Improving physician communication through an automated, integrated ...

Original Contributions

Improving Physician Communication through an Automated, Integrated Sign-Out System Gary Frank, MD, MS; Stephen T. Lawless, MD, MBA; and Terri H. Steinberg, MD, MBA

A B S T R A C T Communication failures among physicians are a leading cause of medical errors. The resident signout sheet is the primary tool used by house staff to facilitate the sign-out process. The resident signout sheet is a structured report, with patient-specific information including demographics, such a patient’s name, age, sex, room number, and attending physician; problem list; medications; and allergies. Some physicians use handwritten notes to keep track of this information, while others use freestanding word processor or database programs. In a previous study, the authors described serious inaccuracies in a manually updated wordprocessor based resident sign-out sheet used by pediatric residents at a tertiary-care children’s hospital.An automated and integrated sign-out system (AISS) was subsequently developed that retrieves pertinent patient information from a computerized provider order entry (CPOE) system. The AISS generates a resident sign-out sheet, which includes demographic information, weight, current medications, allergies, and diet orders, as well as optional free-text information. The AISS has proven to be enormously popular, increasing physician acceptance of CPOE throughout the organization. This paper discusses lessons learned, including technical, design, and workflow aspects of an integrated resident sign-out sheet. The authors recommend that all future commercial CPOE systems incorporate physician sign-out tools such as the one described in this article.

K E Y W O R D S ■

Sign-out ■

■ Communication ■ Medical errors ■ Systems integration ■ Internship and residency Continuity of patient care

Introduction Communication failures among physicians have been described as “an insidious contributor to medical mishaps,”1 and a large review revealed that communication errors were the leading cause of preventable disability or death in Australian hospitals.2 Numerous other studies have demon-

68

Journal of Healthcare Information Management — Vol. 19, No. 4

strated that effective communication is essential to reduce medical errors and improve the quality of care.3,4,5 In a recent report in The New England Journal of Medicine, the authors note that “improving communication” is one of the primary ways that information technology can reduce medical errors.6

Original Contributions A critical period of communication among physicians caring for hospitalized patients occurs at the time of signout, during which patient care is transferred from the day team to the night team and vice-versa. This essential handoff is especially challenging at teaching hospitals, where rotating teams of resident physicians care for large numbers of hospitalized patients. The resident sign-out process has been described as “a precarious exchange” and “one of the most poorly examined transactions in medicine.”7 Peterson et al. showed that preventable adverse events were associated with cross-coverage by a resident from another team.8 The primary tool used by residents to facilitate the signout process, the resident sign-out sheet is a structured report containing a list of patients and pertinent information relevant to their care. The form and content of the sheet varies by physician and institution, but it generally contains the following elements: patient demographics, such as name, age, sex, room number, and attending physician; problem list, including the reason for hospitalization; medications; and allergies.

“The resident sign-out process has been described as ‘a precarious exchange’ and ‘one of the most poorly examined transactions in medicine.’” In a previous study, the authors described serious errors on a manually updated, word processor-based resident sign-out sheet maintained by pediatric residents at a tertiary care children’s hospital.9 The study found that 8 percent of medications on the resident sign-out sheet were not actually ordered; 22 percent of ordered medications were not listed on the sheet; and 67 percent of patients had at least one content error on the sheet. These errors occurred despite tremendous efforts by the residents to maintain a complete and accurate resident sign-out sheet. These inaccuracies create the potential for serious medical errors, and the significant effort required to maintain the resident sign-out sheet led to considerable frustration on the part of the residents. A project was developed to create an automated integrated sign-out sheet (AISS), in which elements such as demographics, medications and allergies are retrieved from a CPOE system. Additionally, the AISS enables users to enter free-text information about their patients, such as a brief medical history, pertinent results, and pending labs and studies, which can be updated at any time. The AISS was quickly adopted by most residents and many attending physicians at the Alfred I. duPont Hospital for Children. The project has improved physician satisfaction overall with the CPOE system. This article reviews

