Appropriate use of consultation can improve patient outcomes, but inappropriate use may cause harm. Factors affecting the variability of inpatient consultation are poorly understood. We aimed to describe physician-, patient-, and admission-level factors influencing the variability of inpatient consultations on general medicine services. We conducted a retrospective study of patients hospitalized from 2011 to 2016 and enrolled in the University of Chicago Hospitalist Project, which included 6,153 admissions of 4,772 patients under 69 attendings. Consultation use varied widely; a 5.7-fold difference existed between the lowest (mean, 0.613) and highest (mean, 3.47) quartiles of use (P <.01). In mixed-effect Poisson regression, consultations decreased over time, with 45% fewer consultations for admissions in 2015 than in 2011 (P <.01). Patients on nonteaching hospitalist teams received 9% more consultations than did those on teaching services (P =.02). Significant variability exists in inpatient consultation use. Further understanding may help to identify groups at high-risk for underuse/overuse and aid in the development of interventions to improve high-value care.

The COSCE was well received and effective for preparing entering interns with varying experience and training to request consultations. Future work will emphasize consultation communication specific to training environments and evaluate skills via direct observation of clinical performance.