Clinical and Health Affairs
Inaccuracy in Patient Handoffs: Discrepancies between Resident-Generated Reports and the Medical Record
By Michael J. Aylward, M.D., Tyson Rogers, M.S., and Peter G. Duane, M.D.
■ A handoff is the transfer of a patient’s care from one provider to another. It usually involves both a verbal and written exchange of information. Although written handoff sheets are critical to good patient care, there is little data on the quality of information they contain. We conducted a study to assess the accuracy of handoff sheets used in one Minneapolis internal medicine residency program. We compared the accuracy of information about code status, medication allergies, medications, and problems recorded on the handoff sheet with that in the patient’s medical record. We found errors were common in resident handoff sheets. Only 83 (19%) of 428 handoff sheets contained no errors. The most common error was one of omission on the medication list (69% of the handoff sheets contained a medication omission). The percentage of patient handoff sheets with code-status errors was 5.7%, and the percentage with medication allergy errors was 2.8%. Important problems were omitted from the problem list in 22% of cases.
Communication among physicians is central to the safe and effective care of hospitalized patients. The work of caring for these patients is distributed over a number of providers, hence the rotating night call system that is a component of many residency programs. This system has made the transfer of patient information between physicians necessary and important. A handoff is the transfer of care of a patient from one provider to another.
Because of the resident duty work-hour restrictions mandated by the Accreditation Council for Graduate Medical Education (ACGME), a great deal of emphasis has been placed on efficient and effective handoffs among residents. As additional restrictions take effect, the ability to perform thorough, accurate handoffs will become even more critical.
Several studies have shown that poor continuity of care leads to more adverse events among hospitalized patients.1,2 For example, research has shown that mistakes related to medications made during the handoff are potentially harmful.3,4 Furthermore, care provided by physicians other than those on the primary team has been found to prolong hospital stays and increase the cost of care; this may be a consequence poor-quality handoffs.2 The Joint Commission has identified improving handoffs as a national patient safety goal, citing problems with communication as “the single most frequent cause of medical errors.”5 In addition, the ACGME has designated interpersonal and communication skills and professionalism as core competencies of physicians.6 The ability to complete an effective handoff encompasses these competencies.
The term “handoff” is synonymous with “sign out” and “check out.” However, it does not imply cessation of patient care or relinquishing responsibility for the care of a patient. Handoffs usually include a verbal as well as a written component. The verbal component is often done informally in a busy work environment. The written component, the handoff sheet, is a document that is typically passed from the primary team to the covering team. The information in the handoff sheet should be succinct, accurate, legible, relevant, and informative.3
In most residency programs, transfer of care is an ad hoc process. Residents and medical students typically maintain patient lists in a spreadsheet or word processing template that they update daily. Patient information comes from the computerized medical record, a physician’s personal note-taking system (eg, index cards), or a physician’s memory. Updating the written information is often the last thing a resident does before going home for the day.
We conducted a study to assess the accuracy of information contained in the handoff sheets in one Minneapolis internal medicine residency program.
Methods
The University of Minnesota’s internal medicine residency program is ACGME-accredited and includes categorical, medicine/pediatrics, and preliminary residents. There are 23 categorical, 10 medicine/pediatrics, and four preliminary residents in each class. All residents rotate through three tertiary care teaching hospitals—the University of Minnesota Medical Center, Fairview in Minneapolis; Regions Hospital in St. Paul; and the Minneapolis Veterans Affairs Medical Center (VAMC). The Minneapolis VAMC was the site chosen for our study because of its robust electronic medical record system and centralized access to handoff sheets. The Minneapolis Veterans Affairs Internal Review Board approved the study.
The medicine teaching services at the Minneapolis VAMC consisted of five general medicine teams, three cardiology teams, and one intensive care team. Each medicine team, as well as the ICU team, included an attending physician, a senior resident, an intern (a first-year resident), and one or two medical students. One general medicine team had a first-year psychiatry resident who functions as a medicine intern. The three cardiology teams were each composed of a senior resident and an intern; the teams shared an attending physician. The cardiology and ICU teams provided cross-coverage on call nights. Interns typically handed off their team’s patients at the end of a workday and provided cross-coverage duties on a call day. The team that was primarily in charge of a patient’s care updated the handoff sheet.
The handoff sheets they used were electronic spreadsheets. Their format and content were left to the discretion of the residents. A single handoff sheet may have contained information on as many as 16 patients. At the end of the day, each handoff sheet was updated manually, printed out, and given to the resident who was covering that night. The following morning, the resident returned the handoff sheet to the primary team and updated them on events that occurred during their absence.