several different strategies for maintaining and updating a resident sign-out sheet. Design considerations and technical aspects of the AISS are explored. Finally, lessons learned during the design and implementation of the AISS are discussed, and recommendations are made to encourage commercial vendors to incorporate these types of communication tools into future CPOE systems. Strategies for Maintaining a Sheet A number of different strategies have been employed to create and manage the resident sign-out sheet. Moreover, the specific content of the sheet varies by physician preference and institutional culture. The simplest way to maintain a resident sign-out sheet is with pen and paper. More sophisticated systems use a word processor or database program. A few institutions let their residents use a Webbased service. Some physicians use commercial patienttracking products that were developed for mobile platforms, such as the PDA. In the research process, only two other reports were found that described an integrated resident sign-out sheet. Many residents maintain their sign-out sheets the oldfashioned way: pen and paper. Some CPOE or electronic medical record products have the ability to print a simple list of patients with some demographic information, such as room number, medical record number, and date of birth, with space for handwritten notes. The pen-and-paper process permits a great deal of flexibility in terms of the content and level of detail of the sign-out process. Disadvantages include the limited amount of information that is generally recorded, the excessive time required to write that information, and the high likelihood of recording inaccurate information or omitting important information, especially regarding medications and allergies. Perhaps the most common method of maintaining a resident sign-out sheet is with a commercial word processor or database program, such as Microsoft Word or Microsoft Access, from Microsoft Corp., Redmond, WA. Often, the file containing the resident sign-out sheet is stored on a computer in the resident’s lounge or on a disk, which can be handed from resident to resident. In many cases, the resident sign-out sheet is e-mailed among residents or resides on a networked drive. While the use of word processors and databases is somewhat more efficient than keeping a handwritten resident sign-out sheet, there are a number of disadvantages. The information entered into these systems is often inaccurate, which creates the potential for serious medical errors. Moreover, these products raise important privacy concerns because large amounts of protected health information reside in files and on disks that may not be password-protected. Finally, residents often spend an inordinate amount of time maintaining these files, which detracts from time spent in clinical care or


69

Original Contributions educational activities. A few commercial products offer Web-based sign-out systems. One example is WebForMDs’ Resident Patient Signout by MicroMed Strategy, Vienna, VA, which says it is a HIPAA-compliant program to help residents manage, track, and share information about their patients.10 These programs offer a number of advantages, including password protection and easy access from any Internet-accessible computer. However, they still require residents to manually enter information, an error-prone and time-consuming process. Also, the benefits of these products must be weighed against the potential risks of making protected health information available on the Web; physicians contemplating the use of a Web-based commercial sign-out system should get permission from their risk management department. A number of commercial patient-tracking products have been developed for the personal digital assistant. These include PatientKeeper, by PatientKeeper, Brighton, MA, Patient Tracker from Shatalmic, Layton, UT, and WardWatch, by Torlesse Systems, Melbourne, Australia. Some of these products are quite sophisticated and include features such as connectivity with existing information systems, electronic prescription capabilities, and charge capture tools. When utilized to their full potential, these products offer a nice solution for physicians who prefer the PDA platform. However, to integrate these products with existing information systems requires a significant infrastructure and often puts organizations in the position of requiring all physicians to have a similar type of PDA device to use the system. Previous Integrated Sign-out Systems Research for this project identified several efforts to develop integrated sign-out systems. UWCores. Physicians at the University of Washington developed UWCores, a system that “combines the patient sign-out and daily ward work information in one central location.”11 UWCores receives hourly downloads of selected demographic and allergy information from the hospital’s clinical information system into a local database. The remaining data, such as problem list, medication list, and study results, are entered by the residents. These two sets of information then are combined to produce various reports, including sign-out reports, which are Web-based and available through any secure Internet connection. According to the published report, UWCores is wellreceived by physicians at the University of Washington, and it clearly offers many advantages over other systems. However, medication information still is entered manually, creating the potential for error. Mount Sinai Medical Center. In 1999, Kannry and Moore described an innovative Web-based sign-out system called MediSign, which was developed at Mount Sinai Medical Center.12 MediSign used hypertext markup language

