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

The Patient-Centered Medical Home (PCMH) model aims to address primary care challenges such as poor access and quality and rising costs by delivering team-based care, particularly for chronic diseases. Yet little is known about the composition of effective teams to achieve best patient outcomes. Furthermore, how team members communicate, share advice, or help to deliver care or how the resulting social structures (i.e., social networks) affect quality and outcomes has not been studied. Our innovative mixed-methods study will fill this gap. We will combine analysis of team configurations and social networks in PCMH practices with assessment of quality of care and patient outcomes to identify team best practices. We will also collect qualitative data to assess the underlying teamwork dynamics not captured quantitatively. The specific aims are to: 1. Identify factors determining team configurations and the formation of social networks in primary care practices adopting the PCMH model. 2. Investigate how team configurations and social networks impact quality of care and patient outcomes for five chronic conditions. 3. Identify barriers, facilitators, and characteristics of teamwork for teams with different configurations, social networks, and performance. Practices (n=24) with the PCMH model at NewYork-Presbyterian Hospital/Columbia University and Weill Cornell Medical Centers and the University of Pittsburgh Medical Center will participate. We will recruit team members including clinicians and staff (n=1,437) through online surveys who will identify their team members from the clinic's roster and report with who they communicate, share advice and/or support, and trust or approach to solve problems. We expect an 80% response rate (n=1,150). We will obtain patient data on quality of care and outcomes for diabetes, asthma, hypertension, cardiovascular disease, and chronic obstructive pulmonary disease and merge it with survey data. ORA* and R software will be used for data analysis. We will map team configuration and social networks, visualize them, and compute network metrics. We will then build Exponential Random Graph Models to predict factors explaining the observed networks and multilevel models to assess the impact of network variables on quality of care and patient outcomes. Based on Aim 1 findings, we will recruit participants who are highly- (n=~20) and poorly- (n=~20) connected to their team members in social networks. Based on Aim 2 findings, we will recruit participants from high- (n=~20) and low- (n=~20) performing teams. We will conduct individual face- to-face interviews with them using an interview guide. Interviews will be audio-taped and transcribed, and data will undergo content analysis. Multiple researchers will code the data and identify themes. Quantitative and qualitative findings will be triangulated. This study has the potential to show how to facilitate teamwork and identify the most effective team attributes to assure best quality of care and outcomes particularly for patients with chronic diseases (AHRQ's priority population). This application is in response to the Special Emphasis Notice (SEN) NOT-HS-16-011 on AHRQ's interest in applications related to innovative primary care research.