Discontinuity of Care: Further Thoughts on Standardized Processes
"M anaging Discontinuity in Academic Health Centers: Strate- gies for Safe and Effective Resident Sign-Out," by Vidyarthi et al., highlights the complexity of our health care system and identifies some of the challenges conscientious physicians face daily. 1 The article focuses on the growing number of patient handoffs that occur in academic medical centers, now estimated at around 300/resident/ month 2 but obviously varying based on the size and structure of an individual program. Added to this is the inaccuracy of information provided to accepting physicians during such sign-outs. 4-7 The authors' recommendations include the use of computerized tools to standardize sign-outs and having formal verbal sign-out to allow the opportunity for questions and discussion. 1 These suggestions echo JCAHO hospital patient safety goal 2E for 2007. 3 A recent morbidity and mortality conference at our institution demonstrated 2 additional issues related to handoff of care. First, this article and others fail to provide data supporting the use of any specific content during the sign-out process, and indeed we could find no published research to assist us with developing standardized content. Second, the typical sign-out, whether written or verbal, has the potential to create anchoring bias if the information is not subjected to healthy skepticism by the receiving physician.
Because of the lack of data on which elements of sign-out actually help covering physicians care for patients, we are left with recommendations based on opinion or personal experience. We believe rigorous investigation of this issue is essential to creating a safe, effective, and efficient process. Given that most institutions have or are implementing electronic medical records, medication and allergy information as well as laboratory data are available electronically and thus may be safely omitted from printed signout. In fact, when such information is included in sign-out, it is often inaccurate based on the available evidence 6,7 and our own experience. Unless the hospital information system and the signout tool are linked and the information is automatically downloaded, we may be providing false information to each other because it is either outdated or transcribed incorrectly. Supplying this data, then, may result in an increase in errors. As an example, a resident is called to prescribe pain medications for a postoperative patient. Her sign-out lists the patient being on 15mg of sustained-released morphine twice daily, so she increases this to 30 mg twice daily and adds a PRN dose of immediate-release morphine as well. Unbeknownst to her, the 15-mg morphine order had actually been discontinued 36 hours earlier when the patient was noted to be oversedated, so her order for 30 mg twice daily results again in oversedation and, even worse, respiratory depression nearly leading to ICU transfer.
The second issue we identified is anchoring bias, which occurs when preformed ideas about a presenting complaint or diagnosis prevent us from generating new hypotheses, even when the available evidence points away from the initial impression. Another example helps to explain the point. A patient with previously diagnosed COPD is admitted to the hospital with increased dys-