July 2009

Grant Results


Researchers conducted a series of telephone surveys to learn how acute-care hospitals across the United States collect data on race, ethnicity and language and whether and how they use that data to improve quality of patient care. Under Grant ID# 050183, researchers at the National Public Health and Hospital Institute surveyed 500 hospitals, and conducted 64 additional in-depth surveys of public hospitals and health systems with diverse patient populations.

Under Grant ID# 059159, researchers at the George Washington University School of Public Health and Health Services updated the survey, and then interviewed executives at 547 acute-care hospitals. They also conducted in-depth case study interviews with hospitals that reported using race, ethnicity and language data to improve care.

Key Findings
In reports to the Robert Wood Johnson Foundation (RWJF), the project director said that:

  • Most hospitals collected some information on the race, ethnicity or language of their patients, but changes between the two surveys were relatively modest.
    • Four out of five hospitals collected information on race in both surveys.
    • About half the hospitals collected data on preferred language in both surveys.
    • The percentage of hospitals collecting data on ethnicity dropped from about 50 percent in the first survey to just below 42 percent in the second.
  • Less than one in five hospitals that collect data on race, ethnicity or language used it to improve the quality of patient care. Even fewer (about one in seven) used the data to identify health disparities.

RWJF supported this project with an unsolicited grant of $189,039 (ID# 050183) from October 2004 through February 2006 and with a solicited grant of $129,105 (ID# 059159) from April 2007 through June 2008.

 See Grant Detail & Contact Information
 Back to the Table of Contents


In its 2002 report Unequal Treatment, the Institute of Medicine (IOM) found that members of racial and ethnic minorities receive lower-quality medical treatment than Whites, even when insurance status, income, age and the severity of the condition are comparable. Among other recommendations to address these disparities, the IOM called for improved data collection by health care organizations on patients' race and ethnicity.

Meeting the IOM's call for better data is a challenge for the health care industry because it has not yet developed uniform standards for collecting data that identifies race and ethnicity, according to the Health Research and Education Trust, the research arm of the American Hospital Association. In 2005, it released a Disparities Toolkit to guide hospitals in collecting patient race and ethnicity data.

 Back to the Table of Contents


As stated in RWJF's Annual Report 2004, "[e]nsuring that all Americans, especially those with chronic conditions, receive high-quality care is central to the Foundation's mission of improving health and health care."

In 2004, the year when RWJF issued the first of the two grants described in this report, the Foundation "pursued a four-pronged approach to improve the quality of care for chronic disease in outpatient settings:

  • National Measures for Quality. To improve the quality of care, we need to first agree on what "quality" means and how to measure it.
  • Engage Patients and Purchasers in Assessing Quality. Patients and purchasers need to be involved to ensure that the care they receive and pay for meets quality standards.
  • Demonstrate that High Quality is Achievable Where Providers, Purchasers and Patients are Aligned Around Common Quality Standards. Working with multiple partners in selected markets, we launch[ed] demonstration projects that align providers, purchasers and patients around common quality goals to raise the standard of care so that outpatient quality standards are met most of the time rather than only half the time. If successful, consumers in these markets will have access to information about the quality of care, and providers will have demonstrated skills in adopting the principles of the Chronic Care Model.
  • Track Progress. It remains critical to track progress toward achieving better care. The Foundation is supporting work to examine whether more purchasers are making decisions based on quality; whether more patients are becoming engaged in managing their own care and ensuring the care they receive is of high quality; and whether more providers are adopting the tools and systems they need to provide high-quality care.

The grants described in this report fit into the fourth of these approaches.

 Back to the Table of Contents


The goal of these grants was to learn how U.S. hospitals collected data on race, ethnicity and language, and how they used the data to reduce disparities and improve patient care, if at all.

ID #050183: The Initial Surveys

Working with a subcontractor, Braun Research of Princeton, NJ, researchers at the National Public Health and Hospital Institute conducted two sets of telephone surveys from October 2004 through February 2006. The National Public Health and Hospital Institute is a private, nonprofit organization established in 1988 to address health policy issues facing public health and hospital systems and underserved communities.

