Social Networks in Medical Homes and Impact on Patient Care and Outcomes

The Patient Centered Medical Home (PCMH) model aims to address the major challenges facing primary care including poor access, quality of care, and rising costs, particularly for chronic diseases. One key element in this model is the design of high-functioning teams which expand primary care capacity and improve care and outcomes. Yet little is known about how to develop best practices for building PCMH teams. In addition, we do not know how PCMH teams operate in practice, share information, support, or advice to deliver care or how the resulting social structures (called social networks) affect quality of care indicators and patient outcomes. This study will fill the critical gap in evidence. We will combine fine-grained analysis of team configurations and social networks in PCMHs with empirical assessment of quality of care and patient outcomes to determine team best practices. The specific aims are: Aim 1. Analyze team configurations and social networks (i.e., communication, advice, trust, support, and problem-solving) in PCMHs from the perspectives of all team members (e.g., physicians, nurse practitioners, nurses, social workers, administrators, etc.). Aim 2. Examine the factors (e.g., profession, experience, education, etc.) determining team configurations and the structures and properties of social networks. Aim 3. Investigate how team configurations and social networks impact quality of care indicators and outcomes in patients with chronic diseases (i.e., diabetes, asthma, hypertension, cardiovascular diseases (CVD), and chronic obstructive pulmonary disease (COPD)). Primary care practices (n=14) designated as PCMHs at NewYork-Presbyterian Hospital (NYP)/Columbia University and Weill Cornell Medical Centers will participate in the study. We will survey all team members including primary care providers and staff (n=1,172) in these PCMHs through online surveys. Participants will be asked to select their team members from the clinics roster and indicate whether they communicate, share advice and/or support and who they trust or approach for problem solving. We expect at least an 80% response rate (n=937). We will also obtain patient data from NYPs Clinical Data Warehouse. Quality of care indicators and outcome measures for diabetes, asthma, hypertension, cardiovascular disease, and chronic obstructive pulmonary disease in adult patients will be obtained and linked to survey data. ORA* and R software will be used for data analysis. We will map team configuration and each social network, visualize them, and compute network metrics at individual and team levels. We will then build Exponential Random Graph and multilevel models to analyze the factors that may explain the observed networks and predict the impact of individual and team network variables on the quality of care and patient outcome measures controlled for patient, provider, and practice factors. The study has the potential to show how to facilitate teamwork in PCMHs and identify the most effective team structures to assure best quality of care and patient outcomes. This application is in response to the Special Emphasis Notice (SEN) NOT-HS-16-011 on AHRQs interest in applications related to innovative primary care research.