Skip Navigation
Policy Innovation Profile

Medicaid Health Plan Increases Collection of Race, Ethnicity, and Language Data by Using Direct and Indirect Sources, Including Genealogy Analyses of U.S. Census Data


Tab for The Profile
Comments
(0)
   

Snapshot

Summary

As part of a quality improvement effort and in response to a Massachusetts Department of Public Health 2009 mandate that hospitals and health plans collect data from patients and members on race, ethnicity, and preferred language, Boston Medical Center HealthNet Plan (which serves Medicaid beneficiaries) combines direct and indirect information sources to collect and verify these data. Direct sources include gathering data from a health needs assessment form filled out by members at plan enrollment and from telephone calls (both incoming and outgoing) with members. Indirect sources include a State form that must be filled out when applying for Medicaid benefits and a genealogy analysis that predicts members' likely background based on what people with the same last name reported in the U.S. Census. To ensure the accuracy of the information gathered, the plan uses probability analyses to assess the validity of the information and discard data unlikely to be accurate. As a result of these efforts, the plan has collected the information for a high proportion of members, with direct sources initially enabling it to do so for 53 percent of members and the addition of indirect sources subsequently boosting that figure to 73 percent. In addition, plan leaders have a high level of confidence that the data are accurate.

Evidence Rating (What is this?)

Moderate: The evidence consists of trends in the proportion of plan members for whom state-mandated data on race, ethnicity, and preferred language have been collected.
begin do

Developing Organizations

Boston Medical Center HealthNet Plan
end do

Date First Implemented

2009
begin ppxml

Patient Population

The health plan serves primarily low-income individuals, those with disabilities, and other vulnerable populations. Approximately 15 percent of members are African American and roughly 25 percent are Hispanic.Insurance Status > Medicaid; Vulnerable Populations > Non-english speaking/limited english proficiency; Racial minoritiesend pp

Problem Addressed

Disparities related to race and ethnicity remain common in health care. To identify and address these disparities, health plans and provider organizations need accurate data on their patients' racial and ethnic background and preferred written and spoken language(s). However, collecting accurate information directly from members in a culturally sensitive manner can be challenging and costly.
  • Significant racial and ethnic disparities: Despite some recent progress, racial and ethnic disparities persist. For example, the 2012 National Healthcare Quality Report found that for approximately 40 measures of health care quality, African Americans receive worse care than whites, and Hispanics receive worse care than non-Hispanic whites; the same report found that American Indians and Alaska Natives receive worse care than whites for roughly one-third of the measures.1
  • Need for demographic data to identify and address problems: Before they can address these disparities, health care organizations (e.g., municipal health agencies, health plans, hospitals) must be able to evaluate the quality of care provided to various patient segments, including by race, ethnicity, and language spoken. Armed with such information, these organizations can identify significant problem areas, devote resources to interventions to address them, and evaluate the impact of these interventions to inform future strategies, policies, and resource allocation.
  • Many challenges to collecting information from traditional sources: Collecting the requisite information in a culturally sensitive manner from traditional data sources can be quite challenging. To make accurate judgments, organizations need information on a high proportion of members or patients; otherwise it may make biased assumptions based on a small or unrepresentative sample.2 Obstacles to collecting accurate information include confusion among patients and those who collect data about definitions of race and ethnicity; patient discomfort about providing information on race or ethnicity (because of concerns about how the information will be used); the high costs of data collection, especially if staff must call patients to obtain the data; and high turnover among the patient or member base.2

