This section provides an introduction to the chartbook and highlights the importance of addressing the quality of health and health care for children and adolescents in Florida. In addition to brief descriptions of the sponsoring agencies, this section outlines the contents and methods used in the development of this chartbook. What We Know About the Quality of Healthcare for Children While many children in Florida and the nation received effective and high quality care that meets their needs, numerous reports over the last decade have shown that this is not always the case.7 The Institute of Medicine released a landmark report in 2001, Crossing the Quality Chasm, which summarized these findings and called for fundamental change to close the quality gap.8 In 2004, the Commonwealth Fund released Quality of Care for Children and Adolescents: A Chartbook, a report that provides the first detailed look at the quality of children's health care after reviewing over 500 studies.9 That same year, the Agency for Healthcare Research and Quality (AHRQ) released the first National Healthcare Quality and Disparities Reports which included several indicators relevant to children's healthcare. More recently, a study by researchers at RAND Health concluded that overall, children and adolescents received less than 50% of clinically recommended care.10 This chartbook builds on this work to shine a light on what we are learning about the health and healthcare quality of children in Florida. We hope that by describing the quality of care for children in our State compared to national benchmarks will spur action by policymakers and other stakeholders in children's health in the areas where we have an opportunity to improve.
What's Different About This Chartbook?
This is the first in a series of reports from FLICHQ that is expanding the information available to stakeholders. Our goal is to provide user friendly information that is actionable at the state and local level so that individuals and organizations can continue and enhance their work toward improving children's health and healthcare in Florida. This chartbook is intended to complement, not duplicate, existing reports available from other stakeholders by using new datasets and by delving more deeply into the health care experiences of children, youth and their families. Important additional sites that users may want to use to complement this report include the Florida Kids Count,2 data from the Florida Department of Health,3 and the Florida Agency for Healthcare Administration.5
Data Sources and Methods
Seven primary state and national data sources were used for this chartbook: - National Survey of Children’s Health (NSCH), 2003
- National Survey of Children with Special Health Care Needs (NS_CSHCN), 2005/2006
- Youth Risk Behavior Surveillance System (YRBSS), 2003
- State and national hospital discharge data from the Health Care Costs and Utilization Project (HCUP), including the Florida State Inpatient Database (SID) and Nationwide Inpatient Sample (NIS), from the Agency for Healthcare Research and Quality.
- Florida Medicaid Utilization (claims) Data, 2005 from the Agency for Health Care Administration (AHCA)
- Florida Health Insurance Survey, 2004 from the University of Florida
- Mail Survey Component of the Evaluation of Florida’s Medicaid Managed Mental Health Plans in Medicaid Areas 5 and 7, in 2005 from the Louis de la Parte Florida Mental Health Institute.
A full list of data sources, including detailed descriptions of each source and their limitations, and analytical methods used for this report can be found in Appendix A: Technical Notes.
Indicator Selection and Development
CAHMI and FLICHQ staff selected child health indicators from the NSCH, NS-CSHCN and YRBSS. The proposed indicators were presented to Florida child health stakeholders for review in a series of consultation sessions. Input from the stakeholders was used to revise and refine the selected indicators. Additional indicators from HCUP were developed to characterize health care utilization and hospital care.The age range used for children varies by dataset. Most of the indicators refer to children and youth under 18 years of age, but some include children up to age 21. Subsets are drawn from the population of children and youth represented in each dataset to examine specific groups of children and youth (e.g., ages 0 to 5, children who have a personal doctor or nurse). For consistency, African American is used throughout the chartbook to encompass children and youth with African American and/or black race or ethnicity. Users should be aware of the general strengths and limitations of the different data systems. For example, population-based surveys obtain sociodemographic data, data on family characteristics, and information on health behaviors. These data are limited by the amount of information a respondent remembers or is willing to report. Specific medical information may not be known, and if known, may not be reported. Conversely, utilization data represent health services usage but do not always represent the quality of services provided nor the consumers’ satisfaction with said services.Hospital