about patient demographics and visit evaluation and treatment. The pediatricians checked current conditions from a list that included overweight/obesity (OW/OB) and hypertension. For 7 health behaviors, pediatricians reported none, brief, or >/= 2 minutes counseling Population Studied: 55 pediatricians were contacted and 22 returned information from a total of 564 visits. 51.3% of patients were 6-11 years old, 50.1% were male, 47.2% were Caucasian, and 49.2 were African-American. 27.3% of visits occurred in federally-funded health centers, and 54.0% were covered by Medicaid. Principal Findings: Based on height and weight recorded in 554 visits, 17.1% of children were at risk of overweight (BMI 85th-94th percentile) and 22.4% were overweight or obese (BMI >/= 95th percentile). Pediatricians identified OW/OB in 26.1% of children in the at-risk category and 83.1% of children with BMI >/= 95th percentile. Reported physician behavior was examined in three visit categories: identified obesity (BMI >/=95th with OW/OB identified) n=98; missed obesity (BMI >/= 95th with no OW/OB identified) n=20; and normal weight (BMI < 85th percentile) n=335. Cholesterol was ordered in 36.1% identified obesity visits, 15.0% missed obesity visits, and 10.7% normal weight visits (identified vs. missed p=.05). Glucose was ordered in 16.4% identified obesity visits, 0 missed obesity visits and 2.2% of normal weight visits (identified vs. missed NS). Blood pressure was recorded in more than 95% of visits with no difference among the three groups. Pediatricians reported hypertension in 8.5% identified obesity visits, 0 missed obesity visits, and 2.0% normal weight visit (identified vs. missed NS). Diet counseling >/=2 minutes was more common among identified obesity visits compared with normal weight and missed obesity visits (81.3% versus 25.9% and 31.3%, p < .001). Similarly, exercise counseling >/= 2 minutes was more common among identified obesity visits (70.5% versus 20.0% and 24.4%, p < .001). Counseling for behavior management and for growth and development did not differ among the groups, nor did counseling for family planning, HIV/STD, and tobacco Conclusions: Pediatricians missed obesity in 16.9% of obese children. The higher rates of diet and exercise counseling and the trend toward higher cholesterol screening when obesity is identified suggest that improved recognition could improve practice patterns. Implications for Policy, Delivery, or Practice: Helping pediatricians identify all overweight children could lead to improved medical care and more intensive diet and exercise counseling. Further work is needed to assure the quality and effectiveness of the counseling. Primary Funding Source: AHRQ ●The Impact of Gaps in Health Insurance Coverage on Immunization Status for Young Children Lynn A. Blewett, Ph.D., Lynn A. Blewett, Ph.D., Gestur Davidson, Ph.D., Holly Rodin, MPA, Mark L. Messonnier, MS, Ph.D. Presented By: Lynn A. Blewett, Ph.D., Associate Professor, School of Public Health, University of Minnesota, 420 Delaware Street, MMC 729, Minneapolis, MN 55455; Tel: (612)626-4739; Email: blewe001@umn.edu Research Objective: To estimate the impact of gaps in public and private health insurance coverage on young children’s rates of immunization receipt. Study Design: We use multivariate logistic regression models to assess the impact of health insurance coverage on the probability of being up-to-date (UTD) on 10 recommended childhood immunizations. We look specifically at the impact of gaps in coverage by comparing full-year and part-year private and public (Medicaid/SCHIP) health insurance coverage on whether immunizations are UTD controlling for other factors that might affect immunization. Covariates include characteristics of both children and their parents (including employment status, income, race/ethnicity). We worked with CDC to obtain a unique data file that linked responses from 8,886 non-special needs children with insurance coverage data from the 2001 SLAITS/CSHCN with provider verified immunization status on age-eligible children from the 2000-2002 National Immunization Survey (NIS). Population Studied: A nationally representative sample of 8,886 children age 19-35 months. Principal Findings: We found that children with private-partyear coverage were less likely to be up-to-date (UTD) across all 10 immunizations and statistically significantly lower for 6 of the 10 immunizations compared to children with private-fullyear coverage. On average, children with gaps in private coverage were 9% less likely to be UTD than children with fullyear covereage. Children with private-full-year coverage were not statistically different in UTD status from children with public-part-year coverage for any of the immunizations, and in 4/10 immunizations their likelihood of being UTD was marginally higher. We found children with public-full-year coverage were significantly more likely to be UTD in 2/10 immunizations compared to children with private-full-year coverage. Children who were uninsured the whole year were not significantly different in their likelihood of being UTD for any of the immunizations, compared to children on privatefull-year coverage. Conclusions: This study provides important new information on the impact type and duration of health insurance coverage has on childhood immunization rates. Specifically gaps in private coverage increases the likelihood that children are not UTD in their immunizations. Even children with part-year public coverage do as well or better on obtaining recommended immunizations on time than children with gaps in private coverage. Gaps in private health insurance coverage may have an important impact in children delaying or foregoing immunizations. Implications for Policy, Delivery, or Practice: While immunization coverage among children is at an all-time high in the US, 20 percent or more of children between 19 and 35 months of age are still missing one or more recommended immunizations. Gaps in private health insurance coverage may be an important barrier to obtaining recommended immunizations. Policy strategies that enhance continuity of coverage for children may improve immunization coverage. In addition, steps should be taken to ensure that both families and providers understand that the current federal Vaccine for Children Program (VCP) provides for free vaccines to uninsured children including those that lose coverage during the year. Primary Funding Source: CDC ●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., RN, Carolyn Byrne, Ph.D., RN, Jacqueline Roberts, MSc, RN., Amiram Gafni, Ph.D., Susan Whittaker, MSc, BSc, BA Presented By: Gina Browne, Ph.D., RN, Founder & Director, System-Linked Research Unit on Health and Social Service Utilizaton; Professor, Nursing & C.E.&B., School of Nursing, McMaster University, Faculty of Health Sciences, 75 Frid Street, Building T30, Hamilton, Ontario, L8P4M3; Tel: (905) 525-9140 x22293; Fax: (905)528-5099; Email: 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, 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 five-arm prospective randomized trial of 765 single parents on social assistance and their 1330 children (aged 0 - 24 years). Principal Findings: Results from a two-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 exits 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 ●Nurse Staffing and Healthcare-Associated Bloodstream Infections in the Neonatal ICU Jeannie Cimiotti, DNS, Janet P. Haas, MS, Kevin D. Roberts, Ph.D., Elaine L. Larson, Ph.D. Presented By: Jeannie Cimiotti, DNS, Research Associate, School of Nursing, University of Pennsylvania, 420 Guardian Drive 327R NEB, Philadelphia, PA 19104-6096; Tel: (215) 8984989; Fax: (215) 573-2062; Email: jcimiott@nursing.upenn.edu Research Objective: Healthcare-associated infections (HAIs) are a cause of significant morbidity and mortality among infants in the neonatal intensive care unit (NICU). Few studies have examined the association between nurse staffing and HAIs in infants; furthermore, no prospective studies have examined the association between nurse staffing and bloodstream infections (BSIs) in the NICU. Study Design: This prospective correlational study examined the association between nurse staffing and healthcareassociated BSI among infants in two, high risk NICUs in New York City. An experienced epidemiologist collected data on BSIs, defined per CDC criteria. Four risk factors associated with BSI in infants were examined: birthweight, intravascular catheterization, major surgery, and total parenteral nutrition (TPN). Birthweight was stratified into four groups: 1) <1000gm, 2) 1000-1500 gm, 3) 1501-2500 gm, and 4) >2500gm. Catheters were classified as umbilical, tunneled (Hickman or Broviac), peripheral inserted central (PICC) and extracorporal oxygenation membrane (ECMO); TPN was defined as days of TPN to infection or discharge. Registered nurse (RN) hours were RN care hours per day, adjusted for case mix based on DRG. RN staffing data were ranked into quartiles with the first and fourth quartile indicating low and high staffing, respectively. Risk factors with a p-value <0.10 were included in a multivariate model and data were analyzed using a Weibull accelerated failure-time model to determine those covariates associated with BSI. Population Studied: Data were collected on all nurses who provided direct patient care and on all infants admitted to the NICU from March 2001 through January 2003. Principal Findings: During the study 2,675 infants were admitted to the NICU for >48 hours. Infants were 53% male and gestational age ranged from 23 to 42 weeks. A total of 309 (12%) infants had a birthweight <1000gm, 330 (12%) 10001500 gm, 812 (30%) 1501-2500 gm and 1,224 (46%) >2500 gm. Catheters were inserted in 1,603 infants: 783 (49%) were umbilical, 60 (4%) tunneled, 737 (46%) PICC and 23 (1%) ECMO. Additionally, 467 (17%) infants had surgery and 1,284 (48%) received TPN during their hospitalization. A total of 224 (8%) infants had an infection that met the study definition of BSI. In a multivariate analysis, the risk factors associated with BSI were birthweight <1000 (p<.0001) and 1000-1500 gm (p=<0.019), umbilical catheter days (p=0.02) and low hours of care provided by RNs (p<0.0001). Conclusions: Infants had an increased risk of BSI when RN care hours were low. These results are consistent with studies on nurse staffing in other populations and provide evidence of a causal link between nurse staffing and BSI. We hypothesize that inadequate nurse staffing results in poor hand hygiene compliance, breaks in aseptic technique or compromises in practice that might increase the risk of transmitting infection. Implications for Policy, Delivery, or Practice: These findings underscore the importance of active surveillance and adequate nurse staffing in healthcare facilities to potentially reduce HAIs, improve the quality of patient care and reduce the economic burden on our healthcare system. Primary Funding Source: NIH/NINR ●Factors Associated with Children’s Participation in Shared Decision Making Elizabeth Cox, M.D., MS, , , Maureen Smith, MD, MPH, PhD, Roger Brown, PhD, Mary Anne Fitzpatrick, PhD, , , , Presented By: Elizabeth Cox, M.D., MS, Clinical Assistant Professor, Population Health Sciences and The Center for Women's Health & Research, University of WisconsinMadison Medical School, 610 Walnut Street, 634 WARF Building, Madison, WI 53726; Tel: (608)263-9104; Fax: (608)263-2820; Email: ecox@wisc.edu Research Objective: Due to its ability to produce improved health outcomes in adults, shared decision-making (SDM) has become a preferred method for reaching treatment decisions with adults. Among adults, males, minorities, and those less educated are less likely to desire SDM and less likely to participate in their healthcare. No work has examined how such characteristics impact children’s participation in key SDM tasks. We examine how the child, parent and physician share visit talk with children of differing ages and how child and parent factors influence child participation in key SDM tasks (relationship building and information exchange). Study Design: We analyzed videotapes of children’s medical visits as well as physician/participant surveys of demographics, practice characteristics and healthcare utilization. Using the Roter Interaction Analysis System, utterances were coded and aggregated to reflect relationship building and information exchange (information giving and information gathering). Negative binomial models were used to analyze the association between the child’s participation in key SDM tasks and parent factors (education and gender of the accompanying parent) or child factors (age, gender, and race), adjusting for visit length and physician factors, as well as clustering by physician. Population Studied: 100 children and their parent(s) visiting 1 of 15 family physicians or pediatricians for acute concerns. Principal Findings: Children’s mean age was 5.4 years (sd 5.0, range 0-18). Physicians spoke 63% of visit talk, parents spoke 31% of visit talk and children spoke the remaining 6% of visit talk. The amount of physician talk did not change with child age, but parent talk decreased with child age, p<0.001, while child talk tended to increase, p<0.06. For relationship building talk, children of college graduates spoke 5 times more than children of less educated parents, while girls spoke 3 times more than boys, both p<0.02. Children gave 8% more information for each year of child age, p<0.05. Children of college graduates did twice as much information giving than children of less educated parents, p<0.05. Only 15 of the children gathered any information during the visit. Conclusions: In pediatric visits, children talked at the expense of parent talk while the amount of physician talk remained unchanged across child age. Increased child participation in key tasks of SDM is seen with higher parental education and among female children. Implications for Policy, Delivery, or Practice: Patterns of participation in key SDM tasks among children appear similar to those of adults, suggesting that such patterns begin early in life. Future work should examine how these patterns are established as a step toward identifying mutable factors that encourage patient participation in SDM. Primary Funding Source: AHRQ ●Explaining Disparities in Patient Centeredness Between Hispanic and Non-Hispanic Children Denise Dougherty, Ph.D., Frances Chevarley, Ph.D., Darryl Gray, M.D., Sc.D., Lisa Simpson, MB, BCh, MPH, Melissa Romaire, MPH, Marc Zodet, MS Presented By: Denise Dougherty, Ph.D., Senior Advisor, Child Health and Quality Improvement, Office of Extramural Research, Education, and Priority Populations, Agency for Healthcare Research and Quality, 540 Gaither Road, Room 2010, Rockville, MD 20850; Tel: (301) 427-1868; Fax: (301) 4271562; Email: ddougher@ahrq.gov Research Objective: To explain Hispanic versus NonHispanic disparities in the parent-provider communication experience using factors in addition to those used in previous studies of this topic (urbanicity, region, child’s country of birth, and CAHMI special health care need measure). Study Design: We used data from the 2002 Medical Expenditure Panel Survey (MEPS), a nationally representative survey, in a series of regression analyses designed to explain differences in parent-clinician communication between Hispanics and non-Hispanics. The outcome variables were two CAHPS® measures of communication: 1) whether the provider always listened carefully to the parent and 2) whether the provider always explained things clearly to the parent regarding the child’s health. Logistic regression examined whether disparities