Advertisement

Annual Report on Health Care for Children and Youth in the United States: Focus on Trends in Hospital Use and Quality

  • Bernard Friedman
    Correspondence
    Address correspondence to Bernard Friedman, PhD, Agency for Healthcare Research and Quality, 540 Gaither Road, Rockville, Maryland 20850.
    Affiliations
    Department of Health and Human Services, Agency for Healthcare Research and Quality, Rockville, Md (Dr Friedman, Dr Berdahl, Dr Andrews); Academy Health, Washington, DC (Dr Simpson); Department of Society, Human Development and Health, Harvard School of Public Health, Boston, Mass (Dr McCormick); Agency for Healthcare Research and Quality, Rockville, Md (Dr Owens); Department of Internal Medicine, Division of Infectious Disease, Washington University School of Medicine, St. Louis, Mo (Dr Owens); Division of General Pediatrics, and Center for Healthcare Policy and Research, University of California, Davis, Sacramento, Calif (Dr Romano)
    Search for articles by this author
  • Terceira Berdahl
    Affiliations
    Department of Health and Human Services, Agency for Healthcare Research and Quality, Rockville, Md (Dr Friedman, Dr Berdahl, Dr Andrews); Academy Health, Washington, DC (Dr Simpson); Department of Society, Human Development and Health, Harvard School of Public Health, Boston, Mass (Dr McCormick); Agency for Healthcare Research and Quality, Rockville, Md (Dr Owens); Department of Internal Medicine, Division of Infectious Disease, Washington University School of Medicine, St. Louis, Mo (Dr Owens); Division of General Pediatrics, and Center for Healthcare Policy and Research, University of California, Davis, Sacramento, Calif (Dr Romano)
    Search for articles by this author
  • Lisa A. Simpson
    Affiliations
    Department of Health and Human Services, Agency for Healthcare Research and Quality, Rockville, Md (Dr Friedman, Dr Berdahl, Dr Andrews); Academy Health, Washington, DC (Dr Simpson); Department of Society, Human Development and Health, Harvard School of Public Health, Boston, Mass (Dr McCormick); Agency for Healthcare Research and Quality, Rockville, Md (Dr Owens); Department of Internal Medicine, Division of Infectious Disease, Washington University School of Medicine, St. Louis, Mo (Dr Owens); Division of General Pediatrics, and Center for Healthcare Policy and Research, University of California, Davis, Sacramento, Calif (Dr Romano)
    Search for articles by this author
  • Marie C. McCormick
    Affiliations
    Department of Health and Human Services, Agency for Healthcare Research and Quality, Rockville, Md (Dr Friedman, Dr Berdahl, Dr Andrews); Academy Health, Washington, DC (Dr Simpson); Department of Society, Human Development and Health, Harvard School of Public Health, Boston, Mass (Dr McCormick); Agency for Healthcare Research and Quality, Rockville, Md (Dr Owens); Department of Internal Medicine, Division of Infectious Disease, Washington University School of Medicine, St. Louis, Mo (Dr Owens); Division of General Pediatrics, and Center for Healthcare Policy and Research, University of California, Davis, Sacramento, Calif (Dr Romano)
    Search for articles by this author
  • Pamela L. Owens
    Affiliations
    Department of Health and Human Services, Agency for Healthcare Research and Quality, Rockville, Md (Dr Friedman, Dr Berdahl, Dr Andrews); Academy Health, Washington, DC (Dr Simpson); Department of Society, Human Development and Health, Harvard School of Public Health, Boston, Mass (Dr McCormick); Agency for Healthcare Research and Quality, Rockville, Md (Dr Owens); Department of Internal Medicine, Division of Infectious Disease, Washington University School of Medicine, St. Louis, Mo (Dr Owens); Division of General Pediatrics, and Center for Healthcare Policy and Research, University of California, Davis, Sacramento, Calif (Dr Romano)
    Search for articles by this author
  • Roxanne Andrews
    Affiliations
    Department of Health and Human Services, Agency for Healthcare Research and Quality, Rockville, Md (Dr Friedman, Dr Berdahl, Dr Andrews); Academy Health, Washington, DC (Dr Simpson); Department of Society, Human Development and Health, Harvard School of Public Health, Boston, Mass (Dr McCormick); Agency for Healthcare Research and Quality, Rockville, Md (Dr Owens); Department of Internal Medicine, Division of Infectious Disease, Washington University School of Medicine, St. Louis, Mo (Dr Owens); Division of General Pediatrics, and Center for Healthcare Policy and Research, University of California, Davis, Sacramento, Calif (Dr Romano)
    Search for articles by this author
  • Patrick S. Romano
    Affiliations
    Department of Health and Human Services, Agency for Healthcare Research and Quality, Rockville, Md (Dr Friedman, Dr Berdahl, Dr Andrews); Academy Health, Washington, DC (Dr Simpson); Department of Society, Human Development and Health, Harvard School of Public Health, Boston, Mass (Dr McCormick); Agency for Healthcare Research and Quality, Rockville, Md (Dr Owens); Department of Internal Medicine, Division of Infectious Disease, Washington University School of Medicine, St. Louis, Mo (Dr Owens); Division of General Pediatrics, and Center for Healthcare Policy and Research, University of California, Davis, Sacramento, Calif (Dr Romano)
    Search for articles by this author

      Abstract

      Objective

      The aim of this study was to describe selected trends in hospital inpatient care for children between 2000 and 2007.

      Study Design

      Analysis was conducted of administrative data from annual nationwide databases of hospital discharges from the Agency for Healthcare Research and Quality’s Healthcare Cost and Utilization Project, along with survey data from a nationally representative random sample of children from the Medical Expenditure Panel Survey. Hospital utilization rates and expenses, risk-adjusted rates of potentially avoidable hospitalization, and safety indicators in the hospital are calculated and tracked with established and downloadable software.

      Results

      The rate of hospital discharges for children aged 15 to 17 years declined significantly, mainly due to fewer maternity-related discharges. The leading principal conditions by age group were similar to the report for 1995 to 2000; however, the rate of admissions for skin infections doubled to 9 per 10 000. Hospital cost per discharge increased by an annual average of 4.5% per year compared with 2.6% annual growth in the gross domestic product deflator. Medicaid is increasingly important relative to private insurance as a payer for hospital care for children. The rate of potentially preventable hospitalizations for both acute and chronic conditions declined substantially (18%, adjusted for age and gender). Several measures of patient safety improved—the rates of postoperative sepsis, iatrogenic pneumothorax, and selected infections due to medical care declined by 14.2%, 17.8%, and 23.5%, respectively. However, the rate of accidental punctures and lacerations and the rate of decubitus ulcer increased by 25.6% and 34.5%, respectively. The trends in safety indicators varied somewhat by age group, income quartile of zip codes, insurance, region, and type of location without a consistent pattern.

      Conclusions/Implications

      Although teenage pregnancy rates were declining, there was a worsening trend in skin infections. The latter may eventually be impacted by recent publication of new guidelines for treatment by office-based physicians. A gradually increasing role of Medicaid as a payer for hospital care for children will likely put an increasing strain on public resources in advance of the full implementation of the health insurance reforms recently enacted. The decline in potentially avoidable admissions reduces the use of the most expensive resources. For asthma and diabetes, children in the lowest income zip codes had persistently higher rates of admission, but the rate fell by one third during the period. Children in the South and West regions had substantial and significant declines in preventable admissions. Particular indicators of safety were improving, whereas others were worsening. Trends were not the same in all types of hospitals, all regions, and income categories. This is already a rich area for further research on the impact of quality improvement strategies; however, attention is needed to developing more tools to more thoroughly track quality of care for children.

      Keywords

      What’s New
      Rates of hospitalization for children during 2000–2007 continued a gradual decline from the previous decade, largely due to a declining birth rate among teenagers. A remarkable increase in admissions for skin infection occurred. On the other hand, there were significant declines in rates of potentially avoidable admissions. Among the 8 indicators of patient safety for children in hospital care, there were several divergent trends and sub-population differences worthy of more thorough investigation.
      This report is the ninth in a series of descriptive reports summarizing various dimensions of health care for children and youth in the United States. The series has dealt with many issues of access to care, rates of utilization, quality of care, and expenditures for children. The 2002 report focused on trends in the relevant data for these topics for the period 1987 to 2001 and included trends in hospital utilization from 1995 to 2000.
      • Simpson L.
      • Zodet M.W.
      • Chevarley F.M.
      • et al.
      Health care for children and youth in the United States: 2002 report on trends in access, utilization, quality, and expenditures.
      The current report updates those trends with newer data through 2007, with a primary focus on hospital utilization, leading diagnoses, and trends in payer mix and costs.
      In addition to these previously reported trends, data on trends in hospital quality measures for children are included. In recent years, advances in measures of various dimensions of hospital care have resulted in a substantial number of indicators being accepted by review bodies and used for quality improvement and in public reporting,

      NQF-endorsed standards. The National Quality Forum. Available at: http://www.qualityforum.org/Measures_List.aspx. Accessed April 29, 2011.

      for Medicare or Medicaid hospital reimbursement,

      Centers for Medicare & Medicaid Services. Available at: http://www.cms.gov/HospitalAcqCond/06_Hospital-Acquired_Conditions.asp. Accessed April 29, 2011.

      and in projects for national tracking of quality indicators such as the National Healthcare Quality Report.

