Chapter 6. AHRQ Program Objectives (continued)

Goal 3—Efficiency

AHRQ continues to support research on ways to overcome barriers to care, provide more efficient and effective care, and improve access to care for all Americans. Research in these areas has become more important than ever, as health care costs continue to rise.

Expenditures for health care constitute a substantial portion of the Nation's gross domestic product. For example, data from AHRQ's Medical Expenditure Panel Survey (MEPS) show that total outpatient drug expenses for the Nation's community population grew from $65.3 billion in 1996 to $94.2 billion in 1999, an increase of 44 percent.

In addition, MEPS data show that employee contributions for employer-sponsored health insurance are also on the rise. In 2001, average employee contributions were $498 for single health insurance coverage and $1,741 for family coverage. Employee contributions increased 10.8 percent and 7.8 percent, respectively, over the 2000 levels, continuing the trend from previous years.

Given the increasing costs of health care, it is vitally important for us to find ways to become more efficient and effective in providing high-quality health care. AHRQ research is at the forefront of this effort to improve health care efficiency and effectiveness. For example:

• Maimonides Medical Center in Brooklyn, NY, uses the five-level triage instrument developed through an AHRQ grant. The Emergency Severity Index, or ESI, changed both hospital policy and clinical practice in the hospital's emergency department (ED). Using ESI, the hospital is able to more efficiently and effectively channel resources to patients according to the severity of their condition. Maimonides Medical Center is a 704-bed, urban, tertiary-care teaching hospital with approximately 80,000 ED visits annually, all of which are evaluated by ESI triage. AHRQ researchers developed the ESI for stratifying patients into five groups with distinct clinical situations. Triage levels range from Level 1, which includes patients who are unresponsive, not breathing, or intubated, to Level 5, which includes patients who are predicted to require only the taking of a history and a physical exam.
• Philips Medical Systems, a large private-sector electronics company, is using software developed as a result of AHRQ-funded research conducted by Tufts-New England Medical Center. There researchers developed software for use with electrocardiogram machines to assist clinical decisionmaking for heart attack patients seen in hospital emergency rooms. Known as the Acute Cardiac Ischemia Time-Insensitive Predictive Instrument (ACI-TIPI), it helps physicians diagnose cardiac ischemia more quickly and recommends whether thrombolytic drugs should be administered. Philips Medical offers the software as a standard feature on their ECG machines sold worldwide. The AHRQ study published in 1998 estimated that widespread use of this instrument could result in 204,000 fewer hospital admissions each year and 112,000 fewer coronary care unit admissions, for an overall annual savings of $728 million.
• A recent AHRQ-funded study by researchers at the University of California, San Diego, found that consultations by specialists to help resolve ethical conflicts resulted in a significant reduction in life-sustaining treatments that were highly unlikely to be beneficial to dying patients. In this randomized controlled trial involving seven U.S. hospitals, dying patients who were in conflict with family members or their health care team over the course of their treatment were offered ethics counseling. Patients who received counseling and died before discharge spent 3 fewer days in the hospital and nearly 1.5 fewer days in the intensive care unit, as well as 2 fewer days receiving life-sustaining treatments such as mechanical ventilation. There was no difference in overall mortality between the patients who received ethics consultation and those who did not.

Addressing Challenges to Care

The combination of rapid advances in medical knowledge and increased use of evidence-based decisionmaking in medicine holds great promise for improving health care. Developments in genomics, pharmaceuticals, informatics, and other technologies promise increased longevity and better health and functioning. Health care, however, can only be as good as the systems that provide it.

Much of the health care provided in the United States is delivered within large and often fragmented systems with complex funding streams. While the United States has an excellent health care system in many ways, it also exhibits waste and inefficiency which in turn exacerbates health care costs, affects affordability, and creates access problems. Low-income individuals, from both rural and urban areas, and those who lack health insurance are particularly likely to experience these problems. In addition, the current health care system lacks the continuity of services that chronically ill patients need.

One byproduct of the current health care system is an increased incidence of injuries to patients from the care that is intended to help them. Problems with patient safety represent only a small part of the unfolding story of problems with quality in American health care. The current health care system also has an impact on other dimensions of quality, such as efficiency, effectiveness, equity, timeliness, and patientcenteredness.

In this complex and sometimes confusing health care marketplace, all participants in the health care system—employers, insurers, providers, consumers, and Federal and State policymakers—need objective, science-based information they can rely on to help them make critical decisions about health care costs and financing and ways to enhance access to care.

For many years, AHRQ has been supporting research to meet this need. The Agency addresses critical health policy issues through ongoing development and updating of nationally representative databases, the production of public use data products, and research analyses conducted by AHRQ staff and extramural researchers.

Impact of Payment and Organization on Cost, Quality, and Equity

In order to be successful, efforts to improve the quality and efficiency of health care in the United States must be based on a thorough understanding of how the Nation's health systems work and how different organizational and financial arrangements affect health care. AHRQ's research initiatives address these issues by asking questions such as:

• How do different payment methods and financial incentives within the health care system affect health care quality, costs, and access?
• What has been the impact of purchaser and public-sector initiatives on quality, costs, and access to health care and health insurance? For example, what is the impact of employer and coalition efforts on the quality and cost-effectiveness of care in the marketplace? Other important and related questions focus on the impact of State efforts to monitor and improve access and quality and the impact of public and private payment changes on access to health care and health insurance for vulnerable populations.
• What impact do managed care features—such as patient cost or sharing, provider payment method, and use of quality assurance techniques—have on the processes of care in relation to evidence-based care recommendations, clinical outcomes, and patient satisfaction?
• How do different organizational and financial features of health plans affect the care provided to people who have chronic illnesses, such as asthma, diabetes, ischemic heart disease, or glaucoma?

