The Patient Protection and Affordable Care Act (Affordable Care Act), signed into law in March 2010, is expected to have extensive and wide-ranging impacts on health insurance coverage and access to health care in the United States, especially for the low-income uninsured. Beginning in 2014, people with incomes at or below 138% of federal poverty level (FPL) will be eligible for Medicaid; although the impacts will vary by state,this represents a significant expansion of the Medicaid program.1 Other major changes in the Affordable Care Act, such as the individual responsibility requirement, will also have impacts on the Medicaid-eligible population (as well as the population more broadly). In addition, the Affordable Care Act provides funding to expand the availability of services through community health centers (CHCs) and support for a range of initiatives aimed at ensuring sufficient access to care for Medicaid enrollees.2
The purpose of this report is to describe and assess existing data sources that could be used to evaluate the impacts of the Affordable Care Act on the Medicaid-eligible population. This data scan is one component of a larger project to design an evaluation of the Affordable Care Act's impacts on this population.
The data scan and data needs assessment component of this project has four primary goals:
Identifying existing data sources that potentially can be used to evaluate the impacts of the Affordable Care Act on Medicaid-eligible populations.
Assessing the strengths and weaknesses of existing data sources with regard to the specific research questions that will be the focus of the evaluation.
Assessing the strengths and weaknesses of existing data sources with regard to the specific populations and provider groups of interest for the evaluation.
Identifying gaps in existing data available for the evaluation.
The data scan builds on previous work to assess how existing data sources might be used to monitor and understand the impacts of the Affordable Care Act over time.3
For this project, it is particularly important to identify data sources that can support state-level estimates in addition to national estimates, so that state-specific factors can be taken into account in evaluating the Affordable Care Act's impacts. For example, it will be important to understand the impact of the different decisions that states will make about Medicaid benefit levels, outreach strategies, provider payment rates, and other issues that could influence the Affordable Care Act's impacts within a state. In addition, there is substantial variation across states in current eligibility for public coverage and in how health care is delivered to low-income populations, and this variation will need to be taken into account in an evaluation of the Affordable Care Act's impacts.
Consistent with the focus of the evaluation design, the data scan focuses primarily on data sources that support the evaluation of impacts on individuals. Provider-related data sources are included as well, with a focus on ways that these data sources can be used to assess the accessibility of services for Medicaid-eligible populations.
Table 1 provides a list of the data sources that were included in the data scan.4 The table includes eight national population surveys, seven provider surveys, and four sources of administrative data from providers that are collected on a regular basis and that are expected to continue to be available in the future. For each source of data, we compiled technical information, including how the data are collected and from whom; how complete or representative the data are for the populations of interest; and what level of geography is available for analysis. We reviewed the data collection instruments (e.g., survey questionnaires), technical documentation for the data sources, and publicly available reports that use the data. We also consulted with data users, the agencies that collect the data, or both as needed to understand the strengths and weaknesses of the data.
Several data sources were considered but not included in the list of data sources in Table 1. Among the most important of these were Medicaid claims data, multipayer claims databases, and Medicaid administrative data (such as enrollment data). A key reason that these data sources were excluded is that, by definition, the Medicaid databases do not include the population that will be made newly eligible for Medicaid in 2014 under the Affordable Care Act and currently eligible individuals who have not enrolled in Medicaid; similarly, multipayerclaims databases are also likely of limited usefulness for this purpose (because most of the population that is expected to gain Medicaid coverage as a result of the Affordable Care Act lacks health insurance, their use of health care would not show up in a claims database). As a result, these sources cannot be used to understand what happened to the populations of interest as a result of the Medicaid eligibility expansion. In addition, analysis of claims data is extremely resource intensive and was determined to be outside of the scope of the evaluation design.
Because of the importance of timely data for evaluating the impacts of the Affordable Care Act's Medicaid expansions, Table 1 also excludes population surveys that are conducted less than annually. In addition, a number of states conduct regular health insurance surveys that could potentially be used to assess the impacts of Affordable Care Act on Medicaid-eligible populations within a given state; however, many of the state surveys are not conducted annually. Furthermore, a key focus of this evaluation is to determine how variations across states in pre and post-2014 policy decisions influence the Affordable Care Act's impacts. Although these state-specific surveys can have tremendous value in understanding within-state impacts, they are generally not useful for cross-state comparisons.
Finally, in addition to an assessment of the existing data available to assess Affordable Care Act's impacts on Medicaid-eligible populations and health care providers generally, the data scan focuses on specific population groups and provider types of interest. These groups were identified by the Office of the Assistant Secretary for Planning and Evaluation (ASPE) as being of particular interest for the evaluation design. For individuals, the specific population groups of interest include people in extreme poverty (defined as having family income at or below 50% of FPL) and people with substance use disorders or mental illness.5 With regard to health care providers, the analysis focuses on emergency departments, CHCs, public hospitals, and other providers that serve the targeted vulnerable populations (e.g., community mental health centers).
|TABLE 1. Data Sources Included in Review|
|Population Surveys||Provider Surveys|
|American Community Survey (ACS)
Current Population Survey (CPS)
National Health Interview Survey (NHIS)
Medical Expenditure Panel Survey--Household Component (MEPS-HC)
Survey of Income and Program Participation (SIPP)
Behavioral Risk Factor Surveillance System (BRFSS)
National Health and Nutrition Examination Survey (NHANES)
National Survey on Drug Use and Health (NSDUH)
National Hospital Ambulatory Medical Care Survey (NHAMCS)
National Hospital Discharge Survey (NHDS)
American Hospital Association (AHA) annual survey
National Association of Public Hospitals and Healthcare Systems (NAPH) survey
Substance Abuse Treatment Providers
|Healthcare Cost and Utilization Project (HCUP)
Medicare cost reports
Uniform Data System (UDS)
Treatment Episode Data Set (TEDS)