      National Healthcare Quality Report, 2009. Rockville, Md: Agency for Healthcare Research and Quality. AHRQ Publication No. 10–0003. Available at: http://www.ahrq.gov/qual/qrdr09.htm. Accessed April 29, 2011.

      This report explores the time trends of 2 dimensions of quality measures derived from hospital data. The first is rates of avoidable hospitalization of children for selected acute and chronic conditions—overall and for subpopulations characterized by age, income, insurance coverage, and region of the country. These indicators are measured for local populations, adjusted for age and gender, and are an indication of access to high-quality primary care.
      • Billings J.
      • Anderson G.M.
      • Newman L.S.
      The second dimension of quality is the safety of care within the hospital. A number of risk-adjusted indicators are measured at the hospital level. To facilitate the use of both of these sets of child indicators, downloadable software is available.
      The quality indicators have been mostly applied to adult populations. This study uses measures developed specifically for children and thus demonstrates new tools for analysts to monitor and compare quality of care for children across geographic areas, across institutions and over time. The study is timely in view of the Patient Protection and Affordable Care Act legislation to expand insurance coverage and funding for health care with incentives for quality improvement.
      This project is motivated by the particular study questions listed below. The results and discussion in the report will be organized accordingly.
      • 1.
        Hospital utilization
        • With increasing emphasis nationwide on outpatient procedures and home care, did hospital discharge rates overall increase or decrease, and how did these trends differ for subgroups of children (eg, by age, insurance, region, or community income characteristics)?
        • Did hospitalization rates increase or decrease for any specific diagnostic categories, suggesting areas in which ambulatory care may have improved or areas that should be targeted for future improvement?
        • Was there a significant change in length of stay (LOS) over the period? How rapidly did annual cost per case increase?
        • How rapidly did expenditures per child with any hospital use increase or decrease? Was the trend different for some subgroups of children (eg, by age, insurance, or region)?
      • 2.
        Quality of care—potentially preventable admissions
        • What are the trends in rates of preventable admissions, age and gender adjusted, for selected acute and chronic conditions in children? Are these similar to trends for adults as shown in recent reports?
        • Are these changes different for different age groups, income groups, regions of the country, suggesting either increasing or decreasing disparities in access to high-quality primary care?
      • 3.
        Quality of care—patient safety indicators
        • What trends emerge in patient safety? Are they consistent across indicators? Is there evidence that recent efforts to reduce central venous catheter-related infection rates and other hospital-associated infection rates have had salutary effects?
        • How did these trends vary by subpopulation (eg, age, income, insurance, geographic type, or region)?

      Methods

      Data Sources

      As in previous reports, we use 2 complementary Agency for Healthcare Research and Quality (AHRQ)–sponsored data sources: the Healthcare Cost and Utilization Project (HCUP) and the Medical Expenditure Panel Survey (MEPS). The HCUP data include a census of hospital discharge records collected from more than 40 states and are able to provide the hospital-level and discharge-level perspective on trends in the quality of hospital care, avoidable admissions, and overall hospital use for children (eg, rates of hospital admissions for specific conditions per population or rates of specific events per procedures). MEPS data come from a nationally representative survey of US civilian households and are able to provide the patient and family perspective of hospital utilization and expenditures over time.
      The HCUP is a federal-state-industry partnership of state data organizations, hospital associations, and private data organizations. The state and industry organizations collect billing data directly from hospitals and voluntarily provide it to HCUP for uniform formatting, reports, and possible distribution. The HCUP Nationwide Inpatient Sample contains all discharges from a stratified random sample of about 1000 hospitals. These hospitals are drawn from over 40 states that contain 95% of all discharges from nonfederal, community, nonrehabilitation hospitals in the United States.

      AHRQ. Overview of the nationwide inpatient sample. Available at: http://www.hcup-us.ahrq.gov/nisoverview.jsp/. Accessed April 29, 2011.

      In this report, we include annual estimates of hospital utilization and expenditures for children by year by using data from MEPS. MEPS is an ongoing nationally representative family of surveys of medical care use and expenditures. MEPS provides estimates of the health care utilization, expenditures, sources of payment, quality, and insurance coverage of the US civilian noninstitutionalized population from data collected via multiple contacts over a 2.5-year period, which cover a 2-year reference period.

      AHRQ Agency for Healthcare Research and Quality. Available at http://ww.meps.ahrq.gov. Accessed April 29, 2011.

      The majority of data for this report come from the full-year consolidated files for 2000 (HC-050), 2001 (HC-060), 2002 (HC-070), 2003 (HC-079), 2004 (HC-089), 2005 (HC-097), 2006 (HC-105), and 2007 (HC-113), and are based on the experiences of children aged 17 years and younger during the time 2000 to 2007. The individual year sample sizes for MEPS range from 6930 children in 2000 to 8497 children in 2007. Consistent with previous reports, we also provide the MEPS-HC Point-in-Time Files for 2008 (HC-101) in Appendix A.

      Measures

      Hospital Discharge Rates

      The estimated national discharge totals from HCUP are divided by the census population measures for the relevant years in the following inclusive age groups: infants, aged 1 to 4 years; 5 to 9 years; 10 to 14 years; and 15 to 17 years. Rates for pregnancy and delivery discharges in the 15 to 17 age group are divided by the number of females in the group. Discharges can be grouped by principal diagnosis or condition by using the Clinical Classification Software available at the HCUP user support Web site.

      Healthcare Cost and Utilization Project (HCUP). Clinical classifications software (CCS) for ICD-9-CM. Available at: http://www.hcup-us.ahrq.gov/toolssoftware/ccs/ccs.jsp. Accessed April 29, 2011.

      A distinct child may have more than 1 discharge in a calendar year.

      Expected Payer

      HCUP files include information on primary and secondary expected payer, including private insurance, Medicaid, other types of insurance (including Medicare, worker’s compensation, Title V, TRICARE, and other government programs), and finally, uninsured (self pay or no charge). Since HCUP data are based on hospital discharge abstracts, we cannot separate out the children covered by the State Children’s Health Insurance Program (SCHIP). Some programs are managed by a private insurance company, whereas others are part of a state-run Medicaid program or a state-run non-Medicaid program. Therefore, the expected primary payer for SCHIP discharges in the nationwide database may be categorized as private insurance, Medicaid, or other types of insurance.

      Income

      Income in HCUP is assigned to each record based on the median household income of the patient’s zip code by using data from Claritas, Inc.

      Claritas Inc. (now a division of Nielsen, Inc.) The Claritas demographic update methodology. July 2005. Available from the authors. Accessed April 1, 2009.

      We define income groups based on quartiles of median household income. For example, in 2005 the quartiles were $1 to 36 999 (poorest communities), $37 000 to $45 999 (lower-middle income communities), $46 000 to $60 999 (upper-middle income communities), and ≥$61 000 (wealthiest communities). Note that in 2005, the poverty threshold level for a family of 4 in the contiguous states was $19 350.

      The U.S. Department of Health and Human Services, Assistant Secretary for Planning and Evaluation. The 2005 HHS poverty guidelines. Available at: http://aspe.hhs.gov/poverty/05poverty.shtml. Accessed August 30, 2009.

      Hospital Cost

      Total hospital charges were converted to costs by using cost-to-charge ratios based on hospital accounting reports from the Centers for Medicare & Medicaid Services.

      Healthcare Cost and Utilization Project (HCUP). Cost-to-charge ratio files. Available at: http://www.hcup-us.ahrq.gov/db/state/costtocharge.jsp. Accessed April 29, 2011.

      Song, X, Friedman, B. Calculate cost adjustment factors by APR-DRG and CCS using selected states with detailed charges. HCUP Methods Series Report # 2008-04. Online October 8, 2008. U.S. Agency for Healthcare Research and Quality. Available at: http://www.hcup-us.ahrq.gov/reports/methods/2008_04.pdf. Accessed April 29, 2011.

      There is a well-known systematic bias for using hospital-wide cost-to-charge ratios. Although the bias is within 10% for 3 quarters of discharges, stays with a high proportion of ancillary services tend to have cost overestimated, whereas stays with a high proportion of routine bed unit services (eg, psychiatric services) tend to have cost underestimated. Costs are meant to reflect the resource costs of production, whereas charges represent what the hospital billed for the case. For each hospital, a hospital-wide cost-to-charge ratio is used because detailed charges are not available across all HCUP States. Hospital charges reflect the amount the hospital charged for the entire hospital stay and do not include professional (physician) fees.

      Hospital Quality Measures

      Health care quality measures can broadly include indicators of effectiveness, patient centeredness, timeliness, and patient safety, as conceptualized in the Institute of Medicine report.
      Institute of Medicine Committee on Quality of Health Care in America
      Crossing the Quality Chasm: A New Health System for the 21st Century.
      Measures of effectiveness are available for both HCUP and MEPS data, but patient safety for HCUP only. In the HCUP data, effectiveness of ambulatory care is indicated by 2 “area level” AHRQ pediatric quality indicators (PDIs) that measure potentially avoidable hospitalizations—hospitalizations that are believed to be mostly preventable when timely outpatient care of good quality is obtained (4 principal diagnosis groups are used: asthma, diabetes, gastroenteritis, and urinary tract infections. Some cases are excluded for each category).