To answer these and other related questions, AHRQ issued a Program Announcement on the effects of payment and organization on the cost, quality, and equity of health care. Examples of projects funded under this program announcement include:

• Hospital finances and the quality of hospital care. These researchers are examining the relationship between a hospital's financial condition, its operational and resource allocation decisions, and the quality of clinical care provided at the hospital. Many hospitals and provider organizations in the United States are facing significant financial constraints. This project will help to explain how financial conditions affect operational decisions.
• Purchaser/provider evaluation: Hospital quality data. These researchers are comparing the perceptions of health care purchasers (employers) and hospital administrators regarding the relative importance of hospital quality measures and how purchasers use reports of hospital performance when negotiating health coverage. The study also will address key policy questions, such as the impact and utility of public disclosure of hospital patient safety measures; possible pathways for improving hospital performance; and ways to increase the impact of future hospital comparative reports.
• Physician networks and children with chronic conditions. The goal of this project is to examine the effects of offering out-of-network benefits for children with one of two chronic conditions: asthma or diabetes. The researchers will determine the associations between the out-of-network benefits and cost-sharing of health plans and children's likelihood of seeing an out-of-network physician. They also will determine whether quality of care and expenditures are significantly different for children seeing in-network versus out-of-network physicians.

Research on Lowering Health Care Costs

AHRQ has a broad portfolio of research focused on identifying ways to lower health care costs without negatively affecting the quality and safety of care. Particular attention is given to initiatives that have the potential to lower costs to the Medicare and Medicaid programs. Examples of findings from recent AHRQ research on lowering health care costs include:

• Prescription drug formularies. Spending on prescription drugs increased by more than 200 percent between 1990 and 2000, making it the most influential driver of increases in health care spending. A recent AHRQ-funded study found that prescription drug formularies—a select list of covered drugs that an insurer usually obtains at discounted prices from the manufacturer—can reduce drug costs. When the Veterans Health Administration implemented a national closed formulary in 1997, they effectively shifted patients toward the selected formulary drugs, achieved substantial price reductions from drug manufacturers, and greatly decreased drug spending. Overall, the formulary saved an estimated $82.4 million for five classes of drugs over a 26-month period following implementation of the closed formulary.
• Impact of financial policies on patient treatment and outcomes. Two recent AHRQ-funded studies demonstrate that State and Federal health care financial policies have an effect on the treatments patients receive and their outcomes. One study found that Medicare physician fees influence the type of surgery—either mastectomy or breast-conserving surgery—offered to older women with breast cancer. This finding suggests that physicians are responsive to financial incentives when the alternative procedures have clinically equivalent outcomes. The other study found that market reforms instituted in New Jersey led to an increase in the mortality rate among that State's uninsured heart attack patients. The rates of expensive cardiac procedures for these patients decreased as well, which may partly explain this finding.
• Cost of hospital care for diabetes patients. According to this study, half of hospital costs for diabetes patients in 1999 were linked to the subset of patients who had multiple hospitalizations. Hospital costs per patient were nearly three times as high for patients with multiple versus single hospital stays. Elderly Hispanics and blacks were more likely than whites to have multiple hospitalizations, as were patients covered by Medicare or Medicaid and those living in low-income areas. The presence and type of diabetes complications among patients with multiple hospitalizations varied by age, race/ethnicity, and insurance status. Adults were more likely to be hospitalized for chronic complications of diabetes. The acute complication rate was much higher for blacks than for other racial/ethnic groups, and it was more than twice as high for uninsured as for insured patients. Acute complications of diabetes was the primary or coexisting condition for the more than 60 percent of children with diabetes who had multiple hospitalizations. The researchers note that many of these complications and related hospitalizations might be prevented with quality outpatient care.

Improving Access to Care

Identifying ways to improve access to care—particularly for low-income individuals, minorities, and other priority populations—has been a major focus for AHRQ research for many decades. Examples of AHRQ-supported research on access to care include:

• Enhancing access to cervical cancer screening through telecolposcopy for rural women who have abnormal Pap smears.
• Eliminating system-level barriers to delivery of preventive services to adolescents during primary care visits.
• Ensuring that individuals enrolled in Medicaid managed care receive care equal to that of privately insured individuals.
• Providing equal access to life-saving therapies for heart attack, regardless of sex or race.
• Using nontraditional avenues to provide poor and minority women with access to mammography services.

