Child Health Call for Papers Assesing Pediatric Practice Guidelines and Quality Measures Chair: Kevin Dombkowski, University of Michigan Sunday, June 25 • 3:45 pm – 5:15 pm ●Indicators of Their Own: Pediatric Quality Indicators Based on Administrative Data Sheryl Davies, MA, Patrick Romano, M.D., M.P.H., Amy Ku, MHSA, Corinna Haberland, M.D., Jeffrey Geppert, JD, Kathryn McDonald, MM Presented By: Sheryl Davies, MA, Research Manager, Medicine, Stanford University, 117 Encina Commons, Stanford, CA 94305; Tel: 650-723-0820; Fax: 650-723-1919; Email: smdavies@stanford.edu Research Objective: Proliferating quality measurement systems have focused on adults rather than children. Hospitalized children represent an important population, for which most inpatient quality indicators are not applicable. This study aims to assess quality of inpatient pediatric care and access to high quality outpatient care using indicators adapted to a pediatric population, the AHRQ Pediatric Quality Indicators (PedQIs). Study Design: Utilizing literature review, semi-structured clinical panel review, and empirical exploration, we adapted the current AHRQ Quality Indicators, widely used indicators based on administrative data, for a pediatric population. We systematically reviewed the literature for evidence regarding coding and construct validity specific to the pediatric population. In a unique partnership with hospitals currently using the indicators, we also examined feedback from chart reviews of cases flagged as a complication by the original AHRQ QIs. Twenty-three revised indicators were then reviewed by clinical panels of 8-13 physicians, using methodology adapted from the RAND Appropriateness Method. We empirically identified high risk populations for six complications indicators, and developed stratification schemes. We then applied these new pediatric indicators to a nationally representative dataset. Downloadable software was developed for application to administrative datasets. Population Studied: We applied the PedQIs to the 2000 AHRQ Healthcare Cost and Utilization Project Kids’ Inpatient Database (KID) to generate weighted estimates for pediatric hospitalizations in nonfederal acute care facilities. We examined rates for children’s hospitals in addition to overall rates. Principal Findings: Panelists recommended 18 indicators for inclusion in the Pediatric Quality Indicator set based on overall usefulness ratings for internal quality improvement. The indicators included: 13 provider-level indicators, including 11 complications indicators (8 procedural), one mortality, and one volume, and 5 area-level potentially preventable hospitalization indicators. Panelists rated 10 provider-level and two area-level indicators as suitable for inter-hospital comparisons. Indicator rates for specific complications varied from 0.007 to 21.6 per 1000 discharges at risk. Foreign Body Left in During a Procedure and Transfusion Reaction were the least common complications; Postoperative Respiratory Failure and Postoperative Sepsis were the most common. Age trends were heterogeneous for complications indicators, with rates of Postoperative Sepsis and Postoperative Respiratory Failure decreasing with age, and Decubitus Ulcer increasing with age. For most indicators, children’s hospitals demonstrated higher complications rates, even when patients were stratified by clinical risk. Non-risk-adjusted mortality for Pediatric Heart Surgery was similar for children’s and nonchildren’s hospitals. Potentially preventable hospitalization rates varied from 19.6 (diabetes hospitalization) to 177.7 (gastroenteritis hospitalization) per 100,000 population in a given area (e.g., county). With the exception of diabetes, all potentially preventable hospitalizations decreased with age; diabetes admissions increased in the adolescent age category. Conclusions: Pediatric patients are at risk for developing potentially preventable complications and admissions. Tracking these events in children is an important aspect of improving quality of care. While children’s hospitals generally demonstrate higher rates even when stratified by risk, these differences may be mitigated with risk adjustment. Implications for Policy, Delivery, or Practice: The AHRQ Pediatric Quality Indicators are a novel set, widely applicable quality indicators. These indicators also incorporate clinical stratification important to quality improvement efforts. The indicators can be applied to readily available data. Primary Funding Source: AHRQ ●Reassessing the Cost Effectiveness of Vaccines for Young People Daniel Eisenberg, Ph.D., Gary Freed, M.D., M.P.H. Presented By: Daniel Eisenberg, Ph.D., Assistant Professor, Health Management and Policy, University of Michigan, M3517 SPH II, Ann Arbor, MI 48109-2029; Tel: (734) 615-7764; Fax: (734) 764-4338; Email: daneis@umich.edu Research Objective: The cost-effectiveness of some recently developed vaccines targeting children and adolescents, such as those for HPV or meningococcal disease, is controversial. If we are to use cost effectiveness analysis (CEA) as a decisionmaking aid regarding coverage and use of these vaccines, it is important that the method reflect society's preferences accurately. We analyzed empirically the extent to which CEA of vaccines as currently practiced may be biased against young people. In related literature this possibility has been supported by three arguments. First, net resource use, whether or not related to the intervention being studied, should be included in the analysis, as David Meltzer has shown. Second, the discount rate should be adjusted downwards to account for the fact that our society is steadily become wealthier and the marginal utility of health is consequently increasing. Third, several survey studies provide direct evidence that most people, when asked to make hypothetical trade-offs between a life-year for a young person and that of an elderly person, will strongly favor the young person. We synthesized these three strands of literature and applied corresponding adjustments to recently conducted CEAs of vaccines. Study Design: We conducted comprehensive searches of Medline and Harvard's CEA Registry for CEAs meeting the following criteria: 1) the intervention primarily affected people under age 25; 2) an incremental cost effectiveness ratio (ICER) was calculated in terms of $/QALYs or $/LYs; 3) the ICER result was between $50,000 and $500,000 (what we call "dubious cost-effectiveness"); 4) the study was published since 1995. For each CEA we then applied adjustments based on the three arguments described above and recalculated the ICERs. In sensitivity analyses we allowed for a range of assumptions regarding these adjustments. Population Studied: Based on the criteria above we selected 10 CEAs for re-analysis. These studies pertained to a range of diseases, including HPV, meningococcal virus, and hepatitis A, and to a range of age groups, including infants, toddlers, young children, and adolescents. Principal Findings: In our main analysis the proposed adjustments reduced the cost effectiveness ratios by as much as 50%, changing 8 of the 10 conclusions from dubious costeffectiveness to clear cost-effectiveness (well under $50,000/QALY). Even under the most conservative sensitivity analyses, the cost effectiveness of 6 of the 10 studies shifted to under $50,000/QALY. Conclusions: Reflecting societal preferences for prioritizing the health of children more accurately reversed the basic conclusion regarding whether an intervention is cost-effective from no to yes in the case of several recently conducted CEAs pertaining to vaccines. Implications for Policy, Delivery, or Practice: CEA is being used explicitly for rationing decisions in other countries, and some signs suggest that its prominence is increasing in the United States. Important policy decisions regarding child and adolescent vaccines, particularly those which are effective but expensive, will depend on assessments of cost effectiveness, among other factors. In this context, refining and using CEA methods to reflect societal preferences related to age as accurately as possible should be a high priority. Primary Funding Source: No Funding ●Risk Factors for RSV-Related Hospitalizations Among Infants Born at 32 to 35 Weeks of Gestation Norma Gavin, Ph.D. Presented By: Norma Gavin, Ph.D., Senior Research Economist, , RTI International, 3040 Cornwallis Road, Research Triangle Park, NC 27709-2194; Tel: (919) 541-6432; Fax: (919) 990-8454; Email: gavin@rti.org Research Objective: Prophylaxis with palivizumab has been demonstrated to be highly safe and effective in reducing hospitalization caused by RSV. However, given the large number of premature infants born at 32 to 35 weeks of gestation and the cost of prophylaxis, the American Academy of Pediatrics has recommended that use of the product in this cohort of infants be restricted to those infants with additional risk factors. This study examines the predictive utility of a variety of risk factors for identifying infants born at 32 to 35 weeks of gestation who are most likely to be hospitalized for severe respiratory syncytial virus (RSV) and therefore most likely to benefit from prophylaxis. Study Design: We used linked birth certificate and Medicaid enrollment and claims data to identify RSV hospitalizations and a variety of risk factors. We used two different definitions of RSV hospitalizations, one including pathogen-specific diagnoses only and the other including both pathogen-specific diagnoses and diagnoses for bronchiolitis and pneumonia of unspecified origin. The risk factors included the presence of siblings, birth within 6 months of the RSV season, birth stays of 7 or more days, neonatal intensive care unit stays at birth, ventilator assistance at birth, teenaged mothers, unmarried mothers, and mothers who smoked during pregnancy. We estimated unadjusted and adjusted odds ratios for each of the eight different risk factors and their sensitivity, specificity and positive and negative predictive values for predicting RSV hospitalizations in the first year of life as a single predictor and in combination with other factors. Population Studied: We included infants born 32 to 35 weeks of gestation in Texas during 1997 who had continuous Medicaid coverage from birth through the first 12 months of life. We excluded infants with missing or incomplete enrollment or claims records, infants enrolled in capitated managed care during their first year of life, infants with congenital anomalies, and infants who had any claims for RSV immune globulin intravenous or palivizumab. The final sample size was 2,098 infants. Principal Findings: Most of the factors investigated had a statistically significant, positive impact on the probability of RSV-related hospital stays among infants born at 32 to 35 weeks of gestation. However, their predictive power was limited; none of the factors, singly or in combination, was both sensitive and specific. Conclusions: Examined risk factors are not a compelling guide for selecting which premature infants should receive RSV prophylaxis. Implications for Policy, Delivery, or Practice: Current guidelines for decisions regarding which infants born at 32 to 35 weeks of gestation should receive prophylaxis for RSV hospitalizations may not be identifying those infants who would most benefit from the treatment. Primary Funding Source: MedImmune ●Development of Measures of the Quality of Emergency Department Care for Children Astrid Guttmann, MDCM, MSc, Asma Razzaq, M.P.H., Patrice Lindsay, Ph.D., Brandon Zagorski, MSc, Geoff Anderson, M.D., Ph.D. Presented By: Astrid Guttmann, M.D.CM, MSc, Scientist, , Institute for Clinical Evaluative Sciences, G Wing Sunnybrook Hospital, 2075 Bayview Ave, Toronto, Ontario, M4N 3M5; Tel: 416-480-4055 x 3783; Fax: 416-480-6048; Email: astrid.guttmann@ices.on.ca Research Objective: To use a systematic process to develop measures of the quality of Emergency Department (ED) care for children that are 1) based on research evidence and expert opinion, 2) representative of a range of conditions treated in most ED's, 3) based on a link between process and outcomes, and 4) feasible to measure. Study Design: We used a panel of providers and managers from a number of ED settings with reference to populationbased data of ED utilization by children, to identify a series of common conditions, across levels of patient acuity, which could be targets for quality improvement efforts. We used a structured panel process informed by a literature review to 1) identify condition-specific links between processes of care and defined outcomes, and 2) select and define measures to assess these process-outcome links. We determined the feasibility of calculating these measures using routinely collected population-based data of all ED visits for Ontario, Canada. Population Studied: All ED visits for children 0-19 years (1,226,849 visits) in Ontario, Canada for 2003/4. Principal Findings: A panel identified 18 clinical conditions for indicator development. The structured panel process identified 61 condition-specific links between process and outcomes for 14 of these conditions. After two rounds of ratings the panel defined 76 specific clinical performance measures for the following 12 conditions: ankle injury, asthma, bronchiolitis, croup, adolescent depression/mood disorders, diabetes, fever, gastroenteritis, minor head injury, neonatal jaundice, seizures, and urinary tract infections. These conditions covered a range of acuity and age groups, and accounted for 38 % of all pediatric ED visits. Using routinely collected data it was possible to calculate 19 measures, covering 9 conditions, representing 21% of all visits. There was significant variability in the measures across hospitals in Ontario. Conclusions: Performance measures are essential components of public reporting and quality improvement. Using a structured panel process, data on ED utilization and a literature review, it was possible to identify common clinical conditions for which processes in the ED have an impact on outcomes, define and measure some indicators of care using existing administrative data. Developing benchmarks for these measures will require a number of strategies. Implications for Policy, Delivery, or Practice: This is the first set of comprehensive performance measures of pediatric ED care relevant across all ED settings. Locally these measures will be used in the Ontario Hospital Report which routinely reports on the quality of care of all acute care hospitals in Ontario. This list of measures could be used by other jurisdictions interested in reporting ED performance either as is, or as a starting point for similar panel processes with relevant stakeholders. The list also provides ED managers and clinicians with tools to focus quality improvement. Primary Funding Source: Ontario Hospital Report Research Collaborative ●Improving Severity Adjustment Measures for the Quality of Care Provided to Premature Infants by Outpatient Pediatric Practices Scott Lorch, M.D., MSCE, Gabriel J. Escobar, M.D., Andrea Millman, BA, Susan Bakewell-Sacks, RN, Ph.D., Barbara Medoff-Cooper, RN, Ph.D., Jeffrey H. Silber, M.D., Ph.D. Presented By: Scott Lorch, M.D., MSCE, Assistant Professor, Neonatology and the Center for Outcomes Research, 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: With the growth of pay-for-performance plans, severity adjustment for outcomes such as the rehospitalization of premature infants has become an important issue for outpatient pediatric care. Traditional severity adjustment tools for these infants, such as gestational age (GA) and the score for neonatal acute physiology (SNAP), are measured at delivery and may not adjust for lateroccurring differences in neonatal practice. The objective of this study, then, was to (1) evaluate the discrimination of various severity adjustment tools for the risk of rehospitalization of premature infants within 2 and 4 weeks of discharge from the neonatal intensive care unit and (2) determine how the addition of physiologic maturity at discharge, a measure of later-occurring differences in the condition of neonates, influences the rank ordering of 31 outpatient pediatric practices. Study Design: As part of the Infant Functional Status Study, we determined the number of days prior to discharge that an infant was monitored without supplemental heat or oxygen, apnea or bradycardia, methylxanthines, naso-gastric feedings, or weight loss. Logistic regression models were constructed to establish the ability of various sets of factors to predict rehospitalizations within 2 and 4 weeks of discharge; these sets of factors included (1) GA and race; (2) GA, race, and SNAP; (3) GA, race, and complications of prematurity; and (4) GA, race, complications of prematurity, SNAP, and physiologic maturity at discharge. These models were then used to rank order outpatient facilities on the basis of 2 and 4week rehospitalization rates. Population Studied: Premature infants born at a GA <= 32 weeks at 5 Kaiser Permanente hospitals (N=780) between 1998 and 2001. Principal Findings: Of the 780 infants, 280 had a GA <= 28 weeks. The median number of days from attainment of each skill to discharge ranged from 7 days (IQR 5-11 days) for feeding to 20 days (IQR 14-27 days) for methylxanthines. Compared to GA and race (c-statistic 0.71), the addition of complications of prematurity improved model discrimination for 2-week rehospitalizations (c-statistic 0.79, p=0.002). The addition of maturity variables resulted in the highest discrimination for both the 2-week (c-statistic 0.84, p<0.02 versus all other models) and 4-week rehospitalization models (c-statistic 0.78). The addition of the maturity variables to the severity adjustment model with GA, SNAP, race, and complications resulted in 26% of the outpatient facilities changing rank order quintiles for 2-week rehospitalizations and 19% of the facilities changing quintiles for 4-week rehospitalizations. 11 of the 14 hospitals that changed quintiles moved at least 2 ranks. 