INTRODUCTIONWith the expansion of coverage and preventive care facilitated by the Affordable Care Act, primary care facilities are facing increased demand for services. At the same time, new payment incentives linked to patient-centered medical homes (PCMHs) are encouraging the adoption of team-based care (Auerbach et al., 2013; Bitton et al., 2012). Team-based care may help contain costs and improve productivity and patient outcomes primary care facilities (Altschuler, Margolius, Bodenheimer, & Grumbach, 2012; Bodenheimer & Smith, 2013; Ku, Frogner, Steinmetz, & Pittman, 2015). However, most studies of primary care teams examine team-based care or teamwork as a variable associated with certain outcomes, with little attention given to the question of how leaders of primary care facilities determine the composition of primary care teams and the configurations that best serve patient populations given their local policy and market contexts (Chen & Bodenheimer, 2011; Chesluk & Holmboe, 2010).Understanding how leaders make configuration decisions the context of the current system transformations is important because it may help build knowledge about how ground-level choices primary care facilities mediate policy outcomes. In addition, such information may help inform national and state health workforce planners as they grapple with the challenge of projecting demand for different types of healthcare workers, some of whom have overlapping roles. Such projections have traditionally measured supply and demand for just one profession, but interest is increasing more sophisticated models that consider alternative team configurations (Auerbach et al, 2013). Furthermore, organizations such as the Association of American Medical Colleges are particularly interested identifying variables that can be used to develop local workforce planning tools (Dill, 2015).Researchers have argued for a complexity perspective on primary care, with greater attention given to how leaders manage competing goals and values, leam from their environment, and create emergent organizational forms (Felix-Bortolotti, 2009; Sweeney, 2006). In this study, we use a qualitative approach to explore how leaders of community health centers (CHCs) make complex medical staff configuration choices their role as safety-net providers for 22 million low-income patients the United States (Health Resources and Services Administration, n.d.). CHCs are a good venue for studying complexity primary care because they require leaders to use tightly constrained resources to provide high-quality primary care to patients, regardless of their ability to pay (Rosenblatt, Andrilla, Curtin, & Hart, 2006). CHCs are seen as in the vanguard of flexible staffing because their leaders make tradeoffs and adapt to the environment to maximize resources a difficult funding environment (Ku et al, 2015).METHODSParticipantsWe used the 2012 Uniform Data System (UDS), an annual administrative reporting system for CHCs that receive federal funding (Section 330 grantees), to identify a maximum variety sample (Patton, 1990) of CHCs with unusually high proportions of advanced practice providers (APPs) (i.e., nurse practitioners [NPs], physician assistants [PAs], and certified nurse midwives), nurses (registered nurses [RNs] and licensed practical nurses [LPNs] or vocational nurses), medical assistants (MAs), case managers, and community health workers. We selected three to five sites from each of these ftve categories for variety with regard to urban versus rural location, NP scope of practice (SOP), Medicaid expansion states versus Medicaid non-expansion states, and large versus small. We determined size (the last variable) on the basis of whether they reported more or less than 9,844, the median number of patients. This maximum variety sampling approach allowed us to analyze a wide range of ideas pertaining to (Patton, 1990). …