TABLE 2–7, Overview of Three Approaches Used in Studying Disparities in Health Care - Guidance for the National Healthcare Disparities Report

Advantages	Disadvantages
Surveys such as the Medical Expenditure Panel Survey (MEPS) and the Medicare Current Beneficiary Survey (MCBS)
Excellent source for data on race, ethnicity, and socioeconomic status. Information nationally representative. Contains measures of potential access such as insurance coverage and usual source of care. Contains information about services not covered by insurance. Contains information about health status and outcomes. Contains information about different access and satisfaction variables. Contains information about out-of-pocket costs.	Limited source of data about health care services that have relatively low rates of use. Reporting errors and non-responses may bias information. Certain racial and ethnic groups may have small cell sizes.
Administrative databases drawn from the Medicare and Medicaid programs, the Veterans Administration (VA), and hospital discharge records
Excellent source for data about health care services that have relatively low rates of use. Personal identifiers may be available to permit linkages with other data sources such as Surveillance, Epidemiology, and End Results (SEER) files. Personal identifiers permit linkages across different types of services and over time.	Information about services not covered and populations not available. Information about utilization of services for enrollees in managed care plans may be unavailable. Limited data on race, ethnicity, and socioeconomic status.
Information on residence permits linkages with U.S. census data.	Missing claims data or coding changes and errors create inaccuracies. Limited clinical/patient information to assess need for services. Medicaid programs differ across states; eligibility may be terminated. Hospital discharge databases may not have personal identifiers for data linkages.
Clinical data such as medical records and disease registries
Excellent source for patient studies to analyze utilization, process, and outcomes for patients with similar needs. SEER linked to Medicare administrative data. Excellent source of access, utilization, and outcomes data for patients diagnosed with cancer: contains date of diagnosis, site of cancer and stage of cancer at time of diagnosis. Fairly nationally representative.	Cell sizes may be too small for various analyses. Certain racial and ethnic groups may have cell sizes too small for analyses. Information may not be nationally representative. Represents patient population; therefore, is not necessarily representative of population at risk of needing services. SEER/Medicare data source limited to persons enrolled in Medicare.

Advantages

Disadvantages

Surveys such as the Medical Expenditure Panel Survey (MEPS) and the Medicare Current Beneficiary Survey (MCBS)

Excellent source for data on race, ethnicity, and socioeconomic status.
Information nationally representative.
Contains measures of potential access such as insurance coverage and usual source of care.
Contains information about services not covered by insurance.
Contains information about health status and outcomes.
Contains information about different access and satisfaction variables.
Contains information about out-of-pocket costs.

Limited source of data about health care services that have relatively low rates of use.
Reporting errors and non-responses may bias information.
Certain racial and ethnic groups may have small cell sizes.

Administrative databases drawn from the Medicare and Medicaid programs, the Veterans Administration (VA), and hospital discharge records

Excellent source for data about health care services that have relatively low rates of use.
Personal identifiers may be available to permit linkages with other data sources such as Surveillance, Epidemiology, and End Results (SEER) files.
Personal identifiers permit linkages across different types of services and over time.

Information about services not covered and populations not available.
Information about utilization of services for enrollees in managed care plans may be unavailable.
Limited data on race, ethnicity, and socioeconomic status.

Information on residence permits linkages with U.S. census data.

Missing claims data or coding changes and errors create inaccuracies.

Limited clinical/patient information to assess need for services.
Medicaid programs differ across states; eligibility may be terminated.
Hospital discharge databases may not have personal identifiers for data linkages.

Clinical data such as medical records and disease registries

Excellent source for patient studies to analyze utilization, process, and outcomes for patients with similar needs.
SEER linked to Medicare administrative data. Excellent source of access, utilization, and outcomes data for patients diagnosed with cancer: contains date of diagnosis, site of cancer and stage of cancer at time of diagnosis. Fairly nationally representative.

Cell sizes may be too small for various analyses.
Certain racial and ethnic groups may have cell sizes too small for analyses.
Information may not be nationally representative.
Represents patient population; therefore, is not necessarily representative of population at risk of needing services.
SEER/Medicare data source limited to persons enrolled in Medicare.

From: 2, MEASURING THE EFFECTS OF SOCIOECONOMIC STATUS ON HEALTH CARE

Guidance for the National Healthcare Disparities Report.

Institute of Medicine (US) Committee on Guidance for Designing a National Healthcare Disparities Report; Swift EK, editor.

Washington (DC): National Academies Press (US); 2002.

NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.

TABLE 2–7Overview of Three Approaches Used in Studying Disparities in Health Care