Background:
Sign‐out, the transfer of patient information and responsibility to a cross‐coverage physician, is a necessary component of caring for hospitalized patients. The number of sign‐outs in training hospitals has increased with the Accreditation Council for Graduate Medical Education's duty hours limitations. Residents commonly use written or computerized mechanisms to track patient information, and little is known about the accuracy of these written signouts.
Method:
We collected all written signouts of internal medicine residents at UCSF Medical Center for three consecutive days. The residents maintain all sign‐out information using a computerized template. Residents were trained on optimal signout procedures both at orientation, five months prior to data collection, and one week prior to our study. Each patient's name, medical record number, location, room, age, attending physician, language of origin, chief complaint, allergies, code status, and medications were assessed for accuracy by comparing the written signout to the patient's medical record (which was paper‐based at the time). ICU patients and those discharged on the day of signout collection were excluded.
Summary of Results:
All of the 25 residents eligible for the study participated. A total of 173 written signouts for 72 patients were collected. Ten signouts were excluded due to an inability to obtain complete data. Demographic information, such as age and chief complaint, was 100% accurate. Attending physician, patient name, medical record number, code status and allergies were 98%, 98%, 98% 96% and 94% accurate, respectively. The least accurate information were the patient's location, room number, and language of origin (83%, 81% and 79% respectively). Only 79% (range 0‐100%) of scheduled medications and 46% (range 0‐100%) of prn medications were recorded on the signout. Of these, 98% of scheduled medications, and 99% of prn medications were accurate. Eight percent (83/992) of scheduled medications and 8% (11/136) of prn medications were present on the signout, but absent from the medical record.
Statement of Conclusions:
Our study found that most demographic information in residents' written signouts is highly accurate, but information that is likely to change during a patients' hospital stay (such as room number and location) is frequently incorrect. Medications are often omitted from signouts and, when present, may not reflect the official medication list. Such omissions and inaccurate inclusions indicate that residents employ a range of practices when reporting patient medications in their written signouts. Such process variation may contribute to medical errors. Standardization, through intensive training, formal templates, and computerized signout linked to the electronic medical record, may improve the accuracy and reliability of written signout and thus imrpove patient safety.
Author Disclosure Block:
C.J. Olson, None; A.R. Vidyarthi, None; B.A. Sharpe, None; R.M. Wachter, None.