administrative data are limited due to confidentiality and data quality. National Survey of Children’s Health 2003, National Center for Health Statistics The National Survey of Children’s Health is a national telephone survey conducted during 2003-2004 by the National Center for Health Statistics at the Centers for Disease Control and Prevention, funded by the Maternal and Child Health Bureau of the U.S. Department of Health and Human Services. The survey is intended to: 1) estimate national and state-level prevalence for a variety of child health indicators, 2) generate information about children, families, and neighborhoods to help guide policymakers, advocates, and researchers, 3) provide baseline estimates for federal and state performance measures, Healthy People 2010 objectives and states’ needs assessments, and 4) complement the 2005/2006 National Survey of Children with Special Health Care Needs.The survey provides a broad range of information about children’s health and well-being collected in a manner that allows comparisons among states as well as nationally. A total of 102,353 surveys completed nationally by parents/caregivers of children ages of 0 to 17; between 1,483–2,241 surveys collected per state. Survey results are weighted to represent the population of noninstitutionalized children ages 0to17 nationally,and in each state. Specific topics include demographics, physical and mental health status, health insurance, access and use of health care services,medical home, early childhood-specific information (ages 0 to 5), middle childhood and adolescent-specific information (6 to 17), family health and activities, parental health status and perceptions of neighborhood characteristics. Univariate, bivariate and limited multivariate analyses of weighted data were conducted using complex sampling methods. Master tables were constructed from the results and were used in turn to generate charts and tables for the report. A 95-percent confidence interval was constructed for each population and proportion estimate;these are reported in chapter summary tables in the appendices. Items with Relative Standard Error greater than 30 percent were identified and are noted as shaded cells (grey) in charts and tables. The selected child health indicators are stratified by geographic area (Florida or National) as well as by several demographic variables. Age groups used for the National Surveys were 0 to 5,6 to 11, and 12 to 17. Categories for race and ethnicity are Hispanic (all races and multi-racial), White (non-Hispanic), African American (non-Hispanic), and Other (non-Hispanic, but including multi-racial, Asian, Native American, and other races and ethnicities). Additional stratification variables are insurance type (private, public or uninsured), income (as percentage of Federal Poverty Level) and special health care needs status (as assessed by the CSHCNScreener).1,2 Additional information and results from the NSCH are available in the Data Resource Center for Child and Adolescent Health, www.childhealthdata.org. National Survey of Children with Special Health Care Needs (2005/2006), National Center for Health Statistics The 2005/2006 National Survey of Children with Special Health Care Needs (NS-CSHCN) is a nationwide telephone survey funded by the Maternal and Child Health Bureau, U.S. Department of Health Resources and Services Administration and conducted by the National Center for Health Statistics using SLAITS (State and Local Area Integrated Telephone Survey) technology for sampling and administration.The 2005/2006 NS-CSHCN sample was achieved by screening 364,841 children representing 192,083 households nationwide, using the CSHCN Screener.1 Telephone interviews were completed for a total of 40,840 randomly selected children and youth with special health care needs (CYSHCN) ages 0 to 17, approximately 800 in each state and the District of Columbia. The survey data are weighted to reflect the population of non-institutionalized children ages 0 to 17 in each state. NS-CSHCN provides detailed state and national level parent-reported information on the health status and health care system experiences of children and youth with special health care needs (CYSHCN) and their families. Topics covered by the survey include health and functional status, insurance coverage and adequacy of coverage, access to health care services, medical home, impact of children’s special needs on their families, family-centeredness of services, and care coordination.3 Univariate, bivariate and limited multivariate analyses of weighted data were conducted using complex sampling methods. Master tableswere constructed from the results and were used in turn to generate charts and tables for the report. A 95-percent confidence interval was constructed for each population and proportion estimate; these are reported in chapter summary tables inthe appendices. Items with Relative Standard Error greater than 30 percent were identified and are noted in charts and tables as shaded grey cells. The selected child health indicators are stratified by geographic area (Florida or National) as well as by several demographic variables. Age