in communication between Hispanics and non-Hispanics persisted after controlling for individual, family, and contextual variables (age, race, health status, insurance status, usual source of care (USC); child born in the U.S., language of interview, education of interviewee, family income; region, urbanicity, and area density of practicing physicians). Population Studied: The study sample consisted of 11,097 children aged 0-17 for whom there was complete information from the Child Supplement section in the 2002 MEPS. Principal Findings: Unadjusted estimates suggest that Hispanic children were less likely than non-Hispanic children to have parents report that providers always listened carefully to them (68.7% v. 72.8%) and always explained things clearly to them (70.7% v. 76.5%). However, after controlling for other covariates using multiple logistic regression, the differences in reported communication between Hispanic children’s and non-Hispanic children’s parents were no longer significant for either measure [i.e., always listening carefully OR 0.89[0.71,1.11], and always explaining clearly OR 0.96 [0.76,1.22]). Living outside the Northeast, with reports of other than excellent child health status, and of being poor/near poor poverty status are associated with a decreased likelihood of reports that the provider always listened carefully or explained things clearly. Children residing in less populated “micro” areas (10,000-50,000 population) have an increased likelihood of positive responses on both measures of communication. Black children have an increased likelihood of reports that the provider always explained things clearly while multiple races and not having a USC are associated with a decreased likelihood of providers always listening carefully. For age, only <1 year is associated with a positive response for providers always listening carefully. Insurance status, language, born in the U.S., and education had no independent effect. Conclusions: In using CAHPS® communication-related questions as a measure of patient-centered care and a nationally representative survey, factors other than Hispanic ethnicity likely exert a strong influence on the quality of the child and parent health care experience. Implications for Policy, Delivery, or Practice: Initiatives to improve communication for Hispanics may be more similar to those for non-Hispanics than expected. However, our data are limited and unable to test a full range of plausible factors explaining disparities in parent-clinician communication. More research is needed to assess whether cultural competence programs for providers who treat Hispanic children will improve parents’ ratings of communications. To reach Hispanics and non-Hispanics, efforts to improve patient-centeredness might be focused outside the Northeast, in non-micro residential areas, and among poor families. Primary Funding Source: No Funding Source ●Social Cognitive, Behavioral, and Psychosocial Predictors of Young Children’s Oral HealthB Tracy Finlayson, BS, Kristine Siefert, Ph.D., MPH, MSW, Amid Ismail, BDS, MPH, DrPH Presented By: Tracy Finlayson, BS, Research Assistant, Health Management and Policy, University of Michigan School of Public Health, 109 South Observatory, Ann Arbor, MI 481092029; Tel: (734)615-5131; Email: tfinlays@umich.edu Research Objective: Dental caries is the most common disease affecting children and is disproportionately high among lower-income African-American children in Detroit. This study seeks to advance knowledge of the social determinants of oral health, by examining how several specific maternal health beliefs, behaviors, and psychosocial factors relate to young children's dental outcomes and practices in this population. Study Design: Survey and dental exam data are from the Detroit Dental Health Project (DDHP; NIDCR grant U-54 14261-01). Social Cognitive Theory guided the study design and selection of predictor variables. Health belief scales reflecting mothers’ self-efficacy, feelings of fatalism, attitudes about dental disease and knowledge about bottle use were constructed based on exploratory factor analysis of survey questions. Mothers’ brushing habits and psychosocial measures of depressive symptoms (CES-D), parenting stress, availability of instrumental social support and traditional background factors were also predictors. The outcomes were children’s brushing frequency, severity of Early Childhood Caries (ECC), and a subjective oral health rating. Descriptive and multivariable regression analyses were conducted with SUDAAN. Population Studied: A population-based sample of 719 African American children aged 1-5 and their mothers living in Detroit who participated in the DDHP in 2002-2003 were studied. 1-3 and 4-5 year olds were examined separately, to account for developmental differences. Principal Findings: The constructed health belief scales were reliable and internally consistent. Maternal ratings of child oral health status as fair/poor were quite congruent with whether or not the child had disease as determined by dental examinations. The regression analyses revealed that maternal self-efficacy, attitudes, fatalism, brushing behavior and social support were significantly associated with 1-3 year old children’s one-week brushing frequency. Efficacy, attitudes, brushing behavior and dental insurance coverage were significant in the model for 4-5 year old children. Among 4-5 year olds, having dental insurance, more education and less bottle knowledge reduced their likelihood of a fair/poor rating, while negative attitudes about disease and help with errands increased the likelihood. Higher parenting stress was inversely related to disease severity for all children. Fatalistic and negative disease attitudes increased the likelihood of having ECC for 4-5 year olds, while more education reduced the likelihood. Restorative dental treatment among 1-3 year olds increased the likelihood of ECC, while higher income reduced the likelihood. Conclusions: Although some findings are counterintuitive and warrant further exploration, mothers’ social cognitive and psychosocial characteristics are important factors that affect her child’s brushing frequency and dental health status in earliest childhood. Implications for Policy, Delivery, or Practice: Behaviors and cognitions are potentially modifiable, and this study’s findings can inform the design of tailored interventions. Such interventions could enhance mothers’ sense of efficacy and educate her about dental disease and caring for young children’s teeth, thus boosting her motivation to positively affect her child’s oral health. Mothers’ regular brushing could also be encouraged, to model a desirable preventive hygiene habit. Primary Funding Source: NIH NIDCR grant U-54 14261-01 ●Risk Factors for Failing Oral Amoxicillin Treatment of Severe Pneumonia in Children Linda Fu, M.D., Robin Ruthazer, MPH, Ira Wilson, M.D., M.Sc., Archana Patel, MB, DNB, MSCE, Donald Thea, M.D., Patricia Hibberd, M.D., Ph.D. Presented By: Linda Fu, M.D., Clinical Research Fellow, Institute for Clinical Research & Health Policy Studies, TuftsNew England Medical Center, 750 Washington Street Box 63, Boston, MA 02111; Tel: (617) 636-8028; Fax: (617) 636-5025; Email: lfu@tufts-nemc.org Research Objective: Oral amoxicillin was found to be equivalent in a recent trial to intramuscular penicillin, the currently recommended treatment for WHO-defined severe pneumonia (lower chest indrawing [LCI]) in children. Since oral therapy may eliminate the need for hospitalization, our goal was to determine which children were likely to fail oral amoxicillin treatment based on presenting characteristics and early response to therapy. Study Design: This is a retrospective cohort study within a previously completed randomized controlled trial conducted at tertiary care centers in 8 countries. Population Studied: Information was analyzed for the 857 participants ages 3-59 months who received oral amoxicillin for severe pneumonia. Patient information was obtained upon enrollment (T0), as well as 12 hours (T12) and 24 hours (T24) after the first dose of amoxicillin. Treatment failure was defined as persistence of LCI or development of danger signs after 48 hours of therapy. Principal Findings: Treatment failure occurred in 18%. Four multi-level multivariate logistic regression models, each including enrollment site as a random effect, were created to predict failure depending on the duration of observation and whether variables related to blood oxygen saturation were included, as pulse oximetry is not always available. In the first model, which assumed only 12 hours of observation and no access to pulse oximetry, infancy [age<6 months] (OR 4.0, 95% CI 2.4-6.6), very fast breathing [>69 breaths per minute if <12 months and >59 breaths per minute if >=12 months] (6.9, 2.9-16.5), and increasing respiratory rate (1.2, 1.1-1.3) predicted failure with an ROC area of 0.68. In the second model, which assumed 12 hours of observation with access to pulse oximetry, infancy (3.9, 2.2-6.8), very fast breathing at T0 (6.3, 2.4-16.4), low blood oxygen saturation at T0 [<90% in room air for all sites at sea level and <88% for Columbia and Mexico] (4.0, 1.8-8.9), increasing respiratory rate (1.1, 1.0-2.9) and decreasing oxygen saturation (1.2, 1.1-1.2) predicted failure with an ROC area of 0.72. In the third model, which assumed 24 hours of observation and no pulse oximetry, infancy (4.5, 2.6-7.7), very fast breathing at T0 (12.2, 4.8-31.2), higher temperature at T0 (1.8, 1.3-2.6), increasing respiratory rate (1.3, 1.0-1.4), increasing temperature (2.0, 1.4-2.7) and wheezing at T24 (2.2, 1.2-3.9) predicted failure with an ROC area of 0.73. In the fourth model which assumed 24 hours of observation with access to pulse oximetry, infancy (3.9, 2.3-6.6), very fast breathing at T0 (11.2, 4.5-27.6), low oxygen saturation at T0 (3.8, 1.8-8.1), increasing respiratory rate (1.3, 1.2-1.4), decreasing oxygen saturation (1.2, 1.1-1.3) and wheezing at T24 (2.1, 1.2-3.6) predicted failure with an ROC area of 0.76. Conclusions: Age and routine vital signs obtained during the first 24 hours of treatment of severe pneumonia with oral amoxicillin can be used to determine with reasonable accuracy which children are at risk for treatment failure. Data from pulse oximetry is needed to achieve similar predictive ability after only 12 hours of observation. Implications for Policy, Delivery, or Practice: These findings may assist with decision-making about the need for hospitalization or duration of observation, but need to be evaluated in further studies. Primary Funding Source: AHRQ ●Parents’ and Guardians' Influence on Children's Use of Mental Health Care Darrell J. Gaskin, Ph.D., Anthony Kouzis, Ph.D. Presented By: Darrell J. Gaskin, Ph.D., Associate Professor, Health Policy and Management, Johns Hopkins Bloomberg School of Public Health, 624 North Broadway Room 441, Baltimore, MD 21205; Tel: (443)287-5297; Fax: (410)614-8964; Email: dgaskin@jhsph.edu Research Objective: While 16 to 22% of children have a diagnosable mental/addictive disorder during a year, most do not seek or receive treatment. This study attempted to understand how parent’s/guardian’s characteristics influenced children’s use of mental health services. We were interested in how four factors influenced children's use of mental health services: parents’/guardians’ mental health status and general health status, family structure, family’s dependency on government assistance and socio-economic status. Study Design: This is a cross sectional analysis. We used logistic regression models to predict for children the odds of having a mental health visit and the odds of having a needed mental health visit delayed/postponed. We used a negative binomial regression to predict the number of visits during the year. Population Studied: We examined the mental health service use of 20,000 children whose families participated in the 1997 National Survey of American Families. Principal Findings: We found that parents’/guardians’ mental health status seemed to have a paradoxical effect on children’s use of mental health services. Having a parent/guardian with poor mental health status was associated with increased mental health service use. However, it also was associated with a greater likelihood of delaying/postponing needed mental health services. A possible explanation is that parents with poor mental health status were more likely to use mental health services for their children. However, parent’s poor mental health status may have been an access barrier because it made in more difficult for parents to navigate the health care system. We also found that children in blended, single-parent, or foster families were at greater risk of using mental health services. Financial factors increased the odds of using a mental health service and the number of mental health visits during the year but did not affect the odds of a needed mental health service being delayed or postponed. Conclusions: The health status of parents and guardian influences their children's access to care. Any effort to improve children's access to mental health services is incomplete if it does not address the health needs of their parents and guardians. Implications for Policy, Delivery, or Practice: Advocates for expanding access to care for children should not forget about their parents and guardians. Also, future research should explore why children in blended, single-parent, or foster families were at greater risk of using mental health services, and in particular, why those in blended and single-parent families were more likely to experience barriers to care. Primary Funding Source: NIMH ●Population-Based Volumes, Rates and Resource Use for Pediatric Cardiac Procedures Darryl Gray, M.D., Sc.D. Presented By: Darryl Gray, M.D., Sc.D., Medical Officer, Center for Quality Improvement and Patient Safety, Agency for Healthcare Research and Quality, 540 Gaither Road 3rd floor, Rockville, MD 20850; Tel: (301) 427-1326; Fax: (301) 427-1341; Email: dgray@ahrq.gov Research Objective: Congenital malformations of the heart and great vessels affect roughly one million Americans. Acquired conditions (generally of infectious origin) add to the growing burden of pediatric heart disease seen in children and adults. While relevant diagnostic and therapeutic options are expanding, little is known in aggregate about the resulting frequencies of procedures used to treat these conditions. Administrative data were used to estimate population-based volumes, rates, and resource use for inpatient cases, which have been estimated to comprise ~90% of pediatric cardiac interventional procedures Study Design: For a retrospective descriptive study, International Classification of Diseases-Clinical Modification (Ninth Revision) principal procedure codes for therapeutic (rather than purely diagnostic) cardiac procedures performed via surgical or transcatheter approaches were determined. Publicly-available aggregate data on inpatient admissions were obtained from the Kids' Inpatient Database (KID) and the National Inpatient Sample (NIS) of the Agency for Healthcare Research and Quality‘s Healthcare Cost and Utilization Project (HCUP). Hospital charges, which exceed costs but do not include physician fees, were expressed in 2002 dollars based on the Consumer Price Index. Volume figures were combined with Census data to generate population-based rates. Population Studied: Weighted, nationally representative samples of admissions of 0-17 year olds (YOs) Principal Findings: KID data from 1997 indicate that infants (i.e., patients < 1 year old) accounted for 15,024 admissions (384.6/100,000 