To assess novel metrics of e-consult appropriateness and utility.
Retrospective cohort study.
Primary and specialty care practices at 2 large academic and 2 community hospitals of an integrated health system.
Patients with e-consult requests to 5 specialties-hematology, infectious disease, dermatology, rheumatology, and psychiatry-between October 2017 and November 2018.
The appropriateness of e-consult inquiries was assessed by review of medical records and defined as meeting the following 4 criteria: not answerable by reviewing evidence-based summary sources (“point-of-care resource test”), not merely requesting logistic information, having appropriate clinical urgency, and having appropriate patient complexity. Interrater agreement in assessments of e-consult appropriateness was assessed by the κ statistic. Utility of e-consults was assessed by the rate of avoided visits (AVs), defined by the absence of an in-person visit to the same specialty within 120 days.
Overall, 6512 eligible e-consults were made by 1096 referring providers to 121 specialist consultants. Inquiries were characterized as diagnostic, therapeutic, for provider education, or at the request of the patient. Most consultations were answered within 1 day, with variation across specialties (73.1% for psychiatry to 87.8% for infectious disease). Overall, 70.2% of e-consults met all 4 criteria for appropriateness; the frequency of unmet criteria varied among specialties. Raters agreed on the appropriateness of 94% of e-consults (κ = 0.57 [95% CI, 0.36 to 0.79]), indicating moderate agreement. The overall rate of AVs across the 5 specialties was 81.2%; the highest rate was in psychiatry (92.6%) and the lowest in dermatology (61.9%).
Generalizability is unknown outside a single integrated health system, where requesting and consulting providers share a common electronic health record.
Novel metrics to assess the appropriateness and utility of e-consults provide meaningful insight into practice, provide a rubric for comparison in future studies in additional settings, and suggest areas to improve resource use and patient care.