groups used for the National Surveys were 0 to 5,6 to 11, and ages 12 to 17. Categories for race and ethnicity are Hispanic (all races and multi-racial), White (non-Hispanic), African American (non-Hispanic), and Other (non-Hispanic, but including multi-racial, Asian, Native American, and other races and ethnicities). Additional stratification variables are insurance type (private,public or uninsured) and income (as percentage of Federal Poverty Level). Additional information and results from the NS-CSHCN are available in the Data Resource Center for Child and Adolescent Health, www.childhealthdata.org. Youth Risk Behavior Surveillance System 2003, Centers for Disease Control and Prevention The Youth Risk Behavior Surveillance System (YRBSS) was designed to determine the prevalence of health-risk behaviors among high school students;assess whether these behaviors increase, decrease, or stay the same over time; and examine the co-occurrence of health-risk behaviors. Six categories of priority health-risk behaviors among youth and young adults are monitored, including behaviors that contribute to unintentional injuries and violence; tobacco use; alcohol and other drug use; sexual behaviors that contribute to unintended pregnancy and sexually transmitted diseases, including human immunodeficiency virus (HIV) infections; unhealthy dietary behaviors; and physical inactivity. In addition, the YRBSS monitors general health status and the prevalence of obesity and asthma. YRBSS includes a national school-based survey conducted by the Centers for Disease Control and Prevention and state and local school-based surveys conducted by state and local education and health agencies. The 2003 YRBSS encompasses results fromthe 15,214 student respondents in the national survey, as well as 43 state surveys, and 22 local surveys conducted among students in grades 9 to 12. In Florida, 4080 students from 75 public high schools completed the YRBSS in 2003.4 In addition to comparable national, state and local data, the YRBSS also provides data for subpopulations of youth (age and racial/ethnic groups) and monitors progress toward achieving national health objectives for 2010 and other program indicators. The national survey uses a three-stage, cluster sample design to obtain a nationally representative sample of students in grades 9 to 12 in the United States. State and local school-based surveys employ a two-stage, cluster sample design to produce representative samples of students in grades 9 to 12 in their jurisdiction. Univariate, bivariate and limited multivariate analyses of weighted data were conducted using complex sampling methods. Master tables were constructed from the results and were used in turn to generate charts andtables for the report. A 95-percent confidence interval was constructed for each population and proportion estimate; these are reported in chapter summary tables in the appendices. Items with Relative Standard Error greater than 30 percent were identified and are noted in charts and tables as shaded grey cells. The selected child health indicators are stratified by geographic area (Florida or National) as well as three demographic variables: age, gender, and race/ethnicity. The YRBSS data was stratified according to grade level reported by the student respondents. Categories for race and ethnicity are Hispanic (all races and multi-racial), White (non-Hispanic), African American (non-Hispanic), and Other (non-Hispanic, but including multi-racial, Asian, Native American, and other races and ethnicities). Florida Health Insurance Survey (FHIS,2004), Agency forHealthcare Administration (AHCA) Data on the level health insurance for Florida children are from the 2004 Florida Health Insurance Study supported by a grant from HRSA to the Agency for Health Care Administration (AHCA). AHCA contracted with the University of Florida (UF) Bureau of Economic and Business Research (BEBR) and the Department of Health Services Research Management and Policy to conduct the survey. A stratified, random sample was used to estimate uninsurance at the state, district and county level. Various characteristics for uninsured Floridians were assessed including race/ethnicity, poverty level, employment and health status, access and utilization. Telephone interviews were conducted between April and August of 2004 with 17,435 Florida households, collecting data on approximately 46,876 individuals under age 65. The primary goal of the study was to estimate the percent of nonelderly (under age 65) Floridians without health insurance, however, this chartbook only presents findings for children (ages 0 to 18) are presented. Some of the data in this report was drawn directly fromthe FHIS report while additional analyses were provided courtesy of researchers at UF. Additional information and results from the FHIS are available at AHCA’s Office of Medicaid Research and Policy Projects at: http://ahca.myflorida.com/Medicaid/quality_management/mrp/Projects/fhis2004/reports.shtml Data on children and youth enrolled in Florida Medicaid are from utilization (claims) data for 2005 collected from the Agency for Health Care Administration. The sample