Utilizing EHR Big Data and Machine Learning to Define and Measure the Impact of Primary Care Teams on Patient Activation

Background: Team-based care has been linked to key outcomes associated with the Quadruple Aim and a key driver of value-based care. With a perfect storm of events taxing our U.S. health system, it is essential to capture and analyze data about teams in real time. Use of the Electronic Health Record (EHR) and machine learning have significant potential to overcome previous barriers including lack of relevance to daily practice, delays in accessing data to improve teamwork, and invisibility of the contributions of team members not included in billable encounters. Methods: This study drew on a large EHR dataset (n = 316,542) from an urban health system to examine the relationship between primary care team composition and patient activation. Team composition was operationalized using consensus definitions of teamwork from the social science and interprofessional healthcare literatures. Patient Activation was measured using the Patient Activation Measure (PAM), a measure with strong reliability and validity across adult populations and ethnicities. Results from multilevel regression analyses were compared to machine learning analyses using multinomial logistic regression to calculate propensity scores for the effect of team composition on PAM scores. Under the machine learning approach, a causal inference model with generalized overlap weighting was used to calculate the average treatment effect (ATE) of team composition. Results/Discussion: 17 different team compositions were observed in the data from the analyzed sample (n=12,448). Team sizes ranged from 2-5 members. After controlling for confounding variables in both analyses, we found more diverse, multidisciplinary teams (team size 4 and greater) to be associated with improved patient activation scores. This is the first study to examine the relationship between team-based care and patient activation using big data analytics. Implications for further research using EHR data and machine learning to study teams and other patient related outcomes are feasible and promising.