Recent findings from AHRQ research on access to care include:

• Language proficiency independently affects care, according to a study of care assessments completed in 2000 by more than 49,000 adults enrolled in Medicaid managed care plans in 14 States. Racial/ethnic minorities had lower reports of care than whites who spoke English, especially for timeliness of care and staff helpfulness (courtesy and respect). Non-English speaking Asians had the lowest reports and ratings of care of all groups studied. Non-English speaking whites and Hispanics had worse reports and ratings of care than people who spoke English or were bilingual. The researchers suggest potential remedies, including the use of interpreters, cultural competency training for office staff, and use of community health workers.
• Medicaid shifts to widely used mental health carve-out programs may interrupt therapy among the most needy patients, particularly those suffering from schizophrenia and other serious mental illnesses. A disruption in continuity of care for these patients may lead to the loss of medication adherence, putting patients at risk of having acute psychotic episodes that require hospitalization. A study of more than 8,000 adults enrolled in Tennessee's Medicaid program revealed that the sickest patients were most likely to miss more than 60 days of antipsychotic therapy and to experience interruptions in their outpatient care.
• Title I funds from the Ryan White Act help many people with HIV/AIDS obtain medical and dental care, case management, substance abuse treatment, mental health services, and other support services. Title I of the Ryan White Act provides emergency assistance to eligible metropolitan areas that are disproportionately affected by HIV/AIDS to provide health care for low-income, uninsured, and under-insured people with HIV. Although Title I programs received about $604 million through the Act for FY 2001, many programs needed to use waiting lists for some services. Congress appropriated over $1.9 billion for the programs in FY 2002, but this increased funding was not sufficient to provide care for all eligible individuals with HIV disease.
• Early access to specialty care improves survival among patients with end-stage renal disease (ESRD). The annual mortality rate for ESRD patients is 20 percent, with a 5-year survival rate of about 29 percent for patients undergoing renal dialysis. The risk of death is highest during the first year of dialysis, but early nephrologist evaluation is associated with better patient outcomes. Nevertheless, 25 percent of ESRD patients first see a nephrologist just 1 month before beginning dialysis. ESRD patients who consult with a nephrologist more than 3 months before beginning dialysis and those who see the specialist more often are less likely to die during the first year of dialysis than patients with delayed referral. Other factors affecting survival include coexisting disease, black race, and lack of health insurance.

Medical Expenditure Panel Survey

AHRQ's Medical Expenditure Panel Survey (MEPS) is the only national source for annual data on the specific health services that Americans use, how frequently the services are used, the cost of the services, and the methods of paying for these services. In addition to collecting detailed information from American households, MEPS also collects data from medical providers and establishments. As a result, the survey is unparalleled in its degree of detail.

MEPS is designed to help us understand how the growth of managed care, changes in private health insurance, and other dynamics of today's market-driven health care delivery system have affected, and are likely to affect, the kinds, amounts, and costs of health care that Americans use. No other survey provides the foundation for estimating the impact of changes on different economic groups or special populations such as the poor, elderly, veterans, the uninsured, or racial/ethnic groups.

Since 1977 when data from the first expenditure survey became available, AHRQ's expenditure surveys have been an important and unique resource for public and private-sector decisionmakers. Over the years, this rich data source has become more comprehensive and timely. Design enhancements have improved the survey's analytic capacities, allowing for analyses over an extended period of time with greater statistical power and efficiency. The ability of MEPS to examine differences in the cost, quality, and access to care for minorities, ethnic groups, and low-income individuals provided critical data for the National Healthcare Quality Report and the National Healthcare Disparities Report, which present baseline views of the quality of health care and differences in use of services.

Findings Based on MEPS Data

MEPS data have been used by researchers both within and outside the Federal Government to examine issues of importance to policymakers, consumers, and providers. For example, in a recent study published in Health Affairs, researchers used MEPS data to estimate the medical costs attributable to overweight and obesity and found that they may account for as much as 9.1 percent of total U.S. medical expenditures. In 2003, researchers also used MEPS data to:

• Estimate the amount and distribution of medical spending for the 15 most costly conditions.
• Determine how much it would cost to provide health care coverage to the uninsured under various private and public programs.
• Examine the importance of out-of-pocket premium costs in workers' decisions to enroll in employer-sponsored health insurance.

According to new MEPS data released in 2003, expenses for outpatient prescribed medicines increased from $72.3 billion in 1997 to $103 billion in 2000. Specifically:

• Outpatient prescription medications accounted for a greater proportion of total medical expenses, increasing from about 13 percent of total expenses in 1997 to more than 16 percent in 2000.
• Average out-of-pocket medication expenses were more than three times higher for people age 65 and older than for people under age 65 every year from 1997 through 2000.
• Between 1997 and 2000, the average expense for people age 65 and older with any prescription medicine expense increased about 35 percent—from $819 to $1,102. For people under 65, the amount increased about 40 percent, from $347 to $485.
• Each year from 1997 through 2000, the average number of prescriptions for people age 65 and older was more than twice the average number of prescriptions for younger people.

During the last few years, AHRQ has developed a series of statistical briefs using MEPS data. These briefs, which are on the MEPS Web site, offer timely statistical estimates on topics of current interest to policymakers, medical practitioners, and the public at large.

Topics in 2003 included smoking, asthma treatment, trends in antibiotic use among children, expenditures, and insurance. For example, more than 25 million Americans have been told by a physician or other health care provider that they have asthma according to data collected in 2000 by the MEPS. In the 12 months prior to their interview, 6.5 million adults and 3.2 million children had an asthma attack. Asthma attacks—caused by inflammation of the lower airways and obstruction of airflow—can vary from mild to life-threatening. Further, the data showed that:

• Children under the age of 18 are more likely than adults to use asthma medications, but they are less likely than adults to use inhaled steroids. About one-third of asthma patients have a peak flow meter in the home. A peak flow meter is a portable, hand-held device used to measure how air flows from lungs in one "fast blast" to measure the ability to push air out of the lungs.
• Females are more likely than males to use inhaled steroids and are more likely to have a peak flow meter at home.
• Children under the age of 18 are more likely than adults to have asthma. Among children, males are more likely than females to have active asthma. Among adults, females are more likely than males to have active asthma.