infants). A mean length of stay (LOS) of 23.0 days generated charges of $147,091/admission. Among 1-17 YOs, KID estimated that there were 14,607 cases (21.9/100,000), where mean LOS=6.1 days and charges averaged $61,458/admission. For 0-17 YOs, aggregate charges (discharge volumes X mean hospital charges) were $3.1 billion. For that year, NIS projected 12,827 infant admissions, an LOS of 21.9 days and charges averaging $128,359/case. For 1-17 YOs, 14,067 admissions averaging 6.4 days generated charges of $48,984/case. Aggregate charges for 0-17 YOs were $ 2.3 billion. The mean LOS figures and the relative contributions of infants vs older children to admission volumes and to charges remained relatively stable over time and across databases. However, for all ages, NIS absolute volumes for 1998-2002 rose from 20,657 in 1998 to 32,368 in 1999. Subsequent volumes and variation decreased, with 24,691 cases estimated for 2002. Aggregate charges were $1.6 billion in 1998, $3.0 billion in 1999 and $3.1 billion in 2002. In 2000, NIS and KID volumes were 23,400 and 30,861 respectively. However, adjusted aggregate charges were similar at $2.3 billion versus $2.4 billion. Conclusions: Observed pediatric cardiac procedure admission volumes and resource use exceed prior estimates. Infants (including newborns) account for roughly ½ the admissions and >2/3 of the charges. While NIS appears annually, its estimates may fluctuate. KID (released for 1997 and 2000; 2003 pending) may better reflect pediatrics care. Implications for Policy, Delivery, or Practice: The considerable volumes and costs associated with pediatric cardiac prccedures indicate the societal importance of these interventions, and argue for increased population-based research on care patterns and outcomes, especially for infants. Such research should use data from HCUP and other sources to distinguish the effects of sampling variation or artifact from those of major changes in actual procedure volume and/or resource use. Primary Funding Source: National Heart, Lung and Blood Insttiute ●Is Cost-sharing Associated with Seeing Out-of-Network Physicians among Children with Diabetes? David Grembowski, Ph.D., William Trejo, BS, Paula Diehr, PhD, Frederick Connell, M.D., MPH, James Stout, M.D., MPH Presented By: David Grembowski, Ph.D., Professor, Health Services, University of Washington, Box 357660, Seattle, WA 98195; Tel: (206) 616-2921; Fax: (206) 543-3964; Email: grem@u.washington.edu Research Objective: To determine the association between out-of-network cost-sharing and the likelihood of seeing an out-of-network physician among insured children with diabetes. The single, common element across all types of managed care organizations, or MCOs, is a provider network. To control their costs, MCOs often build networks composed of physicians with lower fees and/or lower-cost practice styles, and then route enrollees to network physicians through greater out-of-network cost-sharing. Because children with chronic conditions have greater needs and see multiple physicians, it is unclear whether greater out-of-network costsharing reduces the likelihood of seeing an out-of-network physician, and whether this matters in terms of expenditures and quality of care. Study Design: Cross-sectional cohort study. Population Studied: 225,748 children/adolescents with 2 years of continuous coverage, from July 1999 to June 2001, in two insurance companies in Washington state. Of these, 476 children met eligibility criteria for diabetes: at least one diagnosis for diabetes or 2+ insulin prescriptions. Costsharing was measured by out-of-network and in-network costsharing indexes for the child’s health plan. Other data sources include medical claims, physician network directories, enrollee eligibility information, American Medical Association Masterfile, and sociodemographic characteristics of the child’s Zip code tabulation area. Principal Findings: 224 children had diabetes both years, who are prevalence cases, and 217 children had diabetes only in Year 2, or incidence cases. Average age was 12.6 years. About 49% were female, and 19% had other chronic conditions. Children saw an average of 4.17 physicians, excluding anesthesiologists, radiologists and emergency room physicians; about 43% had emergency room visits and 26% were hospitalized. About 18% of children saw an out-ofnetwork physician, and these children saw an average of 1.06 out-of-network physicians. About 59% of the out-of-network physicians had primary care specialties, 24% had endocrinology, diabetes, neurology, ophthalmology, or podiatry specialties, and 13% had other specialties. Regression analyses revealed that the cost-sharing indexes were not associated significantly with seeing an out-of-network physician. Age, gender, incidence vs. prevalence, and chronic conditions also were not associated with seeing out-ofnetwork physicians. HMOs have greater cost-sharing for outof-network care and smaller provider networks, and we also found that use of out-of-network physicians was similar for children in HMOs vs. other managed plans. Allowed ambulatory expenditures in the 2-year period were higher for children seeing out-of-network physicians, avg $3832, than children seeing only in-network physicians, avg $2966; p = .009. Conclusions: Cost-sharing is not associated with seeing an out-of-network physician among children with diabetes in health plans with provider networks. Implications for Policy, Delivery, or Practice: For children with diabetes, cost-sharing does not appear to be a barrier to seeing out-of-network physicians. Out-of-network utilization may be relatively high because of low access to quality care from in-network physicians, and we are examining the quality of care for children seeing only in-network physicians vs. children seeing both in-network and out-of-network physicians. Primary Funding Source: AHRQ ●Direct Medical Costs of Autism in a Privately Insured Population – 2002 Scott D. Grosse, Ph.D., Tom T. Shimabukuro, M.D., MPH, MBA Presented By: Scott D. Grosse, Ph.D., Health Economist, National Center on Birth Defects and Developmental Disabilities, Centers for Disease Control and Prevention, 1600 Clifton Road, Mail Stop E-87, Atlanta, GA 30333; Tel: (404)4983074; Fax: (404)498-3070; Email: SGrosse@cdc.gov Research Objective: Autism spectrum disorder (ASD) is a group of childhood-onset neurobehavioral disorders that is being diagnosed in the U.S. with increasing frequency. We calculated (1) the magnitude of incremental direct medical costs of ASD, (2) the ratio of direct medical costs for individuals with ASD versus those without ASD, and (3) the structure of the individual components: costs for inpatient, outpatient, and pharmaceutical services. Study Design: We used data from the 2002 Medstat MarketScan® Research Databases. The databases included claims for a diverse group of enrollees from large self-insured U.S. companies. We used International Classification of Diseases, Ninth Revision diagnosis codes 299.0x and 299.8x to identify individuals with ASD. We calculated incremental direct medical costs by subtracting the average direct medical costs for enrollees without ASD from that of those with ASD (i.e., the difference in average direct medical costs); we calculated cost ratios by dividing average direct medical costs for those with ASD by those without ASD; and finally we calculated the individual component costs for inpatient, outpatient, and pharmaceutical services as percentages of total direct medical costs. Population Studied: Privately insured children and adolescents 1-19 years old in the 2002 MarketScan® Research Databases and enrolled in the same health plan for the entire year. Principal Findings: Of 773,653 enrollees aged 1-19 years, 1,336 had a diagnosis of ASD. Average direct medical costs for 2002 were as follows: (a) 1-4-year olds: ASD $6,000, no ASD $1,200, incremental cost of ASD $4,800; (b) 5-9-year olds: ASD $5,200, no ASD $900, incremental cost of ASD $4,300; (c) 10-14-year olds: ASD $5,700, no ASD $1,100, incremental cost of ASD $4,600, (d) 15-19-year olds: ASD $6,700, no ASD $1,500, incremental cost of ASD $5,200. When comparing those with ASD to those without ASD, on average 1-4-year olds with ASD incurred 5.0 times more direct medical costs than those without ASD, 5-9-year olds incurred 5.8 times more, 10-14-year olds 5.2 times more, and 15-19-year olds 4.5 times more. For individuals with ASD, outpatient costs as a percentage of total direct medical costs decreased from 86% in 1-4-year olds to 35% in 15-19-year olds while pharmaceutical costs as a percentage of total direct medical costs increased from 7% in 1-4-year olds to 37% in 15-19-year olds. Individuals without ASD exhibited a similar but smaller trend for outpatient costs while drug costs did not change appreciably. Conclusions: Children and adolescents with ASD incurred greater direct medical costs than children without ASD, with the incremental costs ranging from $4,300 to $5,200. On average, individuals with ASD incurred 4.5 to 5.8 times more direct medical costs than those without ASD. In addition, for individuals with ASD, component costs appeared to shift from outpatient services in younger age groups to pharmaceuticals in older age groups, suggesting that individuals with ASD require increasingly intense pharmacotherapy to control behavioral symptoms as they grow older. Implications for Policy, Delivery, or Practice: ASD imposes a substantial economic burden on the health system with pharmaceutical costs appearing to becoming more prominent with age. Primary Funding Source: No Funding Source ●The Impact of Potentially Avoidable Post-Surgical Complications on Pediatric Patient Outcomes Samuel Hohmann, Ph.D. Presented By: Samuel Hohmann, Ph.D., Senior Research Specialist, Information Architecture, University HealthSystem Consortium, 2001 Spring Road, Suite 700, Oak Brook, IL 60523; Tel: (630) 954-1740; Fax: (630) 954-4730; Email: hohmann@uhc.edu Research Objective: Demonstrate the magnitude of impact that post-surgical complications have on pediatric patient outcomes (length of stay, in-hospital mortality, ICU utilization, cost and use of particular services). Study Design: Retrospective review of hospital inpatient discharge billing abstracts (UB-92). Patient outcomes of cohorts of hospital pediatric inpatients who had no documented and coded potentially avoidable complications were compared with outcomes of cohorts having complications. The University HealthSystem Consortium (UHC) uses 25 categroies of potentially avoidable complications, and cases are flagged based on the presence of the defining ICD-9 codes in the discharge abstract. Only cases who had a surgical procedure within 48 hours of admission are including the process of flagging cases for complications. Complications associated with patients having their first procedure later in their stays may not be the result of the procedure and are therefore excluded from the risk pool. Cases were stratified by UHC-defined product lines, and further stratification was performed by DRG. Population Studied: Pediatric patients having an operating room procedure within 48 hours of admission to an academic medical center (AMC). The study included discharges between July 1, 2003, and June 30, 2004. Of 284,682 pediatric discharges in the UHC database in that period (including newborns), there were 58,577 in the post procedure risk pool. The latter were used for comparison of outcomes with and without complications. Principal Findings: Overall, there were 6,651 discharges that had at least one potentially avoidable complication documented and coded in their discharge billing abstracts. This amounted to an overall complication rate of 11.4 percent. Nine UHC-defined product lines accounted for 77.7 percent of the risk pool cases and 84.9 percent of the cases with complications. The highest complication rates were among ventilator support cases (52.2 percent), cardiothoracic surgery cases (25.3 percent), and trauma cases (24.3 percent). The most frequently occurring complication categories were postoperative pulmonary compromise, post-procedure hemorrhage or hematoma, mechanical complications due to implants, and post-operative pneumonia. (Iatrogenic potentially complications, a less specific category, were excluded from the analysis.) The impact was generally a doubling or more of the outcome measure - average length of stay (ALOS), in-hospital mortality (in percent), ICU average length of stay, etc. For example, among cardiothoracic surgery discharges, ALOS was 6.3 days without complications vs. 16.5 days with complications; cases using ICU - 63.6 percent vs. 79.7 percent; ICU ALOS - 4.5 days vs. 11.7 days; in-hospital mortality - 1.3 percent vs. 6.6 percent. Although cardiothoracic surgery is a more homogenous group than all cases combined, the case mix index (CMI) for the two groups was different - 4.4 without complications vs. 5.0 with compications. CMI was not a factor at the DRG level. An example at the DRG level was DRG 105 (cardiac valve procedures without catheterization). For DRG 105, ALOS without complications was 6.3 days vs. 13.8 days with complicaitons; ICU use was higher for cases with complications (84.7 percent vs. 74.7 percent) and ICU ALOS was longer (11.1 days vs. 3.4 days); mortality was 2.4 percent vs. 10.6 percent. Other analysis demonstrated similar impacts on overall cost and cost of particular services (accommodations, medical/surgical supplies, laboratory, blood, pharmacy, etc.). Conclusions: Complications increase the duration of stays, having an impact on patients and allocation of institutional resources. Interventions should focus on efforts to reduce/prevent post-surgical complicaitons in pediatric populations. Performance improvement teams should review cases with documented and coded potentially avoidable complications and strategize on prevention and management when complications occur. Implications for Policy, Delivery, or Practice: Ascertaining the extent of added cost, resource utilization, degradation of patient safety and patient satisfaction, etc. should drive budget decisions on allocating resources to prevent potentially avoidable post procedure complications. Hospital administrators should embrace performance improvement initiatives to address patient safety, patient satisfaction, cost and resource allocation. Primary Funding Source: No Funding Source logistic regression analysis, adjusting for demographics and months of enrollment in Medicaid prior to the fatal event. Population Studied: All Medicaid children under age 10 and victims of homicide as identified through the cause of death documented in the DC, were included in the analysis (n=138). Principal Findings: The yearly average of violence induced fatalities (VIF) during the study period was 27.6 deaths. The majority of decedents had no prior contact for injury-related services (70%) and 86% were between ages 0 and 4 years old. Fatal child abuse, fatal assault, and death by firearms were listed as the cause of death on the death certificate respectively for 64 (46%), 51 (37%), and 23 (17%) of the victims. Children age 0-4 were more likely than their older counterparts not to have had previous contact with the health care system for injury-related services after controlling for demographics and length of enrollment in Medicaid prior to the fatal event (adjusted odds ratio (AOR): 4.11; 95% Confidence Interval (CI) = 1.03-16.5). Those residing in Metropolitan, Appalachian, and Suburban counties were 14.8 (95% CI = 3.2-69.1) times more likely not to have received injury-related services when compared to those living in rural counties. Conclusions: We estimated that a rather