      Agency for Healthcare Research and Quality. Pediatric Quality Indicators Overview AHRQ Quality Indicators. Available at: http://qualityindicators.ahrq.gov/modules/pdi_overview.aspx. Accessed December, 2010.

      Version 3.1 of the software downloadable from AHRQ is used for all hospital quality indicators in this study. A complete list of the International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) diagnosis codes, as well as the method for calculating variances to allow for stratification design, are available online.

      Coffey R, Barrett M, Houchens R, et al. Methods Applying AHRQ Quality Indicators to Healthcare Cost and Utilization Project (HCUP) Data for the Eighth (2010) National Healthcare Quality Report (NHQR) and National Healthcare Disparities Report (NHDR). Rockville, Md: U.S. Agency for Healthcare Research and Quality; 2010. HCUP Methods Series Report 2010-06. Available at: http://www.hcup-us.ahrq.gov/reports/methods/2010_06.pdf.

      Two composite measures of preventable admissions in children over age 5 are used, one for acute conditions (gastroenteritis, urinary tract infections) and one for chronic conditions (asthma, diabetes with short-term complications). Diagnoses are defined from ICD-9-CM. Cases can be excluded based on past literature and clinical reviews during the historical development of the PDIs. Area-based rates are derived from principal diagnosis codes in all local hospitals. The rates for avoidable hospitalizations were adjusted for age and gender by using the total US resident population of children for 2000 as the standard population. It should be emphasized that the conditions incorporated in measures of avoidable hospitalization are not necessarily highly frequent. For example, as a principal diagnosis, diabetes is not in the top 10 of frequent admissions for children aged 1 to 17 years. Other categories in the top 10 of principal diagnosis, such as mood disorders and epilepsy, probably contain a large number of patients with chronic illness. The Appendix tables, and the HCUPnet Web site,

      AHRQ. Welcome to HCUPnet. Available at: http://hcupnet.ahrq.gov/. Accessed April 29, 2011.

      contain much additional information about the most frequent principal diagnoses.
      For indicators of adverse patient safety events, 8 measures are selected from the PDI software.
      • Lubell K.
      • Kegler S.
      • Crosby A.
      • Karch D.
      Suicide trends among youths and young adults aged 10–24 years—United States.
      Each measure includes a specification of a class of “at risk” patients in the hospital, as well as inclusion and exclusion criteria for cases to be considered as experiencing adverse safety events. The national indicators are risk adjusted for age, gender, diagnosis related group (DRG), and comorbidities. The risk adjustment is made within subpopulations as well, except that age is not used within age groups. For descriptive trends in this report, the safety indicators are divided into 2 groups of 4 each: medical-surgical complications (accidental puncture or laceration, iatrogenic pneumothorax, decubitus ulcer, and selected infections due to medical care), or postoperative complications (postoperative hemorrhage or hematoma, postoperative sepsis, postoperative respiratory failure, postoperative wound dehiscence). Rare safety-related PDIs, including “foreign body left during procedure” and transfusion reaction were not analyzed. Obstetric indicators were also not analyzed.

      Hospital utilization from MEPS

      Child-specific population estimates of hospital utilization from MEPS include 2 definitions. First, for some analyses utilization includes only inpatient hospital stays. Second, for some analyses utilization includes all hospital visits (outpatient, inpatient, and emergency visits).

      Medical expenditures for children with hospital visits

      Child-specific population estimates of medical expenditures in MEPS include all amounts paid for health care services from any source for all services provided during each calendar year (2002–2007). Expenditure estimates are adjusted to 2007 dollars by using the consumer price index. In this paper we present average total expenditure estimates for children with an inpatient hospital stay, an outpatient hospital stay, or an emergency room visit.

      Insurance

      MEPS’ measure of insurance status includes 3 categories: any private, public only, and uninsured. Children are classified as having private insurance if they were privately insured (including coverage through TRICARE) at any time during the reference period (ie, first half of the year or entire year). Children with no private coverage but who had any coverage through Medicaid/SCHIP, Medicare, or any other type of government program providing coverage for both hospital and medical care are classified as publicly insured. Children not covered by any comprehensive hospital and physician insurance program at all during the reference period (ie, first half of the year or entire year) are classified as uninsured.

      Statistical Analysis

      Estimates are reported if the relative standard error is less than 30% and at least 10 cases (before weighting) are observed. Z tests are used to assess statistically significant differences between 2007 and 2000. Sometimes, a weighted regression is used to test whether a time series fits a significant trend despite noise and sampling variance from year to year. The minimal criterion for significant tests with either HCUP or MEPS data is P < .05.
      All MEPS estimates are weighted to be nationally representative, and standard errors are estimated accounting for the complex sample designs. Point estimates and accompanying standard errors are shown in the tables. Descriptive bivariate analyses are performed to compare differences in hospital utilization and expenditures over time. Z tests are used to assess statistically significant differences. Expenditure estimates from MEPS data are adjusted to 2007 dollars by using the consumer price index. HCUP hospital cost data are not adjusted for general inflation. Analysts may well differ about the index used to deflate costs to constant dollars. We report cost trends together with trends for the gross domestic product deflator.
      Hospital safety event rates were examined for significant changes from 2000 to 2007 and significant changes within age, income, insurance groups, geographic region, and type of location. Not all such data can be presented; consequently, there are detailed examinations of 3 of the original 8 safety indicators. The indicators were picked after seeing (below) the changes that were significant. Then subsequent exhibits are offered to illustrate different time paths for different subpopulations.
      Consistent with prior years in this report series, we also provide a set of standard updated tables on overall insurance coverage, utilization, and expenditures in Appendix A, and hospital utilization data in Appendix B.

      Results

      Hospital Utilization

      Changes in Discharge Numbers and Rates

      Table 1 provides national estimates for the number of hospital discharges in each age group in 2000 and 2007, total and by leading principal diagnoses, together with standard errors and rates per 100 000 children. No rates are provided for infants, because the numerator would be mainly births and the population on a particular day is not a population at risk of birth. The number of discharges for infants rose significantly between 2000 and 2007, by 9.4% to 5.124 million in 2007. The only age group with a significant change in discharge rate per 100 000 was the 15 to 17 age group, with a decline of 7.8%. Pregnancy and delivery are the largest category of hospitalizations in that age group. There was a decline of 15.8% in the rate of pregnancy and delivery discharges for females in that age group.
      Table 1Leading Principal Diagnoses for Children in Short-Stay Hospitals, 2007 Versus 2000
      Sources: Weighted national estimates of discharges from Healthcare Cost and Utilization Project Nationwide Inpatient Sample, 2000 and 2007, Agency for Healthcare Research and Quality; population data by age group derived from published US census tables.
      Age Group and Diagnostic Group [1]20002007
      Discharges, No.SE%Rate per 100 000 in Age GroupDischarges, No.SE%Rate per 100 000 in Age Group
      Aged under 1 yearAll discharges4 684 870166 467100.05 124 835191 416100.0
       Newborn infants4 135 295166 50688.34 575 987180 55489.3
       Acute bronchitis or bronchiolitis107 41267932.392 44261001.8
       Other perinatal conditions51 91742551.175 89273881.5
       Hemolytic jaundice and perinatal jaundice32 40323010.743 37122580.8
      Aged 1–4 yearsAll discharges497 51635 157100.03246.41511 30341 466100.03107.74
       Pneumonia69 875337214.0455.9565 642381212.8398.98
       Asthma62 731483412.6409.3356 600443911.1344.02
       Fluid and electrolyte disorders42 26728138.5275.8040 83223618.0248.18
       Acute bronchitis or bronchiolitis27 32917535.5178.3329 78820075.8181.05
      Aged 5–9 yearsAll discharges320 36223 136100.01564.55346 01228 181100.01743.17
       Asthma38 327289212.0187.1836 998293410.7186.39
       Pneumonia24 36112517.6118.9724 58014117.1123.83
       Appendicitis and other appendiceal conditions19 57811816.195.6121 97918336.4110.73
       Fluid and electrolyte disorders13 2148304.164.5314 5548584.273.32
      Aged 10–14 yearsAll discharges355 47321 356100.01723.92381 43228 711100.01877.65
       Appendicitis and other appendiceal conditions32 46239479.1157.4334 92621419.2171.93
       Mood disorders (depression or bipolar)33 31713189.4161.5834 86451589.1171.62
       Asthma24 34818016.8118.0819 06016855.093.83
       Fracture of lower limb10 5825203.051.3211 0427612.954.36
      Aged 15–17 yearsAll discharges504 15816 953100.04179.51500 84520 525100.03855.41†
       Mother’s pregnancy and delivery (per 100 000 females)187 312824737.23199.13170 795724134.12693.39
       Mood disorders47 18849109.4391.1944 16360308.8339.96
       Appendicitis and other appendiceal conditions19 6506953.9162.9022 6308854.5174.20
       Fracture of lower limb90164171.874.7483254861.764.08
       Poisonings by medications and drugs (not psychotropic agents)10 0504502.083.3280805581.662.20
      †Significant change of the rate from 2000; P < .05.
      Sources: Weighted national estimates of discharges from Healthcare Cost and Utilization Project Nationwide Inpatient Sample, 2000 and 2007, Agency for Healthcare Research and Quality; population data by age group derived from published US census tables.

