Our review of data sources for impacts on providers included sources for hospitals, physicians, CHCs, and substance abuse treatment facilities. Data sources reviewed for hospitals include the NHAMCS, the AHA annual survey, NHDS, data on inpatient and emergency room use from the HCUP databases, Medicare cost reports submitted annually to the Centers for Medicare and Medicaid Services (CMS), and the NAPH annual member survey. For physicians and clinics, the data sources included the NAMCS and the AMA Physician Masterfile. CHC data sources included NAMCS and the UDS reports submitted to the Health Resources and Services Administration (HRSA) each year by CHCs receiving federal funding. Finally, data sources reviewed for substance abuse treatment facilities included the TEDS and the N-SSATS. Detailed descriptions of each data source are included in the appendix to this report.
Table 12a and Table 12b summarize the availability of the indicators of interest for provider impacts from each data source. For purposes of this project, the NHAMCS data on indicators of interest is limited to volume of substance abuse and mental health treatment services and emergency department visits. The AHA data is more complete, with information about payer mix, financial margins, and both capacity and utilization of substance abuse/mental health treatment services and emergency departments. The NHDS and HCUP data are fairly limited, with volume of substance abuse/mental health treatment services (and emergency department visits in HCUP) and charges by payer. Data of interest from hospital Medicare cost reports are limited to patient mix by payer and financial margins. Finally, the NAPH survey data include information on payer mix, financial margins, and emergency department volume. None of the data sources includes information on measures of willingness to accept Medicaid patients.
The unit of analysis is an important issue to consider in evaluating the usefulness of NHAMCS and NHDS data for purposes of evaluating the Affordable Care Act's impacts on hospitals. These surveys are designed to produce a representative sample of patient visits, not of individual providers. As such, the number of hospitals that are included in the survey is small (365 and 205 hospitals in 2009 for NHAMCS and NHDS, respectively). If the desired unit of analysis is the hospital, these surveys may not be a good choice of data source. However, they may be very useful for other purposes--for example, to examine how hospital utilization patterns are changing.
For physicians, NAMCS includes an indicator of willingness to accept new Medicaid patients compared with patients from other payer sources, as well as information on payer mix. Similar to the NHAMCS and NHDS, the intended unit of analysis in NAMCS is the patient visit, and the data are of limited usefulness for provider-level analysis; however, the NAMCS is larger (1,293 physicians participated in 2009) and the supplemental mail survey is designed to produce state-level estimates on limited topics. The AMA data do not include any of the indicators of interest for this project.
With regard to CHCs, the NAMCS data include some indication of willingness to accept new Medicaid patients, but the number of CHCs included in the NAMCS sample is small (104 CHCs in 2009, with up to three providers per CHC). The UDS data include information on payer mix and financial margins, as well as staffing to provide substance abuse and mental health treatment services, social support services, and outreach services.
For substance abuse services, the TEDS data provide some information on volume of substance abuse services, although reporting is limited to providers that receive federal and state funding. N-SSATS includes information on payment types accepted, but no specific question about willingness to accept new patients by payer type; it also includes information about substance abuse treatment capacity and utilization.