What They Did

Back to Top

Description of the Innovative Activity

As part of a quality improvement effort and in response to a Massachusetts Department of Public Health 2009 mandate that hospitals and health plans collect data from patients and members on race, ethnicity, and preferred language, Boston Medical Center HealthNet Plan (which serves Medicaid beneficiaries) combines direct and indirect information sources to collect and verify these data. Direct sources include gathering data from a health needs assessment form filled out by members at plan enrollment and from telephone calls (both incoming and outgoing) with members. Indirect sources include a State form that must be filled out when applying for Medicaid benefits and a genealogy analysis that predicts members' likely background based on what people with the same last name reported in the U.S. Census. To ensure the accuracy of the information gathered, the plan uses probability analyses to assess the validity of the information and discards data unlikely to be accurate. A detailed description follows:
  • Direct data-collection efforts: The health plan began its data-collection efforts in 2009 using several methods, including the following:
    • Health needs assessment form: During the enrollment process, members fill out a written health needs assessment that requests their race, ethnicity, and preferred written and spoken language.
    • Incoming calls to trained customer service representatives: Whenever a member calls the member services department for any reason, a trained customer service representative asks the member to provide the required demographic data if not already available in the member's records. (For preferred language, the representative asks if it is not available or has not been verified in the last year.) Because race and ethnicity can be confusing and sensitive issues, the representative explains why the plan is asking for the information and how it will be used, and answers any questions the member may have. (The Planning and Development Process section contains more information on the training given to the representatives.)
    • Outgoing calls by care managers: Care managers (primarily nurses) request relevant missing information whenever they call a member to follow up on the member's treatment or care plan. These care managers have received the same training given to the customer service representatives.
  • Indirect data-collection efforts: Some members neglect to fill out the background information section of the health needs assessment or provide invalid data on the form (e.g., entering their ethnicity in response to a question on race or vice versa). In addition, a significant percentage of members does not have telephone contact with customer service representatives or care managers. In response, the plan decided to add two indirect data sources for the requisite information, as outlined below:
    • Medicaid application form: Before enrolling in HealthNet, members must first qualify for Medicaid through the state's MassHealth program. To do so, they complete a benefits request form, which asks about race, ethnicity, and preferred spoken language. The health plan receives these forms when patients enroll, and began using the self-reported information in its efforts to increase the collection of background data in January 2013.
    • Genealogy data from census: A software program (obtained at no cost from the RAND Corporation) uses U.S. Census data to assess an individual's likely race and ethnicity based on their last name. For example, a high percentage of people with the last name Mueller are white and have German ethnicity.
  • Probability analyses to assess validity, discard questionable information: Given the high volume of race data collected directly from members, the health plan's clinical informatics department was able to assess the positive predictive value of the race data provided by the state and generated by the last name analysis. With the information provided by the state, for example, if a high proportion of members with a particular last name note their race on the state's Medicaid eligibility form as Asian and indicate they are Asian when asked by the plan, then it is highly likely that another individual with the same last name who says he is Asian on the eligibility form but has not yet been contacted directly by the plan is also Asian. As long as there is a high positive predictive value, the plan can use this type of information to supplement its racial data. Alternatively, if the predictive value of the State eligibility data for enrollees noting their race is not high when compared with the data already collected, the health plan does not use the State data to supplement its racial data. With the genealogy data, the race of individuals with some last names can accurately be predicted (e.g., someone named Jefferson is likely to be African American), but in other cases it cannot (e.g., someone with the last name of Smith or Jones could be from any of a number of races and ethnicities). The software tool enables users to set a desired threshold (i.e., a 90-percent probability of being accurate) and then automatically discard data that fail to meet this standard.

Context of the Innovation

A nonprofit managed care organization founded by Boston Medical Center in 1997, Boston Medical Center HealthNet Plan serves more than 275,000 members, making it the largest managed care organization serving Medicaid beneficiaries in Massachusetts. Two developments led to the program's inception. First, as part of its ongoing quality improvement efforts, the health plan sought to collect more comprehensive background data on its members, especially those with limited English proficiency and diverse cultural and ethnic backgrounds. Second, in 2009, the Massachusetts Department of Public Health mandated that hospitals and health plans collect basic background data that could be used to identify and subsequently address health disparities. The state's Health Care Quality and Cost Council established standards for data collection, and the Office of Health Equity coordinates data-collection activities. The health plan formed a Disparities Work Group made up of plan executives, clinicians, and representatives from the care management, customer service, information technology, clinical informatics, and quality improvement departments within the plan. This group took charge of planning and implementing the data-collection efforts needed to adhere to the mandate.

Did It Work?

Back to Top

Results

The health plan has collected the required information for nearly three-quarters of its members, and plan leaders have a high level of confidence that the data are accurate.
  • Requisite data collected for high proportion of members: Using direct data collection methods, the health plan was able to collect the required information for 53 percent of its members. After it started using the indirect sources, the figure jumped to 73 percent (as of January 1, 2013).
  • High confidence in accuracy: Plan leaders have a high level of confidence in the information, due in large part to the use of the probability analyses described earlier.

Evidence Rating (What is this?)

Moderate: The evidence consists of trends in the proportion of plan members for whom state-mandated data on race, ethnicity, and preferred language have been collected.

