Journal of Healthcare Management - September/October 2016 - 368

J o u r n al of H ealt H care M anageMent 61:5 S epte Mber /o ctober 2016 substituted another comparable site from the same outlier site list. Two qualitative researchers (P.P., L.M.) conducted semistructured telephone interviews with CEOs or other leaders designated by the CEO (collectively referred to as CEOs) at each CHC. Open-ended interview questions addressed six topics: 1. Current staff configuration 2. CEOs' assessments of the factors that influence staff configuration 3. Ideal staff configuration 4. Barriers to achieving ideal staff configuration 5. CEOs' assessments of teamwork among current staff 6. CEOs' advice for other CHCs about staff configuration We asked questions without suggesting specific factors to allow participants to share their thoughts based on their own experiences. With participants' permission, we recorded and transcribed the interviews for analysis. Data Analysis We used ATLAS.ti, a qualitative data analysis software package, for coding and iterative thematic analysis. We conducted initial coding using a codebook developed from the interview guide and added emergent codes as they appeared in the data. We used the "code family" feature in ATLAS.ti to collapse codes into broader themes and used memos to compare emerging themes across sites. We used a comparative grid to identify key differences in staffing decisions and to relate them to CHCs' structural and policy contexts. Two researchers (P.P., L.M.) and a research assistant (L.B.) conducted the coding. The George Washington University Office of Human Research reviewed and approved the study design. R E S U LT S CEOs in our sample described medical staffing configuration decisions in two distinct domains: provider choices and clinical support choices. While the relative emphasis on variables in each domain differed, we found remarkable continuity in how CEOs discussed their goals and choices across a wide variety of contexts. CEOs explained their decisions in terms of contextual factors over which they perceived to have little control and their perceptions of the value of different types of medical staff (Table 2). Contextual factors included the institution's history, local labor supply, labor costs, SOP regulations, and payment arrangements. CEOs' preferences regarding the appropriate distribution and roles of staff reflected their perceptions of the different professions' degree of clinical experience, personality attributes, and likelihood of retention. Decisions About Physicians Versus APPs CEOs in our sample consistently described their goals with respect to hiring providers as seeking a balance of physicians and APPs. No respondent described NPs and PAs as "second best," to be hired only when a CHC is unable to attract physicians. They viewed APPs as knowledgeable, well accepted by patients, and having specific attributes 368

Table of Contents for the Digital Edition of Journal of Healthcare Management - September/October 2016