Strategies to Reduce Health Disparities
Collecting & Using Data
Presenters:
Walter Philip Bailey, M.P.H., Chief, Office of Health and Demographics, South Carolina Budget and Control Board, Office of Research and Statistics, Columbia, SC
Richard T. Haverkate, M.P.H., Health Services Director, Health Services Division, Inter-Tribal Council of Michigan, Inc., Sault Ste. Marie, MI
|
Accurate, comprehensive data allow policymakers to better understand what is causing disparities within their jurisdictions. Data analysis identifies target areas and gives the baseline data against which interventions should be monitored to evaluate success.
Data from different State agencies can provide answers to many questions about disparities (in this case comparing black and white persons):
- Prevalence of disease. Do black Americans have a higher prevalence rate of diabetes than white people? Is the progression of diabetes at hospitalization different for black and white persons?
- Patient demographics and/or economics. Are patients with different educational levels treated differently by physicians? Do poorer Medicaid patients have higher hospitalization rates than those less poor? How does household composition affect hospitalization and office visit rates? Do certain geographic areas have consistently higher hospitalization rates? Could this be due to census undercount? How does Medicaid utilization compare with utilization in a commercial insurance plan (e.g., State employee plan)?
- Access issues. Do patients with a medical home have lower hospitalization rates than those without a medical home who use primarily the emergency room or outpatient clinics? Are patients compliant with prescription drug therapy? Are patients who live long distances from physicians hospitalized more than those patients who live closer to their physician? Do patients who receive transportation services have lower hospitalization rates than those who do not?
- Physician practice patterns. Are patient visits to physicians in accordance with standards of care prior to hospitalization? Do black physicians treat black patients differently than white physicians? Do some physicians have much higher hospitalization rates than other physicians? Are accepted clinical standards of patient care applied equally for black and white patients?
A policymaker's ability to answer these questions will be limited to the data available.
To increase the available data, and obtain a fuller picture of the needs and utilization patterns of its residents, South Carolina linked the data systems of multiple agencies and the private sector:
- The Department of Health and Human Services currently provides data from Medicaid claims, the Child Care Voucher system, and Community Long-term Care.
- The Department of Education currently provides demography data and exit exam data.
- The Department of Social Services currently provides data from the Temporary Assistance to Needy Families wage match and work support, food stamps, foster care tracking, and Child Protective Services.
- The Division of Developmental Disabilities and Special Needs provides client data on persons with disabilities and special needs.
- The Division of Vocational Rehabilitation provides client data on persons needing vocational rehabilitation.
- The Department of Mental Health provides client data on persons receiving inpatient and outpatient psychiatric services.
- The Division of Labor License Review provides data on licensed health professionals.
- The Department of Health and Environmental Control provides data from Vital Records, Emergency Medical Service Ambulance Run, BabyNet, and Children's Rehabilitative Services.
- The Department of Public Safety provides data on motor vehicle crashes.
- The Department of Juvenile Justice provides data on juvenile justice referrals.
- Private data systems provide data on inpatient hospitalizations, emergency room visits, hospitalizations from emergency rooms, outpatient surgery center visits, home health visits, and private psychiatric hospital visits.
One key in gaining cooperation among agencies to share data has been that the South Carolina Budget and Control Board, Office of Research and Statistics—the office spearheading this effort—has no control over other agencies' data systems. Because the Budget and Control Board has no programmatic function, no sense of competition exists between it and the other agencies. Through the Data Oversight Council, which includes the private entities and the State agencies, each participating entity has equal access to the data. However, no one can access any entity's data without its permission.
A unique person number replaces personal identifiers on the statistical file to protect confidentiality. The number is not a calculated field, so there is no way to "crack" the number and identify an individual.
The Inter-Tribal Council of Michigan, Inc., which serves seven tribes in Michigan, takes a different data-gathering and analysis approach for its Healthy Start program. The tribes face unique challenges in data analysis based on:
- Difficulty in identifying the population in State data (e.g., race miscoding, tribal affiliation).
- Comparatively small size, conventionally reported in the "Other" category.
For a small organization with few resources to spend on data gathering, several strategies will assist in getting needed information:
- Determining the questions to be answered and what types of data are needed to obtain these answers.
- Finding existing databases and information sources, as well as collecting primary data where none exists.
- Finding some help, particularly technical assistance and training on the use of software, data entry/retrieval, data interpretation, and presentation design.
- Making connections with State Vital Statistics staff, as they have access to very useful information and resources. It will take a great deal of effort at the start to locate the "right person," but the effort pays off. Face-to-face meetings will effectively build relationships.
The State's Vital Records and Vital Statistics matched files showed that infant mortality rates for AI/AN children in Michigan were approximately two times higher than for infants of all races and three times higher than for white infants in the State. The Healthy Start project, with the help of the State's Vital Records staff, compared the project area with national trends on several issues:
- Proportion of post-neonatal versus neonatal deaths (for example, Sudden Infant Death Syndrome accounts for half of all post-neonatal deaths).
- Role of low-birthweight (although an average of 5.8 percent of white low-birthweight infants die, an average of 14-16 percent of AI low-birthweight infants die).
- Prevalence of high birthweight (it is much greater among AI/AN in the project area than for rural Michigan residents of all races and is associated with health problems later in life).
- Cause of death, such as inadequate prenatal care (19 percent of mothers whose babies died had inadequate care versus 6 percent of mothers whose babies lived) and maternal smoking (42 percent of mothers whose babies died versus 36 percent of mothers whose babies lived).
Previous Section
Contents
Next Section ![Next Section](https://webarchive.library.unt.edu/eot2008/20081105005032im_/http://www.ahrq.gov/images/arrow_rt.gif)