26% and 13% of the facilities in the 2 and 4-week rehospitalization models, respectively, moved at least 10% in the final rank order when the model accounted for physiologic maturity. Conclusions: The type of severity adjustment can significantly alter the ranked quality of care provided to premature infants by outpatient pediatric practices. Models that include information near discharge may improve the severity adjustment of outpatient pediatric facilities. Implications for Policy, Delivery, or Practice: Administrative data that includes information about the physiologic maturity of premature infants at the time of discharge could improve the assessment of the quality of care delivered by outpatient pediatricians. Primary Funding Source: Maternal and Child Health Bureau Call for Papers Children in the Context of School, Community & Family Chair: Michael Cabana, University of California, San Francisco Monday, June 26 • 3:45 pm – 5:15 pm ●The Effectiveness of the Olweus Bullying Prevention Program in Public Middle Schools: A Controlled Trial Nerissa Bauer, M.D., Paula Lozano, M.D., M.P.H., Frederick P. Rivara, M.D., M.P.H. Presented By: Nerissa Bauer, M.D., Pediatric NRSA Senior Fellow, General Pediatrics, University of Washington/Child Health Institute, 6200 NE 74th Street, Suite 210, Seattle, WA 98115-8160; Tel: 206-616-2898; Fax: 206-616-4623; Email: nerissa@u.washington.edu Research Objective: The Olweus Bullying Prevention Program, OBPP, is a widely-disseminated school-based intervention designed to reduce student bullying and promote healthy peer relationships. However, only 1 study has tested its effectiveness in US schools. Therefore, our objective was to determine the effectiveness of the OBPP in decreasing student reports of victimization and increasing students’ willingness to intervene if another student is being bullied. Study Design: A controlled trial with 10 middle schools was conducted between 2003 and 2005. We assessed studentreports of relational (spreading rumors, social exclusion) and physical victimization, pre- and post-implementation. Students were asked how they felt when they witnessed bullying and whether other students and teachers intervened. Poisson regression was performed, controlling for baseline prevalence and school characteristics such as size, percent free lunch eligible and reading test pass rate. Population Studied: Seattle public middle school students in 6th through 8th grade. Principal Findings: Seven middle schools implemented the OBPP (intervention group) and 3 did not (comparison group). The OBPP was not associated with significantly different overall rates of relational and physical victimization. Yet when stratified by ethnicity, reports of relational and physical victimization decreased by 28% (RR=0.72, 95% CI 0.53-0.98) and 37% (RR=0.63, 95% CI 0.42-0.97) respectively among white students in intervention schools relative to white students in comparison schools. No similar effect was found for students of other races/ethnicities; there were no differences by gender or by grade. Students in intervention school were more likely to perceive other students as actively intervening in bullying incidents (RR=1.21, 95% CI 1.05-1.40) and 6th graders were more likely to feel sorry and want to help victims (RR=1.21, 95% CI 1.05-1.39). Conclusions: The OBPP had some positive effects varying by race/ethnicity and grade on victimization and feelings towards bullying, but did not show an overall effect. The intervention was associated with greater perception of students actively intervening in bullying incidents. Implications for Policy, Delivery, or Practice: Schools should tailor anti-bullying measures to students of all ethnicities and cultures. Future studies of large-scale violence prevention programs in the community must be rigorously evaluated. Primary Funding Source: HRSA ●Acculturation, Generational Status, Place of Birth, Primary Language at Home, and Overweight: A Case of “When in Rome, Do As The Romans Do?” Jane M. Brotanek, M.D., M.P.H., Hong Ji, MS, Emmanuel Ngui, DrPH, MS, Glenn Flores, M.D. Presented By: Jane M. Brotanek, M.D., M.P.H., Assistant Professor, Pediatrics, Medical College of Wisconsin, 8701 Watertown Plank Road, MS 756, , WI 53226; Tel: (414) 4565778; Fax: (414) 456-6385; Email: jbrotane@mcw.edu Research Objective: Along with the overweight epidemic in US adolescents, there is a dramatic surge in the immigrant population. Little is known about whether acculturation, generational status, birthplace, or primary language are associated with overweight in US adolescents. The study objective was to examine whether acculturation, generational status, place of birth, or primary language are associated with overweight in US adolescents. Study Design: Analyses were performed of the National Health and Nutrition Examination Survey (NHANES: 19992002) for a nationally representative sample of US adolescents 12-18 years old. Measures included acculturation (using the Short Acculturation Scale), generational status, birthplace, primary language spoken at home, and overweight/at risk of overweight (OWARO: BMI greater than or equal to the 85th percentile). Multivariate analyses were done to adjust for gender, poverty, and race/ethnicity. Population Studied: A nationally representative sample of US adolescents 12-18 years old Principal Findings: 92% of the 4,339 adolescents were USborn and 8% were foreign-born; 90% had English as their primary language (EPL) and 10%, a non-English primary language (NEPL). Among Latino adolescents, 73% had high acculturation and 27%, low acculturation. 21% of foreign-born adolescents and 33% of US-born adolescents were OWARO (p=.001). 32% of EPL adolescents and 29% of NEPL adolescents were OWARO (p=0.5). 30% of 1st-generation Latino adolescents, 41% of 2nd-generation Latino adolescents, and 38% of 3rd-generation Latino adolescents were OWARO (p=.4). 35% of Latino adolescents with low acculturation and 39% of those with high acculturation were OWARO (p=.2). In multivariate analyses, US-born (OR, 1.8; 95% CI, 1.1-3.0) and EPL (OR, 1.6; 95% CI, 1.1-2.2) adolescents had higher odds of OWARO; among Latino adolescents, being 2nd-generation was associated with OWARO (OR, 2.0; 95% CI, 1.2-3.2). Conclusions: US-born and EPL adolescents are at significantly higher risk of overweight. Among Latino adolescents, generational status, but not acculturation, is associated with OWARO. Further research is needed to identify how generational status, place of birth, and primary language spoken at home might be protective against adolescent overweight. Primary Funding Source: The Center for the Advancement of Underserved Children and the Department of Pediatrics, Medical College of Wisconsin ●The Relationship of Parental Psychopathology and PainRelated Medical Problems to Child Medical Management and Avoidable Emergency Pediatric Services Joseph Logan, MAE, M.H.S., Ph.D. candidate Presented By: Joseph Logan, MAE, M.H.S., Ph.D. candidate, Doctoral Student, Health Policy and Management, The Johns Hopkins Bloomberg School of Public Health, 695 Americana Drive, #46, Annapolis, MD 21403; Tel: 443-722-2194; Email: jlogan@jhsph.edu Research Objective: Studies in health services research have shown that many children do not receive well-child visits or timely care when they are in need of services. Also, a large percentage of emergency department (ED) visits and hospitalizations among children occur for medical problems that could have been avoided had they received timely nonemergency based outpatient care. These avoidable ED visits and hospitalizations are often referred to as Ambulatory Care Sensitive or ACS-ED visits and ACS hospitalizations. Children of parents with mental health or pain-related conditions may be at greater risk for having medical mismanagement, especially if the parent’s symptoms impair their ability to monitor health, communicate, seek care, provide transportation, and/or navigate the health system. This study investigated which types of parental psychiatric and painrelated conditions are associated with poor use of outpatient pediatric services and increased occurrences of child ACS-ED visits and child ACS hospitalizations. Study Design: This investigation was a secondary data analysis of the 1997-1998 Thomson/Medstat® MarketScan claims and administrative dataset. Three logistic regression models using generalized estimating equations (GEE) to control for children being in the same family were analyzed to assess the odds of a child having at least two well-child visits, at least one ACS-ED visit or one ACS hospitalization during the 2-year study period in relation to parental diagnoses of the various mental health conditions and pain-related conditions including back pain, migraine headaches, and arthritis. This analysis controlled for child and parental health status, personal demographics, family characteristics, and health plan characteristics. Population Studied: A study population of 258,332 children and 262,799 caregivers enrolled in private health insurance plans was included for this analysis. Principal Findings: Maternal depression was inversely associated with a child having two well-child visits (OR: 0.87, p<0.05), and positively associated with a child having at least one ACS-ED visit (OR: 1.24, p<0.001) and a child having at least one ACS hospitalization (OR: 1.24, p<0.10). Maternal depression with back pain was strongly associated with a child having an ACS-ED visit (OR: 1.78, p<0.001) and a child having an ACS hospitalization (OR: 2.5, p<0.001). These results were similar to those of mothers with four or more medical conditions. Paternal schizophrenia was significantly and inversely associated with a child having two well-child visits (OR: 0.25, p<0.10). Paternal depression was also significantly and positively associated with a child having at least one ACSED visit (OR: 1.15, p<0.10) and a child having at least one ACS hospitalization (OR: 1.35, p< 0.10). Both parents diagnosed with depression or each having three or more medical conditions was significantly and positively associated with a child having at least one ACS-ED visit (OR: 1.4-1.88, p<0.05), and a child having at least one ACS hospitalization (OR:2.13.00, p<0.05). Conclusions: Parental depression and schizophrenia and multiple medical conditions are both independently and strongly associated with inadequate well-child care and avoidable emergency pediatric services, especially if cooccurring with pain-related health conditions. Implications for Policy, Delivery, or Practice: Mental health providers, pediatricians, policy makers, and insurers need to consider parental mental health as an influencing factor when seeking to improve medical management of potentially serious conditions in children. Primary Funding Source: NIMH ●Are Non-Citizen Parents More Likely than Citizen Parents to Choose Emergency Departments for Their Children’s Non-Urgent Care? Emmanuel Ngui, Dr.P.H., Glenn Flores, M.D. Presented By: Emmanuel Ngui, Dr.P.H., Assistant Professor of Pediatrics, Epidemiology, and Health Policy, Department of Pediatrics, Medical College of Wisconsin, 8701 Watertown Plank Rd, Milwaukee, WI 53226; Tel: 414 456-4302; Fax: 414 456-6385; Email: engui@mcw.edu Research Objective: Emergency department (ED) use for non-urgent care (NUC) is a significant public health problem. Not enough is known about parental use of EDs for children's NUC, particularly among immigrant parents. The study objectives were to 1) compare parental choice of EDs for children’s NUC among citizen parents (CPs) and non-citizen parents (NCPs); 2) examine whether the choice of EDs for children’s NUC differ across the different NCPs subgroups (legal permanent resident parents [LPRPs], undocumented immigrant parents [UIPs], and temporary resident parents [TRPs]); and 3) identify factors associated with disparities in parental choice of using ED for NUC among CPs and NCPs. Study Design: Cross-sectional survey of parents in predominantly Latino inner-city Boston communities, including supermarkets, hair salons, and Laundromats. Participants were asked 76 questions on topics such as insurance coverage, barriers to care, and choices regarding NUC. Multivariate analyses adjusted for age, insurance, education, poverty status, usual source of care, Limited English Proficiency (LEP), and barriers to care. Population Studied: Inner-city citizen and non-citizen parents. Principal Findings: 1100 parents were interviewed; 40% were CPs; 29%, LPRPs; 24%, UIPs; and 7%, TRPs. Parental choice of the ED for NUC differed significantly by parental citizenship status, with only 20% of CPs choosing EDs for children's NUC vs. 48% of NCPs. Among NCPs, the ED was chosen for children's NUC by 35% of LPRPs, 58% of UIPs, and 60% of TRPs. In multivariate analyses, NCPs were significantly more likely than CPs to choose EDs for children's NUC (OR=1.6 [1.12.3]). Among NCPs, a significant gradient was found with LPRPs the least likely to choose ED for children’s NUC (OR=1.3 [0.8-1.9]), followed by UIPs (OR=2.0 [1.3-3.3]), and TRPs (OR=2.9 [1.6-5.4) at greatest risk. For CPs, the number of siblings (OR=1.6 [1.3-1.9]), uninsurance (OR=6.7 [1.8-25.2]), unemployment (OR=4.6 [2.1-9.9]), and LEP (OR=7.4 [3.7-14.8]) were all associated with greater likelihood of choosing ED for children’s NUC. Among NCPs, uninsurance (OR=1.9 [1.1-3.4], lower educational attainment (OR=1.7 [1.2-2.5]) and poverty (OR=2.7 [1.7-4.3]) were associated with choosing the ED for NUC. Parental unemployment (OR=0.5 [0.3-0.8]) and long waits to see a doctor (OR=0.6 [0.4-0.9]) were associated with a lower likelihood of choosing EDs for NUC among NCPs. Conclusions: NCPs are significantly more likely than CPs to choose EDs for children's NUC, with UIPs and TRPs at highest risk. Poverty, uninsurance, and lower parental educational attainment significantly increase the likelihood of NCPs, particularly UIPs and TRPs, choosing EDs for their children’s NUC. In contrast, unemployment, LEP, and the number of siblings in the household increase the likelihood of choosing EDs for children’s NUC among CPs. Implications for Policy, Delivery, or Practice: Educational programs that increase NCP’s (particularly UIPs and TRPs) awareness, knowledge, and skills in responding to children’s NUC needs and the appropriate use of EDs also are needed. Our findings that suggest efforts to reduce poverty, and increase insurance coverage and education attainment among NCPs are likely to reduce the use of EDs for children’s NUC. Primary Funding Source: No Funding ●School Mental Health Services in the United States, 2002-2003 Judith Teich, MSW, Susan Foster, MSSW, M.P.H., Gail Robinson, Ph.D., Mary Rollefson, Ph.D., Theresa Doksum, Ph.D., M.P.H., Denise Noonan, Ph.D. Presented By: Judith Teich, MSW, social Science Analyst, Center for Mental Health Services, SAMHSA, 1 Choke Cherry Road, Rockville, MD 20852; Tel: (240)276-1967; Fax: (240)2761340; Email: judith.teich@samhsa.hhs.gov Research Objective: This study, “School Mental Health Services in the United States, 2002-2003,” provides the first national survey of mental health services in a representative sample of the approximately 83,000 public elementary, middle, and high schools and their associated school districts in the United States. It provides new information about the role of schools in providing mental health services, and how these services are organized, staffed, funded, and coordinated. Study Design: The survey methodology included two mail questionnaires that were developed for this project with the guidance of an expert panel. Questions concerned services and supports delivered to students who have been referred and identified as having psychosocial or mental health problems. The school questionnaire collected data on the types of MH problems encountered in schools, the MH services provided, the types and qualifications of staff providing services, the type and degree of care coordination, and the arrangements for delivering MH services. The district questionnaire collected data on funding sources for MH services and issues related to funding. The survey also solicited impressions from school administrators and MH personnel concerning issues affecting school MH services. Population Studied: School Mental Health Services in the United States, 2002–2003 is a nationally representative sample of public K–12 schools and their associated school districts. A random sample of 2,125 schools and the 1,595 districts associated with them was drawn from the U.S. Department of Education’s public school data file, the Common Core of Data for 2000–2001. The size of the sample was designed to provide reliable estimates of the universe of regular public schools by level (elementary, middle, and high school) and by size, as measured by student enrollment: small (from 1 to 250 students); medium (251–500 students); large (501–1,000 students); and very large (1,001 and more students). The sampling strategy was also designed to yield estimates by each region (Northeast, Midwest, South, West) and locale (urban/central city, suburban/large town, small town/rural) and to populate the standard table shell used in this report. Ultimately,1,147 schools in 1,064 districts across the country responded to the survey. “Critical items” information was collected from an additional 150 schools. The quantifiable data were weighted to create national estimates for numbers of schools and districts by region and by size. This was done so that the total numbers and the distributions would match those of all schools and districts in the nation in the 2002–2003 school year. The final weighted response rates were 60.5 percent for schools and 59.85 percent for school districts. Principal Findings: • Nearly three-quarters (73%) of the schools reported that “social, interpersonal, or family problems” were the most frequent MH problems for both male and female students. • For males, aggression or disruptive behavior and behavior problems associated with neurological disorders were the second and third most frequent problems, while anxiety and adjustment issues were the next most frequent problems for females. • All students, not just those in special education, were eligible to receive MH services in the vast majority of schools (87%). • On average, one-fifth of students received some type of schoolsupported mental health services in the school year prior to the study. • Virtually all schools reported having at least one staff member whose responsibilities included providing mental health services to students. • The most common types of school mental health providers were school counselors, followed by nurses, school psychologists (non-Ph.D.), and social workers. School nurses spend approximately a third of their time providing mental health services. • More than 80 percent of schools provided assessment for mental health problems, behavior management consultation, and crisis intervention, as well as referrals to specialized programs. A majority also provided individual and group counseling, and case management. • Financial constraints of families and inadequate school mental health resources were