Collecting MEPS Data

AHRQ fields a new MEPS panel each year. Two calendar years of information are collected from each household in a series of five rounds of data collection over a 2-½-year period. These data are linked with additional information collected from respondents' medical providers and employers. This series of data collection activities is repeated each year on a new sample of households, resulting in overlapping panels of survey data.

The data from earlier surveys have quickly become a linchpin for the Nation's economic models and projections of health care expenditures and use. The level of detail these surveys supply permits the development of public and private-sector economic models to project national and regional estimates of the impact of changes in financing, coverage, and reimbursement, as well as estimates of who benefits and who bears the cost of a change in policy.

MEPS establishment surveys have been coordinated with the National Compensation Survey conducted by the Bureau of Labor Statistics through AHRQ's participation in the Inter-Departmental Work Group on Establishment Health Insurance Surveys. Based on the Department's survey integration plan, MEPS linked its household survey and the National Center for Health Statistics' National Health Interview Survey, achieving savings in sample frame development and enhancements in analytic capacity.

AHRQ has moved from conducting a medical expenditure survey every 10 years to following a cohort of families on an ongoing basis. Doing so has four primary benefits. It:

• Decreases the cost per year of data collected
• Provides more timely data on a continuous basis.
• Creates for the first time the ability to assess changes over time.
• Permits the correlation of these data with the National Health Accounts, which measure spending for health care in the United States by the type of service delivered (e.g., hospital care, physician services, nursing home care, and other types of care) and the source of funding for those services (private health insurance, Medicare, Medicaid, out-of-pocket spending, and so on).

Public Use Data Files and Other MEPS Products

AHRQ ensures that MEPS data are readily available, consistent with privacy policies, for use in research and policymaking. MEPS data are released in a variety of ways:

• Data files. MEPS data populate a number of analytical databases. Several public use data files that facilitate national estimates of expenditure for use of health care services are released annually.
• Printed data. AHRQ publishes MEPS data in tabular form on a range of topics. Printed publications include methodology reports, findings, and chartbooks.
• Web site. AHRQ maintains a Web site specific to MEPS. Data files and other MEPS products are made available to the research community and other interested audiences. The site may be accessed at http://www.meps.ahrq.gov. MEPS maintains an E-mail address for technical assistance. In 2003, more than 1,000 inquiries were answered, most within 24 hours.
• MEPSnet. AHRQ has developed a set of statistical tools to allow immediate access to MEPS micro data in a nonprogramming environment. From the MEPSnet section of the MEPS Web site, through a series of interactive queries, the most inexperienced user can generate national estimates in a few seconds. This source may be accessed at http://www.meps.ahrq.gov/mepsweb/data_stats/meps_query.jsp.
• LISTSERV™. The purpose of the MEPS LISTSERV™ is to allow free exchange of questions and answers about the use of the MEPS database. Currently, there are more than 500 subscribers to the MEPS LISTSERV™.
• Training. AHRQ conducts workshops—ranging in length from a few hours to several days—to educate policymakers, researchers, and other users about the range of questions that MEPS can answer and how the data can be properly used.
• Data Center. AHRQ's Center for Financing, Access, and Cost Trends (CFACT) operates an on-site data center to facilitate researchers' access to data that cannot otherwise be publicly released for reasons of confidentiality.

Recent Key Findings from the MEPS Household Component

• In 2001, approximately 12 percent of U.S. women ages 18 to 64 had not received Pap test screening within the previous 3 years. Those who were uninsured or covered by Medicaid or other public health insurance were less likely to have had Pap tests than those with private health insurance.
• In 2001, nearly 25 percent of children under age 18 had not been to a doctor's office or visited a clinic during the previous 12 months. Of the 75.8 percent of children who did receive care, 91.9 percent were reported to have had no problems receiving needed care. While a high percentage of Hispanic children were reported to have no problems receiving necessary care (88.4 percent), the percentage was lower than it was for black (91.7 percent) and white children (92.7 percent).
• In the first half of 2002, 16.4 percent of Americans (46.2 million people) were uninsured.
• Among those under age 65, more than a third of Hispanics (36.1 percent) and 20.4 percent of black non-Hispanics were uninsured during the first half of 2002, compared with 14.6 percent of white non-Hispanics.
• In 2002, young adults ages 19-24 were the age group at the greatest risk of being uninsured, with more than one-third (34.9 percent) of this group lacking health insurance.
• From 1996 to 2002, the percentage of uninsured children declined from 15.7 percent to 12.9 percent. Concurrently, the percentage of children covered only by public health insurance increased from 21.3 percent to 26.3 percent.
• Higher hourly earnings were associated with a greater likelihood of workers having health insurance coverage through their primary place of employment (83.2 percent of workers making $21 or more per hour had such coverage). Only one-third of workers making less than minimum wage ($5.15 per hour) had insurance coverage from their primary employer.
• Employees who belonged to a labor union were much more likely to be covered by health insurance through their main job (88.0 percent) than nonunion workers (57.6 percent).