small proportion (30%) of children may have been in contact with the health care system prior to the fatal event for injury-related services – a statistic that translates into a limited window of opportunity for health care providers to prevent such events from happening. Implications for Policy, Delivery, or Practice: The analysis of patterns of use of injury-related services in the period preceding the fatal event may be used to develop a riskassessment tool for health practitioners to identify children at risk for becoming victims of homicide, and hopefully prevent VIFs. Although our study indicated that a rather small proportion of children have any contact with the health care system for injury-related services, this finding should by no means impede further research efforts in this direction, as all such deaths are preventable. Primary Funding Source: Ohio Department of Public Safety ●Children Victims of Homicide are Unlikely to Use InjuryRelated Health Services Prior to the Fatal Event Angelique Stubblefield, BSN, MPH, Siran Koroukian, Ph.D. ●Georgia's Foster Care Children and Medicaid Glenn Landers, MBA, MHA, Mei Zhou, MS, MA Presented By: Siran Koroukian, Ph.D, Professor, Epidemiology and Biostatistics, Case Western Reserve University, School of Medicine, 10900 Euclid Avenue, Cleveland, OH 44106-4945; Tel: (216)368-3197; Fax: (216)3683970; Email: skoroukian@case.edu Research Objective: Numerous studies have reported increased incidence of violence-induced fatalities (VIF) among socio-economically disadvantaged subgroups of the population, especially among children. However, the extent to which victims of violence could be identified through their pattern of utilization of health services is unknown. This study aims at describing patterns of use of injury-related services by Ohio Medicaid children victims of homicide, in the period preceding the fatal event. Study Design: This was a retrospective study using linked Ohio Medicaid and death certificate (DC) files, 1992-1996. Bivariate associations were tested using chi-square tests to assess statistical significance in the difference between proportions. Odds ratios were obtained using multivariable Presented By: Glenn Landers, MBA, MHA, Senior Research Associate, Georgia Health Policy Center, 14 Marietta Street, Atlanta, GA 30303; Tel: (404)463-9562; Fax: (404)651-3147; Email: glanders@gsu.edu Research Objective: Compare the conditions, Medicaid utilization, and costs of Georgia's foster child children with those of other Medicaid children. Study Design: Retrospective cohort analysis; logistic regression of Georgia Medicaid claims data. Population Studied: All Georgia foster children and other Medicaid children with three years continuous enrollment for the years 2000, 2001, and 2002. Principal Findings: 1. Georgia foster children and other Medicaid children have higher rates of mental health conditions than reported nationally - 60 and 35 percent respectively. 2. Fewer Georgia foster children experience inpatient admissions, ER visits, and prescription drug claims, and more experience dental visits, EPSDT screenings, and outpatient visits than other Medicaid children - all contrary to national reporting. 3. Of those children who utilize services, Georgia foster children experience more outpatient visits, EPSDT screenings, dental visits, and prescription drug utilization than other Mediciad children. 4. About equal numbers of Georgia foster children and other Medicaid children visited physicians, but more foster children visited specialists. 5. More foster children showed evidence of lapses in prescription drug use for mental health conditions over three years, and some children in each group showed evidence of mental health prescription drug use without an associated mental health diagnosis. Conclusions: The results for Georgia foster children described in this analysis, in comparison to the general Medicaid child population, are encouraging in many areas, particularly in preventive care. Implications for Policy, Delivery, or Practice: Policy makers may be concerned with the high proportions of mental health diagnoses, lapses in prescription drug claims for mental health conditions, and the presence of mental health drug claims in the Medicaid data without an associated mental health diagnosis. While the proportion of mental health conditions in the foster child population is not surprising, it indicates that any reduction in mental health services to foster children might place foster children at even greater risk than they are currently. Understanding why the foster child population uses preventive services more frequently may benefit the overall population of Medicaid children. Primary Funding Source: Georgia Health Foundation ●Social Inequalities in Perinatal Mortality in Belo Horizonte, Brazil: The Role of Hospital Care Sonia Lansky, M.D., MPH, Ichiro Kawachi, M.D., Ph.D., Elisabeth Franca, M.D., Ph.D. Presented By: Sonia Lansky, M.D., MPH, Research Fellow, Society,Human Development and Health, Harvard School of Public Health, 26 Harding Avenue, Belmont, MA 02478; Tel: (617)4893801; Email: slansky@hsph.harvard.edu Research Objective: Few studies have examined socioeconomic inequalities in perinatal mortality in Brazil, a country with high infant mortality rates (33 per 1,000 live births). The majority of births (97%) and perinatal deaths take place in hospitals, most of the deaths occurring within the first hours after birth. Yet, quality hospital care has not been assessed. As there are marked socio-economic differences in the types of facilities used by mothers, this study’s main goal is to analyze perinatal outcomes between hospital categories in relation to the Public Unified Health System (SUS), which assists almost 80% of the population in the country. Study Design: We examined all live births and perinatal deaths that occurred in Belo Horizonte City in 1999. Mortality surveillance was carried out in 36 hospitals. Data were collected by hospital chart review and linkage of individual records to the National Live Birth Information System and the National Death Information System, yielding 775 perinatal deaths. Information gathered included maternal education, birth weight, cause of death (Wigglesworth classification), and type of hospital (public, philanthropic, private hospitals contracting their services to SUS versus private non-SUS hospitals). Birth-weight specific mortality rates were compared among hospital categories adjusted for maternal education. Antepartum deaths (before the onset of labor) were excluded, as hospital care could not change birth outcomes in these cases. Population Studied: All 826 perinatal deaths (fetal and early neonatal death) and 40,531 live births of Belo Horizonte residents were assessed. Fetal deaths were stillbirth with birth weight of 500 g or more and/or gestational of 22 weeks or more. Early neonatal deaths were all infant deaths up to 7 days of life, birth weight of 500 g or more and/or gestational age of 22 weeks or more. Principal Findings: Although public hospitals had the highest proportion of low birth-weight babies (20.8%), this group of hospitals showed the lowest birth-specific mortality rates among SUS-hospitals. On the other hand, private-SUS hospitals had the highest mortality rate for each birth weight category, despite having the lowest proportion of low birthweight babies (5.9%). Perinatal mortality rates for babies with normal birth weight were high (2.8 to 5.2/1000) for hospitals contracted to SUS, contrasting to the rate for private hospitals (1.2/1000). After controlling for maternal education and birth weight, private-SUS hospitals showed the highest mortality rates in every situation. Intrapartum asphyxia predominated as the cause of death in private-SUS hospitals. Conclusions: Outcomes for SUS hospitals were worse compared to private non-SUS hospitals, reflecting their inadequate capacity to intervene during labor/child birth (and child care after birth) for low birth-weight and premature babies as well as adequate birth-weight children. Our findings illustrate the inverse equity hypothesis: child health inequities increase with the greater access to medical technology by those of higher socio-economic status. Implications for Policy, Delivery, or Practice: In Brazil more than 50% of the hospitals that constitute the Unified Public Health System are privately contracted to SUS, and account for the majority of births occurring to mothers with low socioeconomic status. Complications of delivery and childbirth are expected – but not predictable – in nearly 15% of all childbirths, even for low risk pregnancies. This means that every setting should be prepared to respond adequately in situations such as perinatal intra-partum asphyxia, which can be controlled by appropriate and timely health care. This study demonstrates the urgent necessity to improve hospital health care that could reduce a significant number of deaths considered preventable - especially in the private-SUS hospitals that showed the worst perinatal outcomes. Primary Funding Source: PAHO/WHO ●Preventable Perinatal Deaths in Belo Horizonte, Brazil, 1999: a Hospital Care Evaluation Sonia Lansky, M.D., MPH, Maria do Carmo Leal, Ph.D., Cibele Comini Cesar, Ph.D., Luiz Costa Monteiro Neto, BcH, Elisabeth Franca, Ph.D. Presented By: Sonia Lansky, M.D., MPH, Research Fellow, Society,Human Development and Health, Harvard School of Public Health, 26 Harding Avenue, Belmont, MA 02478; Tel: (617)4893801; Email: slansky@hsph.harvard.edu Research Objective: Evaluation of perinatal health care, particularly delivery and birth assistance, is necessary to reduce perinatal morbidity and mortality rates in Brazil. Most births take place at hospitals (97%) and with doctors’ assistance (77%), a context that should provide a safe environment for women and children. Nonetheless, perinatal mortality rates and maternal mortality are still high in the country. Although the majority of deaths take place within hospitals, few studies have been conducted to examine the role of hospital health care in reducing perinatal mortality. The main purpose of this study is to analyze the possible association between perinatal mortality and factors related to hospital care during birth, since most of these deaths can be prevented by prompt access to quality health care. Study Design: A population-based case-control was conducted, considering all births (40,351) and all perinatal deaths (826) in 1999 in Belo Horizonte, the third largest city in Brazil with a population of 2 million. Multiple regression analysis was conducted to examine association between selected medical procedures during birth as well as the level of hospital, adjusted by maternal and newborn potential confounders. Population Studied: 118 perinatal deaths (cases) and 492 births (controls) of residents in the city took part in the study. Cases and controls were recruited in hospitals participating in the Unified Health System (SUS) that provides health care for around 70-80% of the population in the country. Data were collected by hospital chart review and linkage of individual records to the National Live Birth Information System and the National Death Information System. Principal Findings: Bivariate analysis showed that male sex, prematurity, pathologies during pregnancy, low birth weight, newborn pathologies, lack of prenatal care, no use of the partograph during labor, and less than one fetus assessment per hour during labor were associated (p< 0,05) with perinatal death. Multiple regression analysis demonstrated that no use of the partograph during labor (adjusted OR 2,79) and low quality level hospitals (adjusted OR 4,96) are independent health care factors associated with perinatal death. Conclusions: Besides the need to build a strong, comprehensive and regionalized perinatal health system, this study indicates that improvement in hospital quality care at childbirth is essential to provide safe care and reduce perinatal mortality due to preventable causes. Implications for Policy, Delivery, or Practice: Improving hospital quality care is essential to reduce the high rates of perinatal preventable deaths in Brazil, as most of them occur at the hospital level. More than 50% of the births in the city, as well as in the country, take place in type 1 hospitals (low quality level),which were associated to perinatal death after controlling for confounders. In addition to birth technology, like neonatal intensive care, basic practices must be improved - such as the use of the partograph and adequate maternal and fetal assessment during labor - to change child and maternal adverse indicators in the country. Primary Funding Source: PAHO/WHO ●The Impact of Family Income Gradients on the Health and Health Care of Young Children Kandyce Larson, M.S.W., Neal Halfon, M.D., MPH Presented By: Kandyce Larson, MS., Research Associate, School of Public Health, UCLA Center for Healthier Children, Families, and Communities, 1100 Glendon Avenue, Suite 850, Los Angeles, CA 90024; Tel: (310)794-0981; Fax: (310)7942728; Email: kandyce@ucla.edu Research Objective: Health is not equally distributed across the income distribution. Despite the significance of the early childhood period in providing the foundation of human health and development, research on social class gradients in childhood is limited compared with that in adulthood. The objective of this study is to examine income gradients in early childhood health and health care in a nationally representative sample. Study Design: The 2000 National Survey of Early Childhood Health is a telephone survey of a national sample of parents with young children. Differences in selected health and health care measures are examined across four categories of household income: less than 17,501, 17,501-35,000, 35,00160,000, more than 60,000 dollars. Bivariate associations are examined using linear polynomial testing. Gradient measures that show significant linear associations with income are examined using multivariate logistic regression. Population Studied: Parents of 2,608 U.S. children aged 4 to 35 months. Principal Findings: The percentage of respondents reporting better health increases across income categories for six indicators of child health. Low birth weight and acute health conditions show no graded relationship with income. Controlling for maternal race/ethnicity, maternal age, child age, and child gender in multivariate logistic regression analyses, a one unit change to a higher income level increases the odds of parent reported excellent child health by 31 percent OR: 1.31, CI: 1.12-1.52; decreases the odds of asthma OR: 0.61, CI: 0.48-0.79, parent concern about child emotional/behavior problems OR: 0.78, CI: 0.67-0.91, parent concern about the child’s preschool skills OR: 0.83, CI: 0.69-0.99, and low maternal mental health OR: 0.75, CI: 0.63-0.88; and has no effect on parent concerns about child speech development. Controlling for insurance status and child and family demographics in multivariate analyses, household income predicts global satisfaction with health care, satisfaction with the health care provider, and the type of usual location for well child care. Conclusions: The results of this study confirm the existence of substantial income gradients in the developmental health of children in the first few years of life. Income gradients are particularly steep for developmental, behavioral, and mental health concerns. Implications for Policy, Delivery, or Practice: Analyzing health and health services outcomes in relation to income gradients leads to different conclusions and policy implications than traditional analyses that dichotomize income impacts