In a pilot study, we compared information listed in all of the domains on the handoff sheets (demographic information, patient location, code status, medication allergies, problem lists, and medication lists) with that in the medical record over one week. Using information from the pilot, a physician focus group identified the most important handoff sheet domains: code status, medication allergies, the problem list, and medications. Criteria for assessing accuracy were established based on clinical importance, feasibility, and a review of the literature.
For the full study, we collected data from the four domains identified by the focus group for two nonsequential complete call cycles for each team during January and February of 2006. The time frame was chosen so that interns would have gained about six months of experience with cross-covering and creating handoff sheets since starting their program. Different residents were working during the two call cycles. Each day of the call cycle, the handoff sheets were printed out in the morning, ensuring that they were identical to those carried by the intern the night before. Information from each of the four domains on the handoff sheet was compared with that in the medical record.
A worksheet was created to facilitate abstraction of data from the handoff sheets and medical record. Unique confidential patient and team identifiers were used. Code status and medications were coded as being correct, having errors of omission, or having errors of content. An error of omission meant that information was in the medical record but not on the handoff sheet. An error of content meant that information on the handoff sheet was not in the medical record or was incorrect. The total number of medications prescribed was also noted. Medication allergies were coded as correct or omitted.
Abstraction was performed every morning during the study period by a physician who was not involved in patient care. Entries regarding code status, medications, and medication allergies on the handoff sheets were compared with those in the computerized order system. The problem list was compared with progress notes made by nurses, residents, interns, and attending physicians during the previous two days. The criteria for a problem-list discrepancy was the omission of a major problem that potentially or actually required intervention. These criteria were designed by the authors to capture omissions with practical relevance to a cross-covering intern. The day of the week and the day of the call cycle were also captured.
Because residents often intentionally omit a number of medications from handoff sheets, we generated a list of medications that, if omitted from the handoff sheet, would not be counted as errors. On that list were vitamins, psyllium, Maalox, acetaminophen, iron, all stool softeners and laxatives, calcium, sublingual nitroglycerin, guafenesin, and all inhaled and nebulized medications. These medications are commonly prescribed and generally do not lead to adverse outcomes. Several of the medications, including sublingual nitroglycerin, are on the hospital’s admission order set.
We determined error rates for each team in each domain. For code status and medication allergy status, we identified the percentage of handoff sheets that were discrepant from the medical record. For the problem list, we used the percentage of days in which there was a discrepancy between the problems listed on the handoff sheet and those in the medical record. We expressed medication errors as a ratio (the number of drugs omitted from the handoff sheet and the number of content errors divided by the correct number of medications). We then determined the mean discrepancy rate for each domain (Table).
Results
A total of 428 patient handoff sheets were created by 36 residents over the course of the two nonconsecutive call cycles. The handoff sheets contained information on 186 patients. One handoff sheet was deleted, and its information was not captured because of a computer error.
Eighty-three (19%) of the 428 patient handoff sheets were completely without error. The percentage of patient days with a code status handoff error was 5.7%, and the percentage of patient days with a medication allergy error was 2.8%. Important problems were omitted from the handoff sheet 22% of the time. Examples of problems omitted from the problem list include changes in mental status, urinary tract infections, knee effusions, pneumonias, congestive heart failure, acute renal failure, line infections, anemia if transfused within 48 hours, alcohol withdrawal, atrial fibrillation or flutter requiring rate control, and spontaneous bacterial peritonitis. The most common error on the handoff sheet was an omission on the medication list (69% of the handoff sheets contained medication omissions). Twenty-six percent had one medication omission, 44% had more than one, and 5.6% had more than six medication omissions. The rate of discrepancies between the handoff sheets and the medical record is shown in the Table.
Handoff sheets were likely to have an error on the first day rather than on subsequent days. Eighty-three percent of code-status errors, 100% of medication allergy errors, 66% of problem-list discrepancies, and 89% of medication discrepancies occurred on Day 1 of observation. When there was a patient handoff error on the first day of observation, it was significantly more likely that there would be additional errors on subsequent days, as compared with when there was no error on the first day: 100% vs. 1.6% for code status errors, 80% vs. 0% for medication allergy errors, 74% vs. 23% for problem-list discrepancies, and 96% vs. 43% for medication discrepancies. The P-value for all comparisons was <0.001 using a chi-square test. Day of the week was not associated with increased errors. Errors tended to occur more frequently later in the call cycle, but this became significant only for medication allergy errors (P=0.005), which were most common for patients admitted on the last day of the cycle.