Recent Key Findings from the MEPS Insurance Component

• In 2001, average health insurance premiums were $2,889 for single coverage, $5,463 for employee-plus-one coverage (a plan covering the employee and one other person), and $7,508 for family coverage. Premiums for single and family coverage increased 8.8 percent and 10.9 percent, respectively, over 2000, continuing a trend of increasing premiums each year since 1996. Data on employee-plus-one premiums were collected in 2001 for the first time.
• In 2001, the percent of employees in the 10 largest States opting for single coverage ranged from 38.2 percent in Ohio to 52.4 percent in Texas. The 10 States are California, Texas, New York, Florida, Illinois, Pennsylvania, Ohio, Michigan, New Jersey, and Georgia.
• Single premiums ranged from $2,777 in California to $3,105 in New Jersey, employee-plus-one premiums ranged from $5,098 in California to $6,055 in New York, and family premiums ranged from $7,162 in California to $8,227 in New York.
• The average employee contribution to health insurance premiums in 2001 was $498 for single coverage, $1,070 for employee-plus-one coverage, and $1,741 for family coverage. Employee contributions increased 10.8 percent and 7.8 percent for single and family coverage, respectively, over 2000.
• Employee contributions for single coverage ranged from $369 in California to $584 in Florida, family coverage from $1,358 in Ohio to $2,217 in Florida, and employee-plus-one coverage from $688 in Michigan to $1,663 in Florida.
• Full-time private-sector employees that work for firms with 50 or more employees are almost always offered health insurance by their employer. By comparison, only 55.3 percent of firms with 10 or fewer employees offer health insurance to their full-time employees.
• In 2001, the average health insurance premiums for State and local government employees were $3,182 for single coverage, $5,463 for employee-plus-one coverage, and $7,336 for family coverage. Premiums for single and family coverage increased 11.5 percent and 10.2 percent over 2000.

Healthcare Cost and Utilization Project

As health care costs in the United States escalate and concerns about quality of care become more pronounced, the need for accurate and timely health care data has increased dramatically. Policy analysts, administrators, and the research community require comprehensive and precise data resources in order to evaluate cost, quality, and access to care. The AHRQ-sponsored Healthcare Cost and Utilization Project (HCUP) is a resource comprising a family of databases that meets this need for reliable data.

HCUP develops and maintains a family of health care databases, related software tools, support services, and products whose information resources are grounded in vital partnerships among Federal, State, and industry associations. HCUP databases integrate the data collected by State governments, hospital associations, private data organizations, and the Federal government (a mosaic of "Partners") to create a national health care information resource of hospital, ambulatory surgery center, and emergency department data. HCUP features the largest collection of hospital care data collected over a period of time in the United States. All-payer, encounter-level information is available beginning in 1988.

The multi-State databases contain discharge-level information in a uniform format designed to ensure patient privacy. The resulting HCUP databases facilitate research on a broad range of health policy and health services issues, including:

• Cost and quality of health services.
• Medical practice patterns.
• Access to health care.
• Hospital costs and use, including use by special populations.
• Diffusion of medical technology.
• Effects of market forces on hospitals.
• Treatment outcomes at the national, regional, State, and local market levels.

Because of their large size and scope, HCUP databases enable analyses, such as investigating specific medical conditions and procedures (including rare events); tracking use for population subgroups, such as minorities, children, women, and the uninsured; and analyzing different geographic levels (national, regional, State, and community) within the United States. To augment the HCUP databases, software tools and Web-based products are publicly available for use by audiences with varying levels of research experience.

The collaboration of Federal, State, and industry partners creates a mutually beneficial opportunity for sharing data and building a national resource. HCUP research benefits extend to a diversity of institutions and individuals:

Policy analysts: HCUP data enable policy analysts and decisionmakers to develop effective and informed recommendations on crucial health care policy issues such as cost, use, quality, and access to health care.

Hospital industry: HCUP provides hospital associations, hospitals, and provider alliances with access to national health care databases. Hospital industry members are able to make national comparisons of efficiency, cost, value, and quality of service.

Statewide data organizations: Participants in this national project can contribute to the health care knowledge base and reinforce the value of their data collection efforts. States are also able to compare their health care statistics to other States and to regional and national indicators. Those States that are just starting data collection programs may benefit from HCUP technical assistance and the experience of other organizations already collecting these data.

Researchers: The comprehensive data available in HCUP databases enable researchers to conduct health services research in many different areas, including but not limited to quality of care, medical practice patterns, and treatment outcomes.

HCUP databases include:

• The Nationwide Inpatient Sample (NIS) with inpatient data from a national sample of more than 1,000 hospitals.
• The Kids' Inpatient Database (KID) is a nationwide sample of pediatric inpatient discharges.
• The State Inpatient Databases (SID) contain inpatient discharge abstracts from non-Federal hospitals in participating States.
• The State Ambulatory Surgery Databases (SASD) contain data from ambulatory care encounters from hospital-affiliated and sometimes freestanding ambulatory surgery sites.
• The State Emergency Department Databases (SEDD) contain data from emergency department encounters from hospital-affiliated emergency departments.

AHRQ expands the HCUP databases each year by adding new States that will improve national and regional representation and by expanding the number of partners that contribute ambulatory surgery and emergency department data. In the last year AHRQ added Minnesota, Nebraska, Rhode Island, and Vermont to HCUP. Each of the four new States supplied data beginning with the 2001 data year. AHRQ also added three new ambulatory surgery databases and three new emergency department databases. Currently, 33 Statewide data organizations participate in HCUP.