as poor vs. non poor. Primary Funding Source: CWF ●Looking Behind HEDIS: The Timing of Childhood Immunization Matters Ann Lawthers, Sc.D., SM, Jianying Zhang, MPH, MS, M.D., Katharine Willrich-Nordahl, MS, Louise Bannister, BSN, JD, Jay Himmelstein, M.D., MPH, Lawrence Kleinman, M.D. Presented By: Ann Lawthers, Sc.D., SM, Assistant Professor, Center for Health Policy and Research, University of Massachusetts Medical School, 222 Maple Avenue, Shrewsbury, MA 01545; Tel: (508)856-1531; Email: ann.lawthers@umassmed.edu Research Objective: This project presents a detailed study of immunization rates offering a “look behind” the Health Plan Employer Data and Information Set (HEDIS) performance in four Medicaid managed care plans in Massachusetts. We assess how well HEDIS measures reflect a child’s immunization status and identify factors affecting completeness rates so that payers and plans may identify potentially fruitful opportunities when allocating scarce quality improvement resources. Study Design: Retrospective chart review and administrative data analyses conducted in 1999 and 2002. The dates of all visits, immunizations, immunization contraindications and diagnoses at each visit were captured. Immunization completeness at age two was assessed using the CDC’s 4:3:1:3:3 combination, HEDIS 2002 combination two, and the Massachusetts Department of Public Health (MDPH) 2001 standard. Population Studied: Randomly selected HEDIS eligible two year-olds (n=1,618) enrolled in four capitated Medicaid managed care plans in Massachusetts. Principal Findings: By 2001, overall completeness at age two had both increased and decreased, depending on the completeness standard and method of counting immunizations. Counting immunizations delivered, changes from 1998 to 2001 were mixed, e.g. HEDIS rates increased from 57% to 74% while the CDC rates dropped from 87% to 79%. Completeness by counting only appropriately timed immunizations yielded completeness rates 13 percentage points lower for all three standards. Logistic regression modeling showed that children with fewer late immunizations, missed opportunity visits, and invalid shots were more likely to be complete by age two as were children at one plan with a major quality improvement initiative on childhood immunization. Conclusions: Childhood immunization status is dynamic and the use of snapshot measurement tools, such as HEDIS, may be less useful for quality improvement purposes than methods that describe the dynamic nature of immunization. The HEDIS Childhood Immunization measure may be best construed as a population health status measure that assesses the level of protection against disease in the community. Implications for Policy, Delivery, or Practice: HEDIS has become the national standard for measuring the adequacy of immunization in managed care. Using this standard, Medicaid managed care immunization rates appear to lag behind those in commercial plans. “Looking behind” HEDIS provides a valuable approach to targeting quality activities. We demonstrate both the feasibility and value of transforming a regulatory mechanism, in this case the annual clinically focused study, into a constructive and actionable quality improvement tool which produces demonstrable impacts on child health status. Primary Funding Source: Other Government Funding ●Factors Associated with Family-Centered Care for Children with Special Health Care Needs (CSHCN) Gregory Liptak, M.D., MPH, Jacalyn Yingling, MS, Monica Serdinow, AuD, Mark Orlando, Ph.D., Karen Nolan, PT, MS, PCS, Amy Luvera, MSPT, Emily Kuschner, MA Presented By: Gregory Liptak, M.D., MPH, Professor of Pediatrics, University of Rochester Medical Center, 601 Elmwood Avenue, Rochester, NY 14642; Tel: (585)275-5962; Fax: (585)275-3366; Email: gregory_liptak@urmc.rochester.edu Research Objective: Family-centeredness is a critical element of health care for all children, and provides the context in which health and health-related services are delivered to the child and family. Care that is family-centered can promote the psychosocial well-being of children and their parents, and increase satisfaction with services. The purpose of the study was to evaluate the family-centeredness of care delivered to CSHCN. Study Design: Secondary analysis of data obtained from the State and Local Area Integrated Telephone Survey (SLAITS), developed by the Centers for Disease Control and Prevention. Data were analyzed using SUDAAN statistical software. Family-centered care (FCC) was evaluated by creating a composite variable, the sum of responses on five questions that focused on attributes of FCC. Each attribute was scored using a Likert-scale that ranged from 1 (never) to 4 (always). Interactions between the FCC score and demographic, financial characteristics, access, and coordination of health care services were assessed. Population Studied: Families of 36,453 CSHCN (weighted sample size of 8,742,837), younger than 18 years, responding to the SLAITS. Principal Findings: The possible range of scores was 5-20. The mean was 17.16, (SE mean 0.03), with a median of 18; 33% lacked =1 component of FCC. The factor rated the lowest was getting sufficient information about the child’s condition (5% said they ‘never’ received specific information). Using multiple linear regression, family-centeredness was significantly associated (all p<0.01) with the providers’ ability to communicate with other providers, measures of access to care (including delay in care, continuity of care, and health insurance), severity of the condition, and care coordination. The overall R2 for the model was 0.32. Conclusions: Although families of CSHCN generally perceive that their care is family-centered, gaps exist. Families often do not receive sufficient information about their child’s care or condition. FCC is associated with (1) providers’ communication skills, (2) access to care, including continuity of care, (3) severity of the condition (including impact on the family), and (4) care coordination. Implications for Policy, Delivery, or Practice: Models of care delivery, such as the Medical Home, may be able to address these issues and improve FCC. Continuing education for providers that includes communication skills and knowledge of CSHCN may be able to improve FCC as well. Primary Funding Source: DHHS: Maternal and Child Health Bureau; Leadership Education for Neurodevelopmental and Related Disabilities ●Short-Term Rehospitalization Following Discharge from the NICU: The Influence of Age at Discharge Scott Lorch, M.D., Scott A. Lorch, M.D., MSCE, Marla N. Gardner, MA, John D. Greene, MA, Yanli Wang, MS, Jeffrey H. Silber, M.D., Ph.D. Presented By: Scott Lorch, M.D., Systems Research Institute and Perinatal Research Unit, Kaiser Permanente Division of Research, 2000 Broadway, 2nd floor, Oakland, CA 94612; Tel: (510) 891-3502; Fax: (510) 891-3508; Email: Gabriel.Escobar@kp.org Research Objective: In the period immediately following discharge home from the NICU, premature infants are two to three times as likely to be rehospitalized as infants born at term gestation. However, rehospitalization rates have been calculated based on the date of discharge without accounting for post menstrual age (PMA) at discharge. Study Design: As part of the Infant Functional Status (IFS) Study, we grouped the population according to PMA at discharge (<35 vs. >35 weeks). For each group we developed 2 separate logit models to predict readmission, one for readmission within 2 weeks of discharge and one for readmission within 3 months. Each model included gestational age (GA), SNAP-II, race, sex, SGA, maternal age, NICU site, and presence of follow-up visit within 72 hours. We also constructed a dummy variable interaction model to test whether the coefficients on these variables differed across PMA group. Population Studied: All infants (N=894) born less than 32 weeks gestational age (GA) at a Northern California Kaiser Permanente hospital between 1998 and 2001 were eligible. To supplement the cohort, 543 infants of 33-34 weeks GA were added. Infants were excluded for congenital anomalies, home ventilation, need for a VP shunt, or loss to follow-up. Principal Findings: Of the 1437 infants, 646 were discharged at a PMA < 35 weeks, and 791 were discharged after 35 weeks. Patients discharged after 35 weeks were more premature, had lower birthweight, had greater severity as measured by SNAPII and were more likely to be SGA. The < 35 week group had a 2-week readmission rate of 1.85% vs. 4.24% in the >35 group (p = 0.006), and similarly, 3-month readmission rates were 8.66% vs.14.79% (P < 0.001). The c-statistic for predicting 2week readmission was 0.868 in the < 35 week group versus 0.777 in the >35 week group, and for 3-month readmission models, the c-statistics were 0.733 vs. 0.683 respectively. Using the interaction model, there were no differences in the coefficients on risk factors between the two groups. Conclusions: There are higher rates of readmission in infants discharged greater than 35 weeks PMA vs. those discharged before because only the healthiest children who are eligible for discharge fall into the < 35 week group. The lower discrimination observed in the >35 week group suggests that unobserved variation in physiologic status plays a larger role in these infants. Implications for Policy, Delivery, or Practice: Understanding the increased risk of rehospitalization in premature infants requires physicians and policy makers to consider factors beyond PMA and other time based variables commonly collected in claims data. Utilizing physiologic state at or around the time of discharge, as well as demographic and health services variables, may improve our ability to predict these undesired events. Primary Funding Source: HRSA, MCHB R40-MC00238 ●Analyzing Pre-Discharge Costs to Predict Subsequent Resource Use in Premature Infants Scott A. Lorch, M.D., MSCE, Gabriel J. Escobar, M.D., Orit Even-Shoshan, MS, Justin I. Mathew, BA, Jeffrey H. Silber, M.D., Ph.D. Presented By: Scott A. Lorch, M.D., MSCE, Assistant Professor of Pediatrics, Division of Neonatology, The Children's Hospital of Philadelphia, 3535 Market Street, Suite 1029, Philadelphia, PA 19104; Tel: (215) 590-1714; Fax: (215) 590-2378; Email: lorch@email.chop.edu Research Objective: The timing of discharge of a premature infant may influence subsequent use of medical resources. Because the average daily cost of care in the last 2 weeks before discharge (L2W-ADC) declines as neonates remain hospitalized, this variable may help to identify sicker infants who are discharged too early from the neonatal intensive care unit (NICU). The objective of this project was (1) to determine the relationship between the L2W-ADC, hospital readmissions and medical costs 1 year after discharge from the NICU in a managed care system and (2) to explore alternative methods for analyzing cost data in this population. Study Design: As part of the Infant Functional Study, all infants born less than 32 weeks gestational age (GA) at a Northern California Kaiser Permanente hospital between 1998 and 2001 were eligible. To supplement the cohort, 543 infants of 33-34 weeks GA were added. Exclusions included congenital anomalies, home ventilation, need for a VP shunt, or loss to follow-up. A readmission was defined as any hospitalization within one year of discharge. Costs were calculated from resources used by the child in 2001 dollars. Logistic regression models determined the association between readmission and L2W-ADC after controlling for GA and common neonatal complications. Three models (OLS, logarithmic-transformed linear regression, and log-linear variance regression) were constructed to determine the effect of L2W-ADC on 1 year medical costs after controlling for medical complications. Population Studied: 1437 infants delivered at one of six Northern California Kaiser Permanente hospitals between 1998 and 2001. 265 (18.4%) were born less than 28 weeks GA and 629 (43.8%) were born between 29 and 32 weeks GA. Principal Findings: The median L2W-ADC was $294.66 (intraquartile range $269.48-$321.95). L2W-ADC was increased with bronchopulmonary dysplasia ($33±5), necrotizing enterocolitis ($38±14), and history of ductus arteriosis ($11±4). After controlling for GA, race, sex, and common neonatal conditions, L2W-ADC was not significantly associated with a hospital readmission; a $50 increase in L2W-ADC was associated with only an odds ratio of 1.12 for readmission (95% CI 0.975-1.287). Of the 3 cost regression models, the log-linear variance regression had the lowest and most consistent residual plots across expected cost, followed by the log-transformed model. The log-linear variance model found that a $50 increase in L2W-ADC was associated with a $46 ± 13 increase in post-discharge costs (P=0.0005). The same increase in L2W-ADC was associated with a 6.6% increase in post-discharge costs (95% CI 1.4%-12.1%). The L2W-ADC decreased with later discharge from the NICU after controlling for medical complications and GA. Conclusions: Increases in the L2W-ADC were associated with increased costs after discharge from the NICU using all three models. However, the log-linear variance model displayed the best residual analysis. L2W-ADC decreased with later discharge from the NICU. Implications for Policy, Delivery, or Practice: As daily costs of a NICU stay fall for each added day a patient remains hospitalized, our results suggest that early discharge of premature infants may lead to increased medical use in the year after discharge. However, the causal pathway for these increased costs remains to be determined. Primary Funding Source: HRSA, MCHB R40-MC00238 ●Building Block to Literacy: Efficacy of a Phonetic, Multisensory Therapy Model Maureen K. Martin, Ph.D. Presented By: Maureen K. Martin, Ph.D., Director, DuBard School for Language Disorders, The University of Southern Mississippi, 118 College Drive #10035, Hattiesburg, MS 39406-0001; Tel: (601) 266-5223; Fax: (601) 266-6763; Email: maureen.martin@usm.edu Research Objective: To determine the efficacy of a specialized phonetic, multisensory therapy approach for children with dyslexia. Currently understood as a language disorder (Catts, 1989; Shaywitz, 2003), dyslexia varies in degree of severity, is often familial, and persists into adulthood. Studies have shown that skill in the phonological system, the sounds of the language, at a preschool age is a strong predictor of later success in developing reading skills (Shaywitz, 2003). Conversely, lack of such skill results in poor ability to decode, the first component required in reading development. As a result, the second component of reading, comprehension, may be greatly diminished or nonexistent. Study Design: Subjects’ skills were measured on the Decoding Skills Test (Richardson & DiBenedetto, 1985), and a criterion referenced checklist of single phonemes, prior to the implementation of a phonetic, multisensory therapy program. The program was delivered by speech-language pathologists in a resource schedule of service delivery averaging two onehour sessions per week. Pretest and post-test scores were compared to evaluate efficacy. Population Studied: The population consisted of 15 children diagnosed with dyslexia, some of whom had a concomitant diagnosis of Attention Deficit Hyperactivity Disorder (ADHD). Principal Findings: Dyslexia is characterized by a deficit in utilizing the phonological system—the system of sounds which are combined to form words for speech and which are also a critical component in developing reading skills. A phonetic, multisensory therapy approach was delivered on a resource therapy schedule in a university-based program and resulted in