the most frequently cited barriers to providing mental health services. • Almost half of school districts (49%) used contracts or other formal agreements with community-based individuals and/or organizations to provide mental health services to students. The most frequently reported community-based provider type was county mental health agencies. • Districts reported that the most common funding sources for mental health services or interventions were the Individuals with Disabilities Education Act (IDEA); State special education funds, and local funds. In 28 percent of districts, Medicaid was among the top five funding sources for mental health services. • One third of districts reported that funding for mental health services had decreased since the beginning of the 2000-2001 school year, while over two-thirds of districts reported that the need for mental health services increased. • Sixty percent of districts reported increased referrals to community-based providers; one third reported that the availability of outside providers to deliver services to students had decreased. Conclusions: This study, School Mental Health Services in the United States, 2002–2003, provides the first source of comprehensive information on the mental health services provided in the approximately 83,000 public elementary, middle, and high schools and their associated school districts in the United States. The study's findings confirm that mental health services currently play an integral role in the school setting. The findings also suggest that needs for mental health services are increasing, and that adequate funding and availability of community resources are essential if schools are to meet the challenge of addressing these needs. Implications for Policy, Delivery, or Practice: This survey is the first comprehensive, nationally representative survey of mental health services in U.S. public schools. Obtaining an unbiased sample of schools and districts was a timeconsuming, labor-intensive process. While the survey provides important baseline information, it also leaves many questions unanswered and raises additional questions. It provides measures of the mental health problems encountered in school settings, but it does not address how many students present with each type of problem and how these problems differ by demographic background characteristics of students. While the survey found widespread eligibility for mental health services and a surprisingly wide array of services provided, the measures of service eligibility and provision are fairly broad. The survey revealed the percentage of schools offering various types of services, but it did not ask about the intensity or duration of different services, which services were provided for which mental health problems, the adequacy or appropriateness of the services to the needs of the students, or the degree to which the need for various services was met. Further, the survey gathered only limited information about prevention. Future research should be conducted to guide school officials regarding the most effective combinations of prevention and intervention services for their schools. The individuals who provide mental health services in schools include specialists with graduate degrees and licensure in their fields as well as nurses and paraprofessionals, but this survey does not address whether these staff have specific qualifications to treat the major presenting problems at each school level. Several open-ended comments also pointed to the inadequacy of staff-to-student ratios. Future research should measure specific staff and service assignments, the qualifications of staff to provide those services, and the professional development needs and experiences of staff. It would be important to learn more about the amount of funding allocated to different types of prevention, assessment, and treatment services according to the number of children served and their presenting problems. Such an analysis would shed light on equity of funding, disparities between wellresourced and under-resourced schools, and would guide policymakers in their decisions about funding allocation for mental health services. The survey identified the top sources of funding for mental health services and which sources were used for prevention versus intervention services. However, additional information is needed about the specific services that various funding streams support. Further, more research is needed to determine the distribution of funding and other resources for mental health services by region, urbanicity, minority enrollment, and other school characteristics. Primary Funding Source: Other Related Posters Child Health Poster Session A Sunday, June 25 • 2:00 pm – 3:30 pm ●Why Do Parents Bring Children to the Emergency Department (ED) For Nonurgent Conditions? Anne Berry, BA, David C. Brousseau, M.D., Jane Brotanek, M.D., Sandra C. Tomany-Korman, MS, Glenn Flores, M.D. Presented By: Anne Berry, BA, Medical Student, Penn State College of Medicine, 321 University Manor West, Hershey, PA 17033; Tel: 717-531-7880; Email: aberry@psu.edu Research Objective: Emergency department (ED) use increased 20% from 1992 - 2001, with children comprising 1/4 of all ED visits, or 22 million visits per year. 58-82% of pediatric ED visits are for nonurgent conditions. Demographic characteristics of families using the ED for pediatric nonurgent conditions are well described, but the reasons parents seek this care in the ED have not been identified, and no published qualitative research has examined this issue. The objective of this study was to determine why families use the ED for nonurgent pediatric care by identifying parents’ reasons for choosing the ED over primary care. Study Design: Ethnographic interviews of parents of children presenting for nonurgent care. Parents were asked 20 questions using a semi-structured interview guide. Interview transcripts were analyzed using grounded theory. Population Studied: Parents of children presenting for nonurgent care on weekdays from 8 am - 4 pm at a children’s hospital ED over a 4-week period Principal Findings: For the 31 families (37 parents) interviewed, the mean parental age was 29 years, and the mean child age was 3 years. English was the primary language for 30 parents and Spanish for 7. All patients were insured (68% publicly insured), and only one child did not have a primary care physician (PCP). Reasons identified by caregivers for choosing the ED over their child’s PCP were: healthcare provider referral (“If my doctor recommended it, that was the only reason I went”), long appointment wait (“He wasn’t able to get him in… today or tomorrow, he has so many patients.”), efficiency (“you can’t beat it for speed”), ED resources (“They can perform a test right there”), convenience, quality of care (“I feel that the doctors [in the ED] are more skilled.”), dissatisfaction with the PCP (“They have bad attitudes there”), communication problems (accents and unhelpful staff at PCP), and ED expertise with children. Some parents said they would like education on the urgency of pediatric problems. When asked if her child’s PCP provided such teaching, one mother replied, “It would be helpful, but no.” Conclusions: Parents bring their children to the ED for nonurgent care because of PCP referral, efficiency, dissatisfaction with PCPs, higher quality of care, long waits to see PCPs, and PCP communication problems. Implications for Policy, Delivery, or Practice: Although parents report that education on the urgency of pediatric conditions would be helpful, substantial reduction of pediatric nonurgent ED use may require improvements in families’ PCP office access, efficiency, experiences, and appointment scheduling. In particular, the study findings indicate that the most useful measures might include enhanced patientprovider communication, extended office hours, access to a "fast-track" option in practices for urgent care, and periodic quality assurance review of reasons for ED referrals. Primary Funding Source: APS/SPR Student Research Program ●The Use of Hospital Discharge Data in the Study of Births and Deliveries: Results from the 2003 Nationwide Inpatient Sample. Thomas M. Brady, Ph.D., Claudia A. Steiner, M.D., M.P.H., Chaya T. Merrill, M.P.H. Presented By: Thomas M. Brady, Ph.D., Lead Health Services Researcher, Thomson Medstat, 4301 Connecticut Ave., Suite 330, Washington, DC 20008; Tel: (202) 719-7871; Fax: (202) 719-7801; Email: thomas.brady@thomson.com Research Objective: The National Center for Health Statistics birth certificate data (BCD) are widely considered the gold standard for information regarding births in the United States. In addition to BCD, hospital discharge data are also useful in the analysis of births and deliveries. This paper compares a national sample of BCD with a national data set of hospital discharge data, the Nationwide Inpatient Sample (NIS) of the Healthcare Cost and Utilization Project (HCUP). The aim of the paper is to evaluate the HCUP NIS and provide a valuable resource for researchers who wish to use national discharge data to study births and deliveries. Study Design: The number of hospital-based births and deliveries presented in the NCHS’ report Births: Final Report for 2003 are compared to the counts of live births sampled in the 2003 NIS. Births, deliveries and associated conditions are identified using the International Classification of Diseases, 9th Revision, Clinical Modification (ICD-9-CM) codes in the analysis of NIS data. Rates of low birth weight and prematurity from both national datasets are compared. Comparisons of obstetrical procedures, labor and/or delivery complications, newborn abnormal conditions and congenital anomalies are also presented. Rates are calculated using 2003 NCHS data as the population denominator. Population Studied: All hospital-based U.S. births and deliveries reported in 2003. Principal Findings: Findings from a preliminary analysis of 2002 data are available. In that year, national birth certificate data systems reported 4,021,726 live births compared to 4,154,343 in the NIS. National estimates from the NIS revealed a greater prevalence of abnormal conditions and congenital anomalies of the newborn compared to birth certificate data in 2002. Rates of anencephalus (0.1/1,000 births in both the BCD and NIS) and spina bifida (0.2/1,000 births in the BCD and 0.3/1,000 births in the NIS) were similar between the two data systems. Reports of birth injuries, however, were more than ten times more prevalent in the NIS than the NCHS data set. The rate of birth injuries in the BCD was 2.7/1,000 births and the rate in the NIS was 30.4/1,000 births. In addition, among newborn congenital anomalies, there were much higher estimates of heart malformations and malformed genitalia in the NIS compared to the NCHS data. Findings using 2003 data, including a comparison of frequencies of obstetrical procedures and labor and delivery complications, will be presented. Conclusions: In comparing the two data systems, the total number of annual births is similar. The preliminary findings suggest that there are significant differences in some estimates of abnormal conditions and congenital anomalies of the newborn between the NIS data as compared to birth certificate data. The paper will discuss these findings and potential explanations for the observed differences. The project will also explore the relative strengths and weaknesses of utilizing hospital discharge data in the study of births and deliveries. Implications for Policy, Delivery, or Practice: Our preliminary findings support the use of hospital discharge data as a powerful and reliable alternative data source to the BCD to monitor and study conditions of the newborn as well as track the utilization of perinatal health care services. Primary Funding Source: AHRQ ●Impact of Family Structure on Children's Ambulatory Care Alex Chen, M.D., MS, Jose Escarce, M.D., Ph.D. Presented By: Alex Chen, M.D., MS, Assistant Professor, Pediatrics, Childrens Hospital Los Angeles, 4650 Sunset Blvd. MS # 30, Los Angeles, CA 90027; Tel: 323-671-7620; Fax: ; Email: achen@chla.usc.edu Research Objective: To examine how family structure impact children's use of ambulatory care services including office visits, emergency room visits, and prescription medications. Study Design: This is a secondary data analysis of the 19962001 Medical Expenditure Panel Survey (MEPS). We used negative binomial regression to model office visits and emergency room visits and logistic regression to model the likelihood of prescription medication use. Our analyses adjusted for demographic and socioeconomic characteristics as well as measures of children’s health and parental education and child-rearing experience. Population Studied: The study consisted of a nationally representative sample of children 0-17 years of age living in single-mother or two-parent families. Principal Findings: Living in single-mother family was strongly associated with decreased ambulatory care services use compared to living in two-parent family; however, the effect exerted by single-mother vs. two-parent status was completely explained away after adjustments for race, insurance coverage, family income, and maternal education. In addition, we found that children living in families with many other children had fewer total and physician office visits and a lower likelihood of using a prescription medication than children living in families with no other children. Children who lived with other adults in addition to their parent (such as cohabitation) also had fewer office visits and a lower likelihood of using a prescription medication. Conclusions: Children living in families with many other children or with other adults use less ambulatory care and prescription medications than their peers. Implications for Policy, Delivery, or Practice: Our study findings suggest that children can be vulnerable to decreased ambulatory care and possibly unmet need as a result of their family structure. It would be helpful to raise awareness among health care providers so that they can monitor children in cohabiting households or children with many siblings more closely for follow-up appointments. In addition, policies that provide support for working parents with many children, such as family leave days, may enable them to take their children for care while alleviating some of the stress and demand on their time Primary Funding Source: AHRQ ●Effectiveness of Kentucky's Mandated Preschool Vision Specialist Exam Julia Costich, J.D., Ph.D., William Reynolds, O.D., Svetla Slavova, M.S. Presented By: Julia Costich, J.D., Ph.D., Chair, Dept. of Health Services Management, College of Public Health, 121 Washington Ave., Lexington, KY 40536; Tel: 859-257-6712; Fax: 859-257-3909; Email: julia.costich@uky.edu Research Objective: To determine whether the comprehensive cycloplegic vision examination mandated by Kentucky law prior to school entry is effective in detecting vision problems that are missed on primary care provider vision screenings. Study Design: A survey was fielded with 46 optometrists and ophthalmologists who agreed to participate, resulting in 1,391 complete responses. Data were encoded and analyzed using SAS to determine correlations among findings. An algorithm was developed to identify children whose vision problems were not identified by screening in preschool physicals. Significance and confidence levels were calculated. Vision problems were identified using widely accepted professional guidelines. Population Studied: Kentucky children ages 3-6 years. Children eligible for publicly funded health coverage were disproportionately represented (56% v. 37% statewide). Principal Findings: Some 300 children were identified as having vision problems, 72 of whom had been screened in the context of preschool exams and had not previously seen an eye care specialist. Only 9 identified vision problems had been found in the preschool physicals (95% confidence interval: .76-.93). Supplementary funding sources for children without coverage for the exams were underutilized. Nearly half of parents reported no vision screening in preschool examinations. Conclusions: Kentucky's mandated preschool vision examination identified a significant number of children with previously undetected vision problems, even among those children whose vision was screened in their preschool physicals. Implications for Policy, Delivery, or Practice: While comprehensive vision examinations were significantly superior to screenings in detecting vision problems, the contribution of early detection to school performance or remediation has not been clearly demonstrated. Nevertheless, these findings indicate that broader adoption of and coverage for preschool examinations by a vision care specialist would help children overcome vision-related barriers to success in school. Primary Funding Source: Vision Council of America ●An Examination of Specialization and Outcomes of Care at Children's Hospitals: Are Children's Hospitals Better Qualified to Care for Children? Jami DelliFraine, Ph.D. Presented By: Jami DelliFraine, Ph.D., Assistant Professor, Health Policy and Administration, The Pennsylvania State University, 116 Henderson, University Park, PA 16802; Tel: (814) 863-2861; Email: jld40@psu.edu Research Objective: Objective. To test the hypotheses that 1) children’s hospitals provide more specialized care than general hospitals, 2) specialization is associated with better outcomes of care for children, and 3) some organizational factors are associated with better outcomes for children. Study Design: Study Design. A structural equation model with latent variable analysis assessed the risk-adjusted outcomes differences for children admitted to children’s hospitals and general hospitals. Key variables in the analysis included children’s hospital status, a latent variable representing organizational specialization, and overall riskadjusted mortality rates for children. Population Studied: Population Studied. This study used the 2000 Healthcare Cost and Utilization Project KID database and the 2000 American Hospital Association database. The study examined 464 hospitals that provide children’s care in nineteen metropolitan areas. Data were extracted from the secondary data sources using SPSS 13.0 and AMOS 5.0 statistical software. Principal Findings: Principal Findings: The results of this study indicate that children’s hospitals are more specialized in children’s care than most general acute care hospitals. However, no significant overall mortality differences