contained claims for children and youth ages 0 to 18 throughout Florida. For emergency department visits, all emergency department (ED) claims were assessed including those with obstetrical or psychiatric diagnoses. ED claims were stratified by race and ethnicity, gender and region. The proportion of children who had at least one well-child visit in the past 12 monthswere examined for two specific groups of children separately (ages 3 to 6 and ages12 to 17). Well-child visits were stratified by race and ethnicity, gender and region. The proportion of children with asthma who received appropriate asthma medications were examined for children ages 5 to 17. Children with asthma were stratified by age group (ages 5 to 9 and ages 10 to 17 years). The technical specifications for Well-Child Visits and Appropriate Asthma Medications were from 2005 Health Plan Employer Data and Information Set (HEDIS) from the National Committee for Quality Assurance. Additional information from the Office of Medicaid Research and Quality at the Bureau of Medicaid Quality Management in Florida’s Agency for Health Care Administration at:http://www.fdhc.state.fl.us/Medicaid/quality_management/index.shtml Evaluation of Florida’s Medicaid Managed MentalHealth Plans (2005), Louis de la Parte Florida Mental Health Institute (FMHI) Data regarding child and youth mental health care are from the Summary of the Areas 5 and 7,2005:Mail Survey Component of the Evaluation of Florida’s Medicaid Managed Mental Health Plans conducted from February through April 2005. The survey was supported by the Office of Medicaid Research and Policy at AHCA and the Louis de la Parte FMHI at the University of South Florida (USF). The goal of the evaluationwas to obtain and monitor a variety of Medicaid beneficiary indicators including health andmental health status, service needs and use, and satisfaction with services, and assess changes in these indicators over time. The findings reflect the status of beneficiaries just as the implementation of managed mental health care was getting underway in both areas.Medicaid beneficiaries were randomly sampled after being stratified by age (children ages 5-21, adults ages 22 and older), eligibility status Supplemental Security Income [SSI], Temporary Assistance to Needy Families [TANF]), gender and race/ethnicity. Findings in this report are limited to child beneficiaries (ages 5 to 21). The sample included parents/caregivers of 475 children residing in areas 5 and 7 (i.e.,Orange, Osceola, Pasco, Pinellas, Seminole,and Brevard Counties).Caregivers self-reported their:1) child’s utilization and access to medical and mental health services and medications, 2) child’s physical and mental health status, 3) satisfaction with medical and mental health services, 4) satisfactionwith their child’sMedicaid health care plan, 5) trust in their child’s health care providers and 6) family’s quality of life. Special descriptive analyses conducted by researchers at the Florida Mental Health Institute for this report served as the basis of the data presented in this report. Additional information and results from Summary of the Areas 5 and 7, 2005: Mail Survey Component of the Evaluation of Florida’s Medicaid Managed Mental Health Plans are available at http://mhlp.fmhi.usf.edu/web/mhlp/rdetail.cfm?prid=245Healthcare Cost and Utilization Project (HCUP,2003), Agency for Healthcare Research and Quality (AHRQ) The data for the hospital related analyses come from the 2003 Florida State Inpatient Databases (SID) and the Nationwide Inpatient Sample (NIS), available for research uses through the USF/All Children’s Hospital (ACH) Pediatric Clinical Research Center (PCRC). The Florida SID is part of the Healthcare Cost and Utilization Project (HCUP) sponsored by the Agency for Healthcare Research and Quality (AHRQ) and AHCA.5 The Florida SID is an all payer, inpatient care database which contains discharge abstracts from all community, non-rehabilitation hospitals in the state.3 The NIS is an all-payer, hospital inpatient database which contains all records from a 20-percent stratified sample of U.S. community, non-rehabilitation hospitals. The 2003 NIS includes information from approximately 8 million hospital discharges that were weighted to obtain estimates that represent the total number of inpatient discharges in the U.S. (38.2 million).7 The NIS sample discharges are weighted based on the population of all community hospitals that were open during the year. The NIS is a sample of these hospitals, stratified by region, location/teaching status, bed size, and ownership;for a total of 60 strata. For each stratum, a random sample of 20 percent of hospitals in that stratum is selected. Discharges in the NIS come from these selected hospitals, with each discharge weighted by the ratio of the total number of national discharges in the population to the number of sample discharges in that stratum.5 Analyses are performed using SPSS and/or SAS. The characteristics of children and their hospital stays are assessed including; age, gender, race and ethnicity, expected payer, disposition at discharge, region of patient residence, charges, length of