      Trends of Leading Diagnostic Categories

      Figure 1, Figure 2, Figure 3, Figure 4 show the time charts for 4 leading conditions by age group. Among younger children, the trend lines in Figure 1, Figure 2 suggest decreasing hospitalization rates between 2005 and 2007, with increases for some conditions in previous years (pneumonia, appendicitis, fluid and electrolyte disorders). Alternatively, the 2005 observations could simply be unusual results that soon revert toward the mean. Asthma hospitalization rates also decreased (across the entire time period) among 10- to 14-year-old children, but not among the 15 to 17 age group. Figure 4 for the 15 to 17 age group does not include rates for pregnancy-related and maternity conditions (shown in Table 1), which have rates much higher than the next 4 conditions displayed in Figure 4. There were slightly declining trends for fracture of the lower limb, and poisoning by medication other than psychoactive drugs. In the poisoning category shown, the hospital data indicate more than 70% are self-inflicted injuries, whereas other sources suggest an even higher proportion.
      • Lubell K.
      • Kegler S.
      • Crosby A.
      • Karch D.
      Suicide trends among youths and young adults aged 10–24 years—United States.
      Figure thumbnail gr1
      Figure 1Hospital discharge rates for leading diagnostic groups, children aged 1 to 4 years. Sources: Weighted national estimates from Healthcare Cost and Utilization Project (HCUP) Nationwide Inpatient Sample (NIS), 2000 to 2007, Agency for Healthcare Research and Quality; population data by age group derived from published US census tables.
      Figure thumbnail gr2
      Figure 2Hospital discharge rates for leading diagnostic groups, children aged 5 to 9 years. Sources: Weighted national estimates from HCUP NIS, 2000 to 2007, Agency for Healthcare Research and Quality; population data by age group derived from published US census tables.
      Figure thumbnail gr3
      Figure 3Hospital discharge rates for leading diagnostic groups, children aged 10 to 14 years. Sources: Weighted national estimates from HCUP NIS, 2000 to 2007, Agency for Healthcare Research and Quality; population data by age group derived from published US census tables.
      Figure thumbnail gr4
      Figure 4Hospital discharge rates for leading diagnostic groups, children aged 15 to 17 years. Sources: Weighted national estimates from HCUP NIS, 2000 to 2007, Agency for Healthcare Research and Quality; population data by age group derived from published US census tables.
      During the study, the category of skin infections emerged as a rapidly rising condition for all age groups, although it did not appear among the top 4 conditions in any years. The time trend is shown in Figure 5, separately for infants and for children aged 1 to 17 years. By 2007, the total number of hospitalizations had leveled off at about 65 000 to 70 000, about 90 hospitalizations per 100 000.
      Figure thumbnail gr5
      Figure 5Principal diagnosis of skin and subcutaneous infections, rate per 100 000. Sources: Weighted national estimates from HCUP NIS, 2000 to 2007, Agency for Healthcare Research and Quality; population data by age group derived from published US census tables.

      Changes in LOS, Costs, and Payer

      Mean LOS and cost per stay for children increased during the period 2000 to 2007. Table 2 shows that for infants, the mean LOS increased by 12.5%, from 3.26 to 3.67 days, but for children aged 1 to 17 years there was no change. Cost per stay for children increased by 46% for infants and 44% for older children. Overall, the average annual average increase was 4.5% per year, from $3144 to $4263 (data not shown), which compares with 2.6% annual growth in the gross domestic product deflator. The annual rate of increase in cost for all ages was 5.3%, indicating that stays for children are becoming relatively less costly than stays for adults. Figure 6 shows trends in the percentage of children’s discharges covered by Medicaid, private insurance, or uninsured. The Nationwide Inpatient Sample covers nonfederal, short-term general, and other special hospitals. The proportion of discharges with Medicaid as the expected payer increased significantly (from 36% to 44%), whereas the percentage with private insurance decreased (from 58% to 50%). The percentage of discharges with no insurance coverage remained stable at 5%.
      Table 2Selected Breakdowns of Pediatric Inpatient Discharges, 2000 and 2007
      Sources: Weighted national estimates of discharges from Healthcare Cost and Utilization Project Nationwide Inpatient Sample, 2000 and 2007, Agency for Healthcare Research and Quality; population data by age group derived from published US census tables.
      Utilization and Costs20002007Growth 2000 to 2007 (% Change)
      No.%95% CI
      CI = confidence interval.
      No.%95% CI
      Aged <1 Year
      Total discharges4 685 0001005 125 0001009.39
      Expected payer
      Expected payers of Medicare and no charge are not reported. Sum of percents will be less than 100.
       Medicaid1 649 33235.432.9–37.92 227 16743.640.9–46.235.0
      Significant difference between years; P < .05.
       Private Insurance2 654 81656.953.8–59.92 448 25847.944.9–50.9−7.8
      Significant difference between years; P < .05.
       Uninsured227 6944.93.9–6.1260 3045.14.2–6.114.3
       Other105 1642.31.8–2.9138 8962.72.2–3.332.1
      Household income (zip code)
       Below national median1 667 92236.033.0–39.22 691 82153.750.1–57.361.4
      Significant difference between years; P < .05.
       Above national median2 963 78064.060.8–67.02 319 06046.342.7–49.9−21.8
      Geographic Area
       Northeast826 90017.715.5–20.1883 13317.214.9–19.96.8
       Midwest1 018 80321.819.4–24.41 101 73021.519.0–24.28.1
       South1 774 79837.934.3–41.62 004 46339.135.4–43.012.9
       West1 064 36922.720.2–25.51 135 50922.219.5–25.16.7
      Average length of stay, d3.263.14–3.383.673.51–3.8312.5
      Significant difference between years; P < .05.
      Average total costs, $25582266–284937373308–416646.1
      Significant difference between years; P < .05.
      Aged 1–17 Years
      Total discharges1 678 0001001 658 000100−1.19
      Expected payer
      Expected payers of Medicare and no charge are not reported. Sum of percents will be less than 100.
       Medicaid640 01738.436.3–40.4746 21445.142.7–47.616.6
      Significant difference between years; P < .05.
       Private insurance884 25153.050.5–55.4751 26245.542.9–48.0−15.0
      Significant difference between years; P < .05.
       Uninsured86 1065.24.1–6.474 7894.53.8–5.4−13.1
       Other48 4092.92.4–3.569 1344.23.3–5.242.8
      Household income (zip code)
       Below national median723 00644.040.2–47.8946 02858.854.6–62.930.9
      Significant difference between years; P < .05.
       Above national median921 74256.052.3–59.8663 66741.237.2–45.4−28.0
      Geographic area
       Northeast294 99617.614.1–21.7291 71517.614.4–21.3−1.1
       Midwest413 24824.620.5–29.3386 03023.317.9–29.8−6.6
       South587 35435.030.4–40.0715 27443.136.5–50.121.8
       West382 18122.817.8–28.7264 84216.012.4–20.4−30.7
      Average length of stay, d3.513.35–3.673.713.50–3.935.8
      Average total costs, $46974209–518667656002–752844.0
      Significant difference between years; P < .05.
      Significant difference between years; P < .05.
      Sources: Weighted national estimates of discharges from Healthcare Cost and Utilization Project Nationwide Inpatient Sample, 2000 and 2007, Agency for Healthcare Research and Quality; population data by age group derived from published US census tables.
      CI = confidence interval.
      § Expected payers of Medicare and no charge are not reported. Sum of percents will be less than 100.
      Figure thumbnail gr6
      Figure 6Payer proportion for children in short-stay hospitals, aged 0 to 17 years. Sources: Weighted national estimates from HCUP NIS, 2000 to 2007, Agency for Healthcare Research and Quality; population data by age group derived from published US census tables.
      Table 2 provides some additional information about insurance coverage, income averages in areas where children are living, and region of the country. In both age groups, the number of discharges billed to Medicaid increased between 2000 and 2007, with a 35% increase for infants and a 17% increase for 1- to 17-year-olds. This was countered by a decrease in discharges billed to private insurance in both age groups (8% for infants and 15% for 1–17-year-olds). Two other developments were noteworthy. In both age groups, the percentage of discharges for children from areas with below-median household income increased dramatically between 2000 and 2007 (from 36% to 54% for infants and from 44% to 59% for 1–17-year-olds).

