Article Title

Tracking Health Care Team Response to Electronic Health Record Asynchronous Alerts: Role of In-Basket Message Burden

Publication Date



electronic health records, health care communication


Background/Aims: Electronic health record (EHR)-based messaging allows communication of important patient information to primary care providers, but high message burdens may limit effectiveness and safety of this communication mode.

Methods: We analyzed EHR data for 1,282 discharges of elderly patients (age ≥ 65) from a Massachusetts hospital in 2010–2011. For 799 of these discharges, time-sensitive automated EHR messages were generated on Day 3 postdischarge. These messages highlighted actionable concerns associated with medication changes and were sent to secure EHR In-Baskets belonging to the patients’ primary care physicians. We tracked timing of message reviews and calculated associations between primary care physician In-Basket message burden (i.e. number of messages in In-Basket at time of message delivery) and time to message opening. We also examined whether messages were opened by the intended physician, by a covering physician or by staff members.

Results: Of the 799 messages, 376 (47.1%) were opened by the intended primary care physician within one business day, with an additional 2.0% (16) opened by covering physicians. In some cases, staff members opened messages and generated follow-up communication. These follow-up communications were read by the intended provider (2.8%, n = 22 messages) or a covering provider (0.9%, n = 7) within 24 hours of delivery. Overall, 41.1% of these time-sensitive messages were not opened by anyone within 24 hours; an additional 6.3% were opened by staff within the first 24 hours with no evidence that communications were passed along within the EHR. Rates of opening within 24 hours were inversely related to size of In-Basket at the time of message delivery, with 61.8% of messages opened within 24 hours among providers whose total number of In-Basket messages were in the bottom quartile (157 messages) (P < 0.0001).

Conclusion: Rates of opening of time-sensitive In-Basket messages were suboptimal; less than two-thirds of messages were opened within 24 hours. In-Basket message burden may impact physicians’ ability to devote timely attention to patient-related messages. Reducing In-Basket volume should be explored as a means of improving quality and safety of patient care.




June 30th, 2016


August 12th, 2016