stay, and hospital characteristics. Region is constructedby aggregating the 11 Florida Medicaid areas into 4 geographic areas (Northwest, Northeast, Southwest and Southeast).9 A child’s region of residence is identified using zip code of residence.See Appendix Table 4-1 for a list of the counties in each region. To be consistent with national reporting standards, expected payer is categorized as Private, Medicaid, Uninsured (i.e., self-pay and no charge) and Other. Other expected payers include all non-Medicaid/non-private sources of insurance which includes Medicare, Champus, Veteran’s Administration, other state and local government programs, Worker’s Compensation and other parts of the KidCare program (i.e., Healthy Kids, MediKids and Children’s Medical Services). In the Florida SID and NIS it is not possible to determine the exact number of children in each of the payer types grouped into the other category. The HCUP dataset uses the term “black”, however, to be consistent with other datasets, “African American” is used in this chartbook. The 2003 Florida Hospital Inpatient data provided directly from the Agency for Health Care Administration (AHCA) can provide estimates of the distribution of the different groups collapsed into the Other category.10 In 2003, there are approximately 3,860 newborn hospitalizations with Other payer which are distributed as follows: Champus (57.0%),other state and local programs (18.4%),other types (19.1%), parts of KidCare (1.7%), and Medicare/Worker’s Compensation/Veteran’s Administration (3.8%). There are about 9,916 child and youth hospitalizations (ages 0 to 19, newborns excluded) with Other payer which are distributed as follows: Champus (29.9%), Other state and local programs (29.0%), parts of KidCare (26.8%), other types (9.2%), and Medicare/Worker’s Compensation/ Veteran’s Administration (5.2%). Hospital characteristics (i.e., location/teaching status, size, and control), as defined by AHRQ, are assessed. In order to determine hospital characteristics, discharge data are linked with hospital data fromthe American Hospital Association’s Annual Survey Database. The definitions of the hospital characteristics in this report are consistent with the specifications used in the NIS.11 A difference in proportion chi-square testwas used to assess for significant differences in proportions or rates of selected characteristics within or across selected groups. T-tests were used to test for mean differences in length of stay or charges. The level of significance was set at .05 for all testing. Additional information about all the HCUP datasets is available from the AHRQ at http://www.ahrq.gov/data/hcup/. Hospital Utilization The sample includes all children and youth (ages 0-19) and newborns discharged from community hospitals and nationwide in 2003. Children and youth were assessed separately from newborns. The characteristics of children and their hospital stays are assessed including; charges, length of stay, age, gender, race and ethnicity, expected payer, disposition at discharge, region of patient residence and hospital characteristics. Significant differences in proportions of selected characteristics were tested using standard difference of proportion tests. T-tests were used to test formeandifferences inlengthof stayor charges. The level of significance is set at 0.05, two-tail, level for all testing. Leading Causes of Hospitalization Leading causes of hospitalizations were grouped using the broad categories used in the Clinical Classification System (CCS) from AHRQ. CCS categories are based on the ICD-9-CM diagnosis codes. The CCS collapses ICD-9-CM codes into a smaller number of clinically meaningful categories.12 The sample includes all children and youth (ages 0 to 19, newborns excluded) discharged from community hospitals in Florida and nationwide in 2003. HCUP Pediatric Quality Indicators AHRQ, in conjunction with researchers at its Evidence-Based Practice Center at the University of California and Stanford University, developed and released in 2006 a revised set of Pediatric Quality Indicators.13 This set of indicators was developed to use hospital discharge data to promote internal review of the healthcare quality of care being delivered to children and spur quality improvement. A subset of those indicators is used in this report. Potentially Avoidable Hospitalizations (PAH): Admissions for asthma, gastroenteritis, short-term complications of diabetes mellitus and urinary tract infection are assessed. Admission rates (area-level indicators) per 100,000 child population are calculated for all area-level, quality indicators combined and each one individually. The numerator includes all eligible admissions of children ages 0-19 years, however, the lower age limit does vary by condition (e.g., admissions of children less than age 6 are excluded from the numerator or diabetes). The denominator includes all eligible children ages 0 to 19 in the population.Inclusion and exclusion criteria are consistent with the technical specifications except the upper-age limit is extended to included children up to age 19 to get a broader picture of potentially avoidable hospitalizations for these conditions.14 Population-based rates were also calculated for subgroups stratified by ages (0 to 4, 5 to 9, 10 to 14 and 15 to 19), gender and region of residence. Population estimates were based on United States Census Bureau data.15-16 Population-based rates could not be calculated for subgroups based on race and ethnicity because definitions were not compatible between discharge and census datasets. Safety of Hospital Care: For patient safety indicaors (provider-level indicators), the most current version of the AHRQ Pediatric Quality Indicator software allows for risk adjustment which provide a set of risk categories to assign patients based on their diagnostic profile and the procedures performed during their hospitalization.14 Because the most severe and complex hospital cases are more likely to have adverse medical outcomes, it is sometimes difficult to separate unambiguously preventable adverse events from those adverse events that are less clearly preventable in more difficult cases.The software uses this information, along with the age, sex, and any co-morbid conditions of patients, to estimate risk-adjusted rates at various levels of aggregation. In this analysis, we calculate aggregate, risk-adjusted rates of adverse events for each of the four regions of the state defined in this study.17 In addition, we compare mean length of stay and hospital charges between cases in which no adverse event occurred, and those in which an adverse event occurred for low-risk discharges (i.e. cases in which the adverse event was most likely preventable). For decubitus ulcer, the low-risk group consists of those discharges in risk category 1 (i.e., all eligible pediatric discharges excluding those quadriplegia, paraplegia, hemiplegia, spina bifida or anoxic brain damage; for Selected Infection Due to Medical Care (Infection), the low-risk group consists of those discharges in risk categories 1 or 2 (i.e., the low risk and intermediate risk group); for Accidental Puncture or Laceration (Laceration), the low-risk group consists of those discharges in risk categories 1,2,3,4,or 5 (i.e.,discharges with no therapeutic, minor therapeutic,one major therapeutic without diagnostic, one major therapeutic with minor diagnostic, one major therapeutic with major diagnostic or two major therapeutic procedures).14, 18 Low Birthweight Hospitalizations: Low birthweight (LBW) is a prevention quality indicator developed by AHRQ.13 The characteristics of newborns and theirhospital stays are compared to those of newborns without low birthweight (NLBW). Inclusion and exclusion criteria to identify all newborns and low birthweight newborns are consistent with the technical specifications provided.19,20 Due to differences inthe availability of data elements and the level of accuracy in identifying birth hospitalizations and birthweight, estimates of low birthweight from administrative, hospital discharge data differ from estimates using vital statistics data.21Inaddition, not all births occur incommunityhospitals. For all hospital discharge data, significant differences in proportions or rates of selected characteristics/indicators and are tested using standard difference of proportion tests. T-tests were used to test for mean differences in length of stay or charges. All testing is done with significance set at the 0.05, two-tail, level.Limitations National Surveys Because the parent-report formats of the NSCH, NSCSHCN and YRBSS, responses are subject to limitations of recall and willingness to disclose information, although every effort was made to assure confidentiality in all surveys. Errors in survey administration resulted in missing data on some questions in the NSCH and NSCSHCN. The NSCH and NS-CSHCN do not include data from households that do not have a phone, nor do they include children and youth who live in institutional settings. The YRBSS does not include youth who are not enrolled in high school.In multivariate and sometimes bivariate analysis of NSCH, NS-CSHCN and YRBSS, the total responses in some categories are occasionally so low that the resulting estimates are not statistically reliable. Relative Standard Error was used to identify these results and they were either noted or not reported. Data for all three surveys are weighted only to the state level, and cannot be broken down by city, county or other sub-state regions. Details about sampling, weighting and administration of the three surveys, is available in their respective methods reports.2, 3, 4 HCUP Hospitalization Data For confidentiality purposes, all data are de-identified; therefore, researchers cannot determine the number of unique individuals discharged. It is impossible to identify how many discharges represent multiple discharges of the same person. Data quality can also be a limitation because of missing data, data entry error and coding practices by those who abstract data fromhospitalmedical records. Individuals who abstract data from medical records can vary in their interpretation of the medical record and data entry accuracy. Variation in medical record abstraction validity and reliability can occur within and across facilities. |