      Child-Specific Hospital Use and Expenditures

      Hospital Use

      Using MEPS data, Figure 7 shows the trend in the percentage of all children with at least 1 inpatient hospital stay, by age group. Similar percentages of children reported inpatient stays across time (2.93% of all children had an inpatient stay in 2000 compared with 2.81% in 2007). When we examine the trend by age group, we find a similar pattern, although the data suggest a decrease among 15- to 17-year-old children. The difference between age groups was not statistically significant.
      Figure thumbnail gr7
      Figure 7Percentage of children with an inpatient (IP) hospital stay over time. Data from the Medical Expenditure Panel Survey Full-year Consolidated files 2000 (HC-050), 2001 (HC-060), 2002 (HC-070), 2003 (HC-079), 2004 (HC-089), 2005 (HC-097), 2006 (HC-105), and 2007 (HC-113). T test comparing estimates in 2000 to 2007 showed no statistically significant differences at P < .05.
      Figure 8 shows the percentage of children with inpatient stays by insurance status. Consistent with the overall finding noted above of no trend, rates of hospitalization by expected payer remained stable throughout the period. Additionally, the pattern of significantly higher rates of inpatient stays among publicly insured children versus privately insured children (3.78% vs 2.23% in 2007) persisted throughout the period. The percentage of uninsured children having at least 1 hospital stay was 2.61% in 2000 and 3.32% in 2007, but this difference was not statistically significant.
      Figure thumbnail gr8
      Figure 8Percentage of children with IP visit by insurance status over time. Data from the Medical Expenditure Panel Survey Full-year Consolidated files 2000 (HC-050), 2001 (HC-060), 2002 (HC-070), 2003 (HC-079), 2004 (HC-089), 2005 (HC-097), 2006 (HC-105), and 2007 (HC-113). T test comparing estimates in 2000 to 2007 showed no statistically significant differences at P < .05. Estimates of use for uninsured children are not included for 2001, 2004 to 2006 due to sample sizes <100. Estimates for any private compared with public insurance in 2000 and 2007 were statistically significant at P < .05.

      Annual Expenditures for Children Using Hospital Services

      Figure 9 shows the average annual total medical expenditures for children of different age groups who had at least 1 visit to the hospital (inpatient stay, ED, or outpatient visit). The calculated growth rate for total expenses was 4.8% per year corrected for general inflation. Expenditures per child with public versus private insurance grew at similar rates over the period (data not shown but available from the authors). Expenditures tend to have a relatively high variance in any year.
      Figure thumbnail gr9
      Figure 9Average annual total expenditures for children with an emergency, inpatient or outpatient hospital visit by age group. Data for this chart come from the Medical Expenditure Panel Survey Full-year Consolidated files 2000 (HC-050), 2001 (HC-060), 2002 (HC-070), 2003 (HC-079), 2004 (HC-089), 2005 (HC-097), 2006 (HC-105), and 2007 (HC-113). T test comparing estimates in 2000 to 2007 showed no statistically significant differences at P < .05. The difference between 2000 and 2007 for children aged 5 to 9 years was statistically significant at P < .10.

      Potentially Preventable Admissions

      Figure 10 shows the trend for the composite rate of the 2 acute conditions—gastroenteritis and urinary tract infection—adjusted for age and gender. The observed rate is significantly lower in 2007 than in 2000, by 18% (124.2 vs 151.7 per 100 000, respectively). However, in 2003 and 2005 the rate is essentially the same as at the beginning. Given the fluctuation in the data, and because the variance is different each year, it is instructive to fit a weighted regression to the data. The regression did not find a significant trend in the data.
      Figure thumbnail gr10
      Figure 10Potentially avoidable conditions, acute (gastroenteritis, urinary tract infections) per 100 000 population aged 5 to 17 years, adjusted for age and gender. Source: HCUP NIS and Quality Indicators, version 3.1, 2000 to 2007, Agency for Healthcare Research and Quality.
      The pattern is different for exacerbations of chronic conditions. Figure 11 shows a declining trend for the composite of potentially preventable admissions for asthma and short-term complications of diabetes, by 18.5% from an aggregate rate of 82.4 per 100 000 in 2000 to 67.2 per 100 000 in 2007. This trend is smooth enough and with small enough yearly standard errors for the weighted trend regression to be significantly negative (P < .05). The detailed data for this weighted regression are available from the authors. Key subpopulations were compared to determine whether their trends were parallel or divergent. (Unfortunately, detailed data for the PDIs in subpopulations were not available for this study during 2001–2003.) We found that the declines were similar by age group (data not shown). Figure 12, however, indicates that children in zip codes with the lowest quartile of family income had persistently higher rates of such admissions, but their rate declined by a third during the period. Both of the lowest 2 income quartiles had significant declines after 7 years, whereas the higher income categories did not have significant changes. Expressed in another way, the disparity in potentially preventable hospitalization rates between the lowest and highest income quartiles narrowed substantially, from 2.55-fold in 2000 to 1.73-fold in 2007. In addition, only children in the South and West had significant declines in these indicators, by 19.4% and 32%, respectively (data not shown). Interestingly, in 2000 the rate per 100 000 was highest in the South (100.5) and lowest in the West (63.7).
      Figure thumbnail gr11
      Figure 11Potentially avoidable conditions, chronic (asthma, diabetes) per 100 000 population aged 5 to 17 years, adjusted for age and gender. ∗Significant downward trend of weighted regression; P < .05. Source: HCUP Nationwide Inpatient Sample and Quality Indicators, version 3.1, 2000 to 2007, Agency for Health care Research and Quality.
      Figure thumbnail gr12
      Figure 12Potentially Avoidable Conditions, Chronic, per 100 000 Population, Age 5 to 17, Adjusted for Age and Gender. Significant decline, 2000 to 2007; P < .05. Source: HCUP NIS and Quality Indicators, version 3.1, 2000 to 2007, Agency for Healthcare Research and Quality.

      Safety Indicators

      Figure 13 shows the trends of the 4 indicators in the category of postoperative complications. The chart is calibrated so each index has a value of 100 in 2000. Below the charts, the actual rate of each indicator is shown per 100 patients in the at-risk group for that rate. Two of the indicators (postoperative hemorrhage or hematoma, and postoperative sepsis), had lower rates in 2007 than in 2000, 1 indicator (postoperative wound dehiscence) had a nonsignificantly higher rate, and 1 was essentially unchanged.
      Figure thumbnail gr13
      Figure 13Trend of patient safety for hospitalized children: postoperative complications calibrated to each begin at 100 in year 2000. Significant decline, 2000 to 2007; P < .01. Source: HCUP NIS and Quality Indicators, version 3.1, 2000 to 2007, Agency for Healthcare Research and Quality.
      Declines in postoperative complications were found for both high- and low-income categories, both insurance categories, for children aged under 4 years and 10 to 14 years, in the Midwest and South, and in large metropolitan areas, but not in smaller metropolitan or rural areas. Differences and trends for ownership categories and teaching status were examined. Teaching hospitals tend to have lower rates of safety events. When an indicator improves or worsens, it tends to move in the same direction for both teaching in nonteaching hospitals. The decline in both highlighted indicators of postoperative complications was significant for investor-owned hospitals. The decline in postoperative sepsis was also significant in not-for-profit (NFP) hospitals (all data not shown in this paragraph are available from the authors).
      Figure 14 shows a divergence in the trends for medical and surgical complications in that 2 rates rose significantly and 2 rates fell significantly. Accidental puncture and laceration increased 25.6% from 2000 to 2007 in its at-risk group of patients. Decubitus ulcer increased 34.5% in its at-risk group (excluding stays of under 5 days, neonates, births, pregnancy-related admissions). Rates of iatrogenic pneumothorax and selected infections due to medical care, both of which have relatively broad at-risk populations, decreased by 17.8% and 23.5%, respectively. (The data for at-risk populations can be obtained via the online query system at http://hcupnet.ahrq.gov. For example, in the measure for iatrogenic pneumothorax, all cases in a number of surgical DRGs involving operations in the chest cavity are excluded from the “at-risk” group. In general, a high proportion of emergency surgeries are excluded from the measures of adverse safety events). Focusing on the divergence, we picked 1 example from the rising indicators, accidental punctures and lacerations, and 1 from the falling indicators, iatrogenic pneumothorax, to examine changes for subpopulations.
      Figure thumbnail gr14
      Figure 14Trend of patient safety for hospitalized children: medical and surgical complications calibrated to each begin at 100 in year 2000. Significant increase, 2000 to 2007; P < .01. ∗∗Significant decrease, 2000 to 2007; P < .01. Source: HCUP NIS and Quality Indicators, version 3.1, 2000 to 2007, Agency for Healthcare Research and Quality.
      Figure 15 shows that the percentage increase in accidental punctures and lacerations was largest for children aged between 5 and 14 years. Otherwise, the trends by income, region, and other variables were similar. By ownership, the increase was significant for NFP hospitals and for teaching hospitals. Figure 16 shows that the northeast region had a much higher rate of iatrogenic pneumothorax than other regions in 2000 and the largest percentage decline, by over 40% by 2007. There were only minor differences in trends by age group, income of the area, and type of location. A significant decline was seen in all ownership categories and for both teaching and nonteaching hospitals (data not shown).
      Figure thumbnail gr15
      Figure 15Accidental puncture or laceration during procedure per 1000 discharges (excluding obstetric admissions, normal newborns, and neonates with a birth weight <500 g). Adjusted by gender within age group. Also excludes admissions specifically for problems such as cases from earlier admissions or from other hospitals. Significant increases for age groups 10 to 14 years and 15 to 17 years; P < .01. Source: HCUP NIS and Quality Indicators, version 3.1, 2000 to 2007, Agency for Healthcare Research and Quality.
      Figure thumbnail gr16
      Figure 16Iatrogenic pneumothorax per 1000 discharges (excludes neonates with a birth weight < 2500 g, and patients with chest trauma or thoracic procedures) Adjusted by age and gender. Also excludes admissions specifically for iatrogenic pneumothorax, such as cases from earlier admissions or from other hospitals. Includes barotrauma (including acute respiratory distress syndrome) and central line placement. Significant decline, 2000 to 2007; P < .01. Source: HCUP NIS and Quality Indicators, version 3.1, 2000 to 2007, Agency for Healthcare Research and Quality.