were found between children treated at children’s hospitals and children treated at general acute care hospitals. Conclusions: Conclusions: Findings support the premise that children’s hospitals offer the most specialized care for children and treat the sickest children. The results suggest that organizational factors influence specialization in children’s care. However, more research is needed examining outcomes and quality differences between hospitals that treat children. Implications for Policy, Delivery, or Practice: Implications: For policymakers, this study demonstrates that children's hospitals offer the most specialized services and that these institutions are magnets for very ill children. Funding cuts that decrease Medicaid payments, decrease Disproportionate Share Hospital payments, and cut Graduate Medical Education funding are likely to cut into the revenue of these facilities, threatening the ability of these institutions to offer specialized services. Primary Funding Source: No Funding ●Impact of Managed Care on Perinatal Regionalization Deborah Dobrez, Ph.D., Peter Budetti, Ph.D., M.D., Susan Gerber, M.D., M.P.H. Presented By: Deborah Dobrez, Ph.D., Assistant Professor, Health Policy and Administration, University of Illinois at Chicago, 1603 West Taylor Street, Chicago, IL 60618; Tel: 312413-8854; Email: ddobrez@uic.edu Research Objective: The goal of this study was to describe trends in regionalization of perinatal care, to identify factors that predict the extent of regionalization, and to test the effect of managed care penetration on regionalization. Study Design: Secondary analysis of birth certificate, managed care enrollment, and American Hospital Association databases was conducted. Bivariate tests compared birth rates within weight categories by state and hospital level over time. Panel data models estimated the impact of managed care enrollment on site of delivery for low, very low, and extremely low birthweight neonates. Population Studied: All deliveries at non-military hospitals with obstetrical services and greater than 50 deliveries annually in California, Washington, North Carolina, and Illinois were included in the study. To capture changes over time in the degree of regionalized perinatal care, data were drawn over a ten year time period, from 1989 to 1998. Principal Findings: Strong evidence for enhanced perinatal regionalization over time was observed for North Carolina and Illinois, with little change in site of delivery in Washington, which was highly regionalized at the beginning of the study period. A shift from level III to level II hospitals, which we consider evidence of de-regionalization, was observed for low and very low birthweight neonates in California. Although managed care enrollment increased substantially in all four states, no relationship between enrollment and site of delivery was detected. Conclusions: Perinatal studies provide strong evidence of the benefits from the delivery of high-risk neonates at tertiary care centers, yet the percentages of very low birthweight neonates delivered at level III hospitals were substantially lower than the goal of 90% set by Healthy People 2010. Nevertheless, increased managed care penetration cannot be cited as the reason for the failure to meet this national goal. Implications for Policy, Delivery, or Practice: Our study demonstrates significant failures in meeting Healthy People 2010 goals for site of delivery of high risk infants that require explanations other than the financial pressures introduced by managed care. Further studies of market and institutional barriers to regionalization are needed to identify steps to improve performance toward meeting the Healthy People 2010 goals. Primary Funding Source: National Institute for Child Health and Human Development ●Can Education and a Preemptive Prescription Prevent ER Visits for Ear Pain? Dawn Finnie, M.P.A., Deborah McWilliams, M.D., James Naessens, M.P.H., Holly Van Houten, BA Presented By: Dawn Finnie, MPA, Health Services Analyst, Health Care Policy & Research, Mayo Clinic, Pavilion 3, Rochester, MN 55905; Tel: (507)538-3346; Fax: (507)284-1731; Email: finnie.dawn@mayo.edu Research Objective: To determine whether providing parent education and a prophylactic prescription of an ear-numbing agent (antipyrine/benzocaine) at a well child visit at 15 months of age affects the utilization of ear pain-related emergency room (ER) visits over the next 12 months. Study Design: A “differences in differences” assessment was performed on a controlled cohort study where parents of patients visiting a pediatric clinic for a 15 month well-child visit March 2003 to June 2003 were given a prescription for earanalgesic drops along with RN-provided education on managing ear pain (N=191). All patients with 15 month well child visits during the same timeframe at other sites caring for children at our institution were identified as controls (N=133). Additionally, cohorts of patients with 15 month well child visits in the same 3 months in the year prior to intervention were used as baseline values (N=168 for intervention clinic and 126 for other clinics). Dependent variables included the percent of patients with ear pain – related visits,ER visits and the ER visit rate per eligible. Population Studied: Primary data (patient satisfaction and medical care knowledge) and secondary data (ER and office visit utilization) were collected from 03/2002 to 06/2004 among patients on primary care panels at our institution. All children with 15 month well child visits to primary care sites in the study and baseline timeframes were identified through administrative data. Those children making visits to the intervention site were considered as receiving the intervention. Visits with a principal or secondary diagnosis of ear pain were identified through administrative data. Principal Findings: After the intervention, significant decreases were observed in the intervention clinic in visit utilization for ear pain to the ER (down 80%. p=0.01), Urgent Care Center (40%, p=0.02) and also primary care (28%, p=0.04), with no significant change in the controls over this time. When using a regression model to compare the change in rates between sites, the decline in ER use in the intervention site remained significant (p=0.04). When surveyed, parents receiving education and the prescription strongly endorsed the practice change at the 15 month visit and remained supportive at the 2 year visit. On the 2 year survey, over 80% of those with ear pain since their 15 month visit felt the intervention helped them to avoid an ER/UCC visit. No safety concerns were identified. Conclusions: RN-provided education at a well child visit (with prophylactic analgesic prescriptions) reduced ER, Urgent Care and primary care visits for ear pain in toddlers and was well appreciated by parents. Implications for Policy, Delivery, or Practice: Pediatric health care utilization is influenced by parents' knowledge and their ability to manage common pediatric symptoms. Providing parents with management tools for typically minor illnesses can result in fewer medically-unnecessary after-hours visits. Primary Funding Source: No Funding ●A Method for Risk Adjustment for the AHRQ Pediatric Quality Indicators Corinna Haberland, M.D., MS, Jeffrey Geppert, J.D., Patrick Romano, M.D., Sheryl Davies, MA, Amy Ku, M.H.S.A., Kathryn McDonald, MM Presented By: Corinna Haberland, M.D., MS, Researcher, Center for Health Policy / Primary Care Outcomes Research, Stanford University, 117 Encina Commons, Stanford, CA 94305; Tel: 650-723-1935; Fax: 650-723-1919; Email: corinnah@stanford.edu Research Objective: When using quality indicators for comparative purposes, adjusting for underlying health status of the patient is crucial. Several tools exist for use with adult administrative data sources; however no comparable approach is available for pediatrics. We analyze the feasibility of using the Clinical Classification Software (CCS) from the Agency for Healthcare Research and Quality (AHRQ) in conjunction with the 2002 State Inpatient Data (SID) to develop risk adjustment measures based on clinical importance and empirical validity. Study Design: We augmented standard methods of developing a risk adjustment system with use of newly available data from two states (NY and CA) that specify whether a condition is present on admission (POA). We identified codiagnoses present in pediatric inpatients with diagnoses flagged as complications of care by the AHRQ Quality Indicators (QIs). Using the NY and CA SID data, we identified those cases where the codiagnoses were POA, i.e. likely comorbidities, and those not POA, i.e. likely complications associated with that hospital stay. We also calculated the relative risks (RR) to compare the rates for patients developing the QI diagnosis when the CCS diagnosis was likely either a complication or comorbidity, versus the overall rate. Finally we developed a measure of estimator bias (EB), a function of the QI rate when the CCS is POA, adjusted for the rate calculated if POA status were unknown (as is the case for most states in the SID 2002), with possible values ranging from 0 (no bias) to 1 (maximum bias). After a clinical review of the CCS codiagnoses list, we assessed the consistency of the clinical expectation with these empiricallygenerated values to determine which CCS categories were candidates for risk adjustment, given a dataset where POA status is unknown. Population Studied: Pediatric inpatients in the SID 2002 database. Principal Findings: Clinically, we expected some CCS categories to have higher risks of a QI complication. Empirical results were confirmatory. For example, Fluid and Electrolyte Disorders had a RR of 2.7 in patients with the CCS POA for the Post-Operative Hemorrhage and Hematoma indicator, and a RR of 3.6 in patients where the CCS was not POA. CCS categories that would not be clinically useful risk adjusters showed RR closer to 1 (e.g. Genitourinary Symptoms in the Post-Operative Respiratory Failure indicator). Comparing clinical and empirical results was most helpful when the CCS category could be associated with either an elevated risk of developing the indicator diagnosis, or occurrence along with the indicator diagnosis. Using the EB, which for the Iatrogenic Pneumothorax indicator ranged from 0 to 0.9 in Intestinal Obstruction and Other Upper Respiratory Infections, respectively, we selected CCS categories that were clinically important, and minimized the potential of misclassifying complications as comorbidities. For example, in the Post-Operative Respiratory Failure indicator the CCS group with Urinary Tract Infections has an overall RR of 2.6, a potentially important risk-adjuster. However, when the RR for POA and not POA are compared (less than 2 and 6.5, respectively) its risks as a comorbidity are much less than as a complication, consistent with the relatively high EB value of 0.6. These numbers comport clinically, given that such an infection is unlikely to increase a patient’s risk of respiratory failure if identified on admission (assuming appropriate treatment during hospital stay). Conclusions: Using enriched data and these metrics along with clinical judgment provides insight about potential usefulness of codiagnoses as risk-adjusters (i.e. avoiding adjusting for a comorbidity that is likely to be a complication arising during a hospitalization). Implications for Policy, Delivery, or Practice: Risk adjustment techniques for pediatrics have become increasingly necessary in quality assessment work. This approach provides a parsimonious method of generating a risk adjustment tool based on administrative data for the forthcoming AHRQ pediatric quality indicators. Primary Funding Source: AHRQ ●Access to Safe Parks Helps Increase Physical Activity Among Teenagers Theresa Hastert, M.P.P., Susan Babey, Ph.D., E. Richard Brown, Ph.D. Presented By: Theresa Hastert, MPP, Research Associate, , UCLA Center for Health Policy Research, 10911 Weyburn Ave., Suite 300, Los Angeles, CA 90024; Tel: (310) 794-2827; Fax: (310) 794-2686; Email: thastert@ucla.edu Research Objective: To examine the relationship between neighborhood characteristics and physical activity among adolescents in California. Study Design: We used data from the 2003 California Health Interview Survey (CHIS), a random-digit dial (RDD) telephone survey of 42,000 households drawn from every county in California. Bivariate and multivariate analyses were used to examine the relationship between family and neighborhood characteristics and level of physical activity. Population Studied: Responses of 4,010 adolescents interviewed for CHIS 2003. Principal Findings: Nearly 30% of California adolescents do not meet physical activity recommendations. Teens with access to a safe park have higher rates of regular physical activity (71.8% vs. 67.3%) and lower rates of inactivity (6.4% vs. 10.3%) than those without access, but more than one in four teens does not have access to a safe park, playground or open space. Access to safe parks is particularly important for teens living in urban areas, those from low-income families, those living in apartments, and those living in neighborhoods perceived as unsafe. Twice as many teens living in urban areas are inactive (9.4% vs. 4.5%) and fewer get regular activity (66.8% vs. 76.1%) compared to those living in rural areas, but fewer urban teens with access to a safe park are inactive (7.8% vs. 14.1%) and more get regular activity (68.6% vs. 61.3%) than urban teens without access. More teens living in multi-unit apartment buildings are inactive (10.2% vs. 6.5%) and fewer get regular activity (63.1% vs. 72.9%) compared to those living in single family homes, but among teens living in apartments, half as many with access to a safe park are inactive (8.0% vs. 16.3%) and more get regular activity (65.6% vs. 56.1%) than teens in apartment buildings without access to a safe park. Fewer teens from lower-income families with access to a safe park are inactive (7.6% vs. 12.0%) and more get regular activity (68.7% vs. 63.1%) than those without access to a safe park. In neighborhoods perceived as unsafe, fewer teens with access to a safe park are inactive than those without access to a safe park (9.0% vs. 16.3%). After adjusting for other factors, adolescents living in apartments are less likely to get regular physical activity than those living in houses and those with no access to a safe park are more likely to get NO physical activity than those with access. Conclusions: These findings suggest that access to safe parks has a positive effect on physical activity behaviors in teenagers, especially those from urban areas, those living in apartment buildings, those from low-income families and those living in neighborhoods their parents perceive as unsafe. Implications for Policy, Delivery, or Practice: Expanding park access and safety for adolescents living in urban areas and less advantaged adolescents promises to have the greatest and most positive impact on adolescent physical activity behaviors. Primary Funding Source: RWJF ●On the Front Line in Children’s Health Care: Caseloads, Scope of Practice, and Satisfaction with Care among Rural and Urban School Nurses in Pennsylvania Marianne Hillemeier, Ph.D., M.P.H., Lisa A. Davis, M.H.A. Presented By: Marianne Hillemeier, Ph.D., M.P.H., Assistant Professor, Health Policy and Administration, The Pennsylvania State University, 116 Henderson, University Park, PA 16802; Tel: 814-863-0873; Fax: 814-863-2905; Email: mmh18@psu.edu Research Objective: To evaluate the scope of nursing services provided in rural and urban public schools in Pennsylvania, and nurses’ perceived satisfaction with the level of care provided. Study Design: Survey of school nurses representing each of Pennsylvania’s 500 active school districts (n=615 out of a total sample of 996; response rate=61.7%). Topics assessed included number and proportion of students with specific health care needs; types of nursing services and procedures provided in the school setting; percent of students screened annually for various health conditions; provision of instruction and counseling related to specific health topics; perceived satisfaction with the current level of nursing services provided; and the school nurse to student ratio. Population Studied: School nurses practicing in rural and urban public elementary and secondary schools in Pennsylvania. Principal Findings: School nurse caseloads contained a sizable proportion of students with special health care needs; the most prevalent health problems included asthma (9.4%), weight disorder (6.0%), severe allergies (5.5%), and ADHD (4.9%). Specialized procedures were often provided including blood sugar testing (75.8%), nebulizer treatments (68.9%), insulin pump management (41.0%), and catheterization (15.3%). Similar types and numbers of medical procedures were provided by both rural and urban school nurses. Over 70% of all secondary school nurses provided education and counseling on pregnancy prevention, STD prevention, and substance use, however education and counseling on other important issues such as nutrition and mental health were less often provided. Among secondary school nurses, those in rural settings were more likely to report being very satisfied or satisfied with the health care provided to students in their schools; no differences in satisfaction with care was found among rural and urban elementary school nurses. There were fewer nurses per school district in rural areas relative to urban, however because rural districts generally have fewer enrolled students their school nurse to student ratios tended to be more favorable (1:936 on average in rural districts vs. 1:1084 in urban districts). Overall, lower school nurse to student ratios were positively and significantly associated with greater perceived satisfaction with the level of health care provided to students (p=0.05). Conclusions: The health care needs of children in elementary and secondary schools are diverse and complex, and many school nurses are called upon to provide skill-intensive procedures. Education and counseling about important health-related issues are provided in some, but not all, school settings. Nurses practicing in schools where school nurse to student ratios are lower are more likely to express satisfaction with the level of health care provided to students. Implications for Policy, Delivery, or Practice: School nurses are the health professionals most consistently involved in dealing with the complex health needs in the school-age