Discussion
Our data show that errors are common on manually created resident handoff sheets. The most common ones involved medications and problem lists. Although code status and medication allergy errors were less common, their potential consequences can be serious. A previous study showed that residents began resuscitation on patients whose code status was DNR based on inaccurate information on handoff sheets.4 More concerning is the case of a patient who wishes to be full code and no attempt at resuscitation is made. It is unlikely that a handoff sheet would be the sole source of advanced directive information in such a situation. However, a patient’s code status may influence how a physician views that patient (eg, providing less aggressive medical care to someone who is DNR/DNI). Regardless, the fact that the information is believed to be important enough to communicate to covering providers dictates that it be correct.
For one of every six medications listed on the handoff sheet, there was a discrepancy with the medical record. In all cases, the patient was either on a different medication or a medication was listed that the patient was not taking. Multiple errors regarding medications were common. For example, 44% of handoff sheets omitted more than one medication the patient was taking, and 5.6% omitted more than six medications. A clinically significant problem was omitted from the problem list 22% percent of the time. The high error rates in the medication and problem lists are likely because the information in these domains often changes during the course of a patient’s hospitalization. Code status and medication allergies rarely change during a hospitalization.
The first day that a patient is handed-off is a critical point. Virtually all of the code-status errors, all of the medication allergy errors, and most of the medication and problem-list errors on the handoff sheets occurred on the first day of admission. The first-day errors are likely the result of the fact that completing the handoff sheet is a low priority for residents on busy admitting days. The first day is also when the bulk of data must be entered into the document.
The persistence of errors and the occurrence of new ones speak to the residents’ failure to update handoff sheets. The reasons residents do not update handoff sheets are myriad. Fatigue, interruptions, and workload all likely contribute to making errors. Also, handing off 16 patients is more complex than handing off only a few. The level of involvement of the senior resident or the attending in the creation of handoff sheets may also affect the error rate.
Few articles in the medical literature describe what constitutes the “best practice” for handing off patient care. The handover process in the aviation and aerospace industries, which also have complex work environments, is more regimented and structured than it is in hospitals. The airline industry has long used checklists for preflight and inflight processes. In the aerospace industry, handovers are scheduled events, and no interruptions are allowed. Audio flight controller loops, checklists, and the involvement of the incoming controller in identifying potential problems facilitate the process.7
Physicians can apply several lessons from these industries to ensure that patient handoffs are accurate, standardized, useful, and succinct. First, the importance of the handoff needs to be recognized, and a formal curriculum to improve handoff skills should be implemented.8 The curriculum should emphasize the importance of the initial handoff and the need for thorough daily updates after that. The majority of internal medicine programs in the United States do not provide any handoff education to their residents.9 Residency programs instead use lectures and coaching by attending physicians to improve resident handoffs. At the University of Minnesota, all internal medicine and medicine-pediatrics residents attend a handoff workshop during their orientation.10-12 Second, sufficient time should be given to the process, and interruptions should be limited.8 Third, handoff sheets should be standardized and their creation automated to decrease human error. Ideally, handoff sheets would be automatically created from within an electronic medical record and then printed out or updated online. More systematic creation of handoff sheets results in fewer adverse events, and standardizing the forms results in better-quality handoffs.13-18 Since this study, the Minneapolis VA has implemented a system for automatic creation of hand-off reports from the medical record. Although a standardized handoff sheet may decrease the likelihood of errors during the handoff, care must be taken to not make them overly prescriptive, as that may inhibit information exchange.8
Our study has several limitations. Patient outcome data were not collected, and, therefore, it is not clear whether errors on handoffs had any effect on patient outcomes. In addition, our audit did not capture the verbal component of the handoff, which may have provided more information than what was available on the handoff sheet. Another limitation was the fact that even though the residents involved in our study are representative of those in the residency program as a whole and the handoff process is consistent with that used at other teaching sites, our audit consisted of a discrete sample of residents at a single institution during a limited time period. The abstraction was done entirely by one of the authors, which may have led to bias or errors. Bias may have played a role in the assessment of problem-list omissions; however, the other measures are objective comparisons between the medical record and the handoff sheet. Errors also may have been introduced during the manual abstraction process. Our data reveal the quantity of errors in written handoffs, but they provide only circumstantial evidence about the severity of those errors.