      Discussion

      This report presents a broad review of trends in hospital care for children during the 8-year period from 2000 to 2007. During this time period, significant increases in enrollment in Medicaid and CHIP occurred, mirrored by decreases in rates of uninsured children.
      • Kenney G.
      • Yee J.
      SCHIP at a crossroads: experiences to date and challenges ahead.

      America’s Children in Brief: National Indicators of Well-Being, 2010. Available at: http://www.childstats.gov/americaschildren/care.asp. Accessed March, 2011.

      In addition, a number of state Medicaid programs moved most of their nondisabled child populations into managed care arrangements.

      Kohn LT, Corrigan JM, and Donaldson MS (editors). To Err is Human: Building a Safer Health Care System, Institute of Medicine. Washington DC: National Academy Press, 2000.

      This was also the time period after the release of the landmark Institute of Medicine report, To Err is Human,

      Centers for Medicare & Medicaid Services. Trends for 2000–2009 in Medicaid managed care are available from the U.S. Centers for Medicare and Medicaid services. Available at: https://www.cms.gov/MedicaidDataSourcesGenInfo/05_MdManCrPenRateandExpEnrll.asp. Accessed December, 2010.

      when significant investments in patient safety were made by the federal government, hospitals, and the Joint Commission. It is thus instructive to look at what, if any, trends emerge in the use and costs of hospital care for children, as this is an area that has been the focus of substantial cost-containment efforts. At the same time, the quality indicators available for use with HCUP provide a window on the quality of care being provided in communities as well as the safety of care in hospitals themselves.

      Overall Hospital Use

      Similar to the 2002 report on trends in the late 1990s, this time period revealed a modest decline in the overall rate of hospitalization. The estimated number of hospitalizations per 100 000 children was 8613 in 2007, 8800 hospitalizations in 2000, and 9340 hospitalizations in 1995. The declining trend is also evident in the data from MEPS for children with at least 1 inpatient stay, declining modestly from 2.93% to 2.81%.
      Also consistent with the earlier report, a significant decline in hospitalization rates was again found among the 15- to 17-year-old group—from 4179 per 100 000 in 2000 to 3855 per 100 000 in 2007 (Figure 4). This decline was also reflected in the hospitalization rates for children estimated from MEPS where children aged 15 to 17 years had a decline of 35% (3.6/100 to 2.3/100) in the proportion with any inpatient care, a suggestive result but not statistically significant (Figure 7).
      Few significant changes were noted in the leading reasons for hospitalization during this time period, by age group (Table 1). For those 1- to 9-years-old, pneumonia and asthma remained leading causes as they were throughout the earlier period, from 1995 to 2000. Mood disorders represented 9.1% of all discharges among the 10- to 14-years age group, a similar proportion to that in 2000, but an increase from earlier periods.
      • Simpson L.
      • Zodet M.W.
      • Chevarley F.M.
      • et al.
      Health care for children and youth in the United States: 2002 report on trends in access, utilization, quality, and expenditures.
      However, the decline in poisonings, which are largely self-inflicted (according to the ICD-9-CM E codes), could indicate a decline in some types of mood disorders. The decline in drug poisonings is parallel to the decline in suicide rates among teenagers aged 15 to 19 years reported for 2000 to 2003. Suicide rates in teenagers have declined since the mid-1990s, except for a preliminary report of an upward reversal in 2004.
      • Lubell K.
      • Kegler S.
      • Crosby A.
      • Karch D.
      Suicide trends among youths and young adults aged 10–24 years—United States.
      Two developments were particularly noteworthy. First, most of the decrease in hospitalization for 15- to 17-year-olds was accounted for by a substantial reduction in pregnancy and maternity related discharges. This finding is consistent with other recent reports on teenage pregnancies.

      Guttmacher Institute. Facts on American teens’ sexual and reproductive health. January 2010. Available at: http://www.guttmacher.org/pubs/FB-ATSRH.html. Accessed April 29, 2011.

      Abma JC, Martinez GM, Copen CE. Teenagers in the United States: sexual activity, contraceptive use, and childbearing, National Survey of Family Growth, 2006–08. U.S. Center for Disease Control, National Center for Health Statistics. Vital Health Stat 23(30). 2010. Available at http://www.cdc.gov/nchs/data/series/sr_23/sr23_030.pdf. Accessed April 29, 2011. June, 2010.

      In 2005, the US teenage pregnancy rate reached its lowest point in more than 30 years, down 41% since its peak in 1990, then increased slightly in 2006 and 2007.
      A second important development was an increase in the rate of skin infections, more than doubling for children of all ages combined (Figure 5). This time period coincides with the emergence of community-acquired, methicillin-resistant staphylococcus aureus (MRSA) as a major cause of abscess-forming infections. One study of adults visiting emergency departments in 11 cities in 2004 found a prevalence of MRSA in 76% of skin and soft tissue infections.
      • Moran G.
      • Krishmadasan A.
      • Gorwitz R.
      • et al.
      Methicillin-resistant S. aureus infections among patients in the emergency department.
      Another study at 9 sites in 2005 found a standardized incidence rate of 31.8 per 100 000, with much lower rates for children. More than half of the cases were community onset infections, with only 26.6% hospital onset.
      • Klevens R.
      • Morrison M.
      • Nadle J.
      • et al.
      Invasive methicillin-resistant Staphylococcus aureus infections in the United States.
      The Infectious Diseases Society of America recently released guidelines for improved treatment of MRSA infections in adults and children.

      Liu C, Bayer A, Cosgrove S, et al. Clinical practice guidelines by the Infectious Diseases Society of America for the treatment of methicillin-resistant staphylococcus aureus infections in adults and children. Clinical Infections Diseases. Available at: http://cid.oxfordjournals.org/content/early/2011/01/04/cid.ciq146.full. Accessed March, 2011.

      For children with any hospital visit (inpatient, emergency, or outpatient), the average annual increase in total expenditures (hospital and other, corrected for inflation) was 4.8% (Figure 9). This annual increase combines the rate of increase in units of service (hospital, ambulatory visits, tests, drugs) and the rate of increase in expense per unit. This was considerably higher than just the annual rate of increase in hospital cost per discharge alone (from Table 2, 4.5% before correction for increase in the gross domestic product deflator of about 2.6% per year).
      Consistent with the 2002 report, the proportion of discharges with Medicaid as an expected payer increased, whereas privately insured discharges decreased (Figure 6). In addition, the proportion of children with an inpatient stay remained substantially higher for publicly insured children compared with those privately insured (3.78% vs 2.23%, respectively, Figure 8). Given that hospital care is often the most expensive type of health services utilization for children, and that the proportion of hospitalizations for Medicaid increased over time, there was an important shift of high-cost care to public payers. Medicaid has long been a significant payer of birth hospitalizations, paying for 41.3% of all births in 2003. However this proportion varies by state, from a high of 60% in Mississippi to a low of 27% in Hawaii.

      Kaiser Family Foundation. Medicaid managed care enrollees as a percent of state Medicaid enrollees, as of June 30, 2009. Available at: http://www.statehealthfacts.org/comparemaptable.jsp?ind=217&cat=4. Accessed November, 2010.

      Quality of Care

      Preventable Admissions

      There was a favorable 18.5% reduction in potentially preventable admissions for exacerbations of chronic conditions over the 7 years (Figure 11). For adults, the decline between 2000 and 2006 was only 9%.
      Chapter 6: Efficiency
      National Healthcare Quality Report, 2009.
      Comparisons of subpopulations of children indicated that the declines in preventable hospitalizations for both acute and chronic conditions were significant only for patients in zip codes with average income in the lowest 2 quartiles. Regionally, the West had significant declines for both acute and chronic avoidable admissions (39% and 32%, respectively). The South had a significant decline for chronic conditions (19.4%). The reasons for these different subpopulation trends are unclear. One might presume that efforts to improve the quality of care for asthmatic children through various initiatives using clinical guidelines, medical homes, and other strategies have played a part.

      Homer CJ. Health disparities and the medical home: could it be that simple? Acad Pediatr. July 2009;9:203–205

      Another factor might be the role of managed care, particularly in the Medicaid program. However, the bases for regional differences are difficult to determine. There was a decline in total enrollment in HMO plans over the period. Declines for the South and West were very close to other regions (these data are taken from runs of MEPS household component public use files). However, the proportion of Medicaid enrollees who were in Medicaid Managed Care may have changed differently in different regions.
      There was an 18% decline in potentially preventable admissions for acute conditions (Figure 10). The difference between the end point years was significant and may reflect guidelines of practice for the included conditions.

