Winning Abstracts from the 2014 Medical Student Abstract Competition: The Clock is Ticking: Improving general medicine discharge communication timeliness
First Author: Katherine Y. Wu, Duke University School of Medicine, Class of 2014
Timely discharge (DC) communication is critical for high-quality care transitions from inpatient to outpatient settings. DC communication usually relies upon discharge summaries. Inadequate DC communication is associated with negative patient outcomes. The purpose of this quality improvement (QI) project was to improve the quality of DC communication via improved timeliness of DC summary completion and transmission from general medicine (GM) inpatient to outpatient providers.
DC summaries from Duke University Hospital GM inpatient services were typically dictated by a resident. An attending physician signed the transcribed DC summary, and the document was automatically faxed to outpatient providers listed by the dictating physician as needing to receive a copy. DC communication timeliness was measured by time between DC date and 1) date of dictating provider signature (DC summary "completion") and 2) date of attending provider signature (DC summary "transmission"). These timeliness metrics and date of first post-DC follow-up (F/U) appointment were abstracted in a pre-intervention manual chart review of GM inpatients discharged to home in January 2012.
From April-May 2013, a multi-modal intervention was implemented consisting of: 1) physician awareness campaign to encourage timely DC communication, 2) establishment of 24 hours post-DC as the benchmark for DC summary completion, and 3) real-time performance feedback. Weekly during this intervention period, timeliness metrics were manually abstracted from patient charts. We determined the proportion of DC summaries completed within 24 of DC, transmitted within 48 hours of DC, and transmitted before the first F/U appointment. All patients discharged to home from resident-led GM teams were included. Intervention-period performance data were stratified by team and provider and fed back in real-time to providers. We analyzed pre-intervention and intervention-period data with an unpaired t-test and Fisher's Exact Probability Test.
58 pre-intervention patients and 277 intervention-period patients were included. Pre-intervention median time to DC summary completion was 1.72 days (95% CI: 1.00-2.18) and to transmission to F/U providers was 2.62 days (95% CI: 1.94-3.86). 36% of summaries were completed within 24 hours, 40% were transmitted within 48 hours, and 45% were transmitted before the first F/U appointment. Intervention-period median time to DC summary completion was 1.14 days (95% CI: 0.97-1.49; p=0.01) and to transmission to F/U providers was 2.84 days (95% CI: 2.27-3.05; p=0.73). 45% of summaries were completed within 24 hours (p=0.24), 37% were transmitted within 48 hours (p=0.77), and 45% were transmitted before the first F/U appointment (p=1.00).
This multi-modal QI project achieved a significant decrease in time to DC summary completion by housestaff. This modest improvement in DC communication timeliness was achieved with low-intensity interventions. Improvements could be augmented and sustained with implementation of an automated system for provider-level, real-time feedback.