Over the past 2 years, AHRQ has implemented a multifaceted effort to make HCUP data more accessible to researchers and other interested users. A centerpiece of this effort is HCUPnet, an interactive tool for identifying, tracking, analyzing, and comparing statistics on hospital care. HCUPnet queries generate statistics in a table format using data from the NIS and SID databases for those States that have agreed to participate. New data have recently been added from the Kids' Inpatient Database (KID). Aggregate statistics from HCUPnet are made available to the public on the AHRQ HCUPnet Web site at http://www.ahrq.gov/data/hcup/hcupnet.htm.

Another means instituted by AHRQ to enhance access to HCUP data is the HCUP Central Distributor, which was developed by AHRQ to prepare and distribute HCUP data for use by researchers outside of AHRQ on behalf of participating HCUP partner organizations. Participation in the HCUP Central Distributor is voluntary, and it includes only the partner organizations that agree to release their HCUP data to public and private users.

Since September 2002, AHRQ has been building an HCUP user support infrastructure that includes telephone and E-mail access to technical support staff and a publicly available Web site that features database documentation and HCUP product information. The goal of HCUP user support is to increase awareness of the strengths and uses of HCUP data and to enhance the skills of those using HCUP data and tools for their work in research, education, and policy analysis.

In 2003, HCUP launched a new toll-free dedicated technical assistance phone line (1-866-290-HCUP) and a new dedicated E-mail service (hcup@ahrq.gov). A new Web site has also been established at http://www.hcup-us.ahrq.gov/home.jsp. This Web resource offers general information regarding the HCUP databases, available technical support, project contact information, and background information on the technical support team.

Other HCUP-related activities currently underway at AHRQ include the development of a new fact book that addresses potentially avoidable hospitalizations using the AHRQ Prevention Quality Indicators. This fact book will describe ambulatory care sensitive conditions—that is, conditions that usually can be avoided through timely and effective outpatient care. The fact book will use graphs and tables to describe these conditions, including priority conditions such as asthma, diabetes, congestive heart failure, and hypertension, as well as the incidence of low birthweight infants. Specifically, the fact book evaluates time trends between 1994 and 2000; variations across regions of the United States; and differences among priority populations, including children, the elderly, women, low-income individuals, and rural residents.

Another new resource is Care of Children and Adolescents in U.S. Hospitals, a fact book that examines hospital care for children and adolescents, providing insight into the types of conditions for which children are hospitalized, the types of procedures they receive, who is billed for the stays, the resource use associated with children's hospital stays, and where children are discharged to when they leave the hospital. The fact book begins with an overview of hospital care for children overall and, to put this care into perspective, compares information about children to information about adults' hospital stays. It then provides more detailed information for three major subgroups of pediatric hospital stays:

• Neonatal stays—newborns and infants 30 days and younger.
• Stays for other pediatric illness—admissions for children and adolescents not related to neonatal and maternal conditions.
• Stays for adolescent pregnancy and delivery—admissions for maternal care. The data source for this fact book was the 2000 Kids' Inpatient Database (KID).

Use of HCUP Data

The HCUP databases are being used by a variety of Federal agencies and national health care organizations to examine practices and trends and guide health care decisionsmaking. For example:

• HCUP NIS and SID (2000) data were used to prepare AHRQ's first National Healthcare Quality Report and National Healthcare Disparities Report. These two reports are flagship projects of the Agency that provide national and State-level indicators of how well our health care systems are working.
• A new AHRQ study published in the October 8, 2003, issue of JAMA showed that medical injuries during hospitalization result in longer hospital stays, higher costs, and a higher risk of death based on 2000 HCUP data. The study, which identified medical injuries in 7.45 million hospital discharges from 994 acute care hospitals across 28 States, found that the impact of medical injuries varies substantially. The study provided, for the first time, specific estimates for excess length of stay, charges, and the risk of death for 18 of the Patient Safety Indicators.
• The March of Dimes used HCUP NIS data to analyze and compare hospital charges for infant stays related to prematurity and typical newborn stays. Using these data, the March of Dimes Perinatal Center found that hospital charges for all infant stays with any diagnosis related to prematurity totaled $11.9 billion in 2000. The $11.9 billion figure was cited in the headline of a March 2003 press release announcing the March of Dimes 5-year, $75 million national campaign to decrease the rate of premature births.
• The Vermont Program for Quality and Health Care produces an annual report that examines quality of care and identifies ongoing improvement efforts within the State. The Vermont Healthcare Quality Report provides information regarding high-volume, hospital-based procedures, as well as services delivered within practitioners' offices, nursing homes, and emergency rooms, as well as home based care. The Vermont Program used the HCUP data to benchmark Vermont Statewide estimates. The project uses statistics from HCUPnet and also performs more in-depth analyses using the NIS data files. The use of HCUP data to measure health care quality underscores that HCUP data can be used by State policymakers not only to examine quality in their own State, but also to compare their State to national benchmarks.
• A General Accounting Office (GAO) report on emergency department overcrowding used HCUP Nationwide Inpatient Sample (NIS) data to compare statistics on the types of DRGs (diagnosis related groups—a patient classification system that provides a way of describing the types of patients a hospital treats) and the type of insurance (e.g., managed care, fee-for-service, Medicare, Medicaid, and so on) for patients admitted to the inpatient setting through the emergency department with those admitted through other routes. HCUP staff conducted special analyses to provide these statistics to GAO, which were used in the March 14, 2003 GAO report: Hospital Emergency Departments: Crowded Conditions Vary among Hospitals and Communities, GAO-03-460.
• HCUP assisted GAO with a report released in 2003 on specialty hospitals. HCUP State Inpatient Databases (SID) for six States—Arizona, California, New Jersey, New York, North Carolina, and Texas—were one of the sources of data used in the report. The recent growth in specialty hospitals that are largely for-profit and owned, in part, by physicians, has been controversial. Advocates of these hospitals contend that the focused mission and dedicated resources of specialty hospitals both improve quality and reduce costs. Critics contend that specialty hospitals siphon off the most profitable procedures and patient cases, thus eroding the financial health of neighboring general hospitals and impairing their ability to provide emergency care and other essential community services. Critics also contend that physician ownership of specialty hospitals creates financial incentives that may inappropriately affect physicians' clinical and referral behavior. For this report, the GAO was asked to examine: State policies and local conditions associated with the location of specialty hospitals, how specialty hospitals differ from general hospitals in providing emergency care and serving a community's other medical needs, and how specialty and general hospitals in the same communities compare in terms of market share and financial health.
• The Centers for Disease Control and Prevention, the Hawaii Health Information Corporation (HCUP partner), and AHRQ conducted a collaborative study on Kawasaki's disease—a childhood disorder that primarily affects children younger than 5. It is the leading cause of acquired heart disease in kids, and it is most prevalent in children of Asian and Pacific-Islander background. This study used 1996-2001 Hawaii SID data and 1997-2000 HCUP KID data to investigate the epidemiology, use of care, and outcomes associated with Kawasaki's disease in Hawaii.
• The National Institute for Child Health and Human Development is using HCUP KID data to help in priority setting for their response to the Best Practices for Children Act (BPCA). The BPCA mandates that pediatric-specific pharmaceutical research be conducted. HCUP data are being used to help identify priority conditions and subpopulations on which to focus.
• The Health Resources and Services Administration (HRSA) is using HCUP KID data to create a pediatric-specific case mix index, which will be used in determining payments to children's hospitals in support of graduate medical education at freestanding children's institutions. The current method is based on the Medicare Prospective Payment System; however, because children's hospitals serve a pediatric population and represent a wider spectrum of children's services than hospitals providing care to Medicare patients, HRSA is exploring alternative payment formulas using HCUP data for discharge-level analyses.
• The National Institute of Allergy and Infectious Diseases (NIAID), in collaboration with AHRQ, used the HCUP SID (1993-2001) data to examine the effects of a rotavirus vaccine on hospital admissions among children. Rotavirus is a common cause of severe diarrhea among children and is responsible for half a million deaths in developing countries each year. After a previous study showed an increased risk of a serious and life-threatening complication (intussusception) following administration of rotavirus vaccine, the vaccine was withdrawn from the market. The NIAID and AHRQ analysis found no evidence of increased hospitalizations for intussusception during the period of vaccine availability, which suggests that the risk may be much lower than previously thought. This new information has implications for future vaccine policy.

Quality Indicators

AHRQ's Quality Indicators (QIs) are measures of health care quality that make use of readily available hospital inpatient administrative data, such as diagnoses and procedures, along with information on a patient's age, sex, source of admission, and discharge status. The first QIs were developed in the early 1990s and came out of HCUP through the data partnership among the Agency, the health care industry, and States. These original HCUP QIs, which were created in response to requests for assistance from State-level data organizations and hospital associations, enabled these entities to use their administrative data for quality improvement and tracking.

In preparation for the National Healthcare Quality Report, AHRQ worked in partnership with the University of California, San Francisco (UCSF)/Stanford Evidence-based Practice Center to refine, expand, and risk adjust these measures. The resulting updated AHRQ QIs include three modules: the Prevention Quality Indicators (PQIs), the Inpatient Quality Indicators (IQIs), and the Patient Safety Indicators (PSIs). The AHRQ QI software analyzes administrative data on inpatient stays to produce information about potentially avoidable adverse hospital outcomes such as postsurgical pneumonia, potentially inappropriate and over used hospital procedures such as cesarean section deliveries, and potentially preventable hospital admissions such as pediatric asthma. The AHRQ QIs are designed to highlight probable quality concerns, identify areas that need further study and investigation, and track changes over time. Users include hospitals, managed care organizations, business-health coalitions, researchers, and others at the Federal, State, and local levels. The software is available free on the AHRQ QI Web site at http://www.qualityindicators.ahrq.gov.

The first three AHRQ QI modules—the Prevention Quality Indicators (PQIs), the Inpatient Quality Indicators (IQIs), and the Patient Safety Indicators (PSIs)—include a total of 73 indicators. These QIs are designed for use with inpatient hospital discharge data and primarily apply to the inpatient setting. However, the PQIs are pertinent to outpatient care, since the indicators represent hospitalization rates for events that potentially could be prevented by improvements in access to and/or delivery of outpatient care.