How They Did It

Back to Top

Planning and Development Process

Selected steps included the following:
  • Training customer service representatives and care managers: Participants from the workgroup attended a 2-day training session provided by the Robert Wood Johnson Foundation on how to collect data on race, ethnicity, and language. The plan used this information to train customer service representatives and care managers to collect the necessary background data in an efficient and culturally sensitive manner. Training included simulated telephone calls in which employees used scripts to help them ask and respond to questions.
  • Recognizing need for additional data sources: Although the initial efforts resulted in data being gathered for more than one-half of members, plan leaders noticed that by 2012 the ability to collect the requisite information based on the direct sources had plateaued. Workgroup members felt it was important to reach a higher proportion of members, since the plan still lacked information for 47 percent of members, and hence any strategies, policies, or decisions made based on this information could potentially be flawed. They were particularly concerned because members for whom the data were missing tended to be those who did not have much contact with the plan and hence may be less engaged in their health and health care. Consequently, the group began looking for other data sources.
  • Expanding data-collection efforts: In 2012, the workgroup identified the two additional data sources to be used. The plan already had access to the background data on the MassHealth benefits request form, but had not used it. The plan also had access to the RAND software that uses genealogy and U.S. Census data. The group decided to add these two data sources, believing they could provide additional, accurate data. As part of the implementation of these two data sources in January 2013, the workgroup began using probability analyses to confirm the accuracy of these new data sources.
  • Ongoing collection, verification, and data analysis: The plan continues to look for new methods of collecting and validating the required data. Now that information is available for nearly three-quarters of members, the plan has begun to analyze it to identify significant health disparities and inform efforts to reduce them.

    Resources Used and Skills Needed

    • Staffing: The program required no new staff. Members of the workgroup devote varying amounts of time to the project as part of their regular job responsibilities. In addition, approximately 50 customer service representatives and 35 care managers contribute to the data-collection efforts as part of their regular job responsibilities. It generally takes only 2 to 4 minutes to collect the requisite information over the telephone; in addition, information on race and ethnicity needs to be collected only once, whereas information on preferred language needs to be verified once a year.
    • Costs: The incremental costs of this program are minimal, consisting primarily of the upfront time spent in training the customer service representatives and care managers.
    begin fs

    Funding Sources

    Boston Medical Center HealthNet Plan
    end fs

    Tools and Other Resources

    Elliot MN, Fremont A, Morrison PA, et al. A new method for estimating race/ethnicity and associated disparities where administrative records lack self-reported race/ethnicity. Health Serv Res. 2008;43:1722-36. [PubMed]

    Adoption Considerations

    Back to Top

    Getting Started with This Innovation

    • Obtain support of senior leaders: An organization's data-collection efforts will likely not succeed unless high-level executives support the project. Executive-level support allows project leaders to devote the time and resources needed to overcome any obstacles that may emerge and lets employees responsible for gathering the required data know that their work is valued.
    • Emphasize training: Employees need training on how to collect data on race and ethnicity in a sensitive manner, including guidance on explaining how the information will be used to members who may be reluctant to provide it. Well-trained employees can allay any member concerns by explaining that the data will be kept confidential and used only to improve the quality of care.

      Sustaining This Innovation

      • Track collection rates: Project leaders should closely monitor data-collection rates by source to figure out which methods are most successful in collecting valid data from members. They can use this information to discard or de-emphasize ineffective or unreliable methods and to expand and enhance effective ones.
      • Share data with employees: Customer service representatives and care managers will remain engaged in data-collection efforts if they see that their work is leading to improvements in collection rates. For example, if they see that their work helped boost collection rates from 50 to 60 percent in the last quarter, they likely will continue their efforts with enthusiasm in the hopes of generating further increases.

      More Information

      Back to Top

      Contact the Innovator

      Ana Berridge, MHA
      Manager of Quality Improvement Operations
      Boston Medical Center HealthNet Plan
      Two Copley Place, Suite 600
      Boston, MA 02116
      (617) 748-6448
      E-mail: ana.berridge@bmchp.org

      Innovator Disclosures

      Ana Berridge reported having no financial interests or business/professional affiliations relevant to the work described in this profile.

      Footnotes

      1 U.S. Department of Health and Human Services. National healthcare quality report 2012. Rockville (MD): U.S. Department of Health and Human Services, Agency for Healthcare Research and Quality, 2012. AHRQ Publication No. 13-0002. Available at: http://www.ahrq.gov/research/findings/nhqrdr/nhqr12/2012nhqr.pdf (If you don't have the software to open this PDF, download free Adobe Acrobat ReaderĀ® software External Web Site Policy.).
      2 Ulmer C, McFadden B, Nerenz D, editors. Race, ethnicity, and language data: standardization for health care quality improvement. Washington (DC): Institute of Medicine of the National Academies, 2009. Available at: http://www.ahrq.gov/research/findings/final-reports/iomracereport/iomracereport.pdf.
      Comment on this Innovation

      Disclaimer: The inclusion of an innovation in the Innovations Exchange does not constitute or imply an endorsement by the U.S. Department of Health and Human Services, the Agency for Healthcare Research and Quality, or Westat of the innovation or of the submitter or developer of the innovation. Read more.

      Original publication: September 25, 2013.
      Original publication indicates the date the profile was first posted to the Innovations Exchange.

      Last updated: July 30, 2014.
      Last updated indicates the date the most recent changes to the profile were posted to the Innovations Exchange.