population. Professional nursing organizations and public health initiatives such as Healthy People 2010 recommend that school nurse to student ratios not exceed 1:750, and study findings suggest that lowering current ratios will result in more optimal health services for children in public school settings. Primary Funding Source: The Center for Rural Pennsylvania ●Developing and Evaluating Networks to Enhance Population Health Timothy Huerta, Ph.D., Ann Casebeer, Ph.D., Janice Popp Presented By: Timothy Huerta, Ph.D., Research Scientist, PHSA Research and Networks, Child and Family Research Institute, 601 West Broadway, Vancouver, V5Z4C2; Tel: 604.707.6398; Fax: 604.707.6399; Email: tim.huerta@ttu.edu Research Objective: This paper presents the lessons that five years of networked experience has produced for CYHNC network members collectively, and puts forward a plan for improving network capacity in support of population health goals. Study Design: Action Research involving academics and practitioners to define the implementation challenges faced by interorganization networks. Population Studied: The Child and Youth Health Networks of Canada (CYHNC) is an inter-organizational network connecting child and youth health services networks across Canada. Principal Findings: Increasingly, health services organizations are being called to take on more proactive roles in promotion and prevention rather than focusing on services alone. Several factors are moving them toward collaboration as a strategy for managing the complexity of the problems they encounter: There is growing consensus that solutions require broader responses than health service agencies can marshal alone; Many organizations find themselves offering services to populations served by others in similar ways; creating duplication and inequity; Clients must self-manage the burden of navigation among inter-organizational partners, often at times when they are least able to cope; Organizations are seeking opportunities increasing their awareness of better practices among agencies sharing similar problems and goals. Conclusions: In response, some agencies have self-organized to create inter-organizational networks. This ‘network of networks’ has developed for a number of reasons and in a variety of ways. Individually and collectively, they have struggled to conceptualize, implement and evaluate ways to organize and develop themselves. Implications for Policy, Delivery, or Practice: It identifies key challenges and opportunities for improving networked capacity to promote organizational and social change and improve health. Primary Funding Source: No Funding Hispanic counterparts. Differences in socioeconomic status and insurance coverage accounted for only a small portion of the overall differences in health care utilization between Hispanic and non-Hispanic children with asthma. Primary Funding Source: Flinn Foundation ●Patterns of Health Care Utilization for Hispanic and Non-Hispanic Children with Asthma Tricia Johnson, Ph.D., Mary Rimza, M.D., William G. Johnson, Ph.D. ●Prescription for Success: How School-Based Health Centers Affect Health Status, Healthcare Use and Cost Kathryn Keller, M.P.A., Terrance Wade, Ph.D., Jeff Guo, Ph.D., Ann McCracken, Ph.D. Presented By: Tricia Johnson, Ph.D., Assistant Professor, Health Systems Management, Rush University, 1700 West Van Buren Street, TOB Suite 126B, Chicago, IL 60612; Tel: (312) 942-7107; Fax: (312) 942-4957; Email: tricia_j_johnson@rush.edu Research Objective: A substantial and increasing share of expenditures for hospitalizations and emergency medical services for children in the U.S. are for asthma-related illnesses. This paper examines the effect of ethnicity and insurance coverage on health care utilization patterns for children who have asthma. Study Design: Data come from the Arizona HealthQuery, a state-wide multi-year community health data system, for Yuma County, Arizona. Yuma County is a low-income, semi-rural agricultural community that borders Mexico. More than 70% of Yuma children were Latino. The study included children with an asthma diagnosis (ICD-9-CM diagnosis code of 493) in 2002 and received medical care in each of the years, 2002, 2003 and 2004. Multivariate regression equations were used to estimate the number of emergency department (ED) services, inpatient hospitalizations and ambulatory physician services for asthma care during this time period for Hispanic and non-Hispanic children separately. Differences in utilization were decomposed into differences due to race/ethnicity and other observed characteristics (age, sex, insurance coverage and geographic region of residence). Population Studied: Children with asthma Yuma County, Arizona. Principal Findings: Both Hispanic and non-Hispanic children with one or more gaps in health insurance used more ED visits, and fewer ambulatory physician visits than Medicaidinsured children, while uninsured Hispanic children used fewer hospitalizations than their Medicaid-insured counterparts. Similarly, children who switched between public and private insurance coverage had more hospitalizations and fewer ambulatory physician visits than children consistently covered by Medicaid. Decomposition results showed that non-Hispanic children experienced more ED visits and hospitalizations, but fewer ambulatory physician visits. Conclusions: Differences in health care utilization for Hispanic and non-Hispanic children were primarily due to unobserved differences associated with race/ethnicity than age, sex, insurance coverage or geographic region of origin. Implications for Policy, Delivery, or Practice: Gaps in insurance coverage markedly increased the utilization of ED and inpatient services. Providing public (e.g., Medicaid) insurance to uninsured children who have asthma could decrease utilization of expensive emergency and inpatient services. In this low income, rural community, Hispanic children were more likely to use outpatient physician care and used fewer ED visits and hospitalizations than their non- Presented By: Kathryn Keller, MPA, Program Officer, The Health Foundation of Greater Cincinnati, 3805 Edwards Road, Suite 500, Cincinnati, OH 45209; Tel: 513-458-6616; Email: kkeller@healthfoundation.org Research Objective: To determine how SBHCs influence student health, healthcare access and costs. Study Design: As part of the SBHC Initiative, the Foundation funded two multi-year studies to evaluate the effectiveness of the model. For the first study, the SBHC Health Outcomes Study, we conducted a 3-year longitudinal process and outcome evaluation of eight SBHCs. Four comparison schools were matched on student characteristics and proportion of students receiving free and reduced lunch. Multiple, linked sources of data included a 3-year panel survey of randomly selected parent-child dyads, yearly student absence records, and SBHC records and health encounters. A yearly survey among SBHC school personnel was also collected. The objective of the second study, the SBHC Cost Study, was to evaluate health care utilization and costs of students in SBHC programs. Using a retrospective quasi-experimental timeseries analysis, four SBHCs and two comparison school districts were studied from 1997 to 2003. Repeated-measure analyses were conducted to assess the changes of student’s health care utilization and costs in hospitalizations, emergency departments, physician encounters, mental health services, dental, prescriptions, well child check-ups, and other outpatient care. Population Studied: The population for SBHC Health Outcomes Study consisted of eight SBHCs and the schools using the centers, the Schools’ students in grades K-8, parents and school personnel. Four of these SBHCS were in Ohio and four were in Kentucky. Four SBHCS were rural and four were urban. This study also included matched schools without SBHCS and their students and parents. The SBHC Cost Study consisted of the students who were in the Ohio Medicaid database and who were enrolled in one of the six Ohio schools in the SBHC Health Outcomes study, four with SBHCs and two without. Principal Findings: Among other findings, our primary findings from the studies showed that SBHCs improved the health-related quality of life of students who used the SBHC, did not change the overall Medicaid costs, decreased primary care costs, increased mental and dental health costs,decreased prescription drug costs,decreased hospitalization and emergency room costs for children with asthma, closed the healthcare utilization gap for African American students, and reduced Medicaid costs for children in the rural Conclusions: SBHCs improve the health of students, help keep kids in the classroom, support working parents and do so at no additional cost to the Medicaid system. Implications for Policy, Delivery, or Practice: SBHCs are a model for providing quality healthcare services for children that eliminates most barriers children face when trying to access healthcare. SBHCs can address problems of transportation, lack of nearby providers, lack of providers accepting public insurance, and parental difficulties getting time away from work to take a child to the doctor. While SBHCs provide access to healthcare for children, funding these centers can be challenging for a variety of reasons. Addressing these reasons at a state level can help create an environment in which SBHCs become an integral part of the healthcare system. At the same time, because SBHCs improve children’s health status, they support strategies that seek to improve the lives of children. Primary Funding Source: The Health Foundation of Greater Cincinnati ●Persistence of High Health Care Expenditures Among Children in Medicaid Caprice Knapp, Ph.D., Elizabeth Shenkman, Ph.D., David Sappington, Ph.D., Bruce Vogel, Ph.D., Des Schatz, M.D. Presented By: Caprice Knapp, Ph.D., Post Doctoral Fellow, Institute for Child Health Policy, University of Florida, 1329 SW 16th St, Room 5130, Gainesville, FL 32608; Tel: (352) 265-7220; Fax: (352) 265-7221; Email: cak@ichp.ufl.edu Research Objective: Our objective was to examine the association between children’s health and sociodemographic characteristics and the persistence of high health care expenditures among Medicaid Health Maintenance Organization (HMO) and Primary Case Management PCCM) Program enrollees over a 3 year period. Because children disenroll from and/or move between the HMO and PCCM Programs, we also examined the relationships between the children’s health and sociodemographic characteristics and their odds of disenrollment and their odds of transferring between programs. Study Design: The Texas Medicaid HMO and PCCM Programs provided child-level enrollment and claims/encounter data for 2002 through 2005. The children’s health status was categorized, using the Clinical Risk Groups (CRGS), into one of five categories: healthy, significant acute, minor chronic, moderate chronic, and major chronic conditions. The sociodemographic characteristics included age, gender, race/ethnicity, and months enrolled in the program. Logistic regressions were used to examine the relationship between the children’s CRG health status category and their sociodemographic characteristics and the odds that they would remain in the top 10% of health care spenders in their program (i.e., HMO or PCCM). Other spending cut-points were examined (i.e., 5%, 20%, 30%, and 40% of spenders). The patterns of results were the same regardless of the cut-point used. The 10% cut-point is presented here. Population Studied: A census of children ages 0 to 18 enrolled in the Texas Medicaid program from 2003 (N= 1,202,294) through 2005 (N=1,509,806). Principal Findings: Mean annual expenditures between the top 10% of spenders varied significantly between eligibility category (e.g., PCCM/SSI - $60,680 and PCCM/TANF $5,842) and program (e.g., HMO/TANF - $4,345 and PCCM/TANF - $5,842). Among TANF HMO enrollees, 17% of those in the top 10% of expenditures in 2003 remained at that level in 2004 and 13% remained at that level in 2005. Similar patterns were seen for HMO SSI enrollees and for PCCM TANF and SSI enrollees. The children’s health status was significantly associated with the odds of remaining enrolled in Medicaid, the odds of transferring from HMOs to PCCM, and the odds of remaining in the top 10% of spenders. For example, children with moderate chronic conditions, like diabetes, were 3.42 times more likely than those classified as healthy to remain in the top 10% of health care spenders from 2002 through 2005; whereas those with major chronic conditions like cystic fibrosis were 13 times more likely to do so. Race/ethnicity also was significant with black nonHispanic and Hispanic children about 50% as likely as white non-Hispanic children to remain in the top 10% of spenders from year to year. The same pattern was observed regardless of program and eligibility status. Implications for Policy, Delivery, or Practice: Children at high risk for persistent high health care expenditures can be identified and proactively placed in care coordination programs designed to facilitate access to care and to minimize or prevent morbidity. Further investigation is needed to better understand the racial/ethnic disparities in the persistence of high health care spending to ensure that minority children are receiving needed care. Primary Funding Source: HRSA ●Evaluating Appropriateness of Antibiotic Use for Children with Otitis Media in Rhode Island Sylvia Kuo, Ph.D., Susan Miller, Ph.D., M.B.A., James Burrill, M.D. Presented By: Sylvia Kuo, Ph.D., Investigator, Center for Gerontology & Healthcare Research, Brown University, Box GST210, Providence, RI 02912; Tel: (401) 863-2060; Fax: (401) 862-9219; Email: Sylvia_Kuo@brown.edu Research Objective: To assess whether antibiotic use among children with otitis media (OM) insured by Blue Cross & Blue Shield of Rhode Island (BCBSRI), the dominant insurer in the state of Rhode Island, is consistent with established evidencebased guidelines, and whether this use varied by visit or demographic characteristics. Evidence-based guidelines recommend that antibiotic therapy (i.e. high-dose amoxicillin) should be the default treatment only for children under age 6 months, with use in more limited circumstances in older children. Little is known of variations in antibiotic practices within a health care market despite the local nature of practice patterns and disease spread. Study Design: Using BCBSRI administrative encounter and pharmacy claims in 2004, the visit-based analysis used logistic regressions on the probability of receipt of any antibiotic, and the probability that the child received amoxicillin given that an antibiotic was received. Covariates included visit location (physician office, urgent care center or hospital emergency room), insurance coverage type (PPO, HMO and Medicaid managed care), age (<6 months, 6 months – 2 years, over 2 years), an indicator for whether the visit occurred on the weekend, and interaction terms. Population Studied: The study used a cohort of children age 13 or under as of January 1, 2004 who had a diagnosis of OM at any time during 2004 and were continuously enrolled with BCBSRI with both medical and drug coverage. Antibiotic prescriptions filled within 2 days of an otitis-related encounter were matched to encounter data. The sample included 12,436 children, who were associated with 26,090 otitis-related encounters and 9,318 matched antibiotic prescriptions. Principal Findings: Results suggest that the use of antibiotics by age is consistent with evidence-based practices; children under age 6 months are more likely to receive an antibiotic, with most children receiving the recommended antibiotic of amoxicillin. Interestingly, no differences were found by insurance coverage type. However, children who visit urgent care centers are over twice as likely to receive an antibiotic, while the antibiotic received is only half as likely to be amoxicillin compared to similar children who visit physician offices. Additionally, children are significantly more likely to receive antibiotics during weekend versus weekday visits for all visit locations. Conclusions: The patterns of antibiotic use in Rhode Island appear consistent with evidence-based guidelines. In particular, antibiotic use appears reasonable at physician offices, where most cases of pediatric OM are seen in Rhode Island. However, antibiotics may be used less judiciously for children with OM when care is sought on the weekend and/or at urgent care centers -- visit characteristics not typically available in other studies. Implications for Policy, Delivery, or Practice: The results suggest that maintaining continuity of care through a regular provider may reduce inappropriate use of antibiotics, and provide justification for health plan incentives for offices to provide extended hours (as is BCBSRI’s policy). Although urgent care centers may play an important role in reducing costly emergency room visits for acute situation, these sites may have practice patterns that deviate from recommended pediatric practices because of less familiarity with both the patients themselves, and pediatric patients in general. Primary Funding Source: Blue Cross & Blue Shield of Rhode Island ●ADHD Medications in US and OECD Countries: Determinates of Quantity and Price Hui-Chu Lang, Ph.D., Richard M. Scheffler, Ph.D., Teh-wei Hu, Ph.D. Presented By: Hui-Chu Lang, Ph.D., Postdoctoral Fellow, School of Public Health, UC Berkeley, 140 Warren Hall, Berkeley, CA 7360; Tel: (510)643-3366; Fax: 510-643-1853; Email: hclang@berkeley.edu Research Objective: Attention Deficit Hyperactivity Disorder (ADHD) is the most commonly diagnosed behavioral/mental health disorder in children. Despite public attention and controversy on the increased use of ADHD medications, we have seen no published report comparing expenditures, pricing, and usage internationally. Hence, the purpose of this study is to investigate trends in ADHD medication expenditures among U.S. and OECD countries over a 10-year period. An analysis of changes in expenditures over time will shed light on whether and to what extent they reflect provider price increases or decreases, variations in quantity, or other factors. Study Design: We used data from the IMS Health MIDAS worldwide database from 1993 to 2003. The three-digit Anatomical Therapeutic Chemical Classification (ATC) system for drugs was adopted to extract study drugs. ADHD medications include those in the ATC=N6B psychostimulants category, along with the non-amphetamine-like stimulants Modafinil (Provigil) and Atomoxetine