70


for form development and was connected to a Microsoft Access database. Residents accessed the password-protected system from the hospital’s intranet and were prompted to enter patient-specific information based on the selected medical service, such as general surgery or internal medicine. Recently, Mount Sinai introduced a new sign-out system that is integrated with the hospital’s enterprise data repository, credentialing system and an external resident scheduling system.13 Patient demographics, attending data, clinical data and scheduling information are automatically transferred from these systems to an Oracle database residing on a R6000 AIX machine. The front end is written in Java and JSP and resides on a Solaris Web server. The A.I. DuPont Hospital for Children The Alfred I. duPont Hospital for Children is a freestanding tertiary-care children’s hospital with approximately 8,500 admissions annually. CPOE has been used at the facility since the 1980s. In 1999, PowerChart, from Cerner Corp., Kansas City, MO, was implemented and is used for all orders, including medication orders. Allergy and weight information are required before placing medication orders. Decision support includes checking for potential allergies, drug-drug interactions and duplicate orders. An integrated pharmacy system produces bar-coded medication labels, which are currently being used on a pilot basis. As is customary at most teaching hospitals, patients at duPont Hospital are assigned to a physician team consisting of an attending physician, two to four residents and several medical students. Each team is responsible for the care of approximately 10 to 20 patients. The hospital’s AISS was designed to replace a word processor-based resident sign-out sheet that had been used for several years that is fairly typical of the resident sign-out sheet maintained by residents around the country. The resident sign-out sheet included a header with team-specific information, such as the name of the team, pager numbers of the residents and current date; patient demographics, such as the name, age, sex, medical record number, room number, and attending physician; other patient-specific information, such as allergies, weight, diagnosis, medications, laboratory, and study results; and a to-do list. This information was printed in a table format with an average of two to three patients per page. All information was updated at least one to two times per day, and the overnight resident spent about 30 to 45 minutes per night maintaining the document. The AISS was designed to mimic the prior resident sign-out sheet in terms of content while improving its accuracy and minimizing maintenance efforts. The system was developed after a series of interviews with the residents to determine the optimal form and content of an AISS. Although the AISS was developed using standard tools


Figure 1.

Figure 2.

“A number of commercial patient-tracking products have been developed for the personal digital assistant.” provided with the CPOE system, duPont Hospital is believed to be the first institution to customize the system to produce an AISS. The AISS contains information about patients on a shared team list (see Figure 1), known as a care team. Providers access the report by launching a password-protected reporting tool directly from the CPOE system. After launching the reporting tool, providers select the report of

interest—in this case, the resident sign-out sheet—and are prompted to enter the name of the care team. The report is generated in approximately 10 to 30 seconds, and can be printed by any networked printer. An electronic form (see Figure 2) can be used to enter patient-specific free-text information, which is stored in the CPOE database. The free-text information can be modified at any time and is automatically included in the report. Figure 3 demonstrates a sample AISS report, which contains the following components: Report title. The title of the report includes the name of the team, the author of the report, and the date and time the report was printed. Report header. This includes a standard list of phone numbers and pager numbers, which are hard-coded into the report. Future versions of the AISS should enable customization of the report header according to the preferences of team members. Patient demographics. This includes patient name, room number, age, sex, weight, medical record number, date of birth, date of admission, date of discharge if applicable, medical service, and attending physician. This information is automatically included in the report. Free-text information. Four categories of free-text information can be included in the report by entering them into an electronic form. They include a brief history of the present illness and past medical history, important results, consulting services, and pending labs and studies. Allergies. Current, active allergies are automatically printed on the report.


71


Figure 3. Medications. Current, active medication orders, including order details such as dose, route, and interval, are displayed in reverse chronologic order. Diet. Current active diet orders are displayed in reverse chronologic order. Signout/to-do list. A space on the report is provided for residents to take handwritten notes and keep their personal to-do list. On-call residents print the AISS at the start of their overnight shifts. The sign-out process from the day team to the night team is facilitated by frequent reference to the AISS. The AISS is generally carried or kept in a coat pocket for easy reference throughout the shift. When a new patient is admitted, the patient is added to the care team list, and free-text information is entered into the electronic form. When the day-team arrives, a new AISS is printed. The report is used as a convenient place to manually record information, such as lab results or vital signs, which may be requested during attending rounds. Thus, the AISS also is used as a rounds report. Free-text information is

72


updated by the daytime resident in the electronic form before the end of the shift and the arrival of the night team. Adoption of the AISS The AISS was introduced on a pilot basis in May 2004, and all pediatric residents received instruction on its use the next month. The residents were encouraged to use the system initially for patients admitted to the general pediatrics service. Use of the AISS was purely voluntary and was not mandated by the administration or medical leadership. An audit of the free-text electronic forms was performed to monitor usage of the AISS in the first year since its introduction. In total, forms were created for 5,208 inpatients, or 64 percent of the 8,103 inpatients admitted from June 1, 2004, to May 31, 2005. On average, each form was modified 6.3 times during patient hospitalization. One form was modified 194 times on a patient who spent much of the year in the hospital. Usage of the AISS was further evaluated according to