      American Academy of Pediatrics, Clinical Practice Guidelines. Available at: http://practice.aap.org/content.aspx?aid=1430. Accessed March, 2011.

      The time path was not smooth enough to fit a significant trend.

      Safety Indicators

      There were several noteworthy improvements in patient safety (eg, postoperative sepsis by 14.2%, iatrogenic pneumothorax by 17.8%, and selected infections due to medical care by 23.5%, Figure 13, Figure 14). The findings on infection correspond to a report from another author about success stories in strategies to reduce health care–associated infections in children.
      • Sandora T.
      Prevention of healthcare associated infections in children: new strategies and success stories.
      Other patient safety indicators worsened: accidental puncture and laceration increased by 25.6% and decubitus ulcer increased by 34.5%. The subpopulation data provide some additional and intriguing, though by no means explanatory, information. For example, declines in postoperative sepsis were significant for large metropolitan areas but not for smaller metropolitan or rural areas, whereas the increased rate of accidental puncture or laceration was greater for children in the 5- to 14-year age group than for others. The decline in iatrogenic pneumothorax was particularly large for the Northeast region, which had a much higher rate in 2000 than for the other regions, so the substantial decline (over 40%) may represent a regression to the mean (Figure 15). It should be noted that most of the safety indicators are relatively infrequent, below 3% as shown in Figure 13, Figure 14. Two are quite substantial: 19% (postoperative respiratory failure) and 22% (postoperative sepsis). But the at-risk populations are relatively smaller than for the others.

      Limitations

      This report is subject to several limitations. First, the sampling of hospitals in the Nationwide Inpatient Sample changes from year to year. The percentage of children discharged from hospitals primarily serving children (“children’s hospitals”) was 6.8% in 2000 and 4.1% in 2007. Between these years, the rate fluctuated up and down, with a high of 12.5%, and in 4 years it was over 9%. This variation might have affected variation in rates of discharge for some of the sickest children and resulted in choppiness of some of the utilization trends or safety indicator trends. A large proportion of children have visits to emergency departments (visits per child were over 40% for noninfants in 2007). In some populations, the emergency department visits may have substituted for inpatient stays, and this substitution may have increased or decreased over time. Likewise, treatment in observation units may reduce admissions in some hospitals and geographic regions. There is a need for standardized data collection of observation stays.
      The set of avoidable admissions cannot be exhaustive. Some other patients with acute or chronic conditions could fall into the category of avoidable admissions, but they have not yet been convincingly delineated with administrative data. Hospital use and LOS can be affected by postdischarge and home care. Changes in use of these services may have occurred in association with the increase in proportion of children covered by Medicaid versus private insurance.
      The safety indicators depend on specification of the at-risk populations for each condition, which in turn depends on the treatment mix for patients in the sample of hospitals. As a check on treatment variation, we examined the proportion of children undergoing major diagnostic and/or therapeutic interventions—these are defined by DRGs involving use of the operating room and anesthesia. Downloadable software for categorizing procedure classes is available from the HCUP user support Web site.

      Healthcare Cost and Utilization Project (HCUP). Available at: http://www.hcup-us.ahrq.gov/toolssoftware/procedure/procedure.jsp. Accessed April 29, 2011.

      We found that the proportion of hospitalized children with major procedures was steady at a rate of about 11% for ages 1 to 4 years, falling from 23% to 21% for ages 5 to 9 years, falling from 28% to 27% for ages 10 to 14 years, and rising from 23% to 24% for ages 15 to 17 years. Thus, it is unlikely that there was substantial change in the rate of major procedures that may be affecting trends in the safety indicators.
      The measures of potentially avoidable admissions and safety indicators represent only a limited subset of the dimensions of the quality of care that children receive.
      • Dougherty D.
      • Meikle S.F.
      • Owens P.
      • et al.
      Children’s health care in the first National Healthcare Quality Report and National Healthcare Disparities Report.
      • Simpson L.
      • Dougherty D.
      • Krause D.
      • et al.
      Measuring children’s health care quality.
      • Miller M.
      • Gergen P.
      • Honour M.
      • Zhan C.
      Burden of illness for children and where we stand in measuring the quality of this health care.
      For example, our measures do not directly address the technical clinical quality for acute or chronic care since they are based on hospital administrative data.
      The subpopulation denominators for rates of discharge and potentially avoidable admissions reflect census-based estimates of the child population of specific areas of the country, yet we know that certain diseases have varying prevalence in populations of children. This limitation affects the rates of potentially preventable admissions being apparently higher in some income groups, regions, and types of location (large metropolitan, rural, etc). Prevalence of chronic problems such as asthma or the incidence of acute problems such as urinary tract infections can be higher in some populations and can change over time, giving rise to cross-section and time series comparisons that may be due to disease that are confounded with improvement in availability or effectiveness of services.
      The HCUP indicators of patient safety do not take into account whether some conditions were present on admission (POA) because the data on POA are not uniformly available from most hospitals during this period. Although the importance of POA on the risk adjustment for rates of patient safety events has been well documented in the adult population,

      Coffey R, Milenkovic M, Andrews RM. The Case for the Present-on-Admission (POA) Indicator. Rockville, Md: U.S. Agency for Healthcare Research and Quality; 2006. HCUP Methods Series Report 2006-01. Available at: http://www.hcup-us.ahrq.gov/reports/2006_1.pdf Accessed April 29, 2011.

      less is known about the importance of POA on rates of patient safety in the pediatric population.
      The 8 conditions selected for this report from the wider set in the AHRQ list of PDIs were chosen on the basis of past reviews by clinical review panels and groups such as the National Quality Forum. As this report demonstrates, the number of indicators relevant to children remains limited, and in response, the Child Health Insurance Program Reauthorization Act of 2009 established a major new program to expand quality measures for children and address the numerous gaps. The Act also established a panel at the Institute of Medicine to identify priorities for measure improvement.

      Suggestions for Research

      There are a number of topics worthy of more detailed investigation. For example, there was a remarkable change in the proportion of hospitalized children from lower income zip codes. This may represent a relative decline in access or quality of ambulatory care in those areas, regardless of insurance coverage, or an increasing severity and complexity of illness. Alternatively, one might find differential patterns of movement between neighborhoods for households with children compared with older households. Another area for study is the possible substitution of emergency department visits for inpatient care. This topic might be examined using the new HCUP Nationwide Emergency Department Sample, 2007.

      Available at: http://www.hcup-us.ahrq.gov/databases.jsp. Accessed April 29, 2011.

      Finally, the divergence of some safety indicator trends from others stands out as an area for more systematic multivariate investigation, by using information about the facility characteristics, including volume of experience with particular types of procedures, the type of insurance plan affecting incentives to the hospitals and physicians (eg, capitated or fee for service), competition in the market for hospital services, and physician characteristics. Ownership categories might be picking up some important volume differences or unmeasured case mix differences, since private NFP hospitals generally improved significantly when state and local government hospitals did not.

      Conclusions

      This report offers a number of trends that serve as baseline information for children prior to the enactment of the Patient Protection and Affordable Care Act of 2010. A key feature of the reforms is an expansion of health insurance coverage. One way expanded coverage might reduce health care cost is by reducing avoidable hospital admissions and readmissions. In our report, we find that the rate of avoidable admissions for selected acute and chronic conditions for children declined significantly from 2000 to 2007, but this decline was concentrated in particular subpopulations for reasons that are not yet clear. In 2007, the composite rate of potentially avoidable admissions for 2 chronic conditions (asthma and diabetes) was still substantial, at 7 per 10 000 children aged over 5 years.
      Although some trends in hospital utilization showed improvements, childrens’ hospitalizations for skin infections increased dramatically between 2000 and 2007. Although skin infections were not a leading condition for hospitalization in any particular age group, the overall rate doubled to 9 per 10 000 after 7 years.
      There were several noteworthy improvements in patient safety in hospitals, specifically, declines in the rates of postoperative sepsis, iatrogenic pneumothorax, and selected infections due to medical care. On the other hand, there were rising rates of decubitus ulcer and accidental puncture or laceration. Aggregated indicators could yield reduced variance over time but could obscure the specific impacts of interventions to improve quality of care. Investigation to clarify these different trends will not be simple and will likely require more attention to the detailed patient, facility, and physician characteristics and practices, as well as contextual influences and published guidelines. Finally, despite increased attention to quality and safety measures, the number of measures relevant to children remains very limited.