significant gains in phonological skills. Conclusions: Pre-test and post-test scores on a standardized measure of decoding skills, the Decoding Skills Test (Richardson & DiBenedetto, 1985), indicated that students challenged with dyslexia made significant improvements through phonetic, multisensory therapy. Paired sample t-tests on three subtests resulted in -4.022, df of 11, and significance at .002, -5.760, df of 14, and significance at .001, and -4.232, df of 9, and .002. In addition, on a criterion referenced measure of oral recall of single phonemes, pre- and post-tests showed statistically significant results on a paired t-test: -2.373, df of 14, and significance at .033. The results indicate that dyslexic students are able to effectively and efficiently acquire critical decoding skills when a phonetic, multisensory structured approach is implemented through a resource model of service delivery. Implications for Policy, Delivery, or Practice: These results illustrate the efficacy of a phonetic, multisensory therapy approach for significantly improving phonological skills necessary for reading success. It may be considered that such an approach delivered to all students, including those considered at-risk, as well as those diagnosed with the language disorder of dyslexia, could potentially reduce the nation’s 20% rate of reading failure. Effective intervention may be expected to reduce the incidence of learning disabilities and potentially the self-esteem and social skill problems which are prevalent in those challenged with this disability. In addition, the incidence of learning disabilities among prison inmates has been estimated at between 30-50% (Corley, 1996). Enhancement of literacy skills may be expected to result in greater academic success, fewer social and self-esteem issues, and potential reduction in criminal activity. Primary Funding Source: Other Foundation Funding ●Efficacy of an Intensive, Multisensory Therapy Model for Children Challenged With Developmental Apraxia of Speech Maureen K. Martin, Ph.D. Presented By: Maureen K. Martin, Ph.D., Director, DuBard School for Language Disorders, The University of Southern Mississippi, 118 College Drive #10035, Hattiesburg, MS 39406-0001; Tel: (601) 266-5223; Fax: (601) 266-6763; Email: maureen.martin@usm.edu Research Objective: To determine the efficacy of an intensive, non-traditional model of speech-language therapy for children challenged with developmental apraxia of speech (DAS). This motor speech disorder (Vogel & Cannito, 2001; Yorkston et al, 1999) may be characterized by a wide range of neurological soft signs; however, a specific neuroanatomical site of lesion has not been identified for DAS. The condition results in a severe articulation disorder which does not respond well to traditional speech-language therapy and often results in the need for assistive communication devices and/or sign language. Study Design: Articulation skills were measured on the Arizona Articulation Proficiency Scale (Fudala & Reynolds, 1989, 1994; Fudala, 2001) prior to the implementation of intensive speech-language therapy in a university-based clinical program. A phonetic, multisensory program was delivered by speech-language pathologists in a full-day nontraditional model of therapy. Pretest and post-test scores were compared to evaluate efficacy. Population Studied: The population consisted of 14 children with the severe motor speech disorder known as developmental apraxia of speech (DAS). In addition, some students' concomitant diagnoses included language disorders, cognitive deficits, autism, and attention deficit hyperactivity disorder. Principal Findings: Children with developmental apraxia of speech (DAS) are a subgroup of those affected with articulation disorders. Often, these students are limited to assistive communication devices and sign language as modes of communication due to a poor response to traditional therapy approaches. Through an intensive, non-traditional therapy approach used in a university-based program, students achieved not only speech intelligibility but additional critical academic skills as well. Conclusions: Pre-test and post-test scores on measures of articulation intelligibility indicated that students challenged with DAS made significant improvements in speech skills when an intensive, non-traditional model and approach to therapy was utilized. A paired sample t-test resulted in -8.910, df of 13, and significance at .001 on standardized measures of articulation intelligibility. The results indicate that severely apraxic children can acquire highly intelligible speech skills through a relatively short period of speechlanguage therapy when a phonetic, multisensory structured approach is implemented on an intensive basis as part of a total educational program. In addition, students acquired reading, spelling, cursive handwriting, and acoustic skills in addition to their improvements in articulation. Through traditional therapy approaches, DAS students may require years of therapy and suffer the challenges of inadequate oral communication skills with the related effects of diminished interpersonal and social skills, as well as critical academic skills such as reading. Implications for Policy, Delivery, or Practice: These results illustrate the efficacy of a non-traditional model for service delivery which resulted in significantly improved speech skills, avoidance of the need for alternative communication strategies such as assistive devices or sign language, and provided the added benefit of enhancing reading, writing, and other academic skills. ●Malaria Prevention and Control Among Primary School Children in Kenya Thadeus Odenyo, BSc,MPH, Orago Alloys, BSc, MSc, Ph.D., Michael Otieno, BSc, MSc, Ph.D., Syprine Otieno, BSc, MSc, Ph.D. Presented By: Thadeus Odenyo, BSC, MPH, Student, Health Sciences, Kenyatta University, PO box 43844, Nairobi, 168; Tel: (254)-722885904; Fax: (254)-2-810901; Email: thadeusoba@yahoo.com Research Objective: a) To establish the knowledge levels of school pupils in Busia District on malaria. b) To investigate the effect of knowledge level on the perceptions and beliefs of school children towards prevention and control of malaria. c ) To investigate the practices of school pupils towards malaria prevention and control. Study Design: Descriptive Crosssectional Study Population Studied: Primary School Children Principal Findings: Results show that in spite of good knowledge among school children, there was a disproportionate balance between knowledge and practice The mean knowledge score was 60.99± 0.6 (95 % CI). Respondent who had good knowledge on the etiology of malaria were 94.14 % (n = 649). While pupils sought treatment from health facilities when having clinical malaria, compliance with prescribed dosage was low. At the same time there was disparity between net ownership (93.83 %, n = 649) and net use (54.69 %, n = 649). Net ownership and use appeared to be positively related. In fact, respondents who had nets were 36 times more likely to use them (c2= 330.8, p < 0.001). In this study it appeared that knowledge on malaria was to a great extent academic. It is hence surmised that this contributed to poor malaria communication at the family level. It was also evident that malaria related instructions failed to alight due to instructor incompetence, in the light of which we recommend instructor capacity building on malaria. Visiting a health facility did not significantly influence the perception that mosquito nets can prevent malaria (c2= 2.7, p > 0.05), at the same time perception that treated nets are better than untreated nets appeared to be independent of visiting a health facility (c2= 0.6, p > 0.05). Respondents who visited a health facility were not likely to undertake environmental sanitation as a way of controlling malaria (c2= 1.04, p > 0.05) and were less likely to perceive insecticides as a control tool for malaria (c2= 0.9, p > 0.05). Class of respondent (c2= 7.3, p <0.01) and knowledge on malaria etiology were related to perceptions on visiting a health facility (c2= 11.95, p < 0.001). Nevertheless, having been taught about malaria (c2= 0.04, p > 0.05), number of bouts experienced by a respondent (c2= 0.34, p > 0.05), and perceived distance from a health facility (c2= 1.63, p > 0.05) failed to have any significant relationship Conclusions: We conclude that knowledge, beliefs and perceptions as well as practices form a labyrinth interdependent body. Ingeniously coordinated approaches are needed to overcome the negative impacts of this interdependence. Implications for Policy, Delivery, or Practice: Recommendations Operational recommendations Information, Education and Communication: There is need to tailor a malaria package for use in malaria endemic zones of the country. The package may be transmitted through the school radio program since pupils seem to regard information coming over the radio positively. Policy Formulation: Education programs aimed at increasing awareness among the pupils on malaria transmission might promote the use of personal protection against mosquitoes. Evaluation and Monitoring: The role of the government in coordination, monitoring and evaluation of malaria prevention and control activities in the context of stakeholder involvement should be strengthened. It is recommended that capacity building should be undertaken among government officials charged with the management of malaria matters. Such training should be undertaken regularly to ensure that extant capacities can cope with emerging malaria challenges. Coordination of Intersectoral collaboration: Paramount of all, the ministry of health should be at the frontier of malaria prevention and control. By mobilising the community through the school as a fundamental unit of the framework, children will appreciate the impacts of malaria on them and consequently conceive the advances that target malaria. Suggestions for future work: i) There is need to carry out research to determine the socio-cultural determinants causing the disparity between ownership and net use among pupils in the District. ii) There is need to investigate the deteminants culminating with non-compliance to prescribed dosage on malaria among pupils in Busia district. iii) There is need to follow-up knowledge evaluation on malaria with practices instituted by respondents to prevent malaria. iv) It is imperative to investigate the relationship between respondents buying nets and their use as compared to recepients of nets as donations. General recommendation: Parents, teachers and children must be sensitized on malaria prevention and control and children charged with the obligation to undertake practices that will protect them and the wider community from malaria. Primary Funding Source: Doris Morgan Trust ●Assessment, Optimization and Use of Mortality Prediction Tools for Admissions to Pediatric Intensive Care in the United Kingdom Gareth Parry, BSc, MSc, Ph.D., Sam Jones, BSc, David Harrison, Ph.D., Stephanie Black, MSc, Gale Pearson, M.D., Kathy Rowan, Ph.D. Presented By: Gareth Parry, BSc MSc Ph.D., Reader in Health Services Research, Health Services Research, University of Sheffield, Regent Court, 30 Regent Street, Sheffield, S1 4DA; Tel: 44 114 2220798; Fax: 44 114 2724095; Email: g.parry@sheffield.ac.uk Research Objective: To assess the published tools currently proposed for use in comparing the risk-adjusted mortality of children following admission for pediatric intensive care. Study Design: The US developed Pediatric Risk of Mortality (PRISM, PRISM III-12 and PRISM III-24) systems which use information up to 24 hours post admission were compared with the primarily Australian Pediatric Index of Mortality (PIM and PIM2) systems which use information up to 1 hour post admission. Predicted probability of PICU mortality was calculated using the published algorithms for PIM, PIM2 and PRISM and compared to observed mortality. These scores along with PRISM III-12 and PRISM III-24 (whose algorithms are not published) were optimized for a UK setting. Statistical process control charts in the form of funnel plots of the resulting risk-adjusted mortality were constructed to assess the impact of comparing performance using each of the tools. Population Studied: Data were collected from 23 of 27 (85%) pediatric intensive care units identified in the United Kingdom on 10197 (98%) of 10,385 admissions between 1st March 2001 and 28th February 2002. Principal Findings: Overall, 45% of eligible admissions had complete data for PIM and only 9% for PRISM. Following the published rules for missing values in each scoring system, all cases were eligible for inclusion in the analysis. PIM2 had the best calibration and discrimination using published coefficients but still showed evidence of significant miscalibration (c-index=0.84, 0.81, 0.82 and Hosmer-Lemeshow Chi square = 39.8, 86.7, 549.1 all P<0.001 for PIM2, PIM and PRISM respectively). Following estimation of UK specific coefficients, PRISM III-24 had the best discrimination (cindex=0.88) and marginally superior calibration (HosmerLemeshow Chi square = 13.2) compared to PIM2 (cindex=0.84, Hosmer-Lemeshow Chi square = 14.5). The funnel plots for all the re-calibrated models indicated that the risk-adjusted mortality for all units was consistent with random variation. Conclusions: In the UK both the PRISM III and PIM2 systems provided very similar assessment diagnostics and provided similar conclusions in assessing the distribution of riskadjusted mortality in pediatric intensive care units. PIM2 benefits from greater ease of data collection compared to PRISM III. Implications for Policy, Delivery, or Practice: This study confirms that risk adjustment methods primarily developed in other countries require validation before being used to provide risk-adjusted outcomes of pediatric intensive care unit mortality for units within a new health care setting. It is also important that the calibration of these tools are reassessed periodically in order to ensure their continued validity. Since similar conclusions on the distribution of risk-adjusted mortality were achieved with all the scoring systems and that good discrimination and calibration are achievable without using information obtained after the first hour in a pediatric intensive care unit, there is evidence to suggest that the PIM2 system may be preferable. However, any variation in riskadjusted pediatric intensive care unit mortality needs to be tempered by an assessment of variation in longer term outcome. Primary Funding Source: United Kingdom Medical Research Council ●The Role of Public/Private Partnerships in Communitybased Efforts to Improve the Effectiveness of School Health Programs MaryAnn Phillips, MPH, Bernette McColley, BA, James Emshoff, Ph.D., Bobbi Cleveland, BS Presented By: MaryAnn Phillips, MPH, Senior Research Associate, Georgia Health Policy Center, 14 Marietta Street, Atlanta, GA 30303; Tel: (404)651-1643; Fax: (404)651-3147; Email: mphillips2@gsu.edu Research Objective: To determine if public/private partnerships are effective in expanding sustained access to care for low income and medically underserved school-aged children. Study Design: To facilitate achievement of model school health programs, Georgia’s Department of Community Health and the Philanthropic Collaborative for a Healthy Georgia pooled financial resources to develop and evaluate a competitive grant program to encourage school districts to establish coordinated school health programs. Population Studied: School-aged children in 13 Georgia communities. Principal Findings: School health grants were funded for one, two, or three years beginning in 2001 and concluding in 