Effects of Percutaneous Coronary Intervention Lab Openings and Closures on Patients, Hospitals, and Communities

Access to cardiac technology such as percutaneous coronary intervention (PCI) labs is critical to improving population health. Yet we and others have documented that the proliferation of PCI labs tends to concentrate among affluent communities, and untimely access tends to concentrate among low-income communities. Such differential diffusion and closure patterns across communities may widen gaps in disparities between vulnerable communities and other communities. Moreover, proliferation of PCI labs might lead to fewer procedures performed at each site, especially in markets that already have high PCI capacity at baseline, which could lead to worse outcomes.
Our primary goal is to identify how openings and closures of PCI labs affect resource allocation and patient health in saturated and low-capacity markets, and whether these market activities disproportionately affect vulnerable populations. Our central hypothesis is that uneven proliferation and closures of PCI facilities affects population health through the following pathways: (1) at the community level, they cause widening gaps in resource allocation between communities that already have high PCI capacity and those that still have unmet need; (2) at the hospital level, they change PCI volume in such a way that could result in worse health outcomes for patients with cardiac conditions; and (3) at the individual patient level, they redistribute patients and widen gaps in access, treatment, and health outcomes between vulnerable and other patients. Aims 1 and 2 address the average effect of PCI lab access change among patients with different conditions (AMI and stable CAD), while Aim 3 addresses the differential effect between vulnerable and other groups.
We will link several major databases to address our research questions. First, the primary data with which we will identify the patient cohort will be the 100% Medicare Provider and Analysis Review and Part B claims between 2005 and 2014. Second, we will identify PCI lab availability using the primary data source of the American Hospital Association (AHA) annual surveys for the same years. Third, we have obtained primary data to identify additional hospital characteristics from the AHA surveys and the Healthcare Cost Report
Information System. Fourth, we will use 2010 U.S. Census and American Community Surveys to identify each ZIP code community’s demographic composition, and the Area Resource File for capturing additional health care market resources at the county level. Last, we will derive two databases that identify (1) baseline PCI capacity for each community and (2) the nearest PCI lab for each patient’s ZIP code and corresponding driving time. By providing direct evidence of how changes in PCI access affect health outcomes of patients with AMI and stable CAD in different communities, this research will inform policymakers about the implications of PCI proliferation and closures at community, hospital, and individual levels.