Statistically, counting the same error multiple times can confound and inflate the percentage of errors. To account for this, we summarized errors across days for each patient. An uncorrected error and a corrected error followed by a new error counted equally. Furthermore, during a call cycle, a given intern will be cross- covering once, so any errors introduced onto the handoff sheet will be novel to that intern. Counting an error multiple times has validity from the perspective of the interns, each of whom will use the information on the handoff sheet to make clinical decisions. Finally, the VAMC’s medical record system might have ameliorated some handoff errors because it provides easy access to patient information.
Conclusion
Patient handoffs are an important part of inpatient care especially in the context of resident duty-hour restrictions. Our study found that manually created handoff sheets provide inaccurate information or even fail to provide critical patient information to covering teams. Standardizing handoffs, automating the process, and emphasizing their importance may improve handoff accuracy and patient care. MM
Michael Aylward is an assistant professor and Tyson Rogers was a biostatistician in the University of Minnesota’s department of medicine. Peter Duane is an associate professor in the department of medicine at the Veterans Affairs Medical Center.
REFERENCES
1. Petersen LA, Brennan TA, O’Neill AC, Cook EF, Lee TH. Does housestaff discontinuity of care increase the risk for preventable adverse events? Ann Intern Med. 1994;121(11):866-72.
2. Brennan TA, Localio AR, Leape LL, et al. Identification of adverse events occurring during hospitalization. A cross-sectional study of litigation, quality assurance, and medical records at two teaching hospitals. Ann Intern Med. 1990;112(3):221-6.
3. Arora, V, Johnson J, Lovinger D, Humphrey HJ, Meltzer DO. Communication failures in patient sign-out and suggestions for improvement: a critical incident analysis. Qual Saf Health Care. 2005:14(6):401-7.
4. Arora V, Kao J, Lovinger D, Seiden SC, Meltzer D. Medication discrepancies in resident sign-outs and their potential to harm. J Gen Intern Med. 2007;Dec;22(12):1751-5.
5. Joint Commission on Accreditation of Healthcare Organizations National Patient Safety Goals. Available at: www.jointcommission.org. Accessed April 8, 2011.
6. Accreditation Council For Graduate Medical Education Outcomes Project, Available at: www.acgme.org/outcome/comp/compMin.asp. Accessed April 8, 2011.
7. Patterson ES, Woods DD. Shift changes, updates, and on-call architecture in space shuttle mission control. Comput Support Coop Work. 2001;10(304): 317-46.
8. Patterson ES. Structuring flexibility: the potential good, bad and ugly in standardisation of handovers. Qual Saf Health Care. 2008;17(1):4-5.
9. Horwitz LI, Krumholz HM, Green ML, Huot SJ. Transfers of patient care between house staff on internal medicine wards: a national survey. Arch Intern Med. 2006;166(11):1173-7.
10. Arora VM, Johnson JK, Meltzer DO, Humphrey HJ. A theoretical framework and competency-based approach to improving handoffs. Qual Saf Health Care. 2008;17(1):11-4.
11. Horwitz LI, Moin T, Green ML. Development and implementation of an oral sign-out skills curriculum. J Gen Intern Med. 2007;22(10):1470-4.
12. Chu ES, Reid M, Schulz T, et al. A structured handoff program for interns. Acad Med. 2009r;84(3): 347-52.
13. Lee LH, Levine JA, Schultz HJ. Utility of a standardized sign-out card for new medical interns. J Gen Intern Med. 1996;11(12):753-5.
15. Ram R, Block B. Signing-out patients for off-hours coverage: comparison of manual and computer-aided methods. Proc Annu Symp Comput Appl Med Care. 1992;16:114-8.
16. van Walraven C, Rokosh E. What is necessary for high quality discharge summaries? Am J Med Qual. 1999;14(4):160-9.
17. Petersen LA, Orav EJ, Teich JM, O’Neil AC, Brennan TA. Using a computerized sign-out program to improve continuity of inpatient care and prevent adverse events. Jt. Comm J Qual. Improv. 1998;24(2):77-87.
18. Van Eaton EG, Horvath KD, Lober WB, Rossini AJ, Pellegrini CA. A randomized controlled trial evaluating the impact of a computerized rounding and sign-out system on continuity of care and resident work hours. J Am Coll Surg. 2005;200(4):538-45.