HCL (Strattera). Both branded and generic drugs within those categories were included. The quantity measure was a standard unit (SU), defined as the smallest common dose unit included by the IMS. Price was defined as the standard unit average price at the ex-manufacturer level. To analyze the global and country trends in expenditures over the time period 1994-2003, we disaggregated the expenditures into three components: change in quantity; change in price; and the residual product of the change in quantity and change in price. Principal Findings: The global ADHD medication expenditures were $332 million, $720 million, and $2.26 billion in 1994, 1999, and 2003, respectively. In the U.S. market, the real ADHD medication expenditures were $304 million, $658 million, and $2.11 billion in 1994, 1999, and 2003, respectively. For OECD countries, the real ADHD medication expenditure was $17 million, $50 million, and $121 million in 1994, 1999, and 2003, respectively. During the study decade, the U.S. market had higher rates of expenditure and price increase than OECD nations. U.S. expenditures increased 221% in real dollars; sales volume increased 2%; and price increased 214%. For OECD countries during this period, expenditures increased 141% in real dollars; sales volume rose 63%; and price increased 48%. The rate of price increase in the U.S. market was far higher during the later years of the study (1999-2003) than in the earlier years (1994-1999); the increase rate was 214% in the latter period compared to 23% in the prior period. The OECD countries saw lower rates of price increase: 48% in 1999-2003 compared to 15% in 1994-1999. Conclusions: During the study decade, ADHD medication expenditures have dramatically increased in the United States and the OECD countries. The U.S. increase is primarily a reflection of rising prices, while the OECD countries’ increase reflects mainly higher quantities. Primary Funding Source: UC Berkeley ●Can Real-Time Decision Support Increase Appropriate Influenza Immunization at two Urban Children's Hospitals? James Levin, M.D., Ph.D., Sivakumaran Raman, MBBS, MS, Patsy Stinchfield, MSN, CPNP Presented By: James Levin, M.D., Ph.D., Medical Director, Informatics, Infectious Disease and Information Technology Services, Children's Hospitals and Clinics, 345 Smith Ave N, MS 70-504, Saint Paul, MN 55102; Tel: (651) 220-6444; Email: jim.levin@childrensmn.org Research Objective: Young infants and children with chronic diseases are at as high a risk of death from influenza as any population group. We evaluated whether a combination of family and provider targeted interventions could to increase appropriate immunization rates for these groups at two urban Children’s hospitals. Study Design: Analysis of diagnosis and demographic information in a clinical data warehouse was used to identify high priority patients. Patients were flagged in the health system's electronic medical record online problem list. Reminder/recall postcards were mailed to families. Automated real-time alerts were delivered to clinicians each time unvaccinated patients presented for care. The primary outcome measure was the immunization rate for targeted patients as compared to the two previous yearly seasons. Additional measures included the timeliness of the immunization and the number of immunized patients with missed opportunities (prior encounters during the season without vaccine administration). Multivariate logistic regression was used to evaluate the association with immunization uptake of insurance type (government-funded versus others), language spoken in the home, patient gender, patient home nearest census-tract poverty level, and whether the mailed postcard was undeliverable. Population Studied: Patients with outpatient clinic encounters at Children’s Hospitals and Clinics of Minnesota (CHC) in the 12 months prior to the beginning of the 2004-5 influenza season. CHC is a not-for-profit pediatric care system with inpatient, Emergency Department, primary and subspecialty care clinics in Minneapolis and Saint Paul, Minnesota. Principal Findings: During the 2004-5 vaccine season 11,336 postcards were sent to high priority patients. 7.4% of the postcards were undeliverable. 2,178 patients with chronic conditions were vaccinated, a 28% increase over the previous season. Of the 6-23 month olds, 1,191 were vaccinated, a yearly increase of 135%. High priority patients were vaccinated a median of 3 weeks earlier. There was a decrease of 26% (chronic disease patients) and 45% (6-23 month olds) in immunized patients with missed opportunities. For patients with chronic diseases the following were associated with increased immunization: Previous year immunization (Odds Ratio (OR) 4.2, 95% CI [3.5, 4.9]), Spanish speaking (OR 2.4, [1.8, 3.3]), government insurance payer (OR 1.6, [1.3, 1.8]), and census tract poverty (OR 1.1, [1.1, 1.2]). For this group, African language speaking was associated with decreased immunization (OR 0.6, [0.4, 0.9]). In contrast, for 6-23 month olds, significant decrease was associated with government insurance payer (OR 0.8, [0.7, 1.5]) and undeliverable postcard (OR 0.4, [0.3, 0.5]. Conclusions: Using a data warehouse to identify high priority patients and communicating directly to their families (postcards) and their providers (automated real-time alerts) was associated with improved immunization performance during the first year of the intervention. Demographic and socioeconomic factors associated with immunization uptake differed between two targeted groups. Implications for Policy, Delivery, or Practice: New recommendations for universal vaccination of 6-23 month olds were initiated during the 2004-5 season, a year of unprecedented vaccine shortage. Media attention on the shortage created confusing and sometimes contradictory messages to families and health care workers. Even in this fluid environment, our results suggest a role for the use of advanced information technology to identify and promote accurate immunization at two urban pediatric hospitals. The barriers to immunization may be different for the two high priority groups studied. Primary Funding Source: No Funding ●Family Involvement and Children with Disabilities: Myth or Reality Ana Lopez - De fede, Ph.D., Karen Kirtland, Ph.D., George Johnson, Ph.D. Presented By: Ana Lopez - De fede, Ph.D., Research Associate Professor, Institute for Families in Society, Universty of South Carolina, 1600 Hampton Street - 5th Floor, Columbia, SC 29208; Tel: 803-777-5294; Fax: 803-777-1120; Email: adefede@sc.edu Research Objective: The research literature has documented the extent of minority student disproportionality in special education, citing broad inadequacies across a variety of educational fronts. This study addressed two primary questions. First, are there significant differences in family involvement for children disproportionaly represented in special education? Second, do these factors differ for children in special education based on their disability classification? Study Design: Data were collected through telephone surveys of randomly-selected households in South Carolina with children in special education placements. For the purpose of this survey, special education services were defined as extra reading or math instructions in a resource room, speech or therapies, or special education. Response categories were collapsed and stratified by race (whites and African Americans). Chi square tests were used to detect significant differences in demographics between whites and African Americans as well as to determine significant differences in survey responses between whites and African Americans. Odds ratios were calculated to determine if significant associations existed between having a disability and demographic (individual) characteristics, special education scores, school characteristics scores, or family involvement scores. Population Studied: Households were randomly selected from a representative state sample of 30,670 children in special education in 9 counties. 453 interviews were completed for a response rate of 44.8% with a margin of error of ± 4.61% at the 95% confidence level. Principal Findings: Race, health status, and school characteristics were significantly associated with children having both a learning disability and a serious emotional disability. These children were 1.9 times more likely to be African American than white, and were 8.9 times more likely to be in fair or poor health versus excellent health. School characteristics were negatively associated with children having both a learning disability and a serious emotional disability. For each point increase in the school characteristics score, a caregiver was 0.70 times less likely to have a child with both disabilities. Conclusions: The findings illustrate that there are significant differences in family involvement for African American children disproportionaly represented in special education. They were more likely to be over represented among children with LD, SED or both classifications. Among students in special education, African American students were more likely to report poorer health compared to White students. Implications for Policy, Delivery, or Practice: The disproportionate representation of racial and ethnic minorities in special education and the inequities in educational opportunities are among the most critical issues faced by the U.S. public school system. This study illustrates the continued need place the involvement of families of children with special education in the education system. Primary Funding Source: No Funding ●Medicaid’s Role in Treating Children in Military Families D. Richard Mauery, M.P.H., Peter Shin, Ph.D., M.P.H., Sara Rosenbaum, J.D. Presented By: D. Richard Mauery, M.P.H., Senior Research Scientist, Department of Health Policy, GWU School of Public Health and Health Services, 2021 K Street NW, Suite 800, Washington, DC 20006; Tel: (202) 530-2376; Fax: (202) 2960025; Email: rmauery@gwu.edu Research Objective: To measure and describe Medicaid’s role in meeting the needs of children with special health care needs (CSHCN) in military families covered by TRICARE, the Department of Defense health insurance program for military service members and their families. Study Design: 1) Quantitative analysis of survey data to identify the size of this population and its characteristics and needs. The data derived from CDC’s 2000-2002 State and Local Area Integrated Telephone Survey National Survey of Children with Special Health Care Needs. Quantitative comparisons of military CSHCN with and without Medicaid were performed. 2) By purposive sampling of four large U.S. military bases, telephone interviews with TRICARE Beneficiary Counseling and Assistance Coordinators to assess individual base experiences with CSHCN. Qualitative analysis completed by cross-referencing responses to identify similarities and differences across bases. Population Studied: 1) CSHCN in military families covered by TRICARE and/or Medicaid; 2) TRICARE Beneficiary Counseling and Assistance Coordinators at four U.S. military bases. Principal Findings: Medicaid covers one in 12 military children and one in 9 military CSHCN. Military children with Medicaid are more likely than those without Medicaid to be nonwhite and twice as likely to have low family incomes, consistent with Medicaid’s financial eligibility rules. Among the 14% of all military CSHCN, 11% have both TRICARE and Medicaid coverage. Military CSHCN are more likely than privately insured civilian CSHCN to have Medicaid (one in 9 vs. one in 11). Military CSHCN with Medicaid are almost five times more likely than those without Medicaid to have fluctuating needs, and ten times more likely to receive SSI. The average disease severity rank of military CSHCN with Medicaid is greater than military CSHCN without Medicaid. Interviews suggested that the military health system views Medicaid as an important supplement to TRICARE. However, military CSHCN likely face the same barriers that commonly confront other Medicaid-enrolled children, particularly in the case of behavioral health provider shortages. TRICARE’s new “Extended Care Health Option” (ECHO) program for persons with disabilities may reduce reliance on Medicaid among military CSHCN. However, ECHO contains limitations that would not apply to children covered by Medicaid. Conclusions: Although Medicaid coverage is critical to military CSHCN, military CSHCN with Medicaid are less likely than those without Medicaid to receive the care they need, particularly mental health services, possibly due to greater severity of their conditions. Its capacity to mitigate needs of military CSHCN may also be lessened by state-to-state mobility, interstate coverage and health resource variations, and residential isolation on non-metropolitan military bases. Implications for Policy, Delivery, or Practice: This analysis underscores a less analyzed aspect of Medicaid: its capacity to supplement basic coverage for CSHCN. Medicaid plays an important role for children with significant health risks in both civilian and military sectors. This hidden side of the program merits attention as an element of longer term Medicaid reform. It also suggests the importance of reconsidering existing Federal SCHIP policy, which prevents under-insured children from qualifying for SCHIP in order to fund supplemental benefits. Primary Funding Source: RWJF ●Laws and Policies for Children Deprived of Parental Care in Least Developed and Middle Income Countries: An International Comparative Analysis Emily Nicklett, M.S. Presented By: Emily Nicklett, M.S., Doctoral Student, Health Services Organization and Policy, University of Michigan School of Public Health, 1781 Cram Circle #7, Ann Arbor, MI 48105; Tel: 917-399-4092; Email: enicklet@umich.edu Research Objective: The objective is to document and analyze the current progress of countries in developing laws and policies to uphold the Convention for the Rights of the Child’s (CRC) standards specific to children without parental care. Study Design: A four-page survey instrument was sent to 131 UNICEF country offices in April-May 2005 to analyze laws and policies addressing: forms of family and child assistance to promote preservation; maintenance of family ties; familybased care; and types, standards, and child participation in family-based and institutional care placement settings. Population Studied: 68 questionnaires were analyzed, 25 of which are Least Developed Countries (LDCs) and 43 of which are Middle Income Countries (MICs). The overall response rate was 52%, with 58% from LDCs and 49% from MICs. Principal Findings: A baseline of existing laws and policies was established for the countries studied in the areas of (a) family assistance and preservation, (b) maintenance of family ties, (c) family-based care, and (d) child participation. While the results indicated that numerous LDCs and MICs espoused CRC standards in laws and policies, improvement is particularly necessary in the following areas: (a) Further development of programs for family assistance (such as school feeding programs, free daycare, and parental education) and improvement of preservation efforts (child abuse prevention, home visits, and parental education) (b) Legislative development articulating grounds for removal, requiring a judicial decision and the approval of a designated authority prior to separation (c) Prioritization of family-based care over institutionalization, with regular monitoring of placements and budgetary assistance allocated to providers (d) Participation of children in placement decisions for institutional and family-based care Conclusions: The analysis provides a mechanism by which legislation on children without parental care can be tracked internationally. Future periodic analyses will track the progress of legislative development internationally. These studies should include existing legislation in industrialized countries as well as the extent to which existing laws and policies are upheld in practice. Implications for Policy, Delivery, or Practice: Further development of laws and policies is necessary in some of the countries studied, particularly those lacking comprehensive legislation for child protection. Cooperation is necessary from NGOs, international organizations, and industrialized countries to assist in the attainment of these goals. Primary Funding Source: UNICEF ●Increasing Asthma Awareness by Integrating Asthma Education into The Public School Literacy and Science Curriculums Betty Perez-Rivera, MS, EdD, CHES, Natalie Langston-Davis, M.D., M.P.H. Presented By: Betty Perez-Rivera, MS, EdD, CHES, Program Director, Childhood Asthma Initiative, Children's Health Fund, 317 East 64 Street, New York, NY 10021; Tel: 212 535-9779; Fax: 212 812-5272; Email: bperez@chfund.org Research Objective: The educational programs focus on creating awareness about the impact of asthma in medicallyunderserved communities of the South Bronx as well as empowering families to take control of their asthma through active management of their environmental triggers. Study Design: Nationwide, the prevalence rates for asthma have risen dramatically in the past 2 decades. New York City has the 3rd highest level of asthma (17% citywide). Poor urban minority children in medically underserved communities are disproportionately affected (20-30%). Asthma impacts child and family productivity, physical activity, self-esteem, & overall quality of life. Lack of health education, quality continuity of care and poor living conditions add to the problem. Using National Heart, Lung and Blood Institute (NHLBI) guidelines, the Children’s Health Fund’s Childhood Asthma Initiative (CAI) uses a comprehensive medical model with clinical, psychosocial, and educational components to address family needs. CAI has made substantial progress in creating an asthma friendly community by offering new and innovative comprehensive programs in public schools, after-school programs and community centers. The CAI has created a number of curricula that are suitable for a variety of audiences based on their time constraints. Families are encouraged to speak with their medical providers about asthma and to follow instructions related to asthma medications. An asthma curriculum for public schools was added to teach children about asthma, incorporating asthma education into the literacy and science curriculum of elementary through high school students. The students worked in groups to answer questions about asthma using self-expression, inquiry, research, and writing to present