Original Contributions medical service. For example, 76 percent of patients on the general pediatrics service were included in the AISS in the first month the system was introduced. By the second month, 96 percent of general pediatrics patients were included, and utilization remained well above 90 percent for the remainder of the year. Soon after introduction of the AISS to the general pediatrics service, many of the subspecialty attending physicians expressed interest in the system. For example, the hematology/oncology service began using the system in July 2004, and by September 2004, the AISS was utilized for more than 90 percent of their patients. The surgery services have been the slowest to adopt the AISS. Fewer than 20 percent of general surgery and orthopedic surgery patients have been entered into the system. In general, these physicians cite their desire to carry a one page sign-out sheet as the reason for not using the AISS. Lessons Learned Numerous residents and attending physicians have expressed great enthusiasm for the AISS. In fact, physician approval of CPOE has increased since the introduction of the AISS. Physicians recognize that the purpose of the AISS is to improve their efficiency and to facilitate a process that can otherwise be haphazard and error-prone. They appreciate that the objective of this project was to improve physician workflow that was time-consuming and ineffectual. Conversely, physicians have become fully reliant on the AISS and have expressed much angst during rare periods of system downtime.

“…physician approval of CPOE has increased since the introduction of the AISS.” An important consideration in designing an AISS is report efficiency and its effect on system resources. These reports access a tremendous amount of data from very large database tables, such as orders, demographics, and results tables. Initial versions of the AISS report took from one to three minutes to run, which was unacceptably long. The average report time was decreased to 10 to 30 seconds by careful attention to the design of the report and the number of database tables accessed. So far a significant effect on the performance or efficiency of the CPOE system has not been observed. However, it is conceivable that there could be an unacceptable effect on system performance in a larger hospital serving more patients and physicians. Another consideration is the tradeoff between quantity of information and length of the report. For example, the AISS does not currently include laboratory and imaging results other than those manually entered into the electronic form. These types of results have not been formatted in a manner that would still allow multiple patients to be printed on a

single page. Because residents tend to carry the report with them throughout their shift, they insist that the entire report not be more than a few pages long. The success of the AISS has led to the implementation of a number of other reports and processes that are based on the same design principles. The concept of printing a report based on a computerized list of patients led to the development of a billing report, which is printed by physicians at the end of each day. The report replaces handwritten billing cards or billing sheets, which tend to be more timeconsuming and less complete. The concept of entering data into an electronic form and then printing a report has led to a computerized discharge process, which replaced handwritten discharge instructions. Essentially, residents enter elements of the discharge instructions into an electronic form. When a discharge order is placed, printed discharge instructions are automatically generated, which incorporate data from the electronic form as well as other data that already resides in the database. Among other benefits, this discharge process has dramatically improved the legibility and completeness of patient discharge instructions. Conclusion Other authors have described the barriers to physician acceptance of medical information systems.14,15 One way to improve physician acceptance is to provide tools that improve efficiency and fix flawed workflows. The sign-out process is an example of an inefficient and error-prone process that can be improved with computerized tools and reports. The AISS gives physicians needed information, most of which already exists in the hospital’s information system. Physicians often complain about the amount of time that it takes to put information into CPOE and EMR systems. Physician acceptance and even endorsement of these systems may be achieved by providing tools and reports that improve their efficiency and workflow. An AISS should be a standard component of commercial CPOE and EMR systems. The AISS should be customizable, based on the specific preferences of each physician. Physicians should be able to select not only the type of data but also the placement of the data within the report. Similar tools can be used to address other inefficient workflows, including physician billing and patient discharges. About the Authors Gary Frank, MD, MS, FAAP, is an attending physician in the Division of Pediatrics and the Physician Expert for Clinical Informatics at the Alfred I. duPont Hospital for Children. Stephen T. Lawless, MD, MBA, is a professor of pediatrics in the Department of Anesthesiology/Critical Care at the Alfred I. duPont Hospital for Children and the Chief


73

Original Contributions Quality and Knowledge Officer for Nemours. Terri Steinberg, MD, MBA, joined Nemours in 2003 as the clinical applications manager at the Alfred I. duPont

Hospital for Children, where she is responsible for inpatient clinical information systems.