• Prevention Quality Indicators (PQIs). These indicators, released in November 2001, represent "ambulatory care sensitive conditions." These are conditions for which evidence suggests that hospital admissions could be avoided with good access to quality outpatient care or illness could be less severe if the condition is treated early and appropriately. There are 16 PQIs covering conditions such as diabetes, asthma, heart disease, pneumonia, and selected pediatric conditions.
• Inpatient Quality Indicators (IQIs). These indicators, released in June 2002, reflect quality of care inside hospitals and include inpatient mortality; use of procedures for which there are questions about overuse, underuse, or misuse; and volume of procedures for which there is evidence that a higher volume of procedures is associated with lower mortality. The 31 indicators are specific to hospital care for heart disease and surgery, hip repair, pneumonia, childbirth, and other conditions and procedures.
• Patient Safety Indicators (PSIs). These indicators, released in March 2003, provide information on potential in-hospital complications and adverse events following surgeries, procedures, and childbirth. There are 26 indicators that address such topics as birth trauma, complications of anesthesia, transfusion reaction, accidental puncture and laceration, and postoperative infections.

The AHRQ QIs are being used by a variety of providers, purchasers, and State agencies as an integral part of quality improvement programs. Their use has exceeded all Agency estimates. Each month AHRQ responds to an average of 30 user requests, supports 1,920 LISTSERV subscribers, and logs between 3,500 and 6,000 visits to the Web site and more than 1,000 downloads of QI documentation and software. The rapid uptake of the QIs has been due to many factors including a demand for standardized measures using readily available data for both quality improvement and comparative quality reporting, an increased emphasis on public reporting, and efforts to link payment to quality. Examples of QI use:

• Providers.
  • A number of State hospital associations have integrated the PSIs into their member quality improvement programs. These users include the Healthcare Association of New York State, Georgia Hospital Association, Missouri Hospital Association, Montana Hospital Association, Oregon Association of Hospitals and Health Systems, West Virginia Hospital Association, and a subsidiary of the Illinois Hospital Association that also serves hospitals affiliated with the Association of Kentucky Hospitals and Health Systems.
  • The Dallas-Fort Worth Hospital Council uses the IQI and PSI software to generate comparative data and reports using data from 1999-2002. They published hospital-specific comparative information for their members with their own patient-level data for detailed evaluation and quality improvement activities, as well as provided members with an analysis of the PQIs for the North Texas region.
  • The National Association of Children's Hospitals and Related Institutions is currently collaborating with AHRQ researchers to examine patient safety events, using the AHRQ Patient Safety Indicators, in children's hospitals compared with discharges from general medical-surgical hospitals using their member data file and the 2000 KID.
• Purchasers. The Niagara Health Coalition used the IQI software and the New York State hospital discharge data file to generate comparative data and reports for all hospitals in New York State. They initially published these data and reports on their Web site in fall 2002 and updated the results by adding a second year of analyses in March 2003.
• State agencies. The Texas Healthcare Information Council uses the IQI software and their hospital discharge data to generate comparative data and reports for all hospitals in the State of Texas. They released their initial findings on their Web site at the close of FY 2002. Their Web site currently contains IQI analyses for 3 years (1999-2001), and 2002 data will be published soon.
• Federal agencies.
  • AHRQ used the QIs to produce the first ever National Healthcare Quality Report and National Healthcare Disparities Report in 2003.
  • The PSIs are being used in the CMS-Premier national pay-for-performance demonstration projects. These projects involve an estimated 300 hospitals across the country. In addition, the PSIs have been integrated into several leading private sector pay-for-performance programs, including those being led by Anthem Blue Cross and Blue Shield of Virginia.
  • The Medicare Payment Advisory Committee used all three QI modules to examine the quality of care provided to Medicare beneficiaries. Their findings will be presented to Congress sometime in 2004.
  • The Department of Veterans Affairs (VA) Patient Safety Indicator (PSI) study is a jointly funded VA-AHRQ study that is validating the PSIs against NSQUIP (a database maintained by the VA and based on medical record review of postsurgical patients). In addition to validating selected PSIs, the study will also compare VA and non-VA experience of potential patient safety events using five HCUP States as the non-VA comparison group.
• Researchers. AHRQ researchers used the PSIs to create estimates of excess deaths, excess days in hospitals, and excess charges associated with adverse events as published in the October 2003 issue of JAMA.
• Hospital and Healthcare Systems. AHRQ's QIs can help hospitals and healthcare systems answer specific questions such as:
  • How does our hospital's cesarean section rate compare with the State or the Nation?
  • Do other hospitals have similar mortality rates following hip replacement?
  • How does the volume of coronary artery bypass graft surgery in our hospital compare with other hospitals?
• State data organizations and community health partnerships. These groups use AHRQ QIs to ask questions that provide initial feedback on clinical areas appropriate for further, more in-depth analysis, such as:
  • What can the AHRQ pediatric QIs tell us about the adequacy of pediatric primary care in our community?
  • How does the hysterectomy rate in our area compare with the State and national averages?
• State hospital associations. State hospital associations use the AHRQ QIs to do quick hospital quality and primary care access screens. Other potential users include managed care organizations, business-health coalitions, State data organizations, and others poised to begin assessments using hospital discharge data to answer questions such as:
  • Can we design and implement community quality improvement interventions in areas surrounding hospitals that have high rates of admissions for diabetes-related complications?
  • Which QIs can be incorporated into performance management initiatives for our member hospitals?
• Federal policymakers. AHRQ QIs help these users track health care quality in the United States over time and assess whether health care quality is improving, for example:
  • How does the rate of coronary artery bypass-graft surgery vary over time and across regions of the United States?
  • What is the national average for bilateral cardiac catheterization (a procedure that is not generally recommended), and how has this changed over time?

Return to Contents
Proceed to Next Section