what they learned. These activities were completed while engaging in normal literacy or science activities with their teachers. The students had two weeks to work on presentations which included posters, poetry, short stories, role-play, interviews, and personal experiences about asthma. Letters were sent home to encourage parents to help their children do the research on asthma. They were invited to attend the student presentations in the classroom. Parent and provider workshops were then invited to attend adult workshops, given at a later date. A “Family Asthma Guide” was used as a teaching tool. Population Studied: Grade school students grades 3,4,9, and 10. Principal Findings: In the past 2 years (2004 and 2005) the curriculum has been incorporated into the literacy curriculum for 3rd grade (n=110) and 4th grade (n= 466) students and the science curriculum for 9th and 10th grade students (n=84 combined). Students successfully presented accurate asthma facts. Pre- and post-tests showed that student knowledge had increased by 59%. In addition, 27 workshops reached 226 parents/providers. CAI plans to market this innovative concept for public schools citywide. This program will be implemented in Washington DC by Spring 2006. More than 4000 children and families participated in CAI’s educational programs. Conclusions: Students can learn about health issues while engaging in daily academic activities. Implications for Policy, Delivery, or Practice: This curriculum increases the number of people with formal asthma education. Teachers, under pressure to improve test scores, can use this new approach without reducing valuable classroom instruction time. It can be used to teach about other health concerns. A health educator, associated with primary care, can help the community by providing the connection to clinical care Primary Funding Source: Picower Foundation ●Interactive Programming to Teach Students About Asthma Through Music Betty Perez-Rivera, MS, EdD, CHES, Natalie Langston-Davis, M.D., M.P.H. Presented By: Betty Perez-Rivera, MS, EdD, CHES, Program Director, Childhood Asthma Initiative, Children's Health Fund, 317 East 64 Street, New York, NY 10021; Tel: 212 535-9779; Fax: 212 812-5272; Email: bperez@chfund.org Research Objective: Focus groups and discussions with parents in adult asthma workshops have revealed parent discomfort with having their children participate in physical activities for fear that it will exacerbate their asthma. This fear together with a lack of understanding about asthma limits children from fully participating in their school activities. The Children’s Health Funds, Childhood Asthma Initiative (CAI) has created new and innovative programs to help dispel myths about asthma and empower families by teaching better control of their asthma through management of environmental triggers. As it is also important for students to have knowledge about asthma, an asthma and music workshop for children was created to help students learn about breathing and the impact of asthma in their lives and the community. Study Design: Asthma and music workshops were created to help increase awareness about asthma characteristics and symptoms while encouraging students to participate in music and recreational activities in a fun and creative way. The interactive workshops had two different components. The asthma component provided information and education about asthma management (approximately 10-15 minutes) while the musical component helped to familiarize the students with different types of musical interpretations and the use of an instrument of choice (approximately 70-75 minutes) to expand their airways. Each workshop consisted of four sessions, lasting about 90 minutes each. CAI health educators taught about asthma and coordinated all sessions within the school setting. The Central Park Jazz, a professional quartet, collaborated by teaching the musical component and integrating the breathing exercises into their musical routines to reaffirm the relevance to benefits for asthma. Population Studied: Elementary school students Principal Findings: One hundred and one elementary school students within three New York City Public Schools completed the 4-session Asthma and Music Workshop that allowed them to learn about asthma characteristics, symptoms, triggers, and medications while also spending time learning how to help expand their lungs for better control through breathing exercises and playing a musical instrument. Students had a jam session with the musicians in a concert held during the fourth session of the workshop. Parents, asthmatic students and teachers in the school were invited. The participants used posters to teach others about asthma. They also reinforced the ability of asthmatics to participate in sports and other recreational activities. All participants were encouraged to participate in community service activities and to continue to teach others about asthma. One Hundred and Fifty “Family Asthma Guides” and other educational materials were distributed. Conclusions: Teachers, parents, and school administrators learn how to engage students to learn about asthma and other health topics, reinforcing the need for parent/provider/teacher communication for better health management. Implications for Policy, Delivery, or Practice: A good understanding of the child’s physical capacity can decrease loss of productivity from lack of participation in physical activities, school absenteeism and parent work days missed. In a time when the budget for music and gym classes has been eliminated from the public schools, this asthma and music program is essential to help provide students with avenues to show their creativity, participate in music, and learn new things. Primary Funding Source: Goodman Memorial Foundation ●Does Treating Maternal Depression Improve Child Health Management? The Case of Pediatric Asthma Cynthia Perry, Ph.D. Presented By: Cynthia Perry, Ph.D., Robert Wood Johnson Scholar in Health Policy Research, Harvard University, 1730 Cambridge Street, S-412, Cambridge, MA 02138; Tel: 617-3593828; Fax: 617-496-6085; Email: cdperry@rwj.harvard.edu Research Objective: Past studies have demonstrated an association between maternal depression and poor management of pediatric asthma, but have not answered the causal question of whether maternal depression causes a mother to manage her child’s asthma poorly. This project uses an instrumental variables strategy to assess whether treating maternal depression causes an improvement in pediatric asthma management as measured by child's health care utilization. Study Design: I use differences in physician propensity to diagnose depression as an instrument for whether a woman’s depression is identified and she is therefore offered treatment through the Florida Medicaid system. This is the variable of interest in a set of equations that estimate various types of health care utilization by the asthmatic child in the six months following the offer of depression treatment for the mother, relative to the six months prior. Population Studied: I use Florida Medicaid claims data for mothers and children who qualify for Medicaid because of their cash welfare status. The data cover the time period of July 1996 through June 1999. Principal Findings: I find that emergency department visits for asthma care are significantly reduced following the offer of treatment for maternal depression. I also estimate a reduction in total asthma costs in the six month period following diagnosis of $347 per asthmatic child whose mother is diagnosed with, and offered treatment for, depression. Conclusions: There is a significant improvement in the management of pediatric asthma, manifested in reduced emergency room visits, when maternal depression is diagnosed and treatment is offered. Cost savings from reduced asthma costs for children can offset the cost of screening for maternal depression in this population. Implications for Policy, Delivery, or Practice: The findings suggest that there could be gains in the efficiency of health care delivery with increased screening for maternal depression, especially for mothers of children with chronic diseases such as asthma. Primary Funding Source: RWJF ●Utilizing Logic Models to Create Community Ownership of School Health Programs MaryAnn Phillips, BS, M.P.H., Bernette Sherman, BA, M.P.A., James G. Emshoff, Ph.D. Presented By: MaryAnn Phillips, BS, M.P.H., Senior Research Associate, Georgia Health Policy Center, Georgia State University, 14 Marietta Street, Atlanta, GA 30308; Tel: (404) 651-1643; Fax: (404) 651-3147; Email: mphillips2@gsu.edu Research Objective: To determine whether logic models can be utilized to facilitate ownership of community school health programs and measure program success. Study Design: A public-private partnership created a mechanism to enhance the capacity of select school health programs in Georgia. Researchers and individuals representing six community grantees collaborated to design and implement a participatory evaluation by using logic models as the conceptual framework for the evaluation and as a method for engaging the communities in understanding and buying into the evaluation process. Each model addressed the identified need and target group, specified outcomes and relevant activities, and depicted the relationship between activities and objectives. Activities, outputs, and outcomes identified through the logic model created the framework through which success of the programs would be measured. In addition to the site-specific logic model, a meta-logic model was developed to tie together similarities in values and outcomes among the six sites. Common measures were identified at the meta-level which drove the development of evaluation tools used across the six sites. Population Studied: Six community collaboratives in Georgia funded to enhance capacity, measure impact, and sustain their school health programs. Principal Findings: Outside researchers and local community representatives can work together to evaluate school health programs. Through a participatory evaluation process, the researchers and the local community developed logic models which identified evaluation questions, determined appropriate data collection methods, documented interventions, and analyzed results. Although logic models can be utilized to document best practices and measure prior and post levels of success, these models cannot guarantee that the programmatic activities are completed by each program. For some of the programs, time, resources, and political constraints limited the completion of certain activities. Conclusions: Partnership throughout the process was critical for the successful ownership of the logic models and their successful implementation by the grantees. The logic model is a tool to facilitate ownership of community programs, maximize their sustainability and get results. Created by the community partners– not by the research team -- the individual and meta-logic models are representative of the value and worth that community partners place on specific aspects of their programs. Through this public/private partnership, grantees were able to be very hands on in developing the measures needed to evaluate their programs. Their work, individually and collectively, yielded several tools for measuring success, one of which was a student survey, used by all six grantees to gather data at two points in time for student health behaviors. Implications for Policy, Delivery, or Practice: Community ownership of any local program is essential to its long-term sustainability. A tool used to secure this ownership of school health programs is the logic model. A well thought out, logical, theory-based program design decreases the risk of flaws in the program design. The logic models, interviews, site visits, and survey results will help link what is learned to theory, outcomes, and indicators. The evaluation will provide information on questions/issues that cut across all sites and highlights the strengths, differences, challenges, and best practices found at both the individual and meta levels of school health programs. Primary Funding Source: Other Foundation ●The Effect of Medicaid/SCHIP Disenrollment on Pediatric Health Care Utilization in an Urban County Mary Rimsza, M.D., Megan Jehn, Ph.D., Richard Butler, Ph.D., William G Johnson, Ph.D. Presented By: Mary Rimsza, M.D., Professor and Co-Director, WP Carey School of Business, Arizona State University Center for Health Information & Research, 9885 N 131 Place, Scottsdale, AZ 85259; Tel: 480 9651622; Fax: 480 657 0635; Email: mrimsza@asu.edu Research Objective: To predict changes in health care utilization if children insured by Medicaid/SCHIP are disenrolled in response to changes in enrollment processes, eligibility and/or cost sharing. Study Design: A model has been developed to determine the anticipated changes in community health care utilization using a community wide administrative health database Arizona Health query (AZHQ) which includes healthcare transactions on over 5 million Arizona residents including all Medicaid participants. AZHQ links patients and claims across health care systems over multiple years. The model combines the estimated coefficients of a multivariate model of health care utilization by currently uninsured children with children who are currently enrolled in Arizona’s Medicaid and SCHIP programs if 10% of these children become uninsured. A system of logistic regression equations is used to estimate the probability of emergency department (ED) transactions for ambulatory care sensitive (ACS) conditions, inpatient transactions and ambulatory care transactions, and a system of three nonlinear regression equations is used to estimate the quantity of services for each of the three types of services if 10% of the children became uninsured. We use the Oaxaca decomposition, modified to fit health care comparisons, to separate differences in utilization between Medicaid and SCHIP-insured children and uninsured children into differences due to the characteristics of the children and differences due to insurance. Population Studied: Children living in Maricopa County, Arizona (the greater Phoenix area who are included in the AZHQ database are either uninsured (n = 38,380) or insured by Medicaid/SCHIP (n = 118,244) in 2004. Principal Findings: In 2004, the uninsured children had 83,654 ED transactions for ACS conditions, 70,731 inpatient ransactions, and 110,802 ambulatory transactions. Approximately 37% of all health care transactions for uninsured children were for ACS ED services compared to 21% of ACS ED services for Medicaid/SCHIP children. Using the model, if 10% of the children insured by Medicaid/SCHIP became uninsured, there would be 17,521 more ED visits for ACS conditions in Maricopa County annually. Using MEPS median cost estimates and AZHQ data, the cost to the community of this increase in utilization will be discussed. Decomposing the difference in the probability of use between Medicaid/SCHIP-insured and uninsured children revealed that 100% of the difference in ED utilization was due to the differences in insurance. Conclusions: Disenrollment of children from Medicaid/SCHIP can be expected to increase pediatric care at expensive sites including EDs for conditions which could be managed in physician offices and also increase hospitalizations. These differences in utilization can be attributed almost entirely to changes in insurance status. Implications for Policy, Delivery, or Practice: Policy changes which result in disenrollment of children from Medicaid/SCHIP will not only increase health care costs but will also aggravate current community problems of ED overcrowding and inpatient bed shortages Primary Funding Source: St Lukes Health Initiatives ●Health Care Expenses and Out-of Pocket Costs for Children with Special Health Care Needs (CSHCN) and the Role of Health Insurance Coverage Parag Shah, M.D. Presented By: Parag Shah, M.D., Postdoctoral Fellow, Institute for Healthcare Studies, Northwestern University, 339 E. Chicago Avenue Rm 712, Chicago, IL 60601; Tel: 312-5030448; Fax: ; Email: p-shah7@northwestern.edu Research Objective: Between 6-30% of children in the US have a disability. These children utilize more health care than the rest of the population, and therefore have increased health care costs. Various health insurance characteristics provide different levels of cost sharing for these families. The purpose of this study is to explore the relationship between health insurance status and out-of-pocket costs for various health care services among families with children with special health care needs (CSHCN). Study Design: Secondary data analysis was conducted using the MEPS from 2000-2002. CSHCN were categorized into one of three insurance categories: public all year, private all year, and uninsured all year. Total and out of pocket expenditures for various health care services were computed by insurance category and weighted to represent a national sample. A ratio of out of pocket expenses to total expenditures, representing the level of protection from costs was estimated. Linear regression analyses were used with the percent out of pocket as the dependent variable and insurance status as the independent variable, while controlling for perceived health status, gender, age, and poverty level. The analyses were repeated stratifying by poverty level to examine the effects of insurance for families with varying incomes. Population Studied: All children under age 18 were selected from the MEPS database between the years 2000-2002. A total of 27,256 children were identified and of these 4787 met the Maternal and Child Health Bureau definition of CSHCN. Principal Findings: Overall, 18.5% of U.S. children were classified as CSHCN. Most CSHCN were privately insured (59.3%); with fewer uninsured CSHCN (4.6%) compared to uninsured, non-CSHCN (8.2%). Publicly insured CSHCN paid a significantly lower proportion of total expenditures out of pocket compared to privately insured CSHCN (10% versus 32.5%, p<0.01). Uninsured CSHCN paid the highest percentage out of pocket expense ratio at 73.2%. This trend held true for all CSHCN regardless of family income. All insurances protected CSHCN well from Emergency Department and inpatient costs but not as well against prescription drug costs and miscellaneous other medical expenditures. Conclusions: Publicly insured CSHCN pay a smaller percentage of total health care costs out of pocket compared to privately insured CSHCN, while having an increased amount of total health care