Researchers first completed 500 telephone surveys of executives at acute-care hospitals not operated by the federal government to determine their data collection practices on race, ethnicity and language.

They then conducted 64 more in-depth surveys of executives at public hospitals and health systems that treat disproportionate numbers of uninsured, underinsured or publicly insured patients in their communities. These respondents generally had diverse patient populations and were presumed to have experience collecting data related to race and ethnicity.

ID# 059159: Follow-Up Survey and Case Studies

From April 2007 through June 2008, researchers conducted an updated survey of executives at 547 hospitals to determine whether data collection practices had changed. They also conducted in-depth case study interviews with executives of hospitals that reported using race, ethnicity and language data to improve patient care in order to identify "best practices." None of the hospitals in the first survey were included in the second.

The second grant was made to the George Washington University School of Public Health and Health Services after Marsha Regenstein, the project director on both grants, joined its faculty as associate research professor of public health policy. Braun Research again conducted the survey under a subcontract.


Project staff has produced a report on the first survey, Race, Ethnicity and Language of Patients: Hospital Practices Regarding Collection of Information to Address Disparities in Health Care.

A report on the follow-up survey and case studies, Hospital Practices Regarding Collection and Use of Patient Race, Ethnicity and Language Data: An Update to the Field, was drafted and submitted to RWJF for review.

 Back to the Table of Contents


In their survey reports, the researchers found:

  • Most hospitals collected some information on the race, ethnicity or language of their patients, but changes over time were relatively modest.
    • The percentage of hospitals that collected information on the race of their patients rose from 78.4 percent in the surveys conducted under the first grant to 81 percent in the follow-up grant, an increase that is not statistically significant.
    • The percentage of hospitals collecting information on ethnicity dropped significantly, from about 50 percent to just below 42 percent. Researchers noted that further staff education and training may be needed in ways to present questions of ethnicity.
    • About half the hospitals surveyed under both grants collected data on patient language preference.
    Hospitals that do collect data usually do so in patient registration areas.
  • Investor-owned, teaching and very large hospitals—those with at least 250 beds—were more likely to collect race, ethnicity and language data than either public or nonprofit hospitals. They were also more likely to use electronic hospital records, which allow data to be readily shared across an institution.
  • Only about one in five hospitals surveyed had formal policies governing the collection of racial, ethnic or language data. Even fewer had specific policies governing the categories or methods that should be used to collect data.
  • Fewer than one in five hospitals that collected data on race, ethnicity and language used it to assess or compare quality of care, use of health services, health outcomes or patient satisfaction. See Appendix 1 for examples of hospitals that use data on race, ethnicity and language to improve quality.
  • Even fewer hospitals used the data to identify health disparities. Only about 7 percent of respondents said they had identified disparities based on race or ethnicity, and fewer than 3 percent had identified disparities based on language.
  • The most common barrier to collecting data, among hospitals that do so, was patient reluctance to provide information, reported in about 45 percent of the surveys. Staff inability to communicate in patients' preferred languages (42 percent) and staff reluctance to ask for the information (38 percent) were the next most common barriers.

    Hospitals that trained registration clerks to ask for racial, ethnic and language information reported staff reluctance as a barrier only about 40 percent of the time, compared to 58 percent of those that did not offer training.


In reports to RWJF, project staff recommended that hospitals:

  • Invest time, money and staff training to obtain good data. "Bring everyone who will be involved in data collection to the table early in the process," said the project director. A champion for the cause helps to win the support of hospital leadership, but buy-in from front-line registration staff is also essential, as is making any necessary changes in information systems.
  • Devote substantial up-front time to preparing survey instruments. Project staff had many people review the survey instruments and used their comments to make adjustments, and then field tested the surveys to make sure they provided the sought-after information.
  • Standardize data collection to ensure that it is useful. It is important that data accurately capture the hospital's patient population. Creating meaningful ethnic and racial categories, asking questions in ways that obtain useful and relevant information, and determining how to analyze the data are among the challenges.
  • Use data already being collected to improve quality of care. The project director believes that most hospitals already collect useful data, despite small sample sizes and concerns among physicians and hospital administrators that data collection is not uniform. "Efforts to refine and improve data collection should continue but should not take center stage in the struggle to identify and address health care disparities," according to a grantee report.