References 1. Sutcliffe KM, Lewton E, Rosenthal MM. Communication Failures: An Insidious Contributor to Medical Mishaps. Acad Med 2004;79(2):186-94. 2. Wilson RM, Runciman WB, Gibberd RW, et al. The Quality in Australian Healthcare Study. Med J Aust 1995;63:458-71. 3. Leonard M, Graham S, Bonacum D. The Human Factor: The Critical Importance of Effective Teamwork And Communication Inproviding Safe Care. Qual Saf Health Care 2004;13(suppl 1):i85-90. 4. Levine SR. Talk Is Cheap: Communication Reduces Costly Medical Errors. Mater Manag Health Care 2004:13(7):27-8. 5. Southwick LM. Communication Misadventures and Medical Errors. Jt Comm J Qual Improv 2002;28(8):461-2. 6. Bates D, Gawande AA. Patient Safety: Improving Safety with Information Technology. N Engl J Med 2003;348(25):2526-34. 7. Muhkerjee S. Becoming A Physician: A Precarious Exchange. N Engl J Med 2004;351(18):1822-4. 8. Peterson LA, Brennan TA, O’Neil AC, et al. Does Housestaff Discontinuity Of Care Increase The Risk For Preventable Adverse Events? Ann Intern Med 1994;121(11):866-72. 9. Frank G, Lawler LA, Jackson AA, Steinberg TH, Lawless ST. Resident Miscommunication: Accuracy Of The Resident Sign-Out Sheet. J Healthc Qual Online March/April 2005;W2-10. 10. http://www.webformds.com/resident_signout.htm, accessed May 16, 2005. 11. Van Eaton EG, Horvath KD, Lober WB, Pellegrini CA. Organizing The Transfer Of Patient Care Information: The Development Of A Computerized Resident Sign-Out System. Surgery 204;136(1):5-13. 12. Kannry J, Moore C. MediSign: Using A Web-Based Signout System To Improve Provider Identification. AMIA Annu Symp Proc 1999;550-4. 13. Kushniruk A, Karson T, Moore C, Kannry J. From Prototype to Production System: Lessons Learned From The Evolution Of The Signout System At Mount Sinai Medical Center. AMIA Annu Symp Proc 2003;381-4. 14. Johnson KB. Barriers That Impede the Adoption Of Pediatric Information Technology. Arch Pediatr Adolesc Med 2001;155(12):1374-9. 15. Lowenhaupt M. Removing Barriers to Technology. Physician Exec 2004;30(2):12-4.

Congratulations CPHIMS Recipients! Congratulations to the following individuals who have achieved the Certified Professional in Healthcare Information and Management Systems (CPHIMS) credential from April 23 through July 28, 2005. The CPHIMS credential recognizes those individuals who have met the eligibility requirements and passed the examination. CPHIMS encourages continued personal and professional growth in the practice of healthcare information and management systems and provides a national standard of knowledge required for certification. In addition, the credential assists employers, the public, and members of the healthcare profession in the assessment of a healthcare information and management systems professional.

Phyllis M. Bachand—Monee, IL Steve Cherry—Brentwood, TN Regina M. Cullen—West Patterson, NJ Howard Z. Davis—San Jose, CA Gary L. Fuller—Cooper City, FL Andrea D. Lott—Waterford, VT Gail M. Maestas—Oak Hill, VA Nympha Meindel—Massapequa, NY Heather L. Nelson—Cedar Springs, MI Audrey Paulman—Lincoln, NE Ken Roderman—Plano, TX Jeff Roggensack—Reno, NV Pamela D. Scott—Oak Point, TX Lisa A. Sherwin—Chandler, AZ Diane J. Spencer—Silver Spring, MD Peter Warchol—Levittown, PA Ron Wheelock—Rochester, MI Jesus Zarate—Luke Air Force Base, AZ

Join this elite group. Go to www.himss.org/getcertified to find out more about the CPHIMS certification program. To see a full listing of certificants go to http://www.himss.org/content/files/cphims_certificants.pdf.

74