2004. Preliminary results indicate that access to health services has improved, school faculty and staff skills have been enhanced, overall attendance has increased, and local community support for school health programs is strong. In addition, the project has advanced the knowledge about successful health efforts at the state and national levels by sponsoring a technical assistance conference, preparing a training video distributed to all school health programs in Georgia and broadcast on the statewide education-related network, and participating in a pilot training program for the National Association of School Nurses on the treatment of asthma in the schools. A thorough evaluation of the effort is currently underway. Conclusions: Georgia has developed an innovative model for improving school health programs that may be replicable in other states. Its success reflects the generosity and commitment of 20 Georgia private and corporate foundations whose initial contribution was matched by the State Department of Community Health and then leveraged even further with local and federal dollars. There are advantages to public/private partnerships that accrue to all elements of the program: pooling resources allows foundations to have greater reach and impact, local communities receive funding to improve access to care for low income children, and technical assistance and best practice information receive statewide and national dissemination. Implications for Policy, Delivery, or Practice: The evaluation of this effort will provide best practice information that will impact state policies regarding the provision of health care in Georgia schools; identify new methods of collaboration for providing access to care for low income children; and promote additional opportunities for public/private partnerships around common health care priorities. Primary Funding Source: Georgia Department of Community Health and the Philanthropic Collaborative for a Healthy Georgia ●Utilization of Nasal Airway for EGD in Children George Russell, M.D., Alex Flores, M.D., Pacifico Tuason, M.D., William Denman, M.D., Audrius Zibaitis, M.D. Presented By: George Russell, M.D., Fellow, Pediatric GI/Clinical Care Research, Institute for Health Policy Studies and Clinical Research, Tufts-New England Medical Center, 750 Washington Street, NEMC #213, Boston, MA 02111; Tel: (617) 636-3083; Fax: (617) 636-8718; Email: grussell@tufts-nemc.org Research Objective: In a patient population undegoing endoscopy (EGD), general anesthesia is commonly required to provide an optimal working environment for successful completion of endoscopic examination. We have revisited an old technique utilizing nasal airways to deliver oxygen and inhaled anesthetics for EGD and evaluated its effectiveness and safety profile. Study Design: Retrospective Case Series Population Studied: Two hundred charts of pediatric patients who have undergone EGD in our intitution between 20012003 were identified. We selected 112 charts where EGD alone or in combination with another procedure (e.g., colonoscopy) were performed. Patients were anesthetized by either inhaled anesthetics alone, total intravenous anesthetics (TIVA), or a combination of both. The emphasis of our data collection was to compare the safety profile and efficiency of the nasal airway technique to other techniques by looking at the peri-procedure complication rates and operating room (OR) turnover times. Principal Findings: The age of patients ranged from 7 months to 20 years (mean, 9.8 years). Fifty-four were female (48%). 56 patients used the nasal airway with a mean time to leave the OR at 8.2 min +/- 4.8 min and a mean time to discharge of 85.4 min +/- 35 min. 44 patients used the mask airway with a mean time to leave the OR at 7.6 min +/- 4.1 min and a mean time to discharge of 88.4 min +/- 55 min. 12 patient underwent endotracheal intubation with a mean time to leave the OR at 22.6 min +/- 14 min and a mean time to discharge at 91.6 min =/- 56 min. The average time to leaving the OR was significantly lower in the nasal airway group compared to endotracheal intubation (p value <.001). Discharge times between three comparison groups did not significantly differ. There were a total of 11 peri-operative adverse events among the three groups. No aspiration episodes occured and no event had long-lasting sequelae. The complication rates were 5% for nasal airway, 14% for mask airway, and 17% for endotracheal intubation. Conclusions: In our study, nasal airway provided us with a satisfactory means of maintaining patent airway during EGD. The complication rate in the nasal airway group was lower than in other evaluated groups. The nasal airway group exhibited no increased risk of aspiration. Our finding suggest that a nasal airway method is a safe, non-invasive alternative to endotracheal intubation when used for EGD. OR turnover times were superior in the nasal airway group in comparison to patients being intubated. Implications for Policy, Delivery, or Practice: More study is needed but instituting the nasal airway as standard of care in place of endotracheal intubation for EGD could substantially lower the cost of time spent in the OR and reduce patient waiting times for important diagnostic procedures. Primary Funding Source: AHRQ ●Just One More Visit: Achieving >90% Vaccination Coverage and Reducing Racial Disparities with one Catchup Visit Tom T. Shimabukuro, M.D., MPH, MBA, Elizabeth T. Luman, Ph.D., MS, Richard A. Schieber, M.D.,. MPH, Carla A. Winston, Ph.D., MA Presented By: Richard A. Schieber, M.D., MPH, Medical Epidemiologist, Health Services Research and Evaluation Branch, Immunization Services Division, National Immunization Program, Centers for Disease Control and Prevention, 1600 Clifton Road, MS E-52, Atlanta, GA 30333; Tel: (404)639-6228; Fax: (404)639-8614; Email: RSchieber@cdc.gov Research Objective: Achieving and maintaining the Healthy People 2010 objective of at least 90% coverage for universally recommended vaccines among young children and reducing racial/ethnic disparities in vaccination coverage are important national health priorities. We projected the potential improvement in up-to-date (UTD) coverage achieved by one simulated additional provider visit for catch-up vaccinations. Study Design: We used data from the 2003 National Immunization Survey, a nationally representative randomdigit-dial survey of households with children 19-35 months of age. We calculated crude baseline UTD vaccination coverage levels and stratified levels by race/ethnicity and state among 24-month olds for the 4:3:1:3:3:1 (four diphtheria-tetanuspertussis, three poliovirus, one measles, three Haemophilus influenza type b, three hepatitis B, and one varicella) and 4:3:1:3:3:1:4 (includes four pneumococcal vaccinations recommended since 2000) combined series. We then simulated coverage if incompletely vaccinated children were to receive one additional provider visit to obtain up to four missing vaccinations. Population Studied: Children 24-months of age in the 2003 National Immunization Survey Principal Findings: Overall baseline UTD coverage for the 4:3:1:3:3:1 series was 69.1% (95% confidence interval [CI], 67.9%-70.4%). Among incompletely vaccinated children, 79.0% became UTD with only one additional visit, of which 63.5% required only one vaccination during that visit. With that simulated visit, overall UTD coverage increased 24 percentage points to 93.5% (95% CI, 92.8%-94.2%). Coverage increased significantly within each racial/ethnic group: nonHispanic whites, from 71.7% to 94.6%; non-Hispanic blacks, from 62.9% to 92.2%; and Hispanics, from 66.8% to 92.1%. Baseline coverage in the 50 states and District of Columbia ranged from 50.8% to 84.4% (median 68.5%); following the simulated additional visit, all but 6 of these achieved >90% UTD coverage (range 88.9% to 98.8%, median 93.9%). For the 4:3:1:3:3:1:4 series, overall baseline UTD coverage was 35.1% (95% CI, 33.8%-36.3%). Among incompletely vaccinated children, 88.4% became UTD with only one additional visit, of which 68.2% required only one vaccination during that visit. For this series, the simulated visit increased overall UTD coverage 57 percentage points to 92.5% (95% CI, 91.7%93.2%). Coverage increased significantly within each racial/ethnic group: non-Hispanic whites, from 37.8% to 93.8%; non-Hispanic blacks, from 28.9% to 90.4%; and Hispanics, from 31.1% to 91.2%. Although still statistically significant, differences in final coverage levels among racial/ethnic groups following one simulated additional visit, for both vaccination series, are minimal. For the 4:3:1:3:3:1:4: series, baseline coverage in the 50 states and District of Columbia ranged from 13.2% to 56.1% (median 32.3%); following the simulated additional visit, all but 11 of these achieved >90% UTD coverage (range 87.1% to 98.2%, median 92.1%). Conclusions: Achieving at least 90% coverage appears attainable, regardless of race/ethnicity, with a single additional provider visit for catch-up vaccinations among incompletely vaccinated 24-month olds. Most such children would require only one vaccination during this visit. Furthermore, providing a catch-up visit has the potential to substantially reduce racial disparities in vaccination coverage. Implications for Policy, Delivery, or Practice: Strategies to promote such a catch-up vaccination visit should be developed. Primary Funding Source: No Funding Source ●Injury prevalence among children and adolescents with mental retardation Elspeth M. Slayter, MSW, MA, Deborah W. Garnick, Sc.D., Joanna M. Kubisiak, MPH, Christine E. Bishop, Ph.D., Daniel M. Gilden, MS, Rosemarie B. Hakim, Ph.D. Presented By: Elspeth M. Slayter, MSW, MA, The Heller School for Social Policy and Management, 415 South Street MS 035, Waltham, MA 02454-9110; Tel: (617)686-6594; Email: eslayter@brandeis.edu Research Objective: Medicaid is a major source of health insurance coverage for children and adolescents with mental retardation and is estimated to pay almost a third of their injury care costs. To better understand these basic statistics, we examined the prevalence of specific types of injuries among children and adolescents with mental retardation. Study Design: Using eligibility and claims data collected by the Centers for Medicare and Medicaid Services’ Medicaid Statistical Information System for calendar year 1999, injury prevalence for 8.4 million Medicaid-eligible children in 26 states was measured through the use of an episode algorithm for the identification of injury using ICD-9-CM diagnosis and CPT procedure codes. Data from inpatient, outpatient and long-term care settings were examined in order to provide a broader picture of injury prevalence than is allowed by single- site-of-care studies. We calculated the odds ratio (OR) that children and adolescents with mental retardation were treated for injuries compared with a non-mental retardation comparison group, controlling for age and gender. Population Studied: This project studied Medicaid beneficiaries aged one through twenty with a diagnosis of mental retardation and a comparison group of Medicaid beneficiaries without mental retardation. Principal Findings: In total, 36.9 percent (N=18,344) of children and adolescents with mental retardation had at least one injury claim, compared with 23.5 percent (N=1,978,501) among those without mental retardation. Overall, males were more likely than females to experience injury in this population, a finding that differed from previous research results. Controlling for age and gender, children and adolescents with mental retardation were more likely to be treated for any injury (OR: 1.74), and were much more likely than the comparison group to be treated for poisoning (OR: 3.72), foreign body injuries (OR: 3.45), dislocations (OR: 2.74) and internal injuries (OR: 2.49). Conclusions: Children and adolescents with mental retardation in this sample, especially males, are at higher risk for experiencing injury than are their counterparts in the comparison group without mental retardation. Although most types of injury were prevalent at a greater rate in this population, four types of injuries may be especially important to target for prevention intervention, poisoning, foreign body injuries, dislocations and internal injuries. Implications for Policy, Delivery, or Practice: The higher risk of injury among this sample of youth with mental retardation may result in increased disability and morbidity. This has implications for community inclusion, a central policy goal for this population. Childhood injuries in this population result in significant costs for public insurance programs in the United States. Because a majority of youth with mental retardation are covered by the Medicaid health insurance program, targeted prevention interventions should be formulated and implemented to support this vulnerable population, an effort that has the potential to reduce costs for state and federal Medicaid programs. Using Medicaid data, an important new resource for childhood injury surveillance for the population with mental retardation, planning prevention interventions can be explored further by focusing on the specific injury types for which this population is at greatest risk. Primary Funding Source: CMS, American Association of University Women ●Using Linked State and Local Data to Understand a State-Sponsored Home Visiting Program for Medicaid Insured Families Wm. Thomas Summerfelt, Ph.D., Leeanne Roman, MSN, Ph.D., Lynette Biery, PAC, Denise Holmes, MS, Susan Moran, BSN, MPH, Doug Paterson, MPH Presented By: Wm. Thomas Summerfelt, Ph.D., Director of Research & Innovation, Research & Innovation, Grand Rapids Medical Education & Research Center, 1000 Monroe Avenue, NW, Grand Rapids, MI 49503; Tel: (616) 732-6225; Fax: (616) 732-6255; Email: tom_summerfelt@grmerc.net Research Objective: Infant mortality and children’s health are of public concern and many states have responded by sponsoring home visiting services aimed at those most at risk for poor birth, parenting, and infant outcomes. The primary research objective of this project was to determine if a state sponsored home visiting program was reaching the highest risk women (demographic, behavior, and health risks) with greater intensity of services, and better participation in maternal and child health services (prenatal care, immunization, WIC, EPSDT). A second objective was to develop an outcomes monitoring system using linked state and local data to inform policy and program development and to evaluate these services continually. Study Design: The study cohort was comprised of 46, 617 women who gave birth to 47,300 infants in Michigan during 2001. Individual mother and child data were linked across state and county agencies. Data sources included Medicaid, vital records, WIC, lead screening records, and local data from one county’s home visiting programs. Population Studied: Pregnant women and children who were Medicaid insured Principal Findings: The state sponsored home visiting program did not consistently engage high-risk women and their infants (e.g., of the 46,617 women in the sample, 31% smoked during pregnancy and only 28% of those smokers were enrolled). Further, even for those that were enrolled, service intensity did not vary by risk. However, enrolled families did achieve better participation in child health services (i.e., well child visits, immunization rates, lead screening, and breastfeeding initiation). Conclusions: Comprehensive risk screening of the population with clear linkages to targeted interventions around those risks is critical in developing and maintaining effective programs. Linked data becomes crucial