 Back to the Table of Contents



Study of hospital practices in the collection of race and ethnicity data


National Public Health and Hospital Institute (Washington,  DC)

  • Amount: $ 189,039
    Dates: October 2004 to February 2006
    ID#:  050183


George Washington University School of Public Health and Health Services (Washington,  DC)

  • Amount: $ 129,105
    Dates: April 2007 to June 2008
    ID#:  059159


Marsha Regenstein, Ph.D.
(202) 994-8662

Web Site


 Back to the Table of Contents


Appendix 1

(Current as of the time of the grant; provided by the grantee organization; not verified by RWJF.)

Hospitals that use data on race, ethnicity and language to improve quality

In-depth case study interviews highlighted some of the ways hospitals are using data on race, ethnicity and language to monitor and improve quality care and reduce disparities:

  • The University of Wisconsin Hospital and Clinics, a 471-bed academic medical center in Madison, uses interpreters in some two dozen languages to provide services at 1,500 to 2,000 patient appointments every month. A dedicated staff member became the champion of an effort to standardize and expand data collection in order to measure patient outcomes more effectively.

    This person shared research with the hospital leadership team to assuage concerns about whether registration staff and patients would be comfortable asking and answering questions about race and ethnicity. She then developed a rigorous training program for staff, including cultural competency training, to help them recognize how such data could improve care. "Many of them had never heard about disparities," she said.

    The project team decided to wait 18 months after uniform data collection was initiated to begin analyzing the data so that all patients could be re-registered under the new system. The process of analyzing the data for quality measures and satisfaction has begun, and the hospital expects it to lead to specific clinical interventions.
  • The University of Michigan Health System, an academic public health system in Ann Arbor comprising three hospitals, 30 health centers and 120 outpatient clinics, serves patients who speak some 40 different languages. The health system, a participant in the RWJF program Speaking Together: National Language Services Network, sought to improve services for patients with limited language proficiency.

    The health system started by improving its data collection system, training intake workers on the importance of collecting language data and using computerized reminders whenever the language field was left blank. It also revised its admissions questionnaire to ask what language patients preferred to use in communicating with medical staff, rather than just asking about their "primary language." This avoided assigning interpreters to patients who were comfortable in English.

    The health system also developed several chronic disease registries that analyze quality indicators by language to ensure that patients needing language assistance are receiving the same quality of care as patients who do not. The results are shared throughout the system.
  • The University of Mississippi Medical Center is a 722-bed academic medical center in Jackson, a city whose population is 76 percent Black, 23 percent White and 1 percent Hispanic.

    Seeking to reduce disparities in cardiac care, the hospital—one of 10 participants in the RWJF national program Expecting Success: Excellence in Cardiac Care—developed a system to standardize collection of race, ethnicity and language data. The hospital also provided special training to staff members who frequently did not ask questions about language.

    The increased attention to standardizing data collection helped to identify a Spanish-speaking population of children and pregnant women that was larger than staff had recognized. As a result, the hospital is making certain that needed interpreter services are available at appropriate care sites and expanding services to include a bilingual patient educator for Spanish-speaking patients. The hospital also discovered small patient populations who speak dozens of other languages, and has provided them with telephone interpretation services when needed.

 Back to the Table of Contents


(Current as of date of this report; as provided by grantee organization; not verified by RWJF; items not available from RWJF.)


Regenstein M and Sickler D. Race, Ethnicity, and Language of Patients: Hospital Practices Regarding Collection of Information to Address Disparities in Health Care. Washington: National Public Health and Hospital Institute, 2006. Available online.

Survey Instruments

"Survey: Hospital Practices in the Collection of Data on Race, Ethnicity, Country of Origin and Language of Patients." George Washington University, fielded fall 2007.

 Back to the Table of Contents

Report prepared by: Paul Jablow
Reviewed by: Karyn Feiden
Reviewed by: Marian Bass
Program Officer: Pamela S. Dickson (ID# 050183)
Program Officer: Debra J. Perez (ID# 059159)

Most Requested