costs. When these families were stratified by income, families with public insurance continued to pay smaller percentages of total costs out of pocket compared to their privately insured counterparts. Insurance protects well against certain types of health care costs but not prescription drugs, which represent the fastest rising category of health care expenditures. Implications for Policy, Delivery, or Practice: With states struggling to keep up with rising Medicaid costs, new policies are allowing for more cost-sharing on Medicaid recipients, including children. Increased cost sharing for publicly insured CSHCN will increase the amount these families will pay out of pocket. However, this policy may be justified given that privately insured families at the same income levels currently pay a much higher percent of total health care costs out of pocket. Primary Funding Source: National Institute on Disability and Rehabilitation Research ●Health Care for Children and Youth: Patterns of Coverage, Utilization, Expenditures and Quality by Geographic Area Lisa Simpson, MB, BCh, M.P.H., FAAP, Frances Chevarley, Ph.D., Pamela Owens, Ph.D, Marc Zodet, MS, Marie C. McCormick, M.D., ScD, Denise Dougherty, Ph.D. Presented By: Lisa Simpson, MB, BCh, M.P.H., FAAP, ACH Guild Endowed Chair in Child Health Policy, Pediatrics, University of South Florida, 601 4th St South, CRI 1008, St. Petersburg, FL 33712; Tel: (727) 553-3672; Fax: (727) 553-3666; Email: lsimpso1@hsc.usf.edu Research Objective: Health care availability and other conditions are know to vary between urban and rural areas. Most analyses of geographic differences in children's health care have distinguished between metropolitan (aka urban) and non-metropolitan (aka rural) areas. We aimed to examine differences in child and hospital demographics and children's health care coverage, use, expenditures, and quality using the 2003 Urban Influence Codes. A collapsed version of these codes enable us to make finer distinctions: for metropolitan areas, we could distinguish between the set of counties labeled large or small; for non-metropolitan areas, the UIC enabled us to distinguish between the newly labelled "micropolitan" counties and so-called "non-core" counties, which are typically lower in population density and often remote from large and sophisticated health care settings. Study Design: The 2002 Medical Expenditure Panel Survey and the 2002 Healthcare Cost and Utilization Project Nationwide Inpatient Sample and State Inpatient Databases were used to distinguish sociodemographic patterns, health status, and patterns of insurance coverage, health care utilization, health care expenditures, and health care quality by collapsed versions of the 2003 Urban Influence Codes. In addition, we were able to compare child residence to location of hospitalization by UIC code. Population Studied: U.S. children ages 0-17; U.S. hospitals with discharges for children ages 0-17. Principal Findings: In large metro counties, greater percentages of the child population are Hispanic or black nonHispanic than in small metro, micropolitan and noncore counties; in micropolitan and noncore counties, higher proportions of children live in families with incomes below 200% of the federal poverty level than in large metro and small metro counties. Noncore areas have a greater percentage of children in fair or poor health compared to small metro and micropolitan counties, but not large metro counties. In general, there were no overall differences by child residence in the proportions uninsured, having private health insurance, or public insurance, although differences emerged in subpopulations within UIC types. Noncore low-income children in the South were most likely to be covered by public insurance. The proportion of children with at least one dental visit was generally higher in small metro counties compared to large metro and noncore counties. Children in noncore areas were more likely to have a hospital inpatient stay and any emergency department use compared with children in large metro areas. Fewer medicines were prescribed in large metro areas than all other areas. Over half of hospitalizations for noncore children occurred in county types other than noncore; rates of nonconcordant hospitalizations were higher for conditions require subspecialty care such as mental and emotional disorders and high risk newborn care. Children from noncore counties tended to be hospitalized at greater rates for ambulatory sensitive conditions. Pregnant and postpartum adolescents in noncore areas suffered more complications of pregnancy than their cohorts in other areas. Children from large metro counties were 1.2 times as likely to experience selected nosocomial infections as children from micropolitan areas. Conclusions: Application of the UIC codes allowed us to identify differences not previously seen in grosser comparisons of metropolitan to non-metropolitan areas. In addition, we observed differences within UIC county types by subpopulation. Further exploration of the codes and requirements for adequate sample sizes are needed. Implications for Policy, Delivery, or Practice: In examining children's health care by geography, it seems worthwhile to move beyond the dichotomous notion of urban or metropolitan and rural or non-metropolitan. Primary Funding Source: AHRQ ●Health Beliefs and Health Behaviors of Adolescents with Type 2 Diabetes and their Parents. Asheley Skinner, BS, Shelagh Mulvaney, Ph.D., David Schlundt, Ph.D., Russell Rothman, M.D. Presented By: Asheley Skinner, BS, Health Policy and Administration, University of North Carolina at Chapel Hill, CB 7411, Chapel Hill, NC 27599-7411; Tel: 336-392-5021; Email: asheley@unc.edu Research Objective: The emerging “epidemic” of children and adolescents with type 2 diabetes, which is attributed to increases in childhood overweight, has led to substantial concerns about its impact on young lives. It is likely that both parents’ and children’s perceptions of weight and health beliefs influence adolescents’ self-care behaviors and health care utilization. We hypothesize that: (1) parents and adolescents will underestimate the severity of the adolescent’s overweight and (2) parents and adolescents who perceive greater barriers to, and fewer benefits of, care will engage in fewer self-care behaviors and be more likely to miss clinic appointments. Study Design: Cross-sectional interviews and chart abstraction. Telephone interviews were conducted separately with parents and adolescents in which respondents reported: the adolescents’ estimated weight; their own beliefs regarding diet, exercise, and clinic visits; and perceptions of severity of the adolescents’ overweight (e.g., a little overweight, about right). Adolescents’ actual clinic weights (categorized as normal, at-risk for overweight, or overweight based on bodymass index), clinical parameters, and missed appointments were extracted from chart audits. Both self-reported adherence to recommended regimens and beliefs (e.g., barriers) were assessed separately for diet and exercise using multiple item scales. Multivariate analyses use standardized ordinary least squares regression. Population Studied: Adolescents aged 10-20 years with type 2 diabetes who receive care from an academic health center and their parents (103 pairs) Principal Findings: The adolescents were 85% overweight, 47% white, 69% female, and 42% privately insured. Both parents and adolescents underestimated the adolescent’s weight (mean=7 pounds) and showed poor agreement with actual clinic weights (kappa=0.12 and kappa=0.09). In multivariate analyses, adolescents perceiving more barriers to diet and exercise regimens had poorer adherence to both (b = –0.28; b = –0.48), while parents’ perceptions of greater weight severity are related to better adherence (b = 0.27; b = 0.54). When excluding adolescents’ beliefs, parent self-efficacy is related to poorer adherence to diet (b = –0.38), contrary to expectations. When parents perceived their adolescent’s overweight as more severe, the adolescent had fewer missed appointments (b = –0.36); the opposite was observed for adolescents’ perceptions (b = 0.36). Conclusions: Despite being engaged in the health system and having a serious chronic condition related to their weight, adolescents and their parents underestimated the severity of the adolescent’s overweight. Parents and adolescents’ perceptions regarding the severity of overweight appeared to have competing effects on health behaviors and missed appointments. Implications for Policy, Delivery, or Practice: While virtually all patients would be classified as overweight, these adolescents with type 2 diabetes and their parents both underestimated the actual weight and its severity. Our findings indicate the need to address inconsistencies in parents’ and adolescents’ perceptions of overweight severity. Interventions to enhance self-care must target both adolescents and parents, and must be tailored to their own perceptions of adolescents’ overweight. Primary Funding Source: Vanderbilt Physician Scientist Development Award ●The Impact of State Medicaid Policy on Pediatric Patient Safety in Hospitals Richard Smith, Ph.D., Lisa Simpson, MB, BCh, M.P.H., FAAP, Robyn Cheung, Ph.D., Pamela Owens, Ph.D., Aileen Sedman, M.D., R. Mark Wilson, Ph.D. Presented By: Richard Smith, Ph.D., Assistant Professor, College of Business, University of South Florida St. Petersburg, 140 Seventh Avenue South, St. Petersburg, FL 33701; Tel: (727) 553-4943; Fax: (727) 553-4192; Email: smithrb@stpt.usf.edu Research Objective: Evidence is growing that pediatric patient safety events do not affect all populations similarly, and in particular, that publicly insured patients may be at greater risk of adverse events. Moreover, while Medicaid pays for approximately 17 percent of adult inpatient hospitalizations, it is primary payer for approximately 40 percent of pediatric hospitalizations. The purpose of this study is to examine the relationship of Medicaid characteristics at the patient, hospital, and market level, to adverse medical events for hospitalized children. Study Design: Three of the Agency for Healthcare Research and Quality (AHRQ) patient safety indicators (decubitis ulcer, infection due to medical care, and laceration) are selected for analysis based on their high probability of preventability. Hospital discharge data from 1999-2001 State Inpatient Database from the AHRQ HCUP is combined with hospitallevel data (e.g. nurse staffing, technology) through linkage to the American Hospital Association database. Case-mix variables at the individual level (co-morbidities) as well as at the hospital level (using APR-DRG severity indices) are also included, along with Medicaid variables (% discharges Medicaid and total DSH payments) and measures of both hospital and Medicaid managed care competition using Herfindahl indices. These latter variables are obtained from CMS and state Medicaid offices, as well as derived from the HCUP. For each of the three patient safety indicators, a logistic regression is used to estimate the probability of an individual patient experiencing an adverse event. Population Studied: All pediatric hospital discharges between 1999 and 2001 for patients of Medicaid-eligible age in Florida, New York, and Wisconsin, and for which any of the three patient safety indicator (PSI) events could occur. Principal Findings: Medicaid recipients have significantly higher rate of decubitis ulcers (OR=1.30) and infections due to medical care (OR=1.19), but significantly lower rates of laceration (OR=0.76) than non-Medicaid recipients. In addition, relative to markets that are competitive for both hospitals and payers, concentrated hospital markets (Herfindahl>1800) significantly reduce PSI rates in 2 of 3 cases (Infection, OR=0.76; Laceration, OR=0.71), while an inconsistent patttern emerges in concentrated payer markets (Infection, OR=0.45; Laceration, OR=1.62). Conclusions: While we find significant associations at all three levels of Medicaid characteristics, the more significant and largest effects generally appear at the market level. Markets with only high hospital concentration tend to be associated with the lowest incidence of adverse events. The most compelling explanation, given our results, is that hospital markets with high concentration produce high profits that are then invested in quality-control mechanisms. This explanation is consistent with other recent findings on the relationship between hospital profitability and patient safety (Encinosa and Bernard, 2005). Implications for Policy, Delivery, or Practice: Recent trends are for greater concentration in both hospital and payer markets. From that perspective, and from our results, markets with high payer concentration have a relatively greater likelihood of adverse medical events for pediatric patients. Thus, while states have yet to begun playing a significant role in hospital patient safety, one important role they can play is by ensuring sufficient competition of commercial Medicaid managed care payers in local hospital markets. Primary Funding Source: HRSA, Pediatric Clinical Research Center of All Children's Hospital and the University of South Florida ●Comparing the Experiences of CHIP and Medicaid Children Jeffery Talbert, Ph.D., Kristine Lykens, Ph.D. Presented By: Jeffery Talbert, Ph.D., Professor and Chair, Health Management and Policy, University of North Texas Health Science Center, 3500 Camp Bowie Boulevard, Fort Worth, TX 76107; Tel: 817-735-5027; Email: jtalbert@hsc.unt.edu Research Objective: This study compares the experiences of children and their families in the Kentucky Medicaid and CHIP programs from 1999-2005. Study Design: The study uses a retrospective analysis of CAHPS surveys administered annually from 1999 to 2005. The sample is a statewide proportional sample stratified by program type and county for 5000 children each year. Multivariate analysis is conducted to compare outcomes between children enrolled in CHIP and Medicaid, focusing on quality, access, and health outcomes. Controlls include age, gender, year, rural location, eligibility type, and managed care type. Population Studied: The population is an annual grouping of all children enrolled in the Kentucky Medicaid program and the Kentucky Children Health Insurance Program for at least 11 months during each year. Principal Findings: The professional literature and much legislative discussion has debated the impact of allowing states to run CHIP programs through Medicaid. Many analysts predicted that Medicaid run programs would provide lower levels of care due to stigma and provider disinterest in supporting a Medicaid based system. Conclusions: Our findings indicate few differences in the experiencces of children enrolled in CHIP and Medicaid. Using measures of satisfaction, wait times for appointments, and health outcomes there are no robust differences between services provided to CHIP children and Medicaid children. Implications for Policy, Delivery, or Practice: The SCHIP and Medicaid programs are up for reauthorization in 2007, where questions about the most appropriate design will be explored. This study provides evidence of the impact of program design for policy makers at the state and federal level who may be considering redesigning or continuing current programs. Efficiencies gained through using current Medicaid infrastructure do not appear to impact quality and health outcomes for children. Primary Funding Source: No Funding ●Options for Introducing a Hospital-Based Newborn Hearing Screening Programme in New Zealand: A Cost Effectiveness Analysis Laura Wilkinson-Meyers, MSc Health Policy, Planning & Financing, Paul Brown, Ph.D. Presented By: Laura Wilkinson-Meyers, MSc Health Policy, Planning & Financing, Research Fellow, Centre for Health Services Research + Policy, University of Auckland, School of Population Health, Tamaki Campus, Auckland, 1009; Tel: +64 9 373 7599 x87659; Email: l.wilkinson-meyers@auckland.ac.nz Research Objective: The National Screening Unit (NSU) commissioned the Centre for Health Services Research and Policy to examine the cost effectiveness of implementing a national newborn hearing screening programme in New Zealand to detect cases of permanent congenital hearing impairment. This paper presents the results of the analysis for a number of hospital-based screening programme options. Study Design: A decision analytic model was constructed for various screening scenarios, including: the current high risk referral programme of detection, introducing hospital-based screening for neonatal intenstive care unit (NICU) infants only, and three different screening technologies to be used for well infants. Technologies considered included combinations of otoacoustic emissions testing (OAE), and automated ABR testing, followed by a diagnostic exam. The main outcome measure for the analysis was the number of infants diagnosed by the age of 6 months. Probabilities for decision nodes in the model were obtained from New Zealand data sources where possible, and international evidence where no information was found. Three categories of cost were used: cost of screening, cost of diagnostic assessment, and the cost of monitoring and administering the programme. Given that a societal perspective was taken for the analysis, the lifetime cost savings that would result from early detection of hearing loss were also included. Population Studied: All infants born in New Zealand. Principal Findings: Using an OAE+aABR screening protocol for all well infants screened in a hospital setting is is the most cost effective screening option when compared to the current high risk referral programme of detection. Conclusions: The incremental cost effectiveness analysis suggests that implementing a hospital-based screening programme results in a greater number of early detections of hearing impairment. The cost of implementing the programme is greater than the current costs associated with high risk referral detection, but the cost savings that can be expected to result over the lifetime of the infant far outweigh the costs of implementing and monitoring a screening programme. Implications for Policy, Delivery, or Practice: Introducing a newborn hearing screening programme in New Zealand would be cost effective and these findings are consistent with previous international studies. However, this research is unique in that it is tailored to the New Zealand health care system, and it includes national monitoring and oversight costs. Primary Funding Source: New Zealand National Screening Unit