in identifying the most efficient (i.e., families had home visits by various state and local programs that may have duplicated services) and the most effective interventions. Implications for Policy, Delivery, or Practice: Linked state and local community data can be used to inform practioners and policymakers regarding efficient and effective programming. A population management approach that includes uniform risk screening, assessment, early identification, and stratification into appropriate services is presented. Such an approach is indicated to improve outcomes in Medicaid insured women and infants and help to reduce costs. Primary Funding Source: CMS ●Addressing Early Childhood Behavior Problems Carole Upshur, EdD, Carole Upshur, EdD, Melodie WenzGross, Ph.D., Presented By: Carole Upshur, EdD, Professor, Family Medicine and Community Health, University of Massachusetts Medical School, 55 Lake Ave North, Worcester, MA 01655; Tel: (508)334-7267; Fax: (508)856-1212; Email: carole.upshur@umassmed.edu Research Objective: To pilot an early childhood mental health intervention with children identified in child care programs with behavior problems. Study Design: Teachers were taught how to screen children using a standardized behavioral checklist. Families whose children scored above clinical cutoffs were referred to individualized short term intervention services and reassessed. A mean of 21 hours of services was delivered, including individual child sessions, and parent and teacher guidance on managing the child’s behavior. Waiting list controls were also reassessed in a quasi-experimental repeated measures design. Population Studied: Preschool children, ages 3-5 in 5 low to moderate income urban and suburban community child care centers serving 440 children Principal Findings: The screening methodology identified 119 children, representing approximately 27% of all children, as meeting criteria for clinically significant behavioral issues (81 externalizing and 38 internalizing; 63% boys). Intervention outcome data are available on 40 target children and families enrolled in brief services, with 79 children serving as waitlist controls. The ethnic backgrounds of target children were diverse, 30% Latino, 40% White, 17.5% Black, and 12.5% other. Family education and income also varied, with 70% of caregivers completing high school and 40% having at least some college. Target family incomes ranged from under $5,000 to over $60,000; median incomes were under $25,000. Assessment of outcomes revealed significant improvements in child behavior from baseline to follow-up, including pre- to post- decreases in aggressive F(1,54)=23.80, p<.001 and maladaptive behaviors F(1,82)=43.84, p<001, and increases in adaptive behavior F(1,83)=26.65, p<.001. Further, there was a significant interaction effect for both aggressive F (1,54)=11.28, p=.001, and maladaptive behavior F (1,16)=7.7, p<.013, showing that the target children improved significantly more than the control children in these areas. In addition for the target children, there was a trend toward improvement in the areas of social development t=-1.865, df=28, p=.07; and communication t=-1.80, df=28, p=.08. However, we found no significant change in either parenting stress or parenting skills. Nevertheless, most parents (84%) reported being satisfied or very satisfied with the intervention. Conclusions: Significant improvement in child behavioral outcomes was found after brief individualized mental health interventions targeted to children identified has having clinically important behavioral problems in child care settings. Unfortunately, little change was found in parenting stress or discipline skills. The potential for longitudinal change in the intervention children needs to be assessed. Implications for Policy, Delivery, or Practice: A significant proportion of children in child care in typical moderate to low income communities exhibit behavioral and social skills issues that put them at risk of early stigmatization, expulsion from child care, lack of school readiness, and longer term behavioral problems. Child care centers are important community settings were these children can be identified and primary and secondary intervention can be initiated. Our model tested an approach where child care centers engaged parents in initiating child and family services before more major interventions may be necessary. Modest levels of service and cost produced impression child outcomes. Systems for connecting child care centers to mental health and developmental services need to be developed to address this growing need. Primary Funding Source: Health Foundation of Central Massachusetts ●Cost Effectiveness of Comprehensive Eye Exams for PreSchool Children Alan White, Ph.D. Presented By: Alan White, Ph.D., Senior Associate, Abt Associates, 55 Wheeler Street, Cambridge, MA 02138; Tel: (617) 349-2489; Fax: (617) 349-2675; Email: alan_white@abtassoc.com Research Objective: Comprehensive eye exams and vision screening assessments are two methods for detecting amblyopia and other visual disorders in children. The purpose of this study is to estimate the cost effectiveness of comprehensive exams. Comprehensive exams cost more than screenings, but have higher rates of detection and treatment. Their cost effectiveness depends on whether the benefits resulting from the higher rates of detection and treatment of amblyopia offset the higher costs. Study Design: We assessed cost effectiveness using costutility analysis, comparing the marginal benefits of comprehensive exams (based on quality-adjusted life years (QALYs)) to marginal costs, including both exam and treatment costs. The model focuses on amblyopia, the only vision disorder typically identified during exams/screenings for which there was sufficient information for determining model parameter estimates. The model considers the relative performance of exams and screenings, the probability of successful treatment, costs, utility values associated with impaired vision, and patterns of treatment under usual care. Model parameters were based on an extensive literature review and input from a panel of experts. While the literature is extensive, there are significant gaps in the literature and differences across studies in their estimates of model parameters. We developed a set of base values, but these are subject to uncertainty given the range found in the literature. Sensitivity analyses were used to examine how results change using different values of model parameters. Population Studied: The model tracks individuals from age 3 (the earliest age at which children could receive a preschool exam) through their lifetime. Principal Findings: Using base case parameters, the incremental cost effectiveness (in terms of cost per QALY) of comprehensive exams was $12,796 relative to preschool vision screening. This is well below the conventional thresholds used to determine whether a particular intervention is a reasonable expenditure, indicating that investment in comprehensive eye exams for preschool children produces a greater return than other interventions. This basic conclusion was robust across changes in the values of most model parameters, although results were particularly sensitive to assumptions made about the quality of life associated with untreated amblyopia, prevalence, and the probability of successful treatment. The study’s conclusions are largely driven by the cost effectiveness of treatment for amblyopia. While data are limited, we estimate that treating amblyopia (not including detection) costs about $1,800 per QALY. As a result, spending additional dollars on interventions that detect and treat large numbers of children with amblyopia are highly cost effective. Conclusions: Comprehensive exams almost certainly detect more cases of amblyopia than do vision screenings, and, under a wide range of scenarios, comprehensive eye exams appear to be highly cost effective relative to vision screening programs, given existing standards of cost effectiveness. Implications for Policy, Delivery, or Practice: Policymakers should give consideration to programs that would increase the number of preschool children who receive a comprehensive eye exam. There are, however, important gaps in our knowledge, and a need for further research to make possible more precise estimates of the costs and benefits of comprehensive exams. ●Do Early Intervention Programs Alleviate Behavioral Problems In Children With Learning Disabilities? Jennifer Yu, Sc.D., Stephen Buka, Sc.D., Marie McCormick, M.D., Sc.D., Garrett Fitzmaurice, Sc.D., Alka Indurkhya, Ph.D. Presented By: Jennifer Yu, Sc.D., Research Fellow, Institute for Health Policy Studies, University of California, San Francisco, 3333 California Street, Suite. 265, San Francisco, CA 94143-0936; Tel: (415)514-0244; Fax: (415)476-0705; Email: jyu1@itsa.ucsf.edu Research Objective: Although early interventions such as Early Head Start programs demonstrate positive behavioral findings in children typically at risk for psychosocial maladjustment, it is unclear if such programs would produce similar results in other high-risk populations. This paper aims to fill this gap in the literature by examining the influence of an early intervention program received at ages 0-3 on the behavioral development of 8-year-old children with verbal (VLD) and nonverbal (NVLD) learning disabilities. Study Design: Secondary data analysis was conducted on a longitudinal study of a randomized clinical trial to examine the effects of early intervention on elementary-school aged children with learning disabilities. Logistic regression models were employed in which learning disability status predicted externalizing and internalizing behavior problems. Children were identified at age 8 as VLD or NVLD based on their discrepancy between cognitive performance and academic achievement in reading and arithmetic, respectively. Clinicallydefined externalizing and internalizing behavior problems were identified using the Achenbach Child Behavior Checklist. Population Studied: The study sample was derived from the Infant Health and Development Program, a multi-site, randomized clinical trial examining the effects of an early intervention on the cognitive, social, and emotional development of low birthweight infants. Among the 713 children included in these analyses, 11% were identified with VLD and 10% were identified with NVLD. Principal Findings: There was a significant interaction between VLD and the intervention when predicting internalizing behavior. Specifically, VLD children who received the early intervention were three times more likely to exhibit internalizing behavior problems when compared to VLD children who did not receive the intervention (p=0.01). This interaction was not significant for externalizing behavior problems in VLD children. In addition, the intervention did not significantly modify the relationships between NVLD and externalizing or internalizing behavior. Conclusions: The findings suggest that the early intervention had little impact on most clinical behavior problems in learning disabled children. The findings also suggest that early intervention programs may have iatrogenic effects on the behavioral outcomes of VLD children. Before drawing such conclusions, there are a number of factors to consider, such as methodological issues or the potential consequence of early intervention services without subsequent educational support. Implications for Policy, Delivery, or Practice: A number of implications are associated with these findings. For instance, most evaluations of early interventions recommend additional “booster” resources, such as counseling and family support programs, in order to extend the positive effects of the intervention into later life. This paper discusses how incorporating similar services into special education programs may benefit learning disabled children by reducing the risk of future behavioral problems. Primary Funding Source: RWJF ●Social Capital and Self-reported Negative Parenting Behaviors Adam Zolotor, M.D., MPH, Desmond K. Runyan, M.D., DrPH Presented By: Adam Zolotor, M.D., MPH, NRSA Primary Care Research Fellow, Family Medicine, University of North Carolina, CB# 7595, Chapel Hill, NC 27599-7595; Tel: (919) 843-4817; Fax: (919) 966-6125; Email: ajzolo@med.unc.edu Research Objective: The World Bank defines social capital as the glue that holds a society together. Social capital includes collective efficacy, psychological sense of community, neighborhood cohesion and parental investment in the child. In this study we develop a scale to assess social capital from a population-based survey of parents. We then evaluate social capital as a protective factor from harsh physical punishment, failure to meet a child’s basic needs, and psychologically harsh parenting. Study Design: This study is a secondary analysis of crosssectional telephone survey data. The survey assessed a range of parenting behaviors, demographic factors, and neighborhood characteristics. We factor-analyzed 22 questions designed to assess social capital. We then assessed the relationship of social capital to indices of parenting behavior representing harsh physical punishment, failure to meet a child’s basic needs, and psychologically harsh parenting. Population Studied: North and South Carolina Households were selected by random digit dial including only households with children under 18. A complementary sample of households known to have children was also surveyed. 1435 households participated in the survey for an adjusted response rate of 52%. The data is weighted to reflect the socioeconomic, racial, and ethnic composition of the Carolinas. Principal Findings: Three social capital variables were created from the factor analysis. These variables represent different aspects of social capital: neighborhood characteristics, willingness to take personal action to intervene on the behalf of neighbors, and recent history of helping a neighbor. We found that a willingness to take personal action was the strongest protective factor for harsh physical punishment (p=0.05). Neighborhood characteristics (p=0.04) and willingness to take personal action (p=0.02) were both protective from psychological harsh punishment (p=0.04 and 0.02 respectively). The variable for neighborhood characteristics was most protective against failure to meet a child’s basic needs (p=0.02). There was no significant relationship between a recent history of helping a neighbor and the three indices of parenting behavior examined. Conclusions: In this factor reduction of questions related to social capital, we are able to identify three factors most reflective of social capital. Both neighborhood characteristics and willingness to take personal action are protective factors against the use of harsh physical punishment, failure to meet a child’s basic needs, and psychologically harsh parenting. Implications for Policy, Delivery, or Practice: There is increasing interest among government and private agencies on the impact of social capital on health and well-being. In this study, we demonstrate an effective method of measuring three dimensions of social capital. We also show that neighborhood characteristics and a willingness to take personal action to intervene on the behalf neighbors are protective factors against negative parenting behaviors. These findings support the development of social capital as a resource in communities. Proposed methods for augmenting social capital include programs such as neighborhood centers, clean neighborhoods, social organizations, and community empowerment. Primary Funding Source: National Research Service Association