      Acknowledgments

      We thank the statewide data organizations that participate in the Nationwide Inpatient Sample of the Healthcare Cost and Utilization Project: Arizona Department of Health Services; Arkansas Department of Health; California Office of Statewide Health Planning and Development; Colorado Hospital Association; Connecticut Hospital Association; Florida Agency for Health Care Administration; Georgia Hospital Association; Hawaii Health Information Corporation; Illinois Department of Public Health; Indiana Hospital Association; Iowa Hospital Association; Kansas Hospital Association; Kentucky Cabinet for Health and Family Services; Louisiana Department of Health and Hospitals; Maine Health Data Organization; Maryland Health Services Cost Review Commission; Massachusetts Division of Health Care Finance and Policy; Michigan Health & Hospital Association; Minnesota Hospital Association; Missouri Hospital Industry Data Institute; Nebraska Hospital Association; Nevada Department of Health and Human Services; New Hampshire Department of Health & Human Services; New Jersey Department of Health and Senior Services; New Mexico Health Policy Commission; New York State Department of Health; North Carolina Department of Health and Human Services; Ohio Hospital Association; Oklahoma State Department of Health; Oregon Association of Hospitals and Health Systems; Pennsylvania Health Care Cost Containment Council;Rhode Island Department of Health; South Carolina State Budget & Control Board; South Dakota Association of Healthcare Organizations; Tennessee Hospital Association; Texas Department of State Health Services; Utah Department of Health; Vermont Association of Hospitals and Health Systems; Virginia Health Information; Washington State Department of Health; West Virginia Health Care Authority; Wisconsin Department of Health Services; Wyoming Hospital Association. Finally, we thank Andy Mosso and Zhengyi Fang at Social and Scientific Systems, Inc, for programming support.

      Supplementary Data

      References

        • Simpson L.
        • Zodet M.W.
        • Chevarley F.M.
        • et al.
        Health care for children and youth in the United States: 2002 report on trends in access, utilization, quality, and expenditures.
        Ambul Pediatr. 2004; 4: 131-153
      1. NQF-endorsed standards. The National Quality Forum. Available at: http://www.qualityforum.org/Measures_List.aspx. Accessed April 29, 2011.

      2. Centers for Medicare & Medicaid Services. Available at: http://www.cms.gov/HospitalAcqCond/06_Hospital-Acquired_Conditions.asp. Accessed April 29, 2011.

      3. National Healthcare Quality Report, 2009. Rockville, Md: Agency for Healthcare Research and Quality. AHRQ Publication No. 10–0003. Available at: http://www.ahrq.gov/qual/qrdr09.htm. Accessed April 29, 2011.

        • Billings J.
        • Anderson G.M.
        • Newman L.S.
        Recent findings on preventable hospitalizations. Health Aff. 1996; 5: 240-249
      4. AHRQ. Overview of the nationwide inpatient sample. Available at: http://www.hcup-us.ahrq.gov/nisoverview.jsp/. Accessed April 29, 2011.

      5. AHRQ Agency for Healthcare Research and Quality. Available at http://ww.meps.ahrq.gov. Accessed April 29, 2011.

      6. Healthcare Cost and Utilization Project (HCUP). Clinical classifications software (CCS) for ICD-9-CM. Available at: http://www.hcup-us.ahrq.gov/toolssoftware/ccs/ccs.jsp. Accessed April 29, 2011.

      7. Claritas Inc. (now a division of Nielsen, Inc.) The Claritas demographic update methodology. July 2005. Available from the authors. Accessed April 1, 2009.

      8. The U.S. Department of Health and Human Services, Assistant Secretary for Planning and Evaluation. The 2005 HHS poverty guidelines. Available at: http://aspe.hhs.gov/poverty/05poverty.shtml. Accessed August 30, 2009.

      9. Healthcare Cost and Utilization Project (HCUP). Cost-to-charge ratio files. Available at: http://www.hcup-us.ahrq.gov/db/state/costtocharge.jsp. Accessed April 29, 2011.

      10. Song, X, Friedman, B. Calculate cost adjustment factors by APR-DRG and CCS using selected states with detailed charges. HCUP Methods Series Report # 2008-04. Online October 8, 2008. U.S. Agency for Healthcare Research and Quality. Available at: http://www.hcup-us.ahrq.gov/reports/methods/2008_04.pdf. Accessed April 29, 2011.

        • Institute of Medicine Committee on Quality of Health Care in America
        Crossing the Quality Chasm: A New Health System for the 21st Century.
        National Academies Press, Washington, DC2001
      11. Agency for Healthcare Research and Quality. Pediatric Quality Indicators Overview AHRQ Quality Indicators. Available at: http://qualityindicators.ahrq.gov/modules/pdi_overview.aspx. Accessed December, 2010.

      12. Coffey R, Barrett M, Houchens R, et al. Methods Applying AHRQ Quality Indicators to Healthcare Cost and Utilization Project (HCUP) Data for the Eighth (2010) National Healthcare Quality Report (NHQR) and National Healthcare Disparities Report (NHDR). Rockville, Md: U.S. Agency for Healthcare Research and Quality; 2010. HCUP Methods Series Report 2010-06. Available at: http://www.hcup-us.ahrq.gov/reports/methods/2010_06.pdf.

      13. AHRQ. Welcome to HCUPnet. Available at: http://hcupnet.ahrq.gov/. Accessed April 29, 2011.

        • Lubell K.
        • Kegler S.
        • Crosby A.
        • Karch D.
        Suicide trends among youths and young adults aged 10–24 years—United States.
        MMWR. 2007; 56 (Available at:): 905-908
        • Kenney G.
        • Yee J.
        SCHIP at a crossroads: experiences to date and challenges ahead.
        Health Aff. 2007; 26: 356-369
      14. America’s Children in Brief: National Indicators of Well-Being, 2010. Available at: http://www.childstats.gov/americaschildren/care.asp. Accessed March, 2011.

      15. Kohn LT, Corrigan JM, and Donaldson MS (editors). To Err is Human: Building a Safer Health Care System, Institute of Medicine. Washington DC: National Academy Press, 2000.

      16. Centers for Medicare & Medicaid Services. Trends for 2000–2009 in Medicaid managed care are available from the U.S. Centers for Medicare and Medicaid services. Available at: https://www.cms.gov/MedicaidDataSourcesGenInfo/05_MdManCrPenRateandExpEnrll.asp. Accessed December, 2010.

      17. Guttmacher Institute. Facts on American teens’ sexual and reproductive health. January 2010. Available at: http://www.guttmacher.org/pubs/FB-ATSRH.html. Accessed April 29, 2011.

      18. Abma JC, Martinez GM, Copen CE. Teenagers in the United States: sexual activity, contraceptive use, and childbearing, National Survey of Family Growth, 2006–08. U.S. Center for Disease Control, National Center for Health Statistics. Vital Health Stat 23(30). 2010. Available at http://www.cdc.gov/nchs/data/series/sr_23/sr23_030.pdf. Accessed April 29, 2011. June, 2010.

        • Moran G.
        • Krishmadasan A.
        • Gorwitz R.
        • et al.
        Methicillin-resistant S. aureus infections among patients in the emergency department.
        N Engl J Med. 2006; 355: 666-674
        • Klevens R.
        • Morrison M.
        • Nadle J.
        • et al.
        Invasive methicillin-resistant Staphylococcus aureus infections in the United States.
        JAMA. 2007; 298: 1763-1771
      19. Liu C, Bayer A, Cosgrove S, et al. Clinical practice guidelines by the Infectious Diseases Society of America for the treatment of methicillin-resistant staphylococcus aureus infections in adults and children. Clinical Infections Diseases. Available at: http://cid.oxfordjournals.org/content/early/2011/01/04/cid.ciq146.full. Accessed March, 2011.

      20. Kaiser Family Foundation. Medicaid managed care enrollees as a percent of state Medicaid enrollees, as of June 30, 2009. Available at: http://www.statehealthfacts.org/comparemaptable.jsp?ind=217&cat=4. Accessed November, 2010.

        • Chapter 6: Efficiency
        National Healthcare Quality Report, 2009.
        Agency for Healthcare Research and Quality, Rockville, Md2009 (141. AHRQ Publication No. 10–0003.)
      21. Homer CJ. Health disparities and the medical home: could it be that simple? Acad Pediatr. July 2009;9:203–205

      22. American Academy of Pediatrics, Clinical Practice Guidelines. Available at: http://practice.aap.org/content.aspx?aid=1430. Accessed March, 2011.

        • Sandora T.
        Prevention of healthcare associated infections in children: new strategies and success stories.
        Curr Opin Infect Dis. 2010; 23: 300-305
      23. Healthcare Cost and Utilization Project (HCUP). Available at: http://www.hcup-us.ahrq.gov/toolssoftware/procedure/procedure.jsp. Accessed April 29, 2011.

        • Dougherty D.
        • Meikle S.F.
        • Owens P.
        • et al.
        Children’s health care in the first National Healthcare Quality Report and National Healthcare Disparities Report.
        Med Care. 2005; 43: I58-I63
        • Simpson L.
        • Dougherty D.
        • Krause D.
        • et al.
        Measuring children’s health care quality.
        Am J Med Qual. 2007; 22: 80-84
        • Miller M.
        • Gergen P.
        • Honour M.
        • Zhan C.
        Burden of illness for children and where we stand in measuring the quality of this health care.
        Ambul Pediatr. 2005; 5: 268-278
      24. Coffey R, Milenkovic M, Andrews RM. The Case for the Present-on-Admission (POA) Indicator. Rockville, Md: U.S. Agency for Healthcare Research and Quality; 2006. HCUP Methods Series Report 2006-01. Available at: http://www.hcup-us.ahrq.gov/reports/2006_1.pdf Accessed April 29, 2011.

      25. Available at: http://www.hcup-us